摘要
为了准确检测铝合金机械铰链中存在的裂纹,提出基于图像识别的铝合金机械铰链裂纹检测方法。针对在现场获取裂纹图像时,会受到外界环境、传输过程中等多种干扰因素的影响,需要对裂纹图像进行预处理,利用灰度直方图对图像的灰度情况进行分析,将其作为依据,对图像进行分段线性拉伸,压缩裂纹图像中灰度两端区域上的噪声,在上述基础上,引入边缘检测方法,边缘检测采用一阶微分的方法进行计算,选用索伯尔算子,将2个模板进行组合形成1个新的梯度算子,将模板在图像上进行移动,分别计算不同位置所对应的的梯度值,获取存在异常区域的图像边缘,以达到铝合金机械铰链裂纹检测的目的。实验结果表明,所提方法能够快速、准确检测出铝合金机械铰链的裂纹。
A method to detect cracks in aluminum alloy mechanical hinge based on image recognition is proposed.Because the acquisition of crack image in the scene will be affected by various interference factors in external environment or transmission process,it is necessary to preprocess the crack image.At first,the gray histogram was used to analyze the gray scale of image.On this basis,the piecewise linear stretching method was applied to the image,so as to compress the noise on both ends of gray scale in crack image.In addition,the edge detection method was introduced.The edge detection was performed by the first-order differential method.Then,Sobel operator was used to combine two templates into a new gradient operator.After that,the template was moved on image.The gradient values corresponding to different positions were calculated,so that the edge of image with abnormal region was obtained.Thus,the purpose of detecting the cracks in aluminum alloy mechanical hinge was realized.Simulation results show that the proposed method can quickly and accurately detect the cracks in aluminum alloy mechanical hinge.
作者
李先冲
LI Xian-chong(Rongzhi College,Chongqing Technology and Business University,Chongqing 401320,China)
出处
《计算机仿真》
北大核心
2020年第2期481-485,共5页
Computer Simulation
关键词
图像识别
铝合金
机械
铰链
裂纹检测
Image recognition
Aluminum alloy
Mechanical
Hinge
Crack detection
