基于GSO-Canny算法的输电线路覆冰厚度检测

Detection of ice thickness of transmission line based on GSO-Canny algorithm

基于现有图像识别的覆冰厚度检测方法较少考虑覆冰形状多样性情况,提出了基于人工优化萤火虫算法-Canny算法(GSO-Canny)的边缘检测方法。针对覆冰背景复杂且覆冰形状多样的特点,利用GSO-Canny算法来分割并提取覆冰边缘,解决阈值计算复杂度高、自适应性差等问题;之后,在提取出边界轮廓的基础上通过比较覆冰前后输电线路区域像素宽度,结合实际输电线路直径求取覆冰厚度。实验表明,所提方法可对输电线路带规则覆冰和带冰棱的覆冰图像的进行有效检测,计算覆冰厚度误差在1.2%左右。

The ice thickness detection method based on the existing image recognition less considers the diversity of the ice shape.The paper proposes an edge detection method based on glowworm swarm optimization combined with Canny(GSO-Canny).Aiming at the characteristics of complex icing background and various icing shapes,GSO-Canny is used to optimize the effect of edge contour detection,and solve the problems of inaccurate threshold selection and poor adaptability.After that,on the basis of extracting the boundary contour,the icing thickness is obtained by comparing the pixel width of the transmission line area before and after icing,and combining the actual transmission line diameter.Experiments show that the method proposed in the article can effectively detect the icing images of transmission lines with regular icing and icing edges,and the calculated icing thickness error is about 1.2%.