Research On Infrared Image Preprocessing Method Of Zero Resistance Insulator

Porcelain insulator is widely used in transmission lines as insulation parts. Ifthere is low or zero value insulators in the insulator string, in other words, a portionof the insulation is equivalent to short circuit and correspondingly reduces theinsulator string creepage distance of the insulator string; it will greatly increase theprobability of insulator flashover. Once porcelain insulator string produce flashcollaterals, its short circuit current will through the insulators from interior andproduce the heating effect often makes the insulator steel cap burst open, take off orsteel feet burned off, thus presents insulator string, to break off string and power linewires and serious ground accident, resulting in the local power grid solution and thecollapse of the power grid crack, paralysis, the serious influence of industrial andagricultural production and the life of the people and stability. At the same time alarge number of the existence of low or zero value insulators will increase power lossand cause the loss of the unknown part of power.Therefore, it is very important to accurately detect zero value resistanceinsulators for maintaining the safe of the power system and reducing powerconsumption. Infrared thermal imaging technology is widely used in electric powersystem fault diagnosis and detection equipment with non-contact, safety performance.The feasibility of the method that infrared detection zero value insulators have beenvalidated briefly and have introduced the research of zero value insulator string ofinfrared image preprocessing method in this paper.It is inevitable to merge many kinds of noise to image when the infraredthermography taking images. The noise will cause image quality drop, features tocover up and the contrast is not obvious, directly influence the image segmentation,feature extraction and recognition for the accuracy of the procedure. The method ofimage denoising based on pixel peer groups and neighbor groups is proposed in thepaper. From the image processing results indicate that this method can effectivelyremove the noise and without blurring edges and details.In the course of taking images, it is inevitable to merge many kinds ofbackground information, such as wire, tower and so on. In order maximum limit tosegment zero value insulators from the whole image of insulator string, the thresholdselection is very important. Because of the zero value insulator distribution voltage depends on insulator resistance, so the distribution voltage of zero insulator is low,temperature is low and slight taller than environmental temperature. The generalalgorithm can not effectively segment out zero value insulator images, an infraredthermal imaging segmentation method based on improved immune genetic algorithmhas proposed in this paper. First of all, insulator gray image de-noise using themethod of adaptive median filter and mean filter. Then, using logarithmictransformation enhance gray image and improve image contrast.Finally, the otusfunction for the evaluation function improved immune genetic algorithm to zeroinsulator segmentation, segmentation results show that this method can effectivelysegment the zero insulator, so as to realize the diagnosis of zero insulator.To sum up, the image quality has enhanced by image de-noise, imagesegmentation, for the subsequent zero value of an insulator accurately identify to lay afoundation.