2012最新基于FPGA的Sobel边缘检测应用（The latest FPGA based Sobel edge detection application in 2012）.doc

2012最新基于FPGA的Sobel边缘检测应用（The latest FPGA based Sobel edge detection application in 2012） Edge detection is the basic content of image processing and computer vision. It is an important link in image analysis and identification; It is an important feature of target detection and image segmentation. On the PC, MATLAB is a powerful tool for image processing, and how to move sophisticated algorithms to an embedded system is an urgent problem. In real-time image processing, object using Sobel operator edge detection algorithm is often used, because of the demand is higher and the algorithm to deal with the amount of data processing speed is larger, with the general software implementation will be slow, can not meet the requirement of real-time system. Using the FPGA chip, the hardware implementation can improve the image processing speed and meet the real-time requirements of the system. FPGA technology is the key to using the powerful design tools to shorten the development cycle, improve resource utilization device, provides integrated device selection (for example, in the best size make a choice between speed and design). With the progress of the FPGA chip technology, its price reduced gradually, and the performance increase rapidly, the FPGA chip for use image processing platform provides a possible [1]. 1 Sobel edge detection algorithm Sobel algorithm is a kind of more mature differential edge detection algorithm, and its calculation is simple, and can produce good detection effect, the noise have smooth effect, can provide more accurate edge direction information [2]. The Sobel edge detection operator USES two directional operators (vertical operators and horizontal operators), as shown in figure 1. Its principle is respectively with these two operators for convolution operation, get two matrices, and then calculate the corresponding position of the two matrix two root mean square, get a new matrix, is the gray value of each pixel in image matrix gradient. Finally