基于改进的Canny算子的工艺品裂纹提取方法  

Handicraft Crack Extraction Method Based on Improved Canny Operator

作　　者：张世鹏 牟莉[1] ZHANG Shipeng;MU Li(School of Computer Science,Xi'an Polytechnic University,Xi'an 710600)

机构地区：[1]西安工程大学计算机科学学院,西安710600

出　　处：《计算机与数字工程》2024年第9期2793-2797,共5页Computer & Digital Engineering

摘　　要：针对传统的Canny边缘检测对非线性噪声检测效果不明显,且边缘弱化较为严重的不足,对其进行改进并将其应用于工艺品裂纹边缘提取中。首先对采集到的工艺品缺陷图像进行预处理操作,利用伽马变换对图像进行图像整体增强,突出工艺品缺陷图像表面特征。使用双边滤波器法取代了原来Canny算子中的高斯滤波法实现了对图像边缘信号的保留,使用经过改进的Scharr算子取代了原来的Sobel法计算梯度,以及采用Otsu阈值算法取代了原来的双阈值法得到阈值。与传统方式相较,提出的改进方式可以获得更为清晰完美的图像边缘。The traditional Canny edge detection is not obvious in nonlinear noise detection and the edge weakening is serious,so it is improved and applied to handicraft crack edge extraction.Firstly,the collected handicraft defect images are preprocessed,and the whole image is enhanced by gamma transform to highlight the surface features of handicraft defect images.The two-sided filter method is used to replace the original Gaussian filter method in Canny operator to achieve the retention of image edge signal,the improved Scharr operator is used to replace the original Sobel method to calculate the gradient,and the Otsu threshold algorithm is used to replace the original double-threshold method to get the threshold.Compared with the traditional method,the proposed improved method can obtain clearer and perfect image edges.

关 键 词：伽马变换 非线性双边滤波 改进的Canny算子 改进Scharr算子 Otsu自动阈值法 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]