Patent Application
20050036705
The present invention relates to the field of digital imaging and, more particularly, to systems and methods dealing with streaks and other image artifacts that appear on a printed image and which run parallel to the printer's process direction.
Macro non-uniformity levels have existed in raster scan image output terminals (IOTs) (e.g., xerographic printers) for some time and are a concern for most marking processes. Even small non-uniformity level errors in raster scan IOTs give rise to visually objectionable banding in halftone outputs (e.g., image macro non-uniformity streak artifacts). Such errors typically arise in raster scan image output terminals (IOTs) due to variations in ROS spot size across the field (which is constant in time (print to print)), donor-roll once-around, HSD wire hysteresis, laser diode variations, LED bar power variation, ROS scan line non-uniformity, photoreceptor belt sensitivity variations, and/or ROS velocity non-uniformity. Significantly, many variations occur only in the fast scan (e.g., X) or slow scan (e.g., Y) directions, and they do not interact to first order. Therefore, a correction made in one direction has a negligible effect on artifacts in the other direction. Other printing technologies (e.g. thermal inkjet and acoustical ink printing) also have artifacts that occur in a regular, predictable manner in one or both directions and fall within the scope of this discussion.
Although techniques have been proposed to eliminate such non-uniformity errors by making physical systems more uniform, it is too expensive to control or limit the error to an acceptable level, below which the error will not be detected by the unaided eye. Fixes have been attempted in the marking process, but not enough latitude exists to fully solve the problem. For problem sources such as LED non-uniformity, the correction is sometimes addressed with current control or pulse width control. However, none of the solutions discussed above implements a technique based in digital electronics. With the cost of computing rapidly decreasing, such digital electronics based solutions are becoming more attractive.
Streaks are image artifacts that appear in a printed image and run parallel to the process direction. They are caused by a non-uniform response of at least one of the xerographic components in a given marking engine. Photoreceptor scratches, HSD-wire contamination, and ROS spot size variations are examples of subsystem problems that can give rise to streaks. The current methods used to mitigate streaks are “passive”; i.e., based on demanding subsystem components of high quality and/or tight manufacturing tolerances. For example, the requirements could be a softer toner with low abrasion, a better coating for the HSD-wire, a more accurate optical system for the ROS, etc. As a result, manufacturing (UMC) and maintenance costs for products increases to cope with stringent image quality requirements. This is undesirable.
What is needed in this art is an image processing method to reduce streaking or other image artifacts which appear on a printed image.
What is disclosed is an image processing method to reduce streaking on a printed sheet. The method uses a negative feedback system to reduce streaking. The feedback system consists of a scanner that scans a printed sheet corrected with an error signal. The error signal is obtained from the scanned image of a previous print off the same printer and the desired output. There are a number of error signals in different positions across the process direction which adjust TRC values. The present invention makes use of a spatial filter representing the spatial frequency sensitivity of the human eye such that only differences (errors) of importance in need of correction are those which can be seen by the human eye. More specifically, a spatially distributed controller is utilized to improve sensitivity to measurement and printer noise. The printer adjusts the incoming bitmap so that each column in the image gives the same response for the same gray level. The printer sacrifices the uniformity of the responses at low and high frequencies where the eye is not responsive so that it can achieve better control at mid-frequencies where the eye is more responsive. At each iteration and for a fixed desired gray-level, the controller processes differences between desired output gray-levels and the values of the gray-levels obtained on the printed output image. A spatial filter is used to emphasize ranges of spatial frequencies critical for a human observer. Then, from these filtered differences (errors), TRC corrections to be applied to the output image are computed using simple scalar integral controllers. Advantageously, since the present invention compensates for defects introduced by lower quality components in a printer, printers can be less tightly specified.
The preferred embodiments and other aspects of the invention will become apparent from the following detailed description of the invention when read in conjunction with the accompanying drawings which are provided for the purpose of describing embodiments of the invention and not for limiting same, in which:
What is disclosed is an image processing method to reduce streaking on a printed sheet. The method uses a negative feedback system to reduce streaking. The feedback system consists of a scanner that scans a printed sheet corrected with an error signal. The error signal is obtained from the scanned image of a previous print off the same printer and the desired output. There are a number of error signals in different positions across the process direction which adjust TRC values. The present invention makes use of a spatial filter representing the spatial frequency sensitivity of the human eye such that only differences (errors) of importance corrected are those which can be seen by the human eye.
With reference now being made to
With reference now being made to
A test pattern of a printed image has a uniform gray-level band introduced across the page in the cross-process direction giving rise to the appearance of non-uniformities or streaking defects, i.e., sections that are darker as well as sections that are lighter than expected. From this image an expected gray-level desired to be achieved on the output image is obtained. A global or nominal TRC (Tone Reproduction Curves) is that which gives the gray-level expected. The image intended for correction is then printed and scanned. The local TRCs give the gray-level for each distance x (each pixel) from the edge of this image in the cross-process direction. For each pixel in the cross-process direction, a difference between local TRCs and the global TRC can be measured. The measured difference between local TRCs and the global TRC result in a compensation value computed therefrom. The amount of correction u(k) is the difference between the desired gray-level (ybar) and the scanned gray-level (y), and is given by: y(k)−ybar. This is passed through a multi-input multi-output (MIMO) spatial filter to limit adjustments to only those spatial frequencies to which the human eye can discern. Without this filter, effort may get wasted on frequencies that are unnoticeable to a human observer. Thus, the spatial filter makes sure that the controller's efforts are primarily directed towards critical frequencies only. The MIMO spatial filter can be represented by a constant matrix M specifically relating error vector e(k) at iteration k with the filtered error vector ef(k), such that: ef(k)=Me(k). The filtered output thereof is then preferably sent to individual single-input single-output (SISO) controllers which compute the corrections necessary to drive the filtered errors to zero. A single SISO controller is preferably an integral controller producing output correction: ui(k+1)=ui(k)+aefi(k), where i is the pixel and a is a constant. Each controller gain is preferably tuned so all controllers are the same. This improves speed of response while reducing noise sensitivity.
While particular embodiments have been described, alternatives, modifications, variations, improvements, and substantial equivalents that are or may be presently unforeseen may arise to applicants or others skilled in the art. Accordingly, the appended claims as filed and as they may be amended are intended to embrace all such alternatives, modifications variations, improvements, and substantial equivalents.
