基于SSIM-Sobel与多特征融合的绝缘子缺陷识别方法

Insulator Defect Recognition Method Based on SSIM-Sobel and Multi-feature Fusion

有效检测绝缘子缺陷对于保障输电线路的安全稳定运行至关重要。提出了一种基于SSIM-Sobel与多特征融合的绝缘子缺陷分类识别方法。采用拉普拉斯锐化和伽马校正算法对输电线路绝缘子图像进行对比度增强处理,并通过SSIM-Sobel算法获取绝缘子边缘强化图像;然后提取边缘强化图像的HOG,LBP及GLCM特征,并将其融合后输入随机森林分类器进行训练,实现绝缘子破损、电弧灼伤缺陷的识别。算例结果表明,所提方法在绝缘子缺陷识别上的总体精度为92.67%,绝缘子破损缺陷识别精度为99%,可助力输电线路巡检人员开展绝缘子日常运维工作。

The effective detection of insulator defects is essential to ensure the safe and stable operation of transmission lines.Therefore,the paper proposes an insulator defect classification and recognition method based on SSIM-Sobel and multi-feature fusion.Firstly,the Laplace sharpening and gamma correction algorithms are used to enhance the contrast of the transmission line insulator images,and the SSIM-Sobel algorithm is used to obtain the insulator edge enhancement images.Then the HOG,LBP and GLCM features of the edge-enhanced image are extracted,and these three features are fused and are input into a random forest classifier for training,so as to realize the recognition of insulator breakage and arc burn defects.The example results show that the overall accuracy of the proposed method in the recognition of the insulator defects is 92.67%,and the recognition accuracy of the insulator breakage defects is 99%,helping the transmission line inspectors to do the daily operation and maintenance of the insulators.