Research On Methods Of Image Processing Of The Image Information Processor

Image information processor deals with images with computer to achieve desired results. Its main function is to realize visual information processing and analysis function like human beings on the machine. With the development of technology of information and computer, the image information processor is applied further in the field of aviation and aerospace, E-communication, medicine and industrial control etc. Researches on application technology of image processing of the image information processor are more and more important, too.An algorithm of image processing will be researched in this paper, which is designed to find out the target and its location by information extraction with the image.In this paper, the meaning, research status and development tendency of technology of image processing of the image information processor are discussed comprehensively. Then the characteristics and the image processing results of the relative technology of image processing are analyzed particularly. The improved algorithm, which realizes tracking the target by edge extracted, is raised. In the algorithm noises are removed by the method of threshold denoising, which removes the noise and protects the edge. Then a method based on the Canny operators is used to get the edge image. And the Split-Merger operators is used to decrease the influence of the disaccord of the gray scale in the image, the threshold is selected self-adaptive by the characteristic of the gradient histogram to reserve the true edge and remove the untrue edge to get the preliminary edge image. A method based on the ACO(Ant Colony Optimization) is used to connect the breakpoint of the edge. At last the Chain Code Tracking algorithm is used to remove some untrue edge and get the characteristic value of the target for the sake of finding out the target and its location.The methods raised in the paper get good effects in keeping balance of reserving the true edge and removing the untrue edge. The gotten edge is single-pixel-width edge, have good connectivity and it is positioned accurately,which is favor of finding out the target and its location by edge feature extraction. The threshold is selected self-adaptive which reduce the manual intervention. This method is stable and reliable, can meet the real-time requirements of the image information processor.