The present system relates generally to cameras, and in particular, to a method for determining camera motion.
Detection of camera motion is important in order to be able to compensate for image blur due to movement of the camera, and also to allow for image ‘stitching’ when taking multiple pictures to create a continuous scene. Detecting camera rotation is particularly important when an object to be photographed is not close to the camera. Presently known rotation sensors, such as gyroscopic sensors, are expensive, and marginally sensitive to camera rotation. These rotation sensors are typically based on mechanically resonant structures, and each of these sensors must typically operate at different frequencies to minimize crosstalk.
A method is provided for detecting rotational movement of a camera.
In one embodiment, one to three pairs of accelerometers are located in the camera. The motion sensing axes of each of the accelerometers in each of the pairs are parallel to one another, and the accelerometers are relatively positioned in the camera such that the planes formed by the motion sensing axes of each of the pairs of accelerometers are substantially mutually orthogonal.
In another embodiment, a differencing device is coupled to an output of each pair of the accelerometers, for generating a difference signal with respect to the outputs of each of the accelerometers in the pair. The difference signal is indicative of the angular acceleration of the camera with respect to the angular sensing direction for each accelerometer pair.
In an additional embodiment, an integrator is coupled to the output of the differencing device, for generating an integrated signal with respect to the output of the differencing device. The integrated signal is indicative of the angular velocity of the camera with respect to the angular sensing direction for each accelerometer pair.
The accelerometers in each accelerometer pair A1/A2, B1/B2, and C1/C2 are relatively positioned such that the lines of motion (i.e., lines X1/X2, Y1/Y2, and Z1/Z2, for accelerometer pairs A1/A2, B1/B2, and C1/C2, respectively) of the accelerometer proof mass of the accelerometers in each pair are parallel to one another. The motion sensing axis for a given accelerometer is defined herein as an axis along which movement is detected by the accelerometer, which axis is parallel to the line of motion of the accelerometer's proof mass, and which passes through the center of the accelerometer. For example, the motion sensing axis for accelerometer A1 is the line indicated by arrow(s) X1.
The angular acceleration sensing axis for each accelerometer pair is a line that is orthogonal to the plane formed by the lines of motion of accelerometer proof masses of the two accelerometers in a given pair. For example, the angular acceleration sensing axis for accelerometer pair A1/A2 is a line that is orthogonal to plane X1 X2 (indicated by reference no. 105). Although there is only one instantaneous axis of rotation for a given accelerometer pair at any given time, a single accelerometer pair cannot indicate the specific location of that axis. Rather, an accelerometer pair can only determine the direction of rotation of camera 101. Therefore, the term ‘angular sensing direction’ is used herein to describe the direction of rotational motion of camera 101 about an axis (e.g., line 104 or 104A) that is orthogonal to the plane (e.g., plane X1 X2, reference no. 105) formed by the motion sensing axes of the two accelerometers in a given pair.
As shown in
As the separation between the two accelerometers in a pair is increased, the rotational sensitivity of the pair is increased, when the present system is employed. In an exemplary embodiment, the accelerometers in each of the accelerometer pairs are positioned as far apart from each other as practicable within the camera body 102, each accelerometer preferably located proximate a different side of the camera body. In an exemplary embodiment, an Analog Devices ADXL203 accelerometer may be used for accelerometers A1, A2, B1, B2, C1, and C2. At 5 centimeters typical separation, a pair of this particular type of accelerometers has a sensitivity of 0.2 radians/secˆ2 with a 50 Hz bandwidth, when employed in accordance with the present method.
It is to be noted that alternative embodiments of the present system may employ a linear acceleration-detecting means other than the above-described type of accelerometer, with the requisite condition that the acceleration-detecting means includes a motion sensing axis having an essentially linear ‘line of motion’. Other alternative embodiments of the present system may include devices other than cameras, in which are located one or more pairs of accelerometers positioned in accordance with the method described herein.
As shown in
More specifically, the output signals from accelerometer pair A1/A2 are combined or otherwise processed via differencing function 501 to yield a difference signal (A1−A2) 505. The output signals from accelerometer pairs B1/B2 and C1/C2 are likewise processed via differencing device(s) 301, which perform(s) the signal differencing functions indicated by blocks 502 and 503 to yield difference signals (B1−B2) 506 and (C1−C2) 507, respectively. Each of these difference signals represents a component of the angular acceleration of camera 101 with respect to the corresponding accelerometer pair A1/A2, B1/B2, or C1/C2.
The embodiment shown in
In the present embodiment, at step 415, integrators 511-513 receive difference signals (A1−A2) 505, (B1−B2) 506, and (C1−C2) 507 from differencing device(s) 301. The differenced outputs from accelerometer pairs A1/A2, B1/B2, and C1/C2 are then input to application 500, as indicated by arrows 508, 509, and 510, together with the integrated outputs from integrators 511-513, as respectively indicated by arrows 516, 517, and 518. At step 420, application 500 uses this differenced and integrated information to compute the angular acceleration of camera 101 with respect to the corresponding accelerometer pair A1/A2, B1/B2, or C1/C2, and to thus determine camera motion about the X, Y, and Z axes.
For a more complex application in which motions are too large to be treated independently the differenced outputs from accelerometer pairs A1/A2, B1/B2, and C1/C2 are input directly to application 500, as indicated by arrows 508, 509, and 510. The application then performs the required signal processing, which typically involves double integration of coupled systems of linear equations. In addition, a complex application may also obtain summed signals from each of the accelerometer pairs to obtain positional information in addition to the rotational information.
Changes may be made in the above methods, systems and structures without departing from the scope hereof. It should thus be noted that the matter contained in the above description and/or shown in the accompanying drawings should be interpreted as illustrative and not in a limiting sense. The following claims are intended to cover all generic and specific features described herein, as well as all statements of the scope of the present method, system and structure, which, as a matter of language, might be said to fall therebetween.