Video Moving Object Detection Technology Based on Deep Learning

In the field of computer vision, video intelligent recognition technology is still a research difficulty. With the improvement of power grid automation, video monitoring system has been widely used in power grid. In power application scenarios, real-time monitoring of the scene and equipment can make the power grid truly unattended and make the power grid operation more secure. In order to solve the problem of low accuracy of power scene video intelligent recognition and improve the real-time performance of video processing, a video moving target detection technology based on lightweight convolutional neural network is proposed. Firstly, the image preprocessing technology is used to process the collected image, so as to filter the interference of background and other factors on device recognition; Then, a lightweight convolutional neural network structure mobilenet is designed as the feature extraction module of the target detector; Then, the multi-target detection algorithm based on R-FCN is used to locate the inspection equipment; Finally, the deep neural network compression technology based on network sharing and quantization is used to compress the target detection model to reduce its memory occupation, so as to improve the real-time performance of target detection. The experimental results show that the method can accurately locate and recognize the moving target in the video, and can be applied to the intelligent recognition task in the power scene.