Research on Moving Object Detection Algorithm Based on Improved Three Frame Difference Method and Optical Flow

For optical flow algorithm complexity, computationally intensive, poor real-time drawbacks ,this paper proposes a moving detection algorithm that combine with optical flow method and three frame difference method which be used OTSU to process. Choose the typical Harris corner of extracting area to calculate, it can effectively reduce the complexity of the algorithm. Since moving target area detection was not complete, the three-frame difference image which be processed by OTSU binary image was introduced as a supplement of simplify the optical flow method. Through the experiment, this method make the optical flow method satisfied the real-time request . Detect more complete area of the moving object.