Method for designing freeform surface imaging systems

Abstract

A method for designing freeform surface imaging system comprises: constructing a series of coaxial spherical systems with different optical power (OP) distributions; tilting all optical elements of each coaxial spherical system by a series of angles to obtain a series of off-axis spherical systems; finding all unobscured off-axis spherical systems; and then specifying a system size or structural constraints, and finding a series of compact unobstructed off-axis spherical systems; constructing a series of freeform surface imaging systems based on the series of compact unobstructed off-axis spherical system, and correcting the OP of entire system; improving an image quality of each freeform surface imaging systems and finding an optimal tilt angle of an image surface; and automatically evaluating an image quality of each freeform surface imaging system based on an evaluation metric, and outputting the freeform surface imaging systems that meet a given requirements.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims all benefits accruing under 35 U.S.C. § 119 from China Patent Application No. 202010482358.8, filed on May 29, 2020, in the China Intellectual Property Office, the contents of which are hereby incorporated by reference.

FIELD

The present disclosure relates to a method for designing freeform surface imaging systems, especially relates to a method for freeform surface imaging systems with high degree of automation.

BACKGROUND

Freeform surfaces are optical surfaces that lack translational or rotational symmetry. Freeform surfaces not only improve an overall performance of an optical system but also bring novel functions to the optical system. With advancements in optical processing and testing, the freeform surfaces are becoming popular in an optical design field. The freeform surfaces bring more degrees of freedom to the optical system and increase a dimension of aberration equations.

There are three main ways for solving an initial solution of a freeform surface system. First, an existing design is taken as the initial solution. Second, the initial solution is determined based on aberration theories. For example, a nodal aberration theory (NAT) can be used to estimate a potential capability to correct aberrations in a freeform surface system with different structural forms, and enable full utilization of the degrees-of-freedom of freeform surfaces. Third, the initial solution is constructed based on an object-image relationship using direct design methods, such as methods based on differential equations, simultaneous multiple surface (SMS) methods, and construction-iteration (CI) methods.

In recent years, some procedures in conventional methods for designing the freeform surfaces system have been automated. However, conventional methods for designing the freeform surfaces system still highly rely on manual completion and have less automation. Therefore, conventional methods for designing the freeform surfaces system require higher designer experience, and waste time and human resources.

BRIEF DESCRIPTION OF THE DRAWINGS

Implementations of the present technology will now be described, by way of example only, with reference to the attached figures, wherein:

FIG. 1 is a flow diagram of one embodiment of a method for designing freeform surface imaging system.

FIG. 2 is a schematic diagram of one embodiment of a coordinate system and a description method in a coaxial system.

FIG. 3 is a schematic diagram of one embodiment of a global coordinate system, a local coordinate system, and a description method in an Off-axis spherical system.

FIG. 4 is a schematic diagram of one embodiment of describing a structural form of a three-mirror system.

FIG. 5 is a flow diagram of one embodiment of solving a most compact structural form using dichotomy.

FIG. 6 is an entire framework flowchart of the method for designing freeform surface imaging system in FIG. 1.

FIG. 7 shows a curve of an operating efficiency of a calculation task of Example 1 over time.

FIG. 8 shows some of freeform surface imaging systems obtained of Example 1.

FIG. 9 shows a curve of an operating efficiency of a calculation task of Example 2 over time.

FIG. 10 shows some of freeform surface imaging systems obtained of Example 2.

DETAILED DESCRIPTION

The disclosure is illustrated by way of example and not by way of limitation in the figures of the accompanying drawings in which like references indicate similar elements. It should be noted that references to “another,” “an,” or “one” embodiment in this disclosure are not necessarily to the same embodiment, and such references mean “at least one.”

It will be appreciated that for simplicity and clarity of illustration, where appropriate, reference numerals have been repeated among the different figures to indicate corresponding or analogous elements. In addition, numerous specific details are set forth in order to provide a thorough understanding of the embodiments described herein. However, it will be understood by those of ordinary skill in the art that the embodiments described herein can be practiced without these specific details. In other instances, methods, procedures, and components have not been described in detail so as not to obscure the related relevant feature being described. Also, the description is not to be considered as limiting the scope of the embodiments described herein. The drawings are not necessarily to scale, and the proportions of certain parts have been exaggerated to illustrate details and features of the present disclosure better.

Several definitions that apply throughout this disclosure will now be presented.

The term “substantially” is defined to be essentially conforming to the particular dimension, shape, or other feature which is described, such that the component need not be exactly or strictly conforming to such a feature. The term “comprise,” when utilized, means “include, but not necessarily limited to”; it specifically indicates open-ended inclusion or membership in the so-described combination, group, series, and the like.

Referring to FIG. 1, one embodiment is described in relation to a method for designing freeform surface imaging systems comprises steps of:

S1: constructing a series of coaxial spherical systems with different optical power (OP) distributions, wherein the series of coaxial spherical systems with different OP distributions are defined as P₁, P₂, . . . , P_m, . . . , P_M, and a set of P₁, P₂, . . . , P_m, . . . , P_M is defined as a set {P}; and the OP distribution refers to a combination of radius of curvature and spaces between surfaces of the coaxial spherical system;

S2: tilting all optical elements of each coaxial spherical system in the set {P} by a series of angles to obtain a series of off-axis spherical systems, wherein the series of off-axis spherical systems are defined as C_m,1, C_m,2, . . . , C_m,s, . . . , C_m,Sm, and a set of the C_m,1, C_m,2, . . . , C_m,s is defined as a set {C}_m; finding all unobscured off-axis spherical systems from the set {C}_m, wherein the unobscured off-axis spherical systems are defined as C_m,1, C_m,2, . . . , C_m,r, . . . , C_m,Rm, and a set of C_m,1, C_m,2, . . . , C_m,r, C_m,Rm is defined as a set {C}_m; and then specifying system size or structural constraints, and finding a series of compact unobstructed off-axis spherical systems from the set {C}m, wherein the series of compact unobstructed off-axis spherical systems are defined as {tilde over (C)}_m,1, {tilde over (C)}_m,2, . . . , {tilde over (C)}_m,t, . . . , {tilde over (C)}_m,Tm, and a set of {tilde over (C)}_m,1, {tilde over (C)}_m,2, . . . , {tilde over (C)}_m,t, . . . , {tilde over (C)}_m,Tm is defined as a set {{tilde over (C)}}_m;

S3: constructing a freeform surface imaging system based on each compact unobstructed off-axis spherical system in the set {{tilde over (C)}}_m, to obtain a series of freeform surface imaging systems, and correcting an OP of an entire system in a process of constructing the series of freeform surface imaging systems, wherein the series of freeform imaging systems are defined as {tilde over (F)}_m,1, {tilde over (F)}_m,2, . . . , {tilde over (F)}_m,t, . . . , {tilde over (F)}_m,Tm, and a set of {tilde over (F)}_m,1, {tilde over (F)}_m,2, . . . , {tilde over (F)}_m,t, . . . , {tilde over (F)}_m,Tm is defined as a set {{tilde over (F)}}_m;

S4: improving an image quality of each of the series of freeform surface imaging systems in the set {{tilde over (F)}}_m by calculating a surface shape of each optical element of each freeform surface imaging system in the set {{tilde over (F)}}_m and finding an optimal tilt angle of an image surface, to obtained a series of freeform surface imaging systems with different structural forms and different OP distributions; and

S5: automatically evaluating an image quality of each of the series of freeform surface imaging system with different structural forms and different OP distributions based on an evaluation metric and outputting at least one freeform surface imaging system that meets a given requirements.

Referring to FIG. 2 and FIG. 3, a global coordinate system O-XYZ is defined in a space of a coaxial spherical systems. An incident direction of light in a central field (0°) is defined as a Z axis, and a plane perpendicular to the Z axis is defined as an XOY plane. The number of optical surfaces of the coaxial spherical system is N, and the spherical surfaces of the coaxial spherical system are denoted as S₁, S₂, . . . , S_i, . . . , S_N according to an order of light passing. S₀ is a virtual surface in front of the coaxial spherical system, and S_I is an image surface of the coaxial spherical system. The incident direction of lights in the central field (0°) is set along a direction of a unit axis vector OZ, the lights in the central field (0°) emerge from S₀ and reach S₁. As shown in FIG. 2, the coaxial spherical system is rotationally symmetrical about an optical axis Oz. A centre of the spherical surface S_i is located on the optical axis Oz and defined as O_i. The spherical surface S_i intersects the optical axis Oz at a vertex point V_i. A radius of curvature of the spherical surface S_i is defined as r_i, when a vector V_iO_i and the unit axial vector OZ are oriented in the same direction, the sign of r_i is positive; otherwise, the sign of r_i is negative. A distance between the spherical surface S_i and a spherical surface S_i+1 is defined as d_i, and d_i is equal to |V_iV_i+1|. When a vector V_iV_i+1 and the unit axial vector OZ are oriented in the same direction, the sign of d_i is positive; otherwise, the sign of d_i is negative.

Referring to FIG. 3, since the off-axis spherical system does not have a rotational symmetry axis as a reference, a local coordinate system is established at each surface of the off-axis spherical system, and a surface shape is described using the local coordinate system. A chief light ray of the central field is set as the reference for the off-axis spherical system. An incident point of the chief light ray of the central field on the spherical surface S_i is defined as V_i. A local coordinate system V_i-XYZ is established with V_i as origin. A unit axis vector V_iZ and a surface normal direction of the spherical surface at V_i are parallel. A unit axis vector V_iY is in a plane O-YZ of the global coordinate system O-XYZ and is perpendicular to the unit axis vector V_iZ. A unit axis vector V_iX and a unit axis vector OX have the same direction. The chief light ray of the central field intersects the spherical surface S₀ at point V₀ and intersects the spherical surface S_i at V_I. A direction of the unit axis vector V_IZ of a local coordinate system V_I-XYZ is perpendicular to the image surface, and a unit axis vector V_IY is in the plane O-YZ of the global coordinate system and is perpendicular to the unit axis vector V_IZ. A unit axis vector V_IX and the unit axis vector OX have the same direction.

In step S1, the series of coaxial spherical systems with different OP distributions are constructed according to first-order optics and a known focal length. According to the knowledge of matrix optics, a reflection matrix R_i for reflection of the light ray at the spherical surface S_i is:

$\begin{array}{cc}{R}_i=\left[\begin{array}{cc}1& \left(n_i-n_{i-1}\right)/r_i\\ 0& 1\end{array}\right],& \left(1\right)\end{array}$

where n_i represents a refractive index of a medium between the spherical surface S_i and the spherical surface S_i+1; n_n−1 represents a refractive index of a medium between the spherical surface S_i−1 and the spherical surface S_i; and r_i represents the radius of curvature of the spherical surface S_i.

A transition matrix of a light ray propagating from the spherical surface S_i to the spherical surface S_i+1 is:

$\begin{array}{cc}{D}_i=\left[\begin{array}{cc}1& 0\\ -d_i/n_i& 1\end{array}\right],& \left(2\right)\end{array}$

wherein, d_i is a distance between the spherical surface S_i and the spherical surface S_i+1.

And a transfer matrix T (system matrix) for the entire system is:

$\begin{array}{cc}T=R_N{D}_{N-1}{R}_{N-1}L{R}_2{D}_1{R}_1=\left[\begin{array}{cc}B& A\\ D& C\end{array}\right],& \left(3\right)\end{array}$

wherein A, B, C, and D are functions of r_i, d_i and n_i. “A” can be expressed as A(r_i, d_i, n_i). Therefore, an image focal length of the coaxial spherical system can be expressed as:

$\begin{array}{cc}{f}^{\prime }=\frac{n_N}{A\left(r_i,d_i,n_i\right)}.& \left(4\right)\end{array}$

Because the image focal length of the coaxial spherical system has been given, equation (4) represents an equation that the radius of curvature r_i (i=1, 2, . . . , N) of the spherical surface S_i and the distance between the spherical surface S_i and the spherical surface S_i+1 (i=1, 2, . . . , N−1). In a reflection system, a condition n_i=−n_i−1 applies; therefore, all the refractive indices in the equation (4) can be eliminated. Equation (4) has an infinite number of solutions and thus it is impossible to discuss all solutions to Equation (4). Given that there are manufacturability limits in practice, some solutions to Equation (4) should be disregarded. Related specific parameters include, for example: the size and sign of the radius of curvature of a mirror surface, a vector height of a mirror surface shape, and a distance between mirrors. In a process of solving, constraints on equations should be set according to an actual situation to narrow a solution space.

In Equation (4), there is a total of 2N−1 parameters for the radii of curvature and mirror distances, the radii of curvature are defined as r₁, r₂, . . . , r_N, and the mirror distances are defined as d₁, d₂, . . . , d_N−1. When the focal length of the coaxial spherical system is given, as long as 2N−2 parameters out of the 2N−1 parameters are given, it is possible to solve for a last remaining parameter. After the 2N−1 parameters are obtained, an additional parameter d_N can be determined using the first order optics, the additional parameter d_N represents a distance between a last spherical surface and the image surface. Therefore, there are a total of 2N parameters that describe the coaxial spherical system. Because there are infinite solutions to equation (4), there are infinite possible combinations of the values of the 2N parameters. The 2N parameters are placed together in a vector P=[r₁, r₂, . . . , r_N−1, r_N, d₁, d₂, . . . , d_N−1, d_N]. The vector P can be used to describe the OP distribution of the coaxial spherical system.

In one embodiment, the series of coaxial spherical systems with different OP distributions P₁, P₂, . . . , P_m, . . . , P_M are used as a research object, and the set of P₁, P₂, . . . , P_m, . . . , P_M is defined as the set {P}. The given 2N−2 parameters are assumed as r₁, r₂, . . . , r_N−1, r_N, d₁, d₂, . . . , d_N−2, wherein a range of the radius of curvature r_i (i=1, 2, . . . , N) of each mirror surface of each coaxial spherical system is [r_min, r_max] with an interval Δr. That is, sequences of the radii of curvature r_i (i=1, 2, . . . , N) are given as r_min+Δr, r_min+2Δr, . . . , r_max. A range of the mirror distance of each coaxial spherical system d_i (i=1, 2, . . . , N−2) is [d_min, d_max] with an interval Δd. That is, sequences of mirror distances d_i (i=1, 2, . . . , N−2) are given as d_min+Δd, d_min+2Δd, . . . , d_max. When the values of r₁, r₂, . . . , r_N−1, r_N and d₁, d₂, . . . , d_N−2 are determined, the corresponding d_N−1 and d_N can be solved. In this way, all the parameters can be obtained, and a series of different OP distribution forms {P} satisfying the image focal length focal length f are obtained.

Since there are infinite kinds of OP distribution forms {P} that satisfy the given image focal length f, a value range of the parameters in the series of different OP distribution forms {P} must be limited. When determining the value range of the radius of curvature r_i (i=1, 2, . . . , N), the values of the radii of curvature and the positive/negative state of the OP of each mirror surface of the coaxial spherical system can be controlled. In one embodiment, the range of the radius of curvature of each mirror surface is [−1000 mm, 1000 mm]. In one embodiment, when determining the range of the mirror distance d_i (i=1, 2, . . . , N−2), three factors are considered; first, the values of the mirror distances should not be too large to avoid large system volumes. Second, the values of the mirror distance should not be too small, in order to off-axis the system and avoid blocking in the subsequent phases. Third, the differences between two arbitrary mirror distances should not be too large to guarantee system compactness. Because an overall optical system size is usually comparable to an entrance pupil size, an entrance pupil diameter (EPD) can be used as an unit length to describe the range of the mirror distance. In one embodiment, EPD≤|d₁|≤4×EPD. If the value of Δr and the value of Δd are too high, there will be few output results; however, if the value of Δr and the value of Δd are too small, the number of output results will be increased, but more computation time will be consumed. Therefore, when determining the value of Δr and the value of Δd, a balance should be made between the number of final solutions and the calculation time. In one embodiment, the OP distribution form is P=[r₁, r₂, r₃, d₁, d₂, d₃], wherein a range of a radius of curvature of each of r₁, r₂, r₃ is [−1000 mm, 1000 mm], and Δr=100 mm. A value of Δr is determined according to a percentage of the value range of the radius of curvature, and the value of Δr can be changed according to an absolute value of the radius of curvature. In one embodiment, the value of Δr is 5% of the radius of curvature r; a range of mirror distance d₁ is EPD<|d₁|<4×EPD, and a value interval of the mirror distance Δd₁=EPD/2. The d₂ and d₃ can be solved based on the radius of curvature, the interval of the radius of curvature, the mirror distances and the interval of the mirror distances. Since the series of coaxial spherical systems are total reflection systems, d₁<0, d₂>0, d₃<0. In one embodiment, the constraints are −2≤d₂/d₁≤−½, ½≤d₃/d₁≤2, d₃≤−2×EPD, a volume of the coaxial spherical system should be controlled, to ensure the structure of the system is compact, and facilitate subsequent off-axis.

In step S2, tilting all optical elements of each coaxial spherical system in the set {P} by a series of angles under keeping an incident direction of light in the center field of the coaxial spherical system unchanged, to obtain a series of off-axis spherical systems with different structures. The series of off-axis spherical systems with different structures comprises structural forms with completely different structures, such as offset field or off-axis field, or a combination of the offset field and the off-axis field. Referring to FIG. 4, in one embodiment, the coaxial spherical system is a three-mirror system, the OP distribution form of the three-mirror system is P=[r₁, r₂, r₃, d₁, d₂, d₃], a position of each mirror of the three-mirror system is adjusted to ensure that distances between origins of three local coordinate systems of three mirrors meet V₁V₂=d₁, V₂V₃=d₂, and |V₂V₃=d₃, at the same time, each mirror is tilted by an angle around an unit axis vector V_iX of a respective local coordinate system, to obtain an off-axis spherical system corresponding to the OP distribution form P.

Since a propagation path of main lights of the central field between the surfaces in the off-axis spherical system is a polyline, the propagation path can be used to represent a structural form of the off-axis spherical system. A vector C is used to record a length of each segment of the polyline and an angle between adjacent segments. Referring to FIG. 4, an angle that a vector V_i−1V_i rotates to coincide with a vector V_iV_i+1 is defined as θ_i(i=1, 2, . . . , N), d_i is a distance between two mirrors, and a corresponding structural form can be written as a vector C=[θ₁, |d₁|, θ₂, |d₂|, . . . θ₁, |d_i|, . . . , θ_N, |d_N|]. For off-axis spherical systems with the same OP distribution form, only an angle in the vector C is retained, that is, C=[θ₁, θ₂, . . . θ₁, . . . , θ_N]. θ_i reflects a deflection angle of the chief light ray of the central field at a mirror surface S_i. When the vector V_i−1V_i rotates clockwise to coincide with the vector V_iV_i+1, the sign of θ_i is negative, when the vector V_i−1V_i rotates counterclockwise to coincide with the vector V_iV_i+1, the sign of θ_i is positive. Therefore, θ₁ is more than −360° and less than 0°, or θ_i is more than 0° and less than 360°. In one embodiment, three deflection angles of the chief light ray of the central field at three mirrors of an off-axis three-mirror spherical system are defined as θ₁<0, θ₂>0, θ₃<0. Referring to FIG. 4, in one embodiment, a structural form of a reflective coaxial system with the OP distribution form P is denoted as C₀, for example, a structural form of a coaxial three-mirror spherical system can be written as C₀=[θ₁₀, θ₂₀, θ₃₀]=[−180, 180, −180].

When each angle θ₁ in the structural form C changes continuously within a certain range, there is no obstruction in an off-axis spherical system corresponding to the structural form C. However, since the off-axis spherical system may have many different off-axis structural forms, there are also many possible value ranges for the angle in structural form C that meets the unobstructed condition, that is, a range of possible values of θ_i in the structural form C is not continuous. In order to solve all possible unobstructed structural forms corresponding to the OP distribution form P, first listing as many off-axis structural forms as possible, and then eliminating the obstructed systems from the off-axis structural forms in the subsequent steps, so as to obtain as many unobstructed off-axis spherical systems as possible.

A range of the angle θ_i and an interval of the angle θ_i are determined first. If the structural form C and a structural form C′ satisfy C=−C′, the structural form C and structural form C′ are actually the same. Therefore, in order to avoid re-considering the same structural form, a value range of θ₁ is within (−180°, 0°) or (0°, 180°). For other angles θ_i, i=2, 3, . . . , N, a value range of θ_i is within (−360°, 0°) or within (0°, 360°). The value range of θ_i can be selected according to the requirements of the design structure constraints. In one embodiment, for an off-axis three-mirror spherical system, the value range of θ₁ is (−180°, −120°), the value range of θ₂ is [120°, 240° ], and the value range of θ₃ is (−240°, −120°). If the interval Δθ of the angle θ₁ is too large, the number of output results will be reduced, but a calculation amount of an entire calculation task will also be reduced. For a coaxial spherical system with an OP distribution form P_m, a series of off-axis spherical systems with different off-axis structures can be obtained through the range of the angle θ₁ and the interval of the angle θ₁. The series of off-axis spherical systems with different off-axis structural forms are represented by vectors as C_m,1, C_m,2, . . . , C_m,s, . . . , C_m,Sm, and the vectors C_m,1, C_m,2, . . . , C_m,s, . . . , C_m,Sm are denoted as the set {C}_m. There are S_m elements in the set {C}_m, whether the off-axis spherical system in the set {C}_m is blocked should be determined.

Second, the unobstructed off-axis spherical system is selected from the set {C}_m. When judging whether there is obstruction in the off-axis spherical system, the off-axis spherical system is symmetric about the O-YZ plane is noticed. A projection of a light beam between the mirror surface S_i and the mirror surface S_i+1 in the O-YZ plane is defined as B_i, wherein i=0, 1, . . . , N. An area illuminated by a light beam on the mirror surface is defined as a working area of the mirror surface. A curve segment of a working area of the mirror surface S_j intercepted by the O-YZ plane is defined as E₁^(j)E₂^(j), wherein j=0, 1, . . . , N, I, and S_I means image surface. A condition of no obstruction in the off-axis spherical system is: for any B_i(i=0, 1, . . . , N), any curve segment E₁^(j)E₂^(j) (j=1, 2, . . . , N, I) has no overlap with B_i. After filtering according to the condition, a series of unobstructed off-axis spherical systems with the OP distribution form P_m can be obtained, the series of unobstructed off-axis spherical systems are defined as C_m,1, C_m,2, . . . C_m,r, . . . , C_m,Rm. The set of C_m,1, C_m,2, . . . , C_m,r, . . . , C_m,Rm is defined as the set {C}_m, and there are R_m elements in the set {C}_m.

After obtaining the series of unobstructed off-axis spherical systems with different structural forms corresponding to each OP distribution form, the step of constructing the freeform imaging system can be proceed directly. In one embodiment, after the series of unobstructed off-axis spherical systems are obtained, a plurality of unobstructed off-axis spherical systems with a small folding angle of optical path at each mirror are further found from the series of unobstructed off-axis spherical system; and then the step of constructing the freeform imaging systems can be proceed, to obtain a plurality of compact freeform imaging systems. In one embodiment, the system that meets the given structural requirements are found from the series of unobstructed off-axis spherical systems, by changing the value range of the radius of curvature, the interval of the radius of curvature and setting an appropriate filter conditions. In one embodiment, the unobstructed off-axis spherical system is a three-mirror system, for an unobstructed off-axis structural form C_m,r with the OP distribution form P_m, a new vector ΔC_m,r is defined, and ΔC_m,r=C_m,r−C₀=[Δθ₁, Δθ₂, . . . Δθ_i, . . . , Δθ_N], wherein C₀ is a coaxial structure, and Δθ_i is a folding angle of the optical path at the mirror surface S_i. The set of vectors ΔC_m,r corresponding to the set {C}_m is defined as the set {ΔC}_m. A value of Δθ_i in the vector ΔC_m,r reflects an off-axis degree of a system optical path. The smaller an absolute value of Δθ_i in ΔC_m,r, the smaller an off-axis degree of the system optical path. The symbols of each item of ΔC_m,r reflect the folding form of the system optical path. According to the above regulations on the magnitude and sign of θ₁, the sign of the first term in ΔC_m,r is positive, and the signs of the second item and third term can be positive or negative. Each combination of positive and negative signs corresponds to a folding form of the optical path, there are 1×2×2=4 types of folding forms of all possible optical paths, each folding form includes a variety of structural forms with different optical path off-axis degrees.

For all folding forms, one or several off-axis spherical systems with a small folding angle of optical path can be found through a minimization method. The minimization method is as follows: since the value intervals of the radius of curvature are equal, it does not exist two vectors in the set {ΔC}_m that three items of the two vectors are completely equal, but there are two vectors in the set {ΔC}_m that two terms of the two vectors are equal but the third term of the two vectors are not equal. The vectors in the set {ΔC}_m are classified to a plurality of categories, in each category, the first item Δθ₁ and the second item Δθ₂ are equal, and the third item Δθ₃ is different. A process of minimizing the item Δθ₃ comprises: in each category, a vector with a minimum absolute value of Δθ₃ can be found, and a corresponding off-axis spherical system has the smallest optical path angle at a tertiary mirror. The minimum absolute value of Δθ₃ in different categories of the plurality of categories may be different. Then, for the vector that has been minimized item Δθ₃, the same process of minimizing item Δθ₃ is used to minimize the term Δθ₂ in the vector. And then, for the vector that has been minimized item Δθ₂ and item Δθ₃, the same process of minimizing item Δθ₃ is used to minimize the term Δθ₁ in the vector. Finally, a set of vectors whose term Δθ₁, term Δθ₂ and term Δθ₃ are all minimized are obtained. The structural vectors corresponding to the set of vectors are found from the set {C}_m, and the structural vectors are defined as {tilde over (C)}_m,1, {tilde over (C)}_m,2, . . . , {tilde over (C)}_m,t, . . . , {tilde over (C)}_m,Tm. The set of {tilde over (C)}_m,1, {tilde over (C)}_m,2, . . . , {tilde over (C)}_m,t, . . . , {tilde over (C)}_m,Tm is defined as the set {{tilde over (C)}}_m, and there are T_m elements in the set {{tilde over (C)}}_m. The vectors in the set {{tilde over (C)}}_m represents a series of compact unobstructed off-axis spherical systems with different optical path folding forms and small optical path folding angles.

In one embodiment, further solving a series of off-axis spherical systems with a smaller optical path folding angle at each mirror based on the set {{tilde over (C)}}_m by a dichotomy. Referring to FIG. 5, in one embodiment, the unobstructed off-axis spherical system is an off-axis three-mirror system, {tilde over (C)}_m,t=[θ₁, θ₂, θ₃] is an element in the set {{tilde over (C)}}_m, an optical system corresponding to {tilde over (C)}_m,t is unobstructed; and C₀=[θ₁₀, θ₂₀, θ₃₀]=[−180, 180, −180] corresponds to a coaxial system with obstructions. An upper limit of unobstruction is θ_i,uob=θ_i, a lower limit of obstruction is θ_i,obs=θ_i0, an intermediate value is defined as θ_i,mid=(θ_i,uob+θ_i,obs)/2. If θ_i=θ_i,mid, the system is unobstructed, then let θ_i,uob=θ_i,mid; if θ_i=θ_i,mid, the system is obstructed, then let θ_i,obsθ_i,mid; and the above steps are repeated until |θ_i,uob−θ_i,obs|<θ_max, wherein θ_max is a given threshold. According to the above method, the upper limit θ_3,uob, θ_2,uob and θ_1,uob at each mirror are solved in turn, to obtain more compact unobstructed off-axis structure form vectors {tilde over (C)}′_m,t=[θ_1,uob, θ_2,uob, θ_3,uob]. The vectors in the set {{tilde over (C)}}_m can be replaced by the more compact unobstructed off-axis structure form vectors {tilde over (C)}′_m,t with smaller bending angles of the optical path. A compactness of the optical path folding of the unobstructed off-axis spherical system can be controlled by adjusting the threshold θ_max.

In step S3, since the OP of the entire system has changed after the off-axis process of eliminating the obstruction in step S2, the OP of each mirror should be fine-tuned to correct of the OP of the entire system. An OP correction process of the entire system should follow the following principles. First, the OP correction process can be fully automated; second, the structure of the system remains unchanged during the OP correction process; third, the OP of each mirror is fine-tuned.

In one embodiment, using a point-by-point construction method for freeform surface imaging system to correct the OP of the system. The point-by-point construction method calculates the freeform surface shape based on feature light rays and feature data points. The unobstructed off-axis spherical system is used as a starting point, a plurality of feature light rays corresponding to different aperture positions and different fields is fixed, and a plurality of intersections of the plurality of feature light rays and an optical surface is defined as a plurality of feature data points. Since a field of the plurality of feature light rays is known, according to an object-image relationship, the position coordinates of an image point corresponding to the plurality of feature light rays on the image surface can be determined when the optical system is ideal for imaging (referred to as the position coordinates of an ideal image point). Each of the plurality of feature data points contains two aspects of information, one is a coordinate of the feature data point on the optical surface, and the other is a normal direction of the feature data point on the optical surface. When constructing the freeform surface imaging system, a surface shape of each surface in the freeform surface imaging system is solved one by one in a certain order. In one embodiment, a three-mirror imaging system is solved in an order of tertiary mirror-primary mirror-secondary mirror. When solving a shape of a freeform surface, first, obtaining the starting points and the incident directions of the plurality of feature light rays incident on a surface to be solved through ray tracing. Starting from an initial feature data point on the surface to be sought, according to a principle of the nearest light, and the coordinates and normal directions of the feature data points that have been calculated, determining a next feature light ray and the coordinates of the feature data point corresponding to the next feature light ray. According to the object-image relationship of the system, the plurality of feature light rays starts from the plurality of feature data points on the surface to be solved, and finally reaches the corresponding ideal image points on the image surface. Therefore, according to Fermat's principle, a propagation direction of the next feature light ray leaving the freeform surface to be solved can be obtained. Combining with the incident direction of the next feature light ray on the surface, a normal direction of the feature data point corresponding to the next feature light ray can be obtained. Repeating the above steps until the coordinates and normals of all the plurality of feature data points are solved. Fitting the plurality of feature data points into the freeform surface by a fitting method that simultaneously considers coordinate and normal, to complete solving the shape of a freeform surface. According to the above method, the shape of each surface in the freeform surface imaging system is calculated one by one, therefore, the freeform surface imaging system is constructed.

Each mirror of the unobstructed off-axis spherical system has a certain optical power, and the unobstructed off-axis spherical system meets the first-order imaging conditions. The point-by-point construction method can retain the structure of the system and fine-tune the optical power of each mirror in the system, to complete the correction of the optical power of the entire system. Therefore, the series of freeform imaging systems {tilde over (F)}_m,1, {tilde over (F)}_m,2, . . . , {tilde over (F)}_m,t, . . . , {tilde over (F)}_m,Tm are obtained, and the set of {tilde over (F)}_m,1, {tilde over (F)}_m,2, . . . , {tilde over (F)}_m,t, . . . , {tilde over (F)}_m,Tm is defined as {{tilde over (F)}}_m.

In step S4, on a premise that the OP distribution form and the off-axis structural form of the series of freeform surface systems in the set {{tilde over (F)}}_m remain unchanged, the image quality of each of the series of freeform surface systems in the set {{tilde over (F)}}_m is improved. In one embodiment, a point-by-point iteration is used to calculate the surface shape of the freeform surface imaging system and improve the image quality of the system.

The point-by-point iteration and the point-by-point construction method both solve the surface shape of each surface in the freeform surface imaging system according to the object-image relationship of the system and based on the feature light rays. However, in a process of the point-by-point iteration, the coordinates of the plurality of feature data points obtained by tracing are kept unchanged, only the normal directions of the plurality of feature data points are re-solved, and then a new freeform surface is obtained by fitting the original coordinates and the newly solved normal directions. In a round of iterative calculation, the surface shapes of all freeform surfaces are solved in a certain order, for example, in an order of tertiary mirror-primary mirror-secondary mirror in a three-mirror system. After a round of iterative calculation is completed, the image quality of the freeform surface imaging system needs to be evaluated. In one embodiment, an evaluation index is a root mean square (RMS) of the deviation between a plurality of actual intersections of the plurality of feature light rays with a target surface and the ideal target points. A RMS value is denoted as σ, and σ can be used as an evaluation parameter of the image quality during the process of the point-by-point iteration. A value of σ decreases and gradually converges as the process of the point-by-point iteration progresses. The process of the point-by-point iteration is stopped when the image quality of the system is better than the given threshold σ_itr. When the value of σ does not decrease, that is, the value of σ gradually converges to a certain value, the image quality of the freeform surface imaging system under current condition is difficult to continue to be improved. An improvement rate of image quality in the process of the point-by-point iteration is defined as τ=|σ′−σ|/σ, wherein σ′ is the image quality evaluation parameter of the iteration result of the current round of iterative calculation, and σ is the image quality evaluation parameter of the iteration result of the last round of iterative calculation. The process of the point-by-point iteration is stopped when τ is below a set threshold τ_itr.

In the process of the point-by-point iteration, the normal directions of the plurality of feature data points are solved according to the intersections of the plurality of feature light rays and the image surface are as the ideal image point, which can make the freeform surface imaging system without distortion system. In optical design, the image quality can be improved by sacrificing distortion. In the subsequent stages of the process of the point-by-point iteration, the distortion in the freeform surface imaging system has been controlled to a fairly low level. If the image quality improvement rate T is lower than a given threshold τ_igf, an “image quality priority” mode is activated. Under the “image quality priority” mode, in each round of iterative calculation, every time a surface is calculated, the position coordinates of the target image points corresponding to the plurality of feature light rays of each field are updated. The coordinates of new target image points are obtained by obtaining the actual intersection coordinates of all the plurality of feature light rays of the field on the image surface through ray tracing, and then taking an average of each dimension of the actual intersection coordinates to obtain the coordinates of new target image points. By continuously updating the target image point during each round of iterative calculation, more design results with image quality that meet the requirements can be obtained.

Examples of the method for point-by-point constructing the freeform surface imaging system and the method for fitting the plurality of feature data points into a freeform surface are taught by U.S. Pat. No. 10,101,204 to Zhu et al.

The optical power of the entire system is corrected by calculating the shape of the mirror surface, and the image quality of the freeform surface imaging system is improved by iterating the freeform surface shape of the mirror surface. Therefore, a position and an tilt angle of each mirror, a position of the image surface, and the surface shape of the optical surfaces have been considered. There is another factor that needs to be considered, that is, the tilt angle of the image surface.

The tilt angle of the image surface can be expressed by an angle β between the Y axis of the local coordinate system of the image surface and the Y axis of the global coordinate system. When the image surface of the optical system is perpendicular to the chief ray of the central field, the tilt angle of the image surface is defined as β₀. In one embodiment, a optimal tilt angle of the image surface is found using a one-dimensional search method. One round of one-dimensional search for the optimal tilt angle of the image surface is processed as follow, a same off-axis spherical surface system is taken as a starting point, and only the tilt angle of the image surface is different in the process of constructing the freeform surface imaging system through the step S3 and step S4. When the tilt angle of the image surface is different, the image quality of the freeform surface imaging system is also different, and a tilt angle of the image surface corresponding to a freeform surface imaging system with the best image quality is the optimal tilt angle of the image surface. In one embodiment, multiple rounds of one-dimensional search for the optimal tilt angle of the image surface are performed, to further improve the image quality.

In one embodiment, one round of the one-dimensional search for the best tilt angle of the image plane is processed as follows: starting from β=β₀ and using Δβ as an interval, and taking a value within a range [β₀−β_r, β₀+β_r] as β, wherein βr is a radius size of the range [β₀−β_r, β₀+β_r]. The values of β are defined as β₁, β₂, . . . , β_b, . . . , β_B, and the number of the tilt angle of the image plane is B. For a compact off-axis structure C_m under the OP distribution form P_m, taking the off-axis spherical system corresponding to the compact off-axis structure C_m as the starting point for calculation, and calculating the freeform surface imaging systems when the tilt angle of the image plane is β₁, β₂, . . . , β_b, . . . , β_B through step S3 and step S4 respectively; an image quality evaluation parameter of the freeform surface imaging systems is defined as θ₁, σ₂, . . . , σ_b, . . . , σ_B; finding the best image quality evaluation parameter from σ₁, σ₂, . . . , σ_b, . . . , σ_B and defining as ρ, and a tilt angle β_opt of the image plane corresponding to p is the best tilt angle of the image plane in the current round of one-dimensional search. The only difference between a next round of one-dimensional search and the current round of one-dimensional search is that the value range of β has changed, and a value taken within a range [β_opt−β_r, β_opt+β_r] is as β. During performing multiple rounds of one-dimensional search, the image quality ρ of the freeform surface imaging system is continuously improved and gradually converged. If the image quality ρ of the freeform surface imaging system reaches a given threshold ρ_srh, the one-dimensional search can be stopped. An image quality improvement rate ν is calculated by ν=(ρ′−ρ)/ρ, wherein ρ′ is the best image quality evaluation parameter of the freeform surface imaging system in a previous round of one-dimensional search, and ρ is the best image quality evaluation parameter of freeform surface imaging system in the current round of one-dimensional search. If the image quality improvement rate ν reaches a given threshold ν_srh, the one-dimensional search can be stopped.

In a process of the one-dimensional search, since the tilt angle of the image surface has changed, the obstruction may appear in the freeform surface imaging system, therefore, it is necessary to judge the obstruction in the freeform surface imaging system at any time and eliminate the freeform surface imaging systems with obstruction.

A method of judging whether there is obstruction in the freeform surface imaging system is the same as the method in step S2. Once there is an obstruction in a freeform surface imaging system, the image quality evaluation parameter σ of the corresponding system is set to NaN, and then the next image plane tilt angle state is continue to be calculated. In one embodiment, a field expansion coefficient K_F is used to expand the field of the freeform surface imaging system in the process of judging the obstruction, and an aperture expansion factor K_D is used to expand the aperture of the system to ensure that there is enough space between the components in the freeform surface imaging system.

In one round of one-dimensional search, if except that the tilt angle of the image plane is β₀, the image quality evaluation parameters of the freeform surface imaging systems corresponding to other tilt angles of image planes are NaN; there are obscurations in all freeform surface imaging systems that have been calculated, and the image quality improvement rate ν of the current round of one-dimensional search is zero.

When the image quality improvement rate ν is lower than a given threshold ν_srh, an additional round of one-dimensional search can be added. If the image quality improvement rate of the additional round of one-dimensional search is still lower than the given threshold ν_srh, the one-dimensional search process is terminated, and a freeform surface imaging system with the best image quality is as the final result.

Too many rounds of the one-dimensional search require too long calculation time. In one embodiment, a maximum number of rounds of one-dimensional search Rnd_max is set. When the number of rounds of the one-dimensional search Rnd is greater than Rnd_max, the one-dimensional search is terminated.

In step S5, the evaluation metric can be selected according to actual needs. In one embodiment, the evaluation metric is imaging quality metric. The imaging quality metric comprises a discrete spot size of image points on the image plane, an MTF of the system, a wave aberration of the full field, etc. In one embodiment, an average value of wave aberration RMS in the whole field of the freeform surface imaging system is required to be no less than 0.075λ. According to the imaging quality metric, the freeform surface imaging system meeting the design requirement is selected and outputted.

An entire framework flowchart of the method for designing freeform surface imaging system provided by the present invention is shown in FIG. 6.

The method for designing freeform surface imaging system can be implemented via computer software. In one embodiment, a calculation program is developed using MATLAB. First a debugging of the calculation program is completed on a computer, and then a calculation work is completed on a high-performance computing platform “Exploration 100”.

In one embodiment, the method for designing freeform surface imaging system further comprises a step of manufacturing the freeform surface imaging system according to parameters of the freeform surface imaging system outputted in step S5, to obtain a physical component of the freeform surface imaging system.

The following are two examples of designing a freeform three-mirror imaging system using the method for designing freeform surface imaging system of the present invention.

Example 1

The optical performance indexes of the freeform surface three-mirror imaging system to be designed and the parameters used in the design method are shown in Table 1 and Table 2.

TABLE 1
input parameters
System	field	±4 in X, ±3 in Y
parameters	Full field angles (°)	8° × 6°
	Entrance pupil diameter (EPD)	27.78
	(mm)
	Focal length (mm)	50
	Working band	LWIR
	Primary wavelength	8-14 μm
Step S1	Range and interval of the	r1, 2, 3 ∈ [−1000, 1000],
	curvature radius for spherical	Δr1, 2, 3 = 100
	surfaces (mm)
	Structure constraints	EPD < |d1| < 4 × EPD
		Δd1 = EPD/2
		d1 < 0, d2 > 0, d3 < 0
		−2 ≤ d2/d1 ≤ −1/2,
		1/2 ≤ d3/d1 ≤ 2,
		d3 ≤ −2 × EPD
Step S2	Range and interval of the angle in	θ1 ∈ (−180°, −120°],
	the vector when solving the	θ2 ∈ [120°, 240°],
	off-axis spherical system	θ3 ∈ [−240°, −120°],
		Δθ1, 2, 3 = 5°
	field expansion coefficient KF and	KF = 1.5, KD = 1.2
	aperture expansion factor KD
	the maximum allowable angle	θmax = 5°
	θmax of dichotomy method for
	solving compact structure

TABLE 2
Built-in methods and parameters
Step S3	Optical power correction process	point-by-point construction
		method
	Field angles sampled over the	(0, 0) (0, 1.5) (0, −1.5)
	half field angle	(2, 0) (2, 1.5) (2, −1.5)
	Number of feature light rays	97
	sampled over the full aperture
	for one field angle
step S4	Method to improve the imaging	Point-by-point iteration
	quality	method
	Field angles sampled over the	(0, 0) (0, 1.5) (0, −1.5)
	half field angle	(2, 0) (2, 1.5) (2, −1.5)
	Number of feature light rays	97
	sampled over the full aperture
	for one field angle
	Iteration termination condition:	σitr = 0.003
	imaging quality threshold
	Iteration termination condition:	τitr = 0.5%
	imaging quality improvement
	rate threshold
	Image quality priority mode	τiqf = 3%
	activation conditions: image
	quality improvement rate
	threshold
Step S4	Range and interval for the	β ∈ [β0 − 5, β0 + 5], Δβ = 1
	one-dimensional search of image
	plane tilt angle (°)
	One-dimensional searching	ρsrh = 0.003
	termination condition: imaging
	quality threshold
	One-dimensional searching	υsrh = 0.5%
	termination condition: imaging
	quality improvement rate
	threshold
	One-dimensional searching	Rndmax = 3
	termination condition:maximum
	round number Rndmax

After the system performance indexes and constraints are inputted into a computer, a computing task is completed in approximately 40.4 h without human interaction. Then CODE V is used as a tool to evaluate the image quality of the freeform surface three-mirror imaging system, 11×11 field points are sampled over the 8°×6° field, and the average of the root-mean-square (RMS) values of the wavefront error value at each field point when the wavelength λ=10 m is calculated. If the average value of the RMS value of wavefront error at all field points is no greater than 0.075λ, wherein λ is a primary wavelength, the image quality of the freeform surface three-mirror imaging system is considered to be diffraction-limited or near-diffraction-limited. A total of 127 freeform surface three-mirror imaging systems that satisfy the design requirements are obtained, and an average time to obtain one freeform surface three-mirror imaging system is 19.1 min.

FIG. 7 shows a curve of an operating efficiency of the calculation task of Example 1 over time. A horizontal axis represents a running time (units: h), and a vertical axis represents a percentage of computing resources currently used in the largest computing resources. That is, the vertical axis represents an operating efficiency of the computing task. In one embodiment, each computing task consists of multiple independent computing jobs, and the multiple independent computing jobs are submitted to a distributed system one by one to wait for calculation. Each of the multiple independent computing jobs corresponds to a freeform surface three-mirror imaging system corresponding the structural form of {tilde over (C)}_m,t with the OP distribution of P_m. Since computing tasks of other users are still running on the computing platform, so the operating efficiency of the program cannot reach the maximum after submission. The highest operating efficiency of the calculation task in FIG. 7 is 81.5%. Since the previously submitted tasks have been completed one after another, and all calculation tasks have been submitted; about 7.08 hours after the calculation task started, the operating efficiency began to drop. About 24.5 hours after the calculation task started, the operating efficiency dropped below 2.0%, and more than 99.0% of the computing tasks have been completed. A total workload of the calculation task can be obtained by integrating the operating efficiency with time; therefore, a theoretical calculation time is estimated about 6.17 hours when the operating efficiency is 100%. The current computing time is shared between the computing platform and other users. If computing resources can be monopolized, the computing time can be shorter than the current time.

FIG. 8 shows some of the design results on the same scale of example 1. The freeform surface three-mirror imaging systems of example 1 have different structural forms, different optical power distributions and different volumes. The image quality of the freeform surface three-mirror imaging systems can approach or reach the diffraction limit. The freeform surface three-mirror imaging systems L3-1, L61-1 and L75-1 have similar structures and volumes but different OP distributions. The freeform surface three-mirror imaging systems L14-1, L35-1 and L87-2 have similar structures and volumes but different OP distributions. It can be seen that the volume of each of the freeform surface three-mirror imaging systems in the second row is smaller than the volume of each of the freeform surface three-mirror imaging systems in the first row, and the optical path of each of the freeform surface three-mirror imaging systems in the second row is more compact than the optical path of each of the freeform surface three-mirror imaging in the first row. The structures of the freeform surface three-mirror imaging systems L17-5, L73-2 and L18-1 are not commonly seen and are the most compact among the output results. The freeform surface three-mirror imaging systems L137-3 and L142-2 have spherical contour structures. The primary mirror and the secondary mirror are convex in the freeform surface three-mirror imaging systems L331-3, L306-7 and L303-6; L254-2, L306-2 and L307-2 in the final row, which leads to large system volumes. The structure of the freeform surface three-mirror imaging L162-4 has rarely been seen before.

Example 2

The optical performance indexes of the freeform three-mirror imaging system to be designed and the parameters used in the design method are shown in Table 3 and Table 4.

TABLE 3
input parameters
System	field	±2X, ±2Y
parameters	Full field angles (°)	4° × 4°
	Entrance pupil diameter (EPD) (mm)	50
	Focal length (mm)	450
	Working band	VIS
	Primary wavelength	420-680 nm
Step S1	Range and interval of the curvature	r1, 2, 3 ∈ [−1000, 1000],
	radius for spherical surfaces (mm)	Δr1, 2, 3 = 100
	Structure constraints	EPD/2 < |d1| < 5 × EPD
		Δd1 = EPD/2
		d1 < 0, d2 > 0, d3 < 0
		−2 ≤ d2/d1 ≤ −1/2,
		1/2 ≤ d3/d1 ≤ 2,
		d3 ≤ −2 × EPD
Step S2	Range and interval of the angle in the	θ1 ∈ (−180°, −120°],
	vector when solving the off-axis	θ2 ∈ [120°, 240°],
	spherical system	θ3 ∈ [−240°, −120°],
		Δθ1, 2, 3 = 20o
	field expansion coefficient KF and	KF = 1.5, KD = 1.2
	aperture expansion factor KD
	Volume limit	no
	the maximum allowable angle θmax of	θmax = 5°
	dichotomy method for solving compact
	structure

TABLE 4
Built-in methods and parameters
Step S3	Optical power correction process	point-by-point construction
		method
	Field angles sampled over the half	(0, 0) (0, 2) (0, −2)
	field angle	(2, 0) (2, 2) (2, −2)
	Number of feature light rays sampled	97
	over the full aperture for one field
	angle
Step S4	Method to improve the imaging	Point-by-point iteration
	quality	method
	Field angles sampled over the half	(0, 0) (0, 2) (0, −2)
	field angle	(2, 0) (2, 2) (2, −2)
	Number of feature light rays sampled	97
	over the full aperture for one field
	angle
	Iteration termination condition:	σitr = 0.003
	imaging quality threshold
	Iteration termination condition:	τitr = 0.5%
	imaging quality improvement rate
	threshold
	Image quality priority mode activation	τiqf = 3%
	conditions: image quality
	improvement rate threshold
Step S4	Range and interval for the	β ∈ [β0 − 5, β0 + 5], Δβ = 1
	one-dimensional search of image plane
	tilt angle (°)
	One-dimensional searching	ρsrh = 0.003
	termination condition: imaging quality
	threshold
	One-dimensional searching	υsrh = 0.5%
	termination condition: imaging quality
	improvement rate threshold
	One-dimensional searching	Rndmax = 3
	termination condition: maximum
	round number Rndmax

In one embodiment, the working band is visible light. In a visible light system, a requirement of a accuracy in calculating the shape of the curved surface is high. Moreover, the relative aperture of the freeform surface three-mirror imaging system is relatively small, and the structure of the freeform surface three-mirror imaging system is more variety than the freeform surface three-mirror imaging system with a large relative aperture. A computing task is completed in approximately 34.6 h without human interaction, a total of 59 freeform surface imaging systems that satisfy the design requirements are obtained, and an average time to obtain one freeform surface imaging system is 35.2 min.

FIG. 9 shows a curve of an operating efficiency of the calculation task of Example 2 over time. Compared with the calculation task of embodiment 1, an average operating efficiency of embodiment 2 is higher. That is, although the calculation amount of Example 2 is more than that of Example 1, the calculating time of Example 2 is shorter than the calculating time of Example 1. The highest efficiency in the entire calculation process of Example 2 can reach 99.0%. About 16.4 hours after the calculation task started, the operating efficiency began to drop. About 22.6 hours after the calculation task started, the operating efficiency dropped below 2.0%, and more than 99.7% of the computing tasks have been completed. A total workload of the calculation task can be obtained by integrating the operating efficiency with time; therefore, a theoretical calculation time is estimated about 11.0 hours when the operating efficiency is 100%.

FIG. 10 shows some of the design results on the same scale of example 2. The freeform surface three-mirror imaging systems V9-4, V39-3 and V11-3 in the first row represent three most common structural forms. The three freeform surface three-mirror imaging systems V10-1, V142-2 and V258-1 in the second row have the same structural form, but different sizes. The freeform surface three-mirror imaging V773-4 and V911-3 in the third row have the same structural form, but different sizes. Compared with the freeform surface three-mirror imaging systems V773-4 and V911-3, the freeform surface three-mirror imaging system V912-3 has a more compact structure and a smaller volume. A structure of each of the freeform surface three-mirror imaging systems V26-6, V27-5 and V452-3 in the fourth row has rarely been seen before. The focal power of the primary mirror and the tertiary mirror of each of the freeform surface three-mirror imaging systems V989-4 and V789-3 in the fifth row are negative. A distance between the tertiary mirror and the image plane in the freeform surface three-mirror imaging system V989-4 is shorter than a distance between the tertiary mirror and other mirrors. The distance between the tertiary mirror and the image plane in the freeform surface three-mirror imaging system V789-3 is longer than the distance between the tertiary mirror and the image plane in the freeform surface three-mirror imaging system V989-4. The distance between the secondary mirror and the image plane of the freeform surface three-mirror imaging system V144-5 in the last row of is relatively tight, and the tertiary mirror is located far away from both the secondary mirror and the image plane. The freeform surface three-mirror imaging systems V347-1 and V240-1 have smaller volumes and compact structures. The freeform surface three-mirror imaging system V240-1 is the best design result among all the output results.

The method for designing freeform surface imaging system provided by the present invention has advantages as follows: first, as long as the optical parameters and structural constraints of an optical system are given, a series of freeform surface imaging systems with different structural forms and OP distribution forms can be obtained through automatic calculation without manual intervention. The image quality of the freeform surface imaging systems obtained by the method is close to or reaches the diffraction limit. The freeform surface imaging systems meet the given optical parameters and structural constraints, and the image quality parameters such as wave aberration and modulation transfer function MTF meet the design requirements. Second, in the final stage of optical design, after the image quality of the freeform surface three-mirror system meets the requirements, the designer only needs to select the best one from the many results as the final design result by comprehensively considering factors such as processing difficulty, system structure and shape, therefore, human resources and experience requirements for optical designers are greatly reduced. Third, the method provided by the present invention provides a new idea for liberating and expanding optical design productivity in the future, by realizing high-quality, multi-result, and automated freeform surface imaging system design.

It is to be understood that the above-described embodiments are intended to illustrate rather than limit the present disclosure. Variations may be made to the embodiments without departing from the spirit of the present disclosure as claimed. Elements associated with any of the above embodiments are envisioned to be associated with any other embodiments. The above-described embodiments illustrate the scope of the present disclosure but do not restrict the scope of the present disclosure.

Depending on the embodiment, certain of the steps of a method described may be removed, others may be added, and the sequence of steps may be altered. The description and the claims drawn to a method may include some indication in reference to certain steps. However, the indication used is only to be viewed for identification purposes and not as a suggestion as to an order for the steps.

Claims

1. A method for designing freeform surface imaging system, the method comprises: S1: constructing a series of coaxial spherical systems with different optical power (OP) distributions, wherein the series of coaxial spherical systems with different OP distributions are defined as P1, P2, . . . , Pm, . . . , PM, and a set of P1, P2, . . . , Pm, . . . , PM is defined as a set {P}; and the OP distribution refers to a combination of radii of curvature of surfaces of the coaxial spherical system and distances between surfaces of the coaxial spherical system;S2: tilting all optical elements of each coaxial spherical system in the set {P} by a series of angles to obtain a series of off-axis spherical systems, wherein the series of off-axis spherical systems are defined as Cm,1, Cm,2, . . . , Cm,s, . . . , Cm,Sm, and a set of the Cm,1, Cm,2, . . . , Cm,s is defined as a set {C}m; finding all unobscured off-axis spherical systems from the set {C}m, wherein the unobscured off-axis spherical systems are defined as Cm,1, Cm,2, . . . , Cm,r, . . . , Cm,Rm, and a set of Cm,1, Cm,2, . . . , Cm,r, . . . , Cm,Rm is defined as a set {C}m; and then specifying a system size or structural constraints, and finding a series of compact unobstructed off-axis spherical systems from the set {C}m, wherein the series of compact unobstructed off-axis spherical systems are defined as {tilde over (C)}m,1, {tilde over (C)}m,2, . . . {tilde over (C)}m,t, . . . , {tilde over (C)}m,Tm, and a set of {tilde over (C)}m,1, {tilde over (C)}m,2, . . . , {tilde over (C)}m,t, . . . , {tilde over (C)}m,Tm is defined as a set {{tilde over (C)}}m;S3: constructing a series of freeform surface imaging systems based on the series of compact unobstructed off-axis spherical system in the set {{tilde over (C)}}m, and correcting an OP of each of the series of freeform surface imaging systems in a process of constructing the series of freeform surface imaging systems, wherein the series of freeform imaging systems are defined as {tilde over (F)}m,1, {tilde over (F)}m,2, . . . , {tilde over (F)}m,t, . . . , {tilde over (F)}m,Tm, and a set of {tilde over (F)}m,1, {tilde over (F)}m,2, . . . , {tilde over (F)}m,t, . . . , {tilde over (F)}m,Tm is defined as a set {{tilde over (F)}}m;S4: improving an image quality of each of the series of freeform surface imaging systems in the set {{tilde over (F)}}m by calculating a surface shape of each optical element of each freeform surface imaging system in the set {{tilde over (F)}}m and finding an optimal tilt angle of an image surface, to obtained a series of freeform surface imaging systems with different structural forms and different OP distributions; andS5: automatically evaluating an image quality of each of the series of freeform surface imaging system with different structural forms and different OP distributions based on an evaluation metric, and outputting at least one freeform surface imaging system that meets a given requirements.

2. The method of claim 1, wherein in step S1, the series of coaxial spherical systems with different OP distributions are constructed according to first-order optics and a known focal length.

3. The method of claim 2, wherein an image focal length f of a coaxial spherical system in the set {P} is expressed as:

4. The method of claim 3, wherein there is a total of 2N−1 parameters for the radii of curvature and the distances, the radii of curvature are defined as r1, r2, . . . , rN, and the distances are defined as d1, d2, . . . , dN−1; when the image focal length of the coaxial spherical system is given, after the 2N−1 parameters are obtained, a distance between a last spherical surface and the image surface dN is solved by the first order optics, therefore, there are a total of 2N parameters describing the coaxial spherical system.

5. The method of claim 4, wherein the 2N parameters are placed together in a vector P=[r1, r2, . . . , rN−1, rN, d1, d2, . . . , dN−1, dN], and the vector P is configured to describe the OP distribution of the coaxial spherical system.

6. The method of claim 5, wherein the radii of curvature are defined as ri, and i=1, 2, . . . , N, a range of each ri is [rmin, rmax] with an interval Δr, and ri=rmin+Δr, rmin+2Δr, . . . , rmax.

7. The method of claim 6, wherein the range of each ri is [−1000 mm, 1000 mm], and Δr=100 mm.

8. The method of claim 5, wherein the distances between surfaces of the coaxial spherical system are defined as di, and i=1, 2, . . . , N−2, a range of each di is [dmin, dmax] with an interval Δd, and di=dmin+Δd, dmin+2Δd, . . . , dmax.

9. The method of claim 8, wherein EPD≤|d1|≤4×EPD, wherein EPD is an entrance pupil diameter of the coaxial spherical system.

10. The method of claim 1, wherein a global coordinate system O-XYZ is defined, an incident direction of light ray in a central field (0°) is defined as a Z axis, and a plane perpendicular to the Z axis is defined as an XOY plane; a local coordinate system is established at each surface of each off-axis spherical system in the set {C}m; an incident point of the chief light ray of the central field on a mirror surface Si is defined as Vi, and a local coordinate system Vi-XYZ is established with Vi as origin, a unit axis vector ViZ and a surface normal direction of a surface at Vi are parallel, a unit axis vector ViY is in a plane O-YZ of the global coordinate system O-XYZ and is perpendicular to the unit axis vector ViZ, and a unit axis vector ViX and a unit axis vector OX have the same direction; a projection of a light beam between the mirror surface Si and a mirror surface Si+1 in the O-YZ plane is defined as Bi, wherein i=0, 1, . . . , N, an area illuminated by a light beam on the mirror surface is defined as a working area of the mirror surface, a curve segment of a working area of a mirror surface Sj intercepted by the O-YZ plane is defined as E1(j)E2(j), wherein j=0, 1, . . . , N, I, and SI means the image surface; and a condition of no obstruction in the off-axis spherical system is: for any Bi(i=0, 1, . . . , N), any curve segment E1(j)E2(j) (=1, 2, . . . , N, I) has no overlap with Bi, and the unobscured off-axis spherical systems Cm,1, Cm,2, . . . , Cm,r, . . . , Cm,Rm are obtained after filtering according to the condition.

11. The method of claim 1, wherein after the unobstructed off-axis spherical systems are obtained, a plurality of unobstructed off-axis spherical systems with a small folding angle of optical path at each mirror surface are further found from the series of unobstructed off-axis spherical system, to obtain the series of compact unobstructed off-axis spherical systems.

12. The method of claim 11, wherein with regard to an unobscured off-axis spherical system Cm,r with an OP distribution form Pm, a new vector ΔCm,r is defined, and ΔCm,r=Cm,r−C0=[Δθ1, Δθ2, . . . Δθi, . . . , ΔθN], wherein C0 is a coaxial structure, and Δθi is a folding angle of the optical path at the mirror surface Si, and a set of vectors ΔCm,r corresponding to the set {C}m is defined as the set {ΔC}m.

13. The method of claim 12, wherein the plurality of unobstructed off-axis spherical systems with the small folding angle of optical path at each mirror surface are further found from the series of unobstructed off-axis spherical system by a minimization method.

14. The method of claim 13, wherein the unobscured off-axis spherical system Cm,r with an OP distribution form Pm is a three-mirror system, the vectors in the set {ΔC}m are classified to a plurality of categories, in each category, a first item Δθ1 and a second item Δθ2 are equal, a the third item Δθ3 is different; a process of minimizing the item Δθ3 comprises: in each category, a vector with a minimum absolute value of Δθ3 is found, and an off-axis spherical system corresponding to the vector with the minimum absolute value of Δθ3 has the smallest optical path angle at a tertiary mirror; for the vector that has been minimized item Δθ3, a same process of minimizing item Δθ3 is used to minimize the term Δθ2 in the vector; and then, for the vector that has been minimized the item Δθ2 and the item Δθ3, the same process of minimizing item Δθ3 is used to minimize the term Δθ1 in the vector; and a set of vectors whose term Δθ1, term Δθ2 and term Δθ3 are all minimized are obtained; a plurality of structural vectors corresponding to the set of vectors are found from the set {C}m, and the plurality of structural vectors are the series of compact unobstructed off-axis spherical systems {tilde over (C)}m,1, {tilde over (C)}m,2, . . . , {tilde over (C)}m,t, . . . , {tilde over (C)}m,Tm.

15. The method of claim 14, further solving a series of off-axis spherical systems with a smaller optical path folding angle at each mirror surface based on the set {{tilde over (C)}}m by a dichotomy.

16. The method of claim 15, wherein an upper limit of unobstruction is θi,uob=θi, a lower limit of obstruction is θi,obs=θi0, an intermediate value is defined as θi,mid=(θi,uob+θi,obs)/2, if θi=θi,mid, the off-axis spherical system is unobstructed, then let θi,uob=θi,mid; if θi=θi,mid, the off-axis spherical system is obstructed, then let θi,obs=θi,mid; and above steps are repeated until |θi,uob=θi,obs|<θmax, wherein θmax is a given threshold; and the upper limit θ3,uob, θ2,uob and θ1,uob at each mirror surface are solved according to above method, to obtain more compact unobstructed off-axis structure form vectors {tilde over (C)}′m,t=[θ1,uob, θ2,uob, θ3,uob].

17. The method of claim 1, wherein the optimal tilt angle of the image surface is found using a one-dimensional search method.

18. The method of claim 17, wherein one round of the one-dimensional search for the optimal tilt angle of the image surface comprises: a same off-axis spherical surface system is taken as a starting point, and only the tilt angle of the image surface is different in the process of constructing the freeform surface imaging system through the step S3 and step S4; when the tilt angle of the image surface is different, the image quality of the freeform surface imaging system is also different, and a tilt angle of the image surface corresponding to a freeform surface imaging system with a best image quality is the optimal tilt angle of the image surface.

19. The method of claim 18, wherein multiple rounds of one-dimensional search for the optimal tilt angle of the image surface are performed.

20. The method of claim 1, wherein in step S5, the evaluation metric is an imaging quality metric.