结合梯度差分和Otsu自适应边缘检测算法  

Adaptive edge detection algorithm based on gradient difference and Otsu

作　　者：王静[1] 唐文豪 WANG Jing;TANG Wenhao(School of Communication and Information Engineering,Xi’an University of Science and Technology,Xi’an 710054,China)

机构地区：[1]西安科技大学通信与信息工程学院,陕西西安710054

出　　处：《现代电子技术》2022年第7期41-46,共6页Modern Electronics Technique

基　　金：陕西省教育厅科学研究计划项目(19JK0528);陕西省教育厅科学研究计划项目(19JK0531);西安市科技计划(GXYD13.8);陕西省科技计划项目(2020GY-029)。

摘　　要：图像含有丰富的边缘信息,边缘检测在机器视觉研究领域至关重要。Canny算法边缘检测效果最佳,而传统Canny算法采用高斯滤波,在只考虑像素间空间位置关系的情况下容易造成边缘模糊。文中采用非线性双边滤波代替高斯滤波,滤除噪声的同时,可以有效地保持边缘信息;其次,使用改进的Sobel模板计算梯度,以突出边缘信息;针对传统的Canny边缘检测算法的阈值需要人为设定的问题,提出一种利用一阶、二阶差分梯度直方图并结合Otsu算法的自适应边缘检测算法,采用递归边界跟踪法连接边缘。所提方法不仅解决了Canny算法阈值选择的问题,也优化了Canny算法边缘提取效果。通过信噪比和连通数两种指标对实验结果进行分析,实验结果表明,相比传统Canny算法,改进算法检测出的边缘信息更完整,且边缘连接性更好。Images contain abundant edge information. Edge detection is very important in the field of machine vision research. Canny algorithm has the best edge detection effect. However,the Gaussian filter is adopted in the traditional Canny algorithm,which is prone to causing edge blur when only considering the spatial position relationship between pixels. In this paper,nonlinear bilateral filter is adopted to replace the Gaussian filter,which can effectively keep the edge information clear while filtering the noise. The improved Sobel model is used to calculate the gradient to highlight the edge information. Since the threshold value of the traditional Canny edge detection algorithm needs to be set artificially,an adaptive edge detection algorithm based on first-order and second-order differential gradient histogram and Otsu algorithm is proposed. The recursive boundary tracking method is used to connect the edges. The proposed method can not only solve the problem of Canny algorithm threshold value selection,but also optimize the effect of Canny algorithm edge extraction. The experimental results are analyzed by signal-to-noise ratio and connection number. The results show that,in comparison with the traditional Canny algorithm,the edge information detected by the improved algorithm is more complete and the edge connectivity is better.

关 键 词：边缘检测 CANNY算法 OTSU算法 自适应阈值 双边滤波 边界跟踪 边缘连接 梯度差分 

分 类 号：TN911.73-34[电子电信—通信与信息系统]