Research Of SF₆Leakage Detection Based On Infrared Video

Owing to the superior electronegative and high thermal conductivity at low temperaturefrom sulfur hexafluoride (SF6), it has widely used in high voltage electrical equipment. Butthe following leakage also became a problem that cann’t be ignored, it directly affects thenormal operation of electric power production. The traditional detection methods have someshortcomings, such as accuracy, instantaneity, detecting range and so on. Although infraredimaging SF6leakage detection apparatus has been widely used in power industry, but itsdetection distance is shorter and requiring much time and effort. Withall, this paper studiedthe SF6leakage detection based on infrared video, realized further leakage detection using thevideo processing method research.This paper has analyzed the possibility and potential advantages of SF6leakage detectionusing the similar features of SF6leakage in infrared video and smoke. It has studied themotion area detection methods for SF6leakage detection, choosed the mixed gaussianbackground model method through comparison and analysis after testing, gotten the suspectedleakage area and realized the regional optimization taking mathematical morphology method,completed the preliminary judgment of the SF6leakage.According to the suspected leakage area, using the feature that SF6uninterrupted emittedfrom the leak position, throughing the cumulative model to extract the motion accumulationcharacteristics of SF6; Taking the irregular degree as the irregular characteristics of SF6whichformed by area ratio of leakage and its minimum area bounding rectangle; Taking advantageof the characteristics of SF6leakage spread in all directions, adopting the block matchingmethod at SF6movement direction estimation, and counting the main direction proportion asa characteristic of SF6leakage main movement direction. At the same, setting the decisioninterval, completeing a preliminary classification of dynamic characteristics, using theweighted voting method for eligible classification results, after gaining weight through theexperiment, realizing a combination decision of SF6leakage features.In order to facilitate observation, this paper has designed detecting interface, and carried out SF6video leakage detection experiment, at the same time, taken the comparison andvalidation for interference which has jioned using artificial way. This paper has given out theaverage accuracy through analyzing multigroup experimental results, and compared withinfrared imaging detection methods briefly. It has showed that the SF6leakage detectionmethod based on infrared video has good instantaneity and validity.