CAJ | 학술논문

在不同色温的光源下，CMOS图像传感器采集到的图像颜色会有所不同，可能偏离真实色彩，导致成像质量下降，严重时将影响高清电子内窥镜的诊断结果。提出了一种基于无色差表面识别的自动白平衡算法，使用像素点的RGB三通道之和识别图像中的无色差表面，根据得到的无色差像素集估计光照色度，并用光照色度估计值通过对角变换进行颜色校正，以此对内窥镜图像进行了实时的自动白平衡算法，修正偏色现象。该算法实时、有效地消除了由于光源色温引起的颜色失真现象，使得高清电子内窥镜获得的图像更加接近标准白光照射下人眼所看到的情况。
재불동색온적광원하，CMOS도상전감기채집도적도상안색회유소불동，가능편리진실색채，도치성상질량하강，엄중시장영향고청전자내규경적진단결과。제출료일충기우무색차표면식별적자동백평형산법，사용상소점적RGB삼통도지화식별도상중적무색차표면，근거득도적무색차상소집고계광조색도，병용광조색도고계치통과대각변환진행안색교정，이차대내규경도상진행료실시적자동백평형산법，수정편색현상。해산법실시、유효지소제료유우광원색온인기적안색실진현상，사득고청전자내규경획득적도상경가접근표준백광조사하인안소간도적정황。
Under the different color temperature of light sources, the colors of images obtained by CMOS image sensors will be different, it was likely to happen that it deviate from the real color, resulting in a decline of image quality, seriously affecting the effective of high -definition electronic endoscope. Proposes an auto white-balance algorithm based on achromatic surface detection, using the sum of RGB three-channel of each pixel to detect the achromatic surface. With the achromatic pixels set, the illumination chromaticity was estimated, and then correct the colors of image, through the diagonal transformation of endoscope image in real-time auto white-balance algorithm, modified the color deviation phenomenon. This real time, effectively algorithm eliminates the color deviation phenomenon caused by different illuminate condition, which is in accord with the color with the color characteristic of human visual system.