The Study On Image Processing And Faults Location Of Transmission Line Based On Aerial Image

In the inspection course of the helicopter transmission line, the images are animportant basis to determine transmission line’s running state. Due to the limitationof the helicopter shooting environment, the captured images exist with varyingdegrees of noise and blur, while, noise and blur of the images have a direct impacton the study of the transmission line fault location. So, the pretreatments of denois-ing, deblurring and enhancing to the original image for aerial transmission lines arean essential part.As the aerial images easily mix with Gaussian noise and impulse noise, thispaper combined median filtering and Wiener filtering methods to denoise the image, and then make a comparison with a single median filtering and Wiener filteringnoise. By the PSNR, IEF, SSIM evaluation noising performance indicators to meas-ure the quality of the final results, we can draw the conclusion that the same agegroup combining denoising method is better. As for the motion blur problem ofAerial images, this paper proposes an improved maximum entropy method toprocesse the aerial blurred images’ restoration, and make a comparison with theWiener filter, constrained least squares filtering and Lucy-Richardson filter restora-tion and other conventional recovery methods. Due to the dark image after deblurr-ed, the images’ details are not obvious, this will impact the subsequent imageprocessing, the paper enhance the blurred images, which uses two methods likeLaplace enhance and the fractional order partial differential strengthen. Afterhighlighting, the more detailed information is enhanced. In order to more accuratelyevaluate the effect of rehabilitation and enhancement, this paper uses a combinationof subjective and objective evaluation system, the objective evaluation based on ofthe first-order Markov principle of the processed image quality evaluation; thesubjective evaluation mainly based on the human eyes. The results show thatcompared with the conventional treatment methods, the improved maximumentropy image restoration method improved the images clarity,making the images’details more prominent.In the aerial inspection process, when found out insulator strings, isolation andlack of dislocation rods, wires and towers icing, when power lines sag andexception safe distance to fault, the need for accurate positioning point of failure tofind the point of failure quickly, but GPS positioning is likely to fail by a highvoltage level of complex electromagnetic effects, then the failure of the tower bymeans of signage to locate the point of failure is very effective. According to theuniqueness of identification card number of transmission line towers, the paper proposes automatic positioning and identifying fault-based system according totower signage. This paper mainly research on the extraction and recognition of thetower signage to locate the point of failure. To extract the identification signagecards, the paper based on a priori knowledge of the external shape and colorcharacteristics of the signage and K-means clustering algorithm L*a*b space. Thismethod is better. BP neural network using digital identification system to identifythe card number is a powerful supplement to fault signage location based on GPS.