前段时间在OPENCV中做一个小东西，使用到子图像的提取和拷贝，一时间没有找到好的方法，要么直接访问像素进行赋值，要么通过数组来转换，总是感觉别扭，后来终于弄清了一个比较简单的方法，需要利用一下OPENCV中的ROI方法，该方法对于任何格式的灰度或者彩色图像均适用-Do a small thing, some time ago in the OPENCV in sub-images used to extract and copy, for a time did not find a good way, either direct access to the assignment of pixels, either through the array to convert, and always feel awkward, and later got to know the a relatively simple method, you need to use the OPENCV in the ROI method, the method for grayscale or color images of any format