Image Corner Detection Algorithm Research And Tracking Of Moving Objects Based On The Corner,

In this paper, the methods of comer detection are mainly studied and a new method of object tracking is proposed.In computer vision, machine vision and image procession, feature detection is an important direction. Corner is an important part feature of image, because the shape of the objection is decided by it. So, it is very useful in objection description, motion estimation and objection tracking. Corner includes enough information, so it provides many limitations to image procession, and the speed of computation is improved.First of all, some classical methods of corner detection are introduced and their Characteristics are discussed. Secondly, Based on 'Susan corner detection', an improved method of corner detection is proposed. Meanwhile, we propose a new method of corner detection based on image edge. Experiments have been done respectively. According to the results, we find that the method of improved 'Susan corner detection' has the Characteristic of quickness, accuracy, and the other new method also have good effects. So, the study is meaningful. Lastly, according to the detected corners of image, feature vectors of these corners are built to track the object.