摘要
针对东莞市某钢铁线材公司圆形线材表面缺陷,其图像存在噪声导致检测效果不佳、效率偏低的问题,提出在图像采集后与检测前之间选用多种滤波算法对带有噪声的图像进行降噪处理的方法。结合MATLAB软件,对添加不同噪声类型的图像分别运用不同的图像滤波算法进行去噪处理来解决线材现场实际问题。经实验效果分析,结果表明,不同类型的噪声需要选用不同的滤波算法进行去噪处理,才能为后期的特征提取和线材分类识别提供清晰高质量的图像。结合实际环境等因素,选用对应的滤波算法去除图像中的噪声,对检测系统具有一定的实用价值。
Aiming at the problems in a certain steel material company in Dongguan city that the image noise of the surface defects of round wire rod leads to poor detection results and low efficiency, a method selecting a variety of filtering algorithms to denoise the images is put forward between image acquisition and detection. Combined with matlab, different image filtering algorithms are used to denoise images added different noise types to solve practical problems of the field of wire rod. Through the experimental results analysis, it shows that different filtering algorithms are need to choose to denoise different types of image noise, providing clear and high quality images for later feature extraction and wire classification recognition. According to the actual environment and other factors, that selecting the corresponding filtering algorithms to remove the noise in the images, has a certain practical value for detecting system.
作者
杨培义
张天彤
杨辰飞
王若飞
YANG Pei-yi;ZHANG Tian-tong;YANG Chen-fei;WANG Ruo-fei(Zhengzhou Railway Vocational & Technical College, Zhengzhou 450046)
出处
《现代计算机》
2019年第3期50-54,共5页
Modern Computer
