Application And Research Of Camera Tampering Detection In Surveillance Video Based On Deep Learning

With the rapid development of video surveillance technology,there are more and more needs for functional and high performance video surveillance products.However,due to long time out of repair,equipment is aging and unstable which causes video suffering noise in different levels of quality degradation.To solve this problem,this article has proposed a deep learning based method to detect leaf occlusion and blur video.This work includes the following aspects:In the first part,we have reviewed the theory and development of convolutional neural network.Then the researches of video surveillance distortion detection and the current research status of image quality assessment detection algorithms have been reviewed.At the end of this part,we have summarized the surveillance video distortion types,and specifically analyzed the leaves occlusion and video sharpness assessment problem.The objective of leaf occlusion detection in video is to automatically determine whether the video suffers from leaf occlusion or not.To tackle this challenge,this paper proposes a two-step learning framework for leaf occlusion detection.First,the convolutional neural network is used to learn the discriminative features of leaf particles.Then the trained model is used to detect candidate leaf patches in the image.Second,a probabilistic approach is used to pool decisions of each candidate leaf patch to produce final detection result in the video.To assess the video sharpness,we have first used the ELO rating system to get the ground truth,then tried different convolution neural network structures to get better result to assess sharpness,and introduced the details in implementing project.Experimental results are provided to demonstrate that the proposed approach can effectively detect the leaf occlusion and assess the quality of sharpness in real-world traffic surveillance video.