基于图像处理的钢材表面缺陷检测方法  

Steel surface defect detection method based on image processing

作　　者：肖玲 李国民[1] XIAO Ling;LI Guomin(College of Communication and Information Engineering,Xi’an University of Science and Technology,Xi’an 710054,China)

机构地区：[1]西安科技大学通信与信息工程学院,陕西西安710054

出　　处：《电子设计工程》2024年第21期167-171,共5页Electronic Design Engineering

摘　　要：由于钢材生产过程中因环境光的存在、光源照明不均匀等外界环境的干扰,采集到的图像容易产生光照不均匀、缺陷特征不明显以及噪声大等问题,故提出了一种基于图像处理的钢材表面缺陷检测方法。采用CLAHE算法突出缺陷特征以及双边滤波算法去除噪声;采用改进的Bottom-hat算法,增强缺陷特征与背景的对比度;采用基于LBP与支持向量机SVM相结合的方法进行钢材表面缺陷检测。结果表明,此方法对钢材表面缺陷测试集检测的准确率为81.67%,与改进之前相比提高了9.67%,能够对钢材表面缺陷进行有效检测。Due to the interference of the external environment such as the existence of ambient light and uneven illumination of the light source in the steel production process,the collected images are prone to problems such as uneven illumination,inobvious defect characteristics and large noise,a steel surface defect detection method based on image processing is proposed.The CLAHE algorithm is used to highlight the defect characteristics and the bilateral filtering algorithm is used to remove the noise.The improved Bottom⁃hat algorithm is adopted to enhance the contrast between the defect feature and the background.The method based on LBP combined with SVM is used to detect steel surface defects.The results show that the accuracy of this method for the detection of steel surface defect test set is 81.67%,which is 9.67%higher than before the improvement,and can effectively detect steel surface defects.

关 键 词：缺陷检测 图像处理 LBP算法 支持向量机 

分 类 号：TN948.6[电子电信—信号与信息处理]