Font Size: a A A

The Study Of Algorithm For Fire Detection In Dangerous Chemicals Warehouse Based On Binocular Vision Technology

Posted on:2022-01-23Degree:MasterType:Thesis
Country:ChinaCandidate:J D ZhangFull Text:PDF
GTID:2491306338493504Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
The safety of the chemical industry,especially the storage safety,has always been related to the healthy and stable development of the industry.In recent years,such as the "Tianjin Port 8.12 fire and explosion accident",the "Zhangjiakou 11.28 major explosion accident" in 2018,and the "Jiangsu Xiangshui explosion accident" on March 21,2019,the storage safety accidents of hazardous chemicals occurred frequently,It shows that the storage of hazardous chemicals in China is facing serious challenges in safety management.In the safety management of hazardous chemicals,fire monitoring and early warning are the key issues,among which the most critical link is the fire flame information feature extraction and location.On the basis of full investigation of the research status at home and abroad,based on the characteristics and requirements of the fire in dangerous chemicals warehouse,binocular vision and image processing technology are used to study the detection and ranging positioning technology of fire information in the warehouse.In this paper,based on the simulated stacking fire scene in hazardous chemicals warehouse,Zhang Zhengyou’s calibration method is used to get the camera’s internal and external parameters of the binocular system,so as to realize the original image extraction and distortion correction of the fire scene,and then combined with pixel RGB & HSI color analysis,Otsu classification method and image gray fusion method,After removing the interference information from the original collected fire simulation image,the flame area map and its features are extracted,which reduces the noise interference for the subsequent binocular ranging and positioning of the flame information.BM string matching algorithm is used to match and calculate the flame image extracted from left and right cameras,generate depth information map,calculate the position of the current flame in the library,and realize ranging and positioning.In the 50 ~ 100 cm range of simulated flame detection experiment,this algorithm has a high ranging positioning accuracy,better realizes the detection of dangerous chemicals warehouse stacking fire,and provides strong technical support for warehouse fire warning.
Keywords/Search Tags:Fire detection of dangerous chemicals warehouse, Color analysis, Otsu threshold classification, Gray level fusion, Binocular vision
PDF Full Text Request
Related items