| Detection moving objects in serial images is the mostly research field in computer vision and digital image processing. The subject focuses on the algorithm design and hardware implement of an embedded digital video monitoring system based on DM642 technique.In this paper some common approaches in the field of detecting moving objects are introduced. By analyzing the advantages and weaknesses of each method, our algorithm of background difference basis edge detecting is designed based on the visual information of time and space relationships. The benefits of this algorithm are simple, small amount of calculation, not only can get complete information on the verge of moving targets, but will not be subject to interference from external factors change, such as the lighting changes and changes in the background. According to the algorithm system software is designed that include image filtering, image enhancement, target detection, and several other modules. Image filtering role is to reduce the noise interference in the image, image enhancement of the role to strengthen the image of the edge information, image filtering and image enhancement is also known as the image pre-processing technology, for their follow-up of image processing to provide high-quality image data in order to improve detection of moving targets in the implementation of efficient algorithms. Edge refinement and discrete points removal known as the image after processing technology, they can get through more accurate and better information on the edge of the image display. Then this detection algorithm is achieved on the hardware platform of DM642 as core, including the preparation of bootloader procedures, using of DDK development of video drivers, the design of the main program structure based on DSP / BIOS, process and the distribution of The process simultaneously. Finally, through using of the method, such as intrinsics, cycle expansion and so on, to optimize the operating system in order to speed up the pace. |
PDF Full Text Request