Research On Tracking Of Moving Object Based On Integral Image And Camshift

Video motion object detection and tracking is a very popular direction in the field of computer vision,which includes digital image processing,pattern recognition,and even artificial intelligence,machine learning and so on.It has been well applied in military,industrial,intelligent traffic,Security monitoring and other aspects,and has been in development.Firstly,This thesis introduces the object detection algorithm,including the commonly used object detection algorithm and object detection based on feature.It focuses on the detection algorithm based on machine learning and can be used to object detection through training proper classifiers by plus-minus samples.Then the thesis introduces the more commonly used object tracking algorithm,including Meanshift,Camshift and Kalman filter algo rithm.Camshift algorithm is the improvement of Meanshift algorithm in fact and proves Camshift algorithm has self-adaptability to the dimension through the experiment.Faced with the problem of tracking fast moving object,the thesis proposes the integral image tracking algorithm in histogram similarity matching,studies the time complexity of the integral image,and proves the integral image can be able to track the fast moving object through the experiment.The algorithm can be combined with the Meanshift algorithm and the experiment proves that the tracking effect is better than the previous algorithm.For the problem of using Camshift algorithm failed to track object that similar to the color of background.In this paper,an algorithm is proposed to achieve tracking object of color similarity by subtracting background from each frame image and increasing the color discrimination of object and background.In the process of tracking the fast moving object,the machine learning algorithm is used to detect the object,and then faced with the problem of using the traditional Camshift algorithm to track the fast moving object which will lead to loss the track.After combining with the color space integral image algorithm,can be able to track and also has the self-adaptability to the dimension.For the fast moving object under the same object interference.In case of using the integral image and Camshift algorithm to track lost,combined with the increasing color discrimination algorithm that proposed in this paper,can be able to track effectively.