Research On The Arithmetic And Application Of Zero Value Insulator Identification Based On Infrared Imaging

Insulators are important insulation equipment and widely used in transmissionlines, of which the porcelain insulators are superior in numbers as traditionalelement. In the long term run, the degradation rate of porcelain insulators is highestdue to their own material restriction and being affected by the internal mechanicalstress, mechanical and electrical load, air pollution and other factors. In case zeroresistance insulators are not detected and replaced timely, it will cause accident,which affects the grid security. Infrared image detection technique transforms theinfrared radiation signal of targets into electrical signal, and displays astwo-dimensional thermal images after enlargement processing, equipmentperformance can be diagnosed based on surface temperature changes of theequipment in thermal images. This method has the advantages of non-contact, highsecurity, convenient operation, which conventional detection methods cannot match,and it has been widely used in electrical equipment fault detection. In this paper, thealgorithm of detecting zero resistance insulators based on infrared thermal imagingis discussed, and system is developed. The concrete works are as follows:1. The key of detecting zero resistance insulators by infrared thermal imaging is toobtain the optimal and appropriate infrared thermal image features. At the basis oftheoretical research and experimental validation, an infrared thermal image featureselection method based on logistic regression analysis is proposed. To begin with,the infrared thermal image was denoised with median filtering and wavelet adaptivediffusion, and the contrast of the image is enhanced by means of gray stretching.Then the image is segmented to binary image with two-dimensional maximumentropy threshold. After binary image filling, the minimum enclosing rectanglearound the disc surface of insulator string was intercepted automatically to obtainrectangle image of the insulator string area, and through gray-gradient co-occurrencematrix,13texture features of rectangle area were extracted. Finally,14characteristic parameters consisting of pollution class and texture features werescreened by logistic regression analysis, with7parameters among them havingsignificant influence on the classification result. The experiment shows that theproposed selection method is simple, and can effectively screen cha racteristicparameters and eliminate redundant data. 2. This paper introduces an infrared thermal imaging on-line detection system ofzero value insulators based on WebGIS, aiming at achieving standardized operationand information management of porcelain insulator in power transmission lineinspection as well as improving the accuracy, security and ease of operation of zerovalue insulator on-line detection. The system is designed with WebGIS architecture,which based on the thin client mode realized by J2EE, taking the regionaltemperature distribution variance and temperature maximum difference of insulatorstrings, the average value of maximum temperature differences between plates andthat between plates and background as the characteristic parameters o f the zerovalues, and combining the ambient relative humidity, equivalent salt deposit densityas a total of6inputs of BP neural network, realizes the zero value insulatormap-positioning and intelligent recognition through WebGIS technology and the BPneural network algorithm. The system’s result was verified by the experiment of110kV line insulators. It is suitable for on-line detection and management oftransmission line insulators.