Research On Video Image Enhancement And Personnel Detection Technology In Coal Mine

The intelligentialize of video monitoring is the developing direction of monitoring technology in the production process. Using intelligent video analysis technology to automatically analyze anomalies video surveillance data could help monitor the scene to detect incidents and raise the alarm timely, so it has an important effect on the real-time management of production process. Intelligent video surveillance could avoid failing to report incidents and other problems caused by manual monitoring, and could significantly reduce the workload of monitoring personnel.Taking the staff behavior monitoring in coal mine production as the study objective, aiming at the coal mine video features, the article focuses on how to make a video image clearer and how to detect the target coal mines scene personnel accurately and efficiently. This article includes the following two aspects:(1) The enhancement of coal mine degraded image. Video quality is poor because of the impact of mine dust, artificial lighting, water vapor and other factors, it is impossible to restore severely degraded images only using the image denoising method. In this paper, a prior model of dark colors(DCP) combine with fuzzy theory can complete downhole video image’s enhancement. The latter image illumination distribution is relatively uniform, details is more clear.(2) The detection of coal miners in hazardous area. Aiming at the importance of underground miners safety in coal mine, the method would positioning downhole equipment area, and inclined roadways in which winch is running in real time to make sure there is no miners. To complete the extraction the moving area in video image and updated the image background, the article uses the consecutive frames subtraction to detect underground miners, and completes the extraction of foreground by background subtraction method, then uses the outline feature parameters and HOG + SVM method to detect the underground miners.Experiments show that the image enhancement method used in this paper is better than the histogram equalization and image enhancement techniques based on fuzzy theory. The given miners detection methods is a real-time method, and its detection identification rate is 86.7% which is higher than the recognition rate of detection method based miner helmet, so it is of great practical significance to the improvement of underground coal miners detection technology.