Design Of Real-Time Image Edge Detection System Based On FPGA

With the rapid development of modern technology,image processing has received a great deal of attention in industrial technology.The edge detection technique allows the boundary information to be retained,thus simplifying the image analysis process.The gradual increase in the amount of data processed in images,and even the fact that some data is reused several times,makes the speed and real-time nature of image processing a pressing challenge.FPGA has the advantages of parallelization,pipeline design structure and fast running speed,which can greatly reduce computing latency,so they have become more and more widely used in related fields in recent years.In order to enable the embedded software to process images in real time and to improve the processing of noisy images,this paper improves the traditional Sobel operator and designs a video image edge detection system with an FPGA chip model EP4CE10F17C8 as the core.The system uses the OV5640 camera as the front-end for video image acquisition,and caches the captured 90 fps image data to the video image processing module for the next step of image processing,which is mainly responsible for color space conversion and Sobel edge detection.In order to reduce the influence of ambient light on the results,the image after edge detection is swelled and corroded,which further enhances the details of the image.Then the edge detection image is cached to the SDRAM read and write control module,and then the image data is read from the SDRAM and the VGA display is driven by the VGA driver module.Finally,the edge detection results can be displayed on the VGA display in real time.The video image edge detection system designed in this paper provides a detailed design of the camera image acquisition module,the video image processing module,the SDRAM read/write control module and the VGA driver module.The system features improved matrix coefficients and detection directions for the traditional Sobel edge detection algorithm,and expansive erosion processing of the detected image.The processed image data is cached in offchip SDRAM and then transferred to a VGA monitor for real-time display.The results show that the improved algorithm detects clearer and more continuous image edges than the unimproved conventional algorithm,reduces the effect of noise on the results,and the implementation of threshold adaption provides a great convenience to the detection process.