摘要
针对传统边缘检测算法在提取含噪齿轮边缘过程中,存在难以有效抑制噪声和边缘不连续不清晰的问题,提出了一种融合改进数学形态学和高斯拉普拉斯(LOG)算子的齿轮边缘检测算法。首先,用改进的数学形态学边缘检测算法和改进的LOG边缘检测算法分别对原图像进行边缘检测,得到两幅边缘检测图像。其次,对两幅图像进行4层小波分解且对得到的高、低频信息赋予一定的融合规则进行融合处理。最后,利用小波逆变换重构融合图像。实验结果表明,对比单一使用LOG算子和数学形态学算法,该算法不仅能更好抑制噪声还能得到更加清晰的图像边缘。
In the process of extracting the edges of noisy gears,the traditional edge detection algorithm has the problem that it is difficult to effectively suppress the noise and the edges are discontinuous and unclear.A gear edge detection algorithm based on improved mathematical morphology and Laplacian of Gaussian(LOG)operator was proposed.Firstly,two edge detection images were obtained by the improved mathematical morphology edge detection algorithm and the improved LOG edge detection algorithm.Secondly,the two images were decomposed by four layers of wavelet,and then the high and low frequency information was fused through certain fusion rules.Finally,the fused image was reconstructed by inverse wavelet transform.Experimental results show that compared with LOG operator and mathematical morphology algorithm,the fusion algorithm can not only suppress noise better,but also get clearer image edges.
作者
杜伟
何毅斌
吴林慧
汪强
DU Wei;HE Yibin;WU Linhui;WANG Qiang(School of Mechanical and Electrical Engineering,Wuhan Institute of Technology,Wuhan 430205,China)
出处
《武汉工程大学学报》
CAS
2021年第6期675-680,共6页
Journal of Wuhan Institute of Technology
基金
化工装备强化与本质安全湖北省重点实验室开放基金(2018KA01)
湖北省科技厅重大专项(2016AAA056)
武汉工程大学研究生教育创新基金(CX2020036)。
关键词
齿轮
边缘检测
LOG算子
数学形态学
小波融合
gear
edge detection
LOG operator
mathematical morphology
wavelet fusion
