Research On The Detection And Tracking Of Multiple Moving Objects

The detection and tracking of multiple moving objects has become the one of the hottest research topics in the field of computer vision, and it technology involve digital image processing and pattern recognition. The detection and tracking of the multiple moving objects can be summarized as follows. Firstly, using the image capture technology to extract the images from a video sequences. Secondly, using image processing techniques for each frame image denoising processing. Then, selecting a suitable method of detection to extract the object from the image. Finally, using pattern recognition technology and tracking algorithm to track the object. The traditional method of object detection and tracking have some shortcomings, such as shelter materials, complex background and interlaced moving target. When these conditions are present, the accuracy of the traditional detection method and the stability of the traditional tracking method will decrease. In this paper, the multiple moving objects which in the actual road will be studied in order to further study the detection and tracking technology.This paper mainly focuses on the multiple moving objects detection and tracking technology based on Wavelet Transform. The paper consists of three parts. The first part is researching on image processing technology based on the detection and tracking. The second part is researching on the detection of multiple objects. The last part is researching on the tracking technology of multiple objects. The content of the innovation of this paper are as follows:First of all, In the research of image processing technology, this paper presents an improved wavelet denoising method. Through simulation experiments, make the appropriate analysis, and the experimental results show that the improved method is superior to the traditional wavelet de-noising method.Secondly, In the study of multiple moving objects detection, this paper presents an improved detection algorithm based wavelet transform combine with the mixed Gauss background subtraction, to solve the problem of the traditional detection method which has poor anti-noise and anti-blocking capability. Proved by experimental comparison, improved detection algorithm proposed in this paper has a strong anti-noise and anti-blocking properties, and the accuracy is superior to the traditional detection method.Finally, In the study of multiple moving objects tracking, As the traditional tracking algorithm can’t accurate tracking of multiple moving objects in complex background, this paper presents an improved tracking algorithm based wavelet transform combine with the Meanshift and Kalman. Through simulation experiments, compared with the improved algorithm Meanshift tracking algorithm, the result shows that the accuracy of the improved tracking algorithm is superior to the traditional Meanshift tracking method in complex background, and has a strong robustness. But because of increased calculation amount, the real-time of tracking will be reduced. Further research needs to be done in the future is how to balance robustness and real-time.