电力无人机巡检系统图像处理研究

Study of Image Processing Method in Electric Unmanned Aerial Vehicle Inspection System

随着无人机技术的发展,电力巡线形成了依靠无人机巡线、人工筛选的方法完成,图像处理和识别技术的发展,让无人机自动判断故障点成为可能。canny算子检测法是一种基于二阶高斯函数的平滑处理、梯度计算、非极大值限制和阈值处理的图像检测算法。改进canny算法从分别从45°和135°方向对原图像进行梯度计算,减少原梯度计算中的孤立点,从而增加图像的连续性。另一方面,使用霍夫变换方法对电力线路进行直线识别,以使得电力线路更加清晰。试验结果表明:采用改进canny算法和霍夫变换后的方法,能滤除场景干扰,更加清晰连续地展示电力线路,进而对电力线路进行识别。

With the development of unmanned aerial vehicle technology,power line inspections have been completed by relying on unmanned aerial vehicle line inspections and manual screening methods.The development of image processing and recognition tech⁃nology has made it possible for unmanned aerial vehicles to automatically determine the point of failure.The canny operator detection method is an image detection algorithm based on second-order Gaussian function smoothing,gradient calculation,non-maximum limit and threshold processing.The improved canny algorithm performs gradient calculation on the original image from 45°and 135°direc⁃tions respectively,reducing the isolated points in the original gradient calculation,thereby increasing the continuity of the image.On the other hand,the Hough transform method is used to identify the power line to make the power line clearer.The test results show that the improved canny algorithm and the Hough transform method can filter out scene interference,display the power lines more clearly and continuously,and then identify the power lines.