Video surveillance has been crucial for security in recent years, thus it’s critical to guarantee the accuracy of these recordings. Unfortunately, it is easy to fake surveillance films by erasing an object from a scene and leaving no visual evidence. In this study, an approach for video intra-frame forgery forensics based on the SSIM (Structural Similarity Index Measure) is proposed. This system can recognize forged video frames automatically. In order to extract the steganography features, the method first decompresses the input video into a sequence of frames. It next calculates each frame’s motion residual map. The ability to tell whether or not an object is deleted from a video is a key problem in video security.