基于补色CFA模板的图像重构算法及其后续处理研究

Research on the Image Reconstruction Algorithm Based on Complementary Color CFA Template and the Subsequent Processing

采用双线性补色CCD插值算法对插出的图像再进行色空间转换(从CMYG转换到YCrCb空间),然后分别对Y通道、Cr通道、Cb通道进行滤波处理,最后把图像再转换到RGB原色空间,这样既考虑了C,M,Y,G通道的色彩相关性,又滤掉了部分明显的伪彩。在后续处理中,应用Sobel算子对图像的边缘进行检测,然后对检测出的边缘用改进的中值滤波器滤波,并保持原贝尔模板中的数值不变,使滤波的图像与原始图像更接近,细节比以前的图像更丰富。

By using the bitinear complementary color CCD interpolation algorithm, this paper converts the color space of the interpolated images from CMYG to YCrCb, then filters the Y channel, Cr channel and Cb channel, and finally converts the images into the original RGB color space, which not only considers the color correlation of C, M, Y, G channels,but also filters out some obvious pseudo-colors. In the subsequent processing, this paper detectes the edge of the image with Sobel operator, then filters detected edges by using the improved median filter, and keeps the original numeric value of Bayer pattern, which makes the filtered images closer to the original image and the details of the images richer than that of the original images.