基于Hessian矩阵的改进EDlines输电线识别算法

Improved EDlines Transmission Line Recognition Algorithm Based on Hessian Matrix

随着电力网络的飞速发展,采用无人机搭载高清摄像头来进行输电线巡检已成常态。为提高巡检的实时性和准确性,本研究提出了一种基于Hessian矩阵的改进EDlines输电线识别算法。首先,通过伽马变换对输电线图像进行预处理,利用Hessian矩阵特征值和特征向量以求取像素点主方向和主曲率,并获得输电线主体轮廓,从而摒弃了传统方法中梯度计算锚点和像素方向的繁琐步骤。接着,在主体轮廓基础上连接锚点以得到潜在直线线段像素链,并运用随机抽样一致性(RANSAC)算法来进行线段拟合。最后,根据直线间的距离和角度,迭代拟合以得到最终的输电线。实验结果表明,该方法能应对多种复杂环境下的输电线识别任务,抗干扰能力强,误检率显著降低,为高空输电线巡检提供了可靠的技术支持,具有重要的工程应用价值。

With the rapid development of power grid,the use of unmanned aerial vehicles equipped with high-definition cameras for power line inspection has become routine.To enhance the real-time performance and accuracy of this task,this study proposes an improved EDlines transmission line recognition algorithm based on the Hessian matrix.Firstly,gamma transformation is applied to preprocess the transmission line images.Then,the main direction and principal curvature of pixels are determined using the Hessian matrix eigenvalues and eigenvectors to obtain the main contour of the transmission line.This eliminates the cumbersome steps of gradient computation and pixel direction calculation in traditional methods.Based on the main contour,anchor points are connected to form potential line segment pixel chains.The Random Sample Consensus(RANSAC)algorithm is used to fit these line segments,and the final transmission line is iteratively determined based on the distance and angle between the lines.Experimental results demonstrate that this approach is adaptable to transmission line recognition tasks in various complex backgrounds,significantly improving noise resistance and reducing false detection rates.This research has practical significance and provides more reliable technical support for high-altitude transmission line inspection.