This paper uses the open source computer vision library (OpenCV for short) as basic library and calls library functions to achieve real-time video capture and output. We narrate the process of making the matrix interpretable as an image through analyzing the data structure of IplImage and meaning of its member variable. The library functions to read video take the video frame for unit and the storage form of video frame in memory is the same as the ordinary image, so we use Arnold transformation to scramble each frame of video and accomplish video encryption.Ultimately play the processed video frames successively and achieve video encryption based on OpenCV simply.