使用OpenCV精确比较2张图像的差异 [英] Precisely Compare 2 images for differences using OpenCV

问题描述

我正在一个项目中，以预定的时间间隔通过网络摄像头捕获图像，并将图像与模板(良好图像)进行连续比较，如果差异>公差设置，则会给出错误消息.

I am working on a project to capture images via webcam in a predefined time interval, and continuously compare the images to a template (good image) and give an error if the difference is > tolerance set.

我正在使用OpenCV进行锻炼.想咨询我应该如何做，例如最好的方法，等等.

I working out using OpenCV. Would like to have advice how should I do it, e.g. best method, etc.

如果有人可以帮助我，请感谢.谢谢.

Appreciate if any one can help me on this. Thanks.

推荐答案

一种简单的方法是在图像对之间采用L2-范数:

An easy way is to just take the L2-norm between the image pairs:

double l2_norm = cvNorm( img1, img2 );

您必须尝试设置适当的阈值.当然，这种方法对于光照变化，视点变化等并不稳健，但简单快捷.

You'll have to experiment with setting the appropriate threshold. Of course this method is not robust to lighting changes, viewpoint changes, etc but its simple and fast.

这篇关于使用OpenCV精确比较2张图像的差异的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！