Images Processing System Based On ARM Embedded Platform

Vision information plays an important role in computer information processing systems and their applications. Today, digital image processing attracts more and more attentions and many good, mature algorithms have been proposed. But these algorithms are not used to applications because of the need of application is widely enlarged.The thesis explores a new solution of image processing system based on embedded platform aimed at real time and mobility. Embedded platform has high integration, supporting real time multitask operation system. So it can face more and more complicated image applications because of it meets the demand of real time and mobility.Thesis presents the designing and realization of a mobile vision device that detects and tracks the pedestrians in sequence images (video). The device can use in domains of traffic flux statistical works and active safety of automobiles.In this thesis, we bring forward the function analysis and collectivity design based on the vision device. We analysis the critical problems of the solution of image processing system based on the embedded platform such as the necessity of using embedded platform, design of system framework, layout of development flow and how to select the embedded processor and operating system. These problems are meaningful to the design of image processing system based on embedded platform.Thesis brings forward the design of hardware, customization of OS kernel, the development of boot loader and drivers. In the end of thesis, we discuss the algorithm of pedestrians detecting and tracking.The solution of image processing system based on embedded platform has enormous potentials in order to meet the demand of image process applications, which should be further researched and explored. At the end of this thesis, we prospect and give a future direction of the research of the solution.