A color correction method for color digital cameras with high fault tolerance

Abstract

The invention discloses a color correction method with high error tolerance for the chromaticity estimation result of a light source of an automatic white balance module in a color digital camera. The present invention converts the original RGB response signal of the camera into the device-independent CIE1931 XYZ space by calculating the spectral transformation relationship between the camera spectral sensitivity function and the CIE1931 color matching function, and uses the CAT02 color adaptation transformation model to calculate its color under the reference light source. The corresponding color response value after adaptation, so as to realize the color correction process that does not depend on the calibration parameters in advance.

Description

technical field [0001] The invention relates to a method for realizing color correction of a color digital camera by using a spectral sensitivity function transformation matrix and a CAT02 color adaptation transformation model. The method can enable the color digital camera to realize a color correction process independent of calibration parameters related to light sources. Background technique [0002] The color correction module is one of the important links in the image signal processing flow of a color digital camera. It converts the device-related color signal into a device-independent reference response value under a standard light source, so as to achieve human visual perception. digital image color reproduction. [0003] The color correction module in the current color digital camera usually uses several groups of pre-calibrated color correction matrix parameters as the basis for correction. In the calibration stage, for several typical light sources, the conversion...

AI Technical Summary

Problems solved by technology

Since the estimation of light source color from image content is an ill-posed problem, its estimation accuracy is easily affected by the content of the scene being photographed. Therefore, when using this method, once the estimated light source color is consistent with the real value If there is a large difference between them, the color correction module is very likely to select a wrong set of calibration parameters, resulting in severe color degradation of the color-corrected image

[0004] At the same time, for the existing color correction methods based on several sets of correction matrix parameters, how to determine the appropriate number of calibration light sources is also a big problem

If there are too few calibration light sources, the consistency of the color correction results cannot be guaranteed. At this time, once the light source chromaticity estimation results fluctuate, the appearance of the color-corrected image will also show obvious jumps; if there are too many calibration light sources, the The probability of choosing the wrong calibration parameters is correspondingly increased, and it is also more susceptible to interference from metamerism problems

Images