摘要
焊缝图像处理是焊缝自动跟踪的一个较为重要的环节,而图像二值化在其中起着关键性的作用。传统的Otsu阈值化算法对信噪比较低的图像分割效果不理想,本文通过灰度拉伸增大背景与前景之间的灰度值分布,结合遗传算法,通过编码、选择、交叉、变异等操作对传统的类间方差法进行优化,并将该方法应用于焊缝图像。实际焊缝图像试验证明了该方法的有效性,可以更加准确地提取出适合焊缝图像的二值化阈值,更利于后续的图像处理操作。
Weld image processing is an important part of automatic welding seam tracking where image binaryzation plays a key role.The traditional Otsu thresholding algorithm is not ideal for image segmentation with low signal-to-noise ratio.This paper uses gray-scale stretching to increase the gray-scale value distribution between the background and the foreground and combines genetic algorithms to optimize the traditional between-class variance method through coding,selection,and crossover and apply this method to the weld image.The actual weld image test results prove the effectiveness of the method,which can more accurately extract the binarization threshold suitable for the weld image and is more conducive to subsequent image processing operations.
作者
黄静
蒋泽宁
HUANG Jing;JIANG Zening(School of Information,Zhejiang Sci-Tech University,Hangzhou 310018,China)
出处
《智能计算机与应用》
2021年第9期99-102,107,共5页
Intelligent Computer and Applications
基金
浙江省重点研发计划项目(2021C01048)。
关键词
灰度拉伸
遗传算法
焊缝图像
图像二值化
grayscale stretch
genetic algorithm
weld image
image binaryzation
