Research Overview on Moving Object Detection Methods Based on Video Image Analysis

Moving object detection based on video image analysis is the process of extracting target objects from video sequences, aiming to identify moving target objects and separate them from the background. With the wide application of video surveillance technology, moving object detection has become more and more important in intelligent surveillance, traffic management, security protection, and other fields. This paper reviews the methods of moving object detection based on video image analysis and describes the basic concepts and technical framework of moving object detection under two different conditions: the camera is still and the camera is moving. These methods include Temporal Difference method, background modeling method based on statistical models, robust principal component analysis based on sparse representation, algorithms based on feature point trajectory, background subtraction method, optical flow method, deep learning method and fusion algorithm. The principles, methods, advantages, and disadvantages of these methods are summarized in depth. This paper aims to provide a comprehensive technical overview for subsequent research through comparative analysis.