基于FPGA的多方向阈值自适应Sobel算法实现

Implementation of multidirectional threshold adaptive Sobel algorithm based on FPGA

Sobel边缘检测技术是图像处理中最基本的技术之一。传统算法速度较慢且只有垂直和水平两个方向的梯度,对于一些其他方向的边缘检测效果不明显。文章设计了一种四方向(0°,45°,90°,135°)的Sobel算法来提高边缘检测的准确性,针对传统Sobel边缘检测算法阈值设置会出现偏差的缺点,基于中值滤波,采用自适应算法提高阈值准确性。本设计采用Verilog语言编程。仿真结果表明:多方向阈值自适应Sobel处理速度快,自适应阈值合理,边缘的完整性比传统方法有显著提高。

Sobel edge detection technology is one of the most basic technologies in image processing.The traditional algorithm is slow and has only two gradients in the vertical and horizontal directions.The effect of edge detection in some other directions is not obvious.This paper designs a four-direction(0°,45°,90°,135°) Sobel algorithm to improve the accuracy of edge detection.Aiming at the defect that the threshold setting of the traditional Sobel edge detection algorithm may deviate,based on median filtering,adaptive algorithm is adopted to improve the accuracy of the threshold.The design is programmed in Verilog language.The simulation results show that the multi-directional threshold adaptive Sobel processing speed is fast,the adaptive threshold is reasonable,and the integrity of the edge is significantly improved than the traditional method.