基于深度学习目标检测的电力设备锈蚀缺陷检测算法

Corrosion Defect Detection Algorithms for Electric Power Equipment Based on Deep Learning Target Detection

为提升电网运维智能化水平,充分利用电网巡检图像,提高电力设备锈蚀缺陷检测结果的可信赖性。针对无人机巡检图像质量差异性较大的实际情况,介绍了考虑背景明亮度和Gamma变换的锈蚀检测方法。该算法能够在图片背景复杂、明亮度差别大、噪声污染等情况下,自适应地进行归一化处理,有效地识别锈蚀区域。以锈蚀率、锈蚀置信度和锈蚀热力图3种锈蚀考核指标,为运维人员提供全方位的锈蚀缺陷信息,确保电力设备安全稳定运行。

In order to improve the intelligent level of power grid operation and maintenance,it is necessary to make full use of grid inspection images to improve the reliability of the detection results of rust defects of power equipment.Aiming at the actual situation that the image quality difference of drone inspection is large,the paper introduces the rust detection method considering background brightness and Gamma transformation. The algorithm can adaptively normalize the image in the case of complex background,large difference in brightness and noise pollution,and effectively identify the rust area. The article uses the three rust evaluation indicators of corrosion rate,rust confidence and rust heat map to provide all-round rust defect information for operation and maintenance personnel to ensure safe and stable operation of power equipment.