Key Technology Simulation of Equipment Anomaly Detection Based on Image Processing In Underground Mining

With the deepening of automation of underground mining management, more and more mining area are equipped with the video monitoring system and alarm systems to realize unattended operation. Hence, it is of great practical value for the safe operation of the underground mining system to process the equipment image collected by the monitoring system, detect the equipment abnormality quickly and accurately through the monitoring image, and warn the potential danger. The related algorithms of the image processing are applied to the underground mining equipment image to analyze and study the characteristics of images when the electrical equipment is abnormal and the related detection techniques and methods. At the same time, a detection method of the mining transformer oil leakage is proposed. We first use the difference method on the sample images and monitoring images to get the abnormal region, and the noise of it will be eliminated and it will be segmented. Then the H-S color histogram of the abnormal region corresponding to the sample image and the monitoring image is extracted. The oil leakage condition is detected by observing the change of the gray-brown part of the H-S color histogram. Through experiments, we find that this method can more accurately detect the oil leakage anomalies of the transformer.