Patent Application
20070165963
1. Field of the Invention
This invention relates to image editing, and more particularly to correcting shading noise.
2. Description of the Prior Art
In general, a scanning process of a scanner includes: moving a light source and a scanning module along a scanning direction by means of a stepper motor. Then, the light source illuminates the material of demand and is reflected into the scanning module. An: image is captured by a light sensor such as CCD (Charge-Coupled Device) and is then translated into digital data to be saved.
The digital data of the image is different from the captured data with the fight sensor, due to the light provided by the light source not being an ideal linear light. and the brightness varying along a direction of the arrangement of the light sensor. The brightness is approximately brighter in the center of the light than in the edge thereof, as shown in
Furthermore, a light sensor is a plurality of CCDs arranged in a line and the data detected by different CCDs is different due to each CCD having a different light sensitivity. Moreover, external factors may vary with time, for example: the variant of light resulting from a power supply varying with time, the variant in the light sensitivity of the CCD with time, etc. These factors are generally called “shading distortion”. Therefore, the data of the image is corrected for the elimination of the disproportionate distribution of the light, but the corrected data still has a problem of being uneven, resulting from the shading noise, as shown in
The corrective method of the shading distortion is that the scanner pre-scans a reference white (a white corrective board) or a reference black (covering with the light sensor before capturing image). The data of the reference white or the reference black is the shading distortion corrective curve. Hence, a captured image can be corrected with the corrective curve to obtain a more corrective image data.
Nevertheless, the shading distortion corrective curve has local maximums or local minimums resulting from the noise of the electrical devices, bad CCDs, or inexact deductive methods. The local maximums or minimums resulting from random noise do not appear every time. The inexact deductive method causes the data to be improperly corrected. These results are generally called “shading noise”. If the image data is corrected by a corrective curve that comprises shading noise, the resultant corrected image has some shading noise lines. The shading noise lines in the corrected image are caused by the shading corrective curve. While the shading corrective curve may be one dimensional, the shading noise that appears in the resultant corrected image appears as lines in a two-dimensional image in positions correlating to the shading distortions in the corrective curve.
Hence, the shading corrective curve in the conventional arts can correct the problem of shading distortion, but they easily generate problems with shading noise and reduce the quality of the image.
The conventional arts mentioned above can correct the shading distortion, but generates the problem of the shading noise. In accordance with the present invention, image editing may be used for correcting the shading noise to efficiently improve the problem of the shading noise resulting from the correction of the shading distortion.
It is another object of this invention to employ image editing to correct the shading noise for increasing the image quality.
In accordance with the above-mentioned objects, the present invention provides image editing for correcting the shading noise. In the present invention, it detects the shading noise by means of characteristics of the shading noise, and eliminates the shading noise for an increased image quality.
Some sample embodiments of the invention will now be described in greater detail. Nevertheless, it should be recognized that the present invention can be practiced in a wide range of other embodiments besides those explicitly described, and the scope of the present invention is expressly not limited except as specified in the accompanying claims.
The process of obtaining the shading distortion corrective curve includes measurement with three primary colors RGB (red, green, and blue), respectively. Therefore, the characteristic of the shading noise is a line of one pixel width with one certain primary color in an image. The primary values in the pixels of the lines are lower or higher, than one preselected percentage of pixels (such as one certain percent, for example) of the color values of the primary color, within the two adjacent pixels in the adjacent lines. Therefore, the shading noise of an image can be detected and eliminated by means of the characteristic of the shading noise. The steps of eliminating shading noise are: start, detecting shading noise, eliminating shading noise, and end, as shown in
X(i)>Max[X(i−1),X(i+1)]
or
X(i)<Min[X(i−1),X(i+1)]
wherein X(n) is the nth color value of one certain primary color.
Eq.(1) can be rewritten
[X(i−1)−X(i)][X(i)−X(i+1)]<0 (2)
If the ith pixel of the corrective curve has a shading noise in one certain primary color, the inage of two-dimension M×N corrected writh the corrective line. has a shading noise line ((ij), j=1 to N). Therefore, Eq. (2) in one-dimension can be rewritten to translate into Eq. (3) in two-dimension:
[X(i−1,J)−X(i,j)][X(i+1,j)]<0 (3)
wherein x(ij) is a color value of one certain primary color and i is one certain value among 1 to M ,j=1 to N.
If the Eq. (3) is true dunig j=1 to N, the line of (i,j) j=1 to N) is a shading shading noise line. The primary color values with shading noise of the ith line ansforms into the average of two adjacent same primary color values in adjacent lines for correcting the shading noise.
Xn(i,j)=[X(i−1,j)+X(i+1,j)]/2 (4)
wherein the Xn(i,j) is a new connected color value of (i,j) and j is 1 to N.
If the Eq. (3) is not true during j=1 to N, the line of (i,j: j=1 to N) is not a shading noise line and the color value in the lie is not corrected.
Xn(i,j)=X(i,j) (5)
wherein j is 1 to N.
Referring to
Regarding an image where the color or brightness greatly varies, the variant of the color or brightness may be larger than the variant of the shading noise. Therefore, Eq. (3) may be not true during j=1 to N. Accordingly, if more than one certain percent (e.g. 80%) thereof during j=1 to N is true, the color value may be corrected with Eq. (4). The value of the one certain percent can be determined according to different conditions. Referring to
Although the present invention has been described in its preferred embodiment, it is not intended to limit the invention to the precise embodiment disclosed herein. Those who are skilled in this technology can still make various alterations and modifications without departing from the scope and spirit of this invention. Therefore, the scope of the present invention shall be defined and protected by the following claims and their equivalents.
This application is a continuation application of U.S. application Ser. No. 10/320,681, filed Dec. 17, 2002. The entire disclosure of U.S. application Ser. No. 10/320,681 is considered as being part of the disclosure of the present application and is hereby incorporated by reference herein in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10320681 | Dec 2002 | US |
| Child | 11668960 | Jan 2007 | US |
