一种激光图像边缘检测方法的研究  

Research on A Laser Image Edge Detection Method

作　　者：张广云[1] ZHANG Guang-yun(Guangdong Polytechnic of Science and Technology,Zhuhai 644000 China)

机构地区：[1]广东科学技术职业学院,广东珠海519090

出　　处：《自动化技术与应用》2020年第5期84-87,共4页Techniques of Automation and Applications

摘　　要：针对激光图像边缘检测问题,提出了一种结合数学形态学滤波和Candy算法的边缘检测算法。首先,设计了一种激光图像处理系统,主要包括FPGA、外部存储器、链路层芯片、物理层芯片以及上位机等部分。针对传统Candy算法的不足,借助数学形态学滤波技术对图像做滤波处理,在领域的4个方向计算一阶偏导数有限差分,进而获取图像灰度值的梯度幅值和方向。根据梯度信息检测和连接边缘点。最后进行了实验研究。实验结果表明,所述激光图像边缘检测方法在保证执行效率的前提下,在去噪、连续性、定位等方面优势比较明显,能够满足激光图像处理要求。Aiming at the problem of laser image edge detection,a new edge detection algorithm based on mathematical morphological filtering and Candy algorithm is proposed.Firstly,a laser image processing system is designed,including FPGA,external memory,link layer chip,physical layer chip and host computer.Aiming at the shortcomings of traditional Candy algorithm,the image is filtered by mathematical morphological filtering technology,and the finite-difference of the first partial derivative is calculated in four directions of the field,so as to obtain the gradient amplitude and direction of the image gray value.Detect and connect edge points based on gradient information.Finally,an experimental study is carried out.Experimental results show that the laser image edge detection method has obvious advantages in denoising,continuity and positioning,etc.Under the premise of ensuring the execution efficiency,and can meet the requirements of laser image processing.

关 键 词：激光图像 边缘检测 数学形态学 CANNY算子 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]