Design Of Image Edge Detection Algorithm Based On FPGA

Digital image processing is becoming more and more important in the fields of communication,management,remote sensing,medical industry automation,seismology,robotics,aerospace and education.When processing image data,the key to image is the information contained in the edge features.In general,edge detection technology can significantly reduce the data operation of the processing process,while the physical characteristics of the image can be saved without change.Regardless of image classification,target recognition,or semantic segmentation,scene understanding,detecting the edge of the target image is the basis of processing for further image processing.Edge detection techniques can be used to assess the quality of an object's appearance,as well as to guide the improvement of the technical route.However,edge detection is a difficult task,especially if the edges are incomplete or not closed.Therefore,the Sobel algorithm is used as the core algorithm to analyze the advantages and disadvantages of the classical first-order edge detection algorithm.For the traditional image edge detection results,the multi-directional edge extraction is weak and the edge location is coarse.A Sobel based on eight directions is proposed.The edge detection algorithm improves the traditional algorithm.The algorithm adds convolution templates in two directions of 22.5°,45°,67.5°,112.5°,135°,and 157.5°based on the original 0°and 90°directions,which can better detect and detect multiple directions edge.The weight of the template is determined according to the Euclidean distance and the direction angle from the central pixel and the neighboring pixel.The closer the distance of the central pixel is,the larger the weight is,and the better the information of different edges of the image can be detected.At the same time,the double threshold judgment method in Canny algorithm is applied to the system,and the improved algorithm is used to realize the edge feature extraction of the object.The optimized Sobel algorithm system generates more calculations due to the change of the number and size of convolution kernels.Reading to the high parallelism of the algorithm,FPGA is used as the main control chip to complete the pipeline operation by using space and time parallelism.Morever,Furthermore the parallel processing technology to design the arithmetic circuit can improve the processing speed.The overall processing flow of the system is as follows:the OV7725 module is selected as the image acquisition module,the image control module of the FPGA is initialized and arranged,the brightness image data stored in the YCbCr422 format is input to the FPGA chip,and then passed through the FPGA filter processing module and the graphic calculation processing module.The result and the threshold are judged,and finally the image features are displayed through the VGA control module.The edge detection processing platform based on FPGA has a high running speed,the image edge information extraction is complete,the contour is clear and has good continuous smoothness,and the edge line width is close to single pixel.It improves the utilization of system resources,reduces the waiting time of data operations,and can quickly acquire the characteristics of target objects,which can be applied to a variety of image processing fields..