Research On Infrared Video Image Detection Method For Flammable And Explosive Gas Leakage

With the advancement of our modernization process,the demand for chemical products is increasing rapidly,and these products are mostly produced from flammable and combustible or toxic and harmful gases such as ethylene,ammonia and hydrogen sulfide.So how to carry out the safe and effective monitoring of the production,storage and transportation of such dangerous gases,so that the disposal of the leak immediately after the discovery,to avoid the occurrence of secondary hazards,as we continue to solve the problem.At present,the leakage detection technology based on infrared imaging has gradually become the development direction of gas leakage detection with its large scene,long distance and all-weather monitoring and other advantages.Through the target gas infrared spectral characteristics,people can select the appropriate filters installed on the infrared collection device,to achieve the leakage detection of target gas.Among them,the passive infrared imaging detection technology has become the mainstream direction of development because of its relatively simple structure,long-range detection capability and wide range of spectral detection.Due to the limitation of the passive infrared imaging detection technology,the relative temperature difference between the leaked gas and the background need to be at least 4k,and collected image signal of noise ratio is poor.Therefore,it is difficult to detect the leakage gas in the complex background.This article,through a large number of gas leakage infrared video data observation and research on the relevant theories,combing with the research results both at home and abroad,put forward a set of highly robustness and real-time flammable and explosive gas leakage infrared imaging detection algorithm.The method which proposed in this article can monitor the fixed scene all-weather,and targeted on the random emergence of pedestrians and vehicles,can-not avoid background cloud,wind and other complex moving targets to remove.The background of the monitoring scene is established,and the background is selectively updated and processed according to the actual situation.The detection of the gas leakage area is succeeded by threshold segmentation,connected domain traversal,feature point extraction and k-mean clustering algorithm.Finally,the image is sharpened and the gas leakage area is subjected to pseudo-color enhancement processing to achieve a better visual effect.