| With the development of high-voltage transmission lines in China and the promotion of the strategy of power transmission from west to east and from north to south,drone inspections have been widely used.In this context,the detection technology operation of aerial images of transmission lines through image processing technology was born.This paper further studies the image denoising technology and edge detection technology,and conducts experimental analysis for the still existing problems in the edge detection of transmission line icing images.In the aspect of image denoising,a new image denoising method based on compressed sensing is studied for the problem of unclear edges after denoising.Firstly,the mathematical model of denoising and reconstruction based on compressive sensing theory is analyzed,then the maximum value function is used to improve the denoising model,and finally the smooth norm is constructed to continuously approach the original function to find the optimal solution to reconstruct the original signal and complete the image removal noise.The effect and peak signal-to-noise ratio index of the improved algorithm were compared with those of median filtering,mean filtering,and wavelet hard threshold denoising methods,and Gaussian noise and salt and pepper noise with different variances were added to verify the effectiveness of the improved algorithm.In the aspect of edge detection,a new edge detection method based on improved Canny operator is studied to realize edge detection of regular and irregular icing.Firstly,the processing effects of different traditional edge detection operators are analyzed and compared,and the Canny operator that is less affected by noise is selected for further improvement.Secondly,in view of the problem of edge discontinuity and adhesion caused by improper selection of Canny operator threshold,an improved firefly algorithm is used to obtain the optimal threshold to accurately separate the foreground and background.Finally,the edge detection of Canny operator is carried out on the basis of obtaining the binary image after threshold segmentation.The experimental analysis is carried out by using the laboratory simulated transmission line icing image and the actual image of a line in Jinzhong to verify the effectiveness and superiority of the algorithm proposed in this paper.In the design of transmission line icing detection system,on the basis of preprocessing and edge detection of icing images,the thickness of icing is calculated according to the number of pixels before and after icing of transmission lines,and four-level icing severity levels are set at the same time.The system integrates the functions of image preprocessing,icing edge detection,thickness measurement and icing severity judgment proposed in this paper.When simulating an icing environment in the laboratory,the icing thickness error is within 1.2%.It is applied to a power transmission in Jinzhong.The error on the line image is about 2%,and the calculation error is optimized compared with the current work,which proves the effectiveness and feasibility of the system. |
