Research On Moving Object Detection Based On Improved Mixture Gaussian Model

Moving target detection technology plays an important role in intelligent video field and have a wide range of applications.However,the current moving target detection algorithm has many deficiencies in the actual application,result in not well adapted to complex scenes.So the dissertation will research on moving target detection algorithm.This dissertation introduces the basic knowledge of graphic image processing at first as well as discussing the common methods of moving target detection and their advantages and disadvantages.Aiming at the traditional mixture Gaussian model is easily affected by the illumination mutations,and it can not detect the slow-moving object.It is not well adapted to complex scene.An improved moving object detection method based on mixture Gaussian model was proposed.It combines the improved mixture Gaussian model and four-frame subtraction.Through introducing a dynamic threshold,it solves the problem affected by the illumination mutations.The improved mixture Gaussian model adjusts the numbers of the Gaussian distribution adaptively and improves the accuracy of the background description.Meanwhile,motion states of objects with different learning rates were discussed in the dissertation where slow-moving object detection was improved.Experiment result shows that the combined algorithm can detect moving object accurately and has better adaptability in complex scenes.