Research On Soccer Robot Vision System Based On Wavelet Analysis

Vision system is an important part of robot system. It has been employed various techniques in different kinds of scientific subjects such as patter recognition, artificial intelligence, computer image processing, etc. to fulfill a function like human eyes to perceive and acquire information from the outside world. As a branch of the machine vision field, soccer-robot vision system is crucial to fulfill its real-time, precise and robust performance, and it provides the information for the decision-making system. This dissertation is devoted to the research on image processing and related techniques for soccer-robot vision system. It represents special theories and algorithms according to the specialties of color images, proposes a whole solution to achieve the goal of defect detection and segmentation.Wavelet analysis is a novel research field in the world. It has a perfect performance on both local time and frequency domain, so it is an effective analytical tool. Digital image processing by the aid of wavelet analysis is one of the hot issues of wavelet research and application. After a brief description of the basic theory of wavelet, this paper makes a deep research on the algorithm of image denoising and compression.Following theory and technology have been studied and explained in this paper:1 The Robot Soccer competitions and the key technologies and visual system of the robot soccer will be introduced.2 Wavelet analysis will be introduced in detail.3 Wavelet denoising. In the visual system of the Robot Soccer, camera which is two meters above the platform has been adopted to acquire the needed images. During the dissemination process, the images from the camera are subject to varying degrees of interference, so the robots or ball maybe disappeared from the system interface of computer, so competition will fail owing to the reason above. In view of this phenomenon, this paper put forward an improved filtering algorithm using wavelet analysis to study the robot visual image.4 Wavelet image compression has been used in processing image of Robot Soccer in the light of the merits of wavelet image compression. During the process of Robot Soccer competitions, some systems use a "field" way to display the image containing large amounts of data in order to make the speed of the image processing doubled. The result can also be achieved by using the wavelet analysis and at the same time the removal of image redundancy cannot make the useful information lost.5 Window search method has been used to identify the robots in real time. This method can avoid a waste of time in the traditional global search.6 At the same time, it also implements the perfect binding of the parallel offline process of system by VC + + and MATLAB and online processing by VC + +. In a word, this process method simplifies the difficulty of development.In view of the above methods, the paper for simulations and practices has been proved the correctness of this view. With the robot soccer matches, the visual system will be more perfect and the technology will be more mature.