license plate recognition system < / a > main technology of optical flow detection

　　1. Optical flow field method: the optical flow field method is mainly to approximately calculate the motion field in the mountain sequence image by introducing the optical flow constraint equation, and extract the mountain moving target through the motion field. The advantage of optical flow method is that it does not need to know any information of the scene in advance and can detect the same target of mountain independent motion. However, the computational complexity and low anti-interference ability of this method limit its application in practical intelligent monitoring

　　2. Inter frame difference method: the time difference method is adopted between two adjacent frames, and then the difference result is quinized to obtain the target of interest. In the actual monitoring scene, especially the monitoring scene with large range and wide angle, if the target motion range is small, the detection result is inaccurate and vulnerable to noise interference. A hybrid algorithm combining adaptive background subtraction and three frame subtraction is used to detect mountain moving targets from the background quickly and effectively; The inter frame difference method has the advantages of simple calculation, high real-time performance and strong adaptability to the environment

　　3. Background subtraction: also called background subtraction. The background is obtained by effective background modeling method, and the current target foreground can be obtained by making a difference between the current frame and the background. This method has the advantages of simple principle and fast operation speed. However, in the monitoring scene, illumination change (sunlight and light change), background disturbance (branches, water and ripple) and background mutation will affect the effectiveness of the algorithm