基于以太网与FPGA图像处理的边缘检测系统

Edge Detection System Based on Ethernet and FPGA Image Processing

针对传统的边缘检测系统提取物体边缘信息少、噪声多、图像显示与传输不稳定的情况,文中以FPGA硬件平台为基础提出一种结合对比度增强、多方向与自适应阈值的Sobel边缘检测算子,强度可调的形态学滤波与以太网UDP传输的实时边缘检测系统。该系统具有较强的去噪能力,能有效提取物体更多的边缘信息,通过SDRAM进行实时数据缓存,以太网UDP传输和上位机实时显示,能确保图像传输和显示的实时性和稳定性,能为后续的实时图像配准、图像融合与图像拼接奠定基础。

Aiming at the condition that traditional edge detection systems extract less object edge information,generate more noise,and result in unstable image display and transmission,this article proposed a real-time edge detection system based on FPGA hardware platform,which combined contrast enhancement,multi-directional and adaptive threshold Sobel edge detection operator,intensity adjustable morphological filtering,and Ethernet UDP transmission.The system has strong denoising capability,can effectively extract more edge information of objects,and can carry out real-time data caching through SDRAM,Ethernet UDP transmission and real-time display on upper computer,which can ensure the real-time and stability of image transmission and display,and can lay the foundation for subsequent real-time image registration,image fusion and image mosaic.