Smoke Detection In Coal Mine Based On Video Image

Coal mine fire is one of the major disasters affecting mine safafety production,which seriously threatens human health.Automatic fire detection technology based on video image has become an important means of monitoring and fire early warning,which is of great significance to the safety of coal mine work.However,the underground environment is wet and dusty,and the light beam generated by artificial patrol is extremely similar to the smoke generated in the early stage of fire,which easily causes computer misjudgment.In view of the complex environment of coal mine,this paper,studies the smoke detection method,and then judges the mine fire.For the fog of the underground roadway,the collected image will be blurred,and the down-light intensity is low and uneven.In this paper,the dark primary color dehazing enhancement algorithm is improved.The fog in the roadway is filtered first,and then the CLAHE algorithm is used to image.Enhance the brightness,restore the original appearance of the roadway,increase the contrast of the coal mine image in video surveillance,clear outline of smoke.and richer details.The moving target detection method is studied,and the moving targel basetd on the mixed Gaussian model is extracted to extract the moving target in the video sequence,and the smoke is extracted together with other moving targets.Then,the extracted motion area is color-analyzed.the color law of the undrerground smoke image is calculated,the judgment condition of the underground smoke color is established,and the moving object without the smoke color feature is further excluded,thereby segmenting the suspected smoke area.The characteristics of smoke are analyzed to further improve the accuracy of smoke detection.In this paper,the main motion direction threshold is used to filter the moving beams which are inconsistent with the direction of smoke motion.Then,the SVM smoke classifier is trained to eliminate the interference of pseudo-smoke by using the texture,irregular shape and average gradient of smoke.Through the detection and analysis of a large number of smoke and pseudo-smoke interference videos,it is shown that the proposed algorithm can effectively filter out the fog in the underground image,restore the face of the roadway,and detect the smoke in the video image timely and effectively,and has a certain anti-interference ability.