ABL去紫边分析和过程

过程"/>

ABL去紫边分析和过程

思路：如果是蓝边，让b>statution&&b-g>thd则认为是蓝边，倒时有冲突景再细分
因为调整完的图像会黑一块正常的一块，我们就需要在黑的那块乘上一个gain，但是如果r&g差不多的话需要考虑一下，避免误抬高，thd就是gain。

	// 计算蓝色通道和绿色通道的差值cv::Mat diff_g_b;cv::Mat diff_g_b_1;cv::Mat diff_g_r_1;uchar b_thd=160;uchar diff_thd = 70;float diff_gr_thd = 1.3;float thd=1;cv::subtract(channels[0], channels[1], diff_g_b);cv::split(image, channels);for (int i = 0; i < image1.rows; ++i){		for (int j = 0; j < image1.cols; ++j){			if (channels1[0].at<uchar>(i, j)> b_thd){ if(channels1[0].at<uchar>(i, j)- channels1[1].at<uchar>(i, j)>diff_thd){blue_pixel_org = channels1[0].at<uchar>(i, j);green_pixel_org = channels1[1].at<uchar>(i, j);red_pixel_org = channels1[2].at<uchar>(i, j);float result3 = static_cast<float>(blue_pixel_org);float result4 = static_cast<float>(green_pixel_org);float result5 = static_cast<float>(red_pixel_org);float green_blude_pixel_add1 = blue_pixel_org + green_pixel_org;channels1[0].at<uchar>(i, j) = green_pixel_org - red_pixel_org + channels1[1].at<uchar>(i, j);if(result4/ result5> diff_gr_thd){float result2 = (green_blude_pixel_add1) /(static_cast<float>(channels1[0].at<uchar>(i, j))+ static_cast<float>(channels1[1].at<uchar>(i, j)));channels1[0].at<uchar>(i, j) = channels1[0].at<uchar>(i, j) * result2*thd;channels1[1].at<uchar>(i, j) = channels1[1].at<uchar>(i, j) * result2*thd;channels1[2].at<uchar>(i, j) = channels1[2].at<uchar>(i, j) * result2*thd;}}				 				}				}}cv::subtract(channels1[0], channels1[1], diff_g_b_1);// 将通道合并回去cv::merge(channels1, 3, image1);