最近愚蠢的我碰到的问题：
一、如下面一样写是错的：全输出0 0 0
Mat src = imread(srcfolder); Mat bw = imread(donefolder,CV_LOAD_IMAGE_GRAYSCALE); int rows = src.rows; int cols = src.cols; for (int i = 0; i != rows;i++) { uchar *bwdata = bw.ptr<uchar>(i); for (int j = 0; j != cols;j++) { Scalar bwvalue = bwdata[j]; if (bwvalue==Scalar(0)) { continue; } else { Vec3b values = src.at<uchar>(j, i); int blue = values.val[0]; int green = values.val[1]; int red = values.val[2]; cout << red << " " << green << " " << blue << endl; //int index = red + green * 256 + blue * 256 * 256; //rgbcsv[index] = 255; //file << red << " " << green << " " << blue<<"\n"; } } }但改成下面就正确了： IplImage* src = cvLoadImage(srcfolder); IplImage* bw = cvLoadImage(donefolder, CV_LOAD_IMAGE_GRAYSCALE); int srcstep = src->widthStep / sizeof(uchar); uchar *srcdata = (uchar*)src->imageData; int bwstep = bw->widthStep / sizeof(uchar); uchar *bwdata = (uchar*)bw->imageData; for (int i = 0; i != src->height; i++) { for (int j = 0; j != src->width; j++) { uchar bwtemp = bwdata[i*bwstep + j + 0]; if (bwtemp == 0) { continue; } else { int tempb = srcdata[i*srcstep + j*src->nChannels + 0]; int tempg = srcdata[i*srcstep + j*src->nChannels + 1]; int tempr = srcdata[i*srcstep + j*src->nChannels + 2]; cout << tempr << " " << tempg << " " << tempb << endl; //file << tempr << " " << tempg << " " << tempb << "\n"; } } } 同样的问题：下面这样错误： Vec3b values = srcrect.at<uchar>(j, i); int blue = values.val[0]; int green = values.val[1]; int red = values.val[2]; int index = red + green * 256 + blue * 256 * 256; int gray = red*0.299 + green*0.587 + blue*0.114; if (rgbcsv[index]==255) { if (gray<17) { continue; } else dst.at<uchar>(j, i) = 255; }改成下面就正确了： int dststep = dstimg->widthStep / sizeof(uchar); uchar* dstdata = (uchar*)dstimg->imageData; int step = testimg->widthStep / sizeof(uchar); uchar* data = (uchar*)testimg->imageData; uchar tempb, tempg, tempr; for (int i = 0; i < testimg->height; i++) { for (int j = 0; j < testimg->width; j++) { tempb = data[i*step + j*testimg->nChannels + 0]; tempg = data[i*step + j*testimg->nChannels + 1]; tempr = data[i*step + j*testimg->nChannels + 2]; int rgbpixels = tempr + tempg * 256 + tempb * 256 * 256; uchar rgbelement = rgbarray[rgbpixels]; if (rgbelement == 255){ uchar gray = tempr*0.299 + tempg*0.587 + tempb*0.114; if (gray<17) { continue; } dstdata[i*dststep + j*dstimg->nChannels + 0] = 255; } } }我知道了： Mat时不能用at！要用.ptr<uchar>这样的方式才行！！！！！！！！！！！！！！！！！！！！！！！！！！！！！！！！！！！！！！
二、将图像像素值保存为csv文件时遇到的问题：
1、uchar类型的像素值保存为csv时无法用excel打开，里面是乱码的。所以我将uchar转成了int才行；
2、将int型的r g b通过file保存为csv时 中间不能只用一个空格，否则r g b只占了excel的一个单元格，改成转义字符\t 竟然也不行，所以我改成了8个空格就可以了；这样就可以后续被CvMLData读出来；
三、关于Alpha通道总所周知，opencv在imread图像时默认是BGR通道，也就是CV_LOAD_IMAGE_COLOR (>0) 那么就是三通道的BGR；如果选择是CV_LOAD_IMAGE_UNCHANGED (<0) 那么就是BGRA ，也就是四通道！！！但我刚在linux下OpenCV3.2运行了下 CV_LOAD_IMAGE_UNCHANGED (<0) 这种情况的，然后立刻输出这个图像的channels是多少，结果发现仍然是3！！！！！！这里与我以前以为的不一样！甚是奇怪？？？？
无奈，我通过：
void createAlphaMat(Mat &mat,Mat primer_src_img) { for (int i = 0; i < mat.rows; ++i) { for (int j = 0; j < mat.cols; ++j) { Vec4b& bgra = mat.at<Vec4b>(i, j); Vec3b bgr=primer_src_img.at<Vec3b>(i,j); bgra[0] = bgr[0]; bgra[1] = bgr[1]; bgra[2] = bgr[2]; //bgra[3] = saturate_cast<uchar>(0.5 * (bgra[1] + bgra[2])); bgra[3]=0; } } }其中，Mat primer_src_img是CV_LOAD_IMAGE_COLOR (>0) 也就是常用的三通道的BGR；但Mat srcimg(2048,2448,CV_8UC4); 是我规定的4通道的，然后通过 createAlphaMat(srcimg,primer_src_img);然后紧跟着打印srcimg的channels，结果发现 现在输出是 4！！！！这样才满足我要的4通道的图像？！
