| Abstract:In view of the high complexity of three types of basic faults of substation equipment, the inability to locate the spatial millimeter level regional faults, the lack of vibration signals, the incomplete fault state analysis and the low detection accuracy, a three-dimensional field monitoring method of substation spatial millimeter level regional faults based on binocular stereo vision is proposed. This method uses binocular stereo vision to establish an imaging model. After the calibration and correction of the model, the model is used to collect the three-dimensional field image of the millimeter level fault in the substation space, and the collected image is filtered to eliminate the image noise pollution and improve the follow-up monitoring effect; According to the processing results, the fault feature vector is extracted based on the improved Hu invariant moment, and it is sent to the regularized limit learning machine model for training. The output result is the final monitoring result, so as to realize the three-dimensional field monitoring of the millimeter level fault in the substation space. The experimental results show that, through the image filtering contrast test and monitoring effect contrast test of this method, it is verified that the image obtained by this method has the highest definition and quality, and the detection accuracy of the six fault states is the highest, maintained at more than 80%, and the highest is close to 100%. It improves the monitoring efficiency and accuracy of substation space fault monitoring, and has a good application prospect. |
点此下载全文
位访问者,今日访问
次