Visible light-thermal infrared based multispectral multi-scale forest fire monitoring method

Abstract

The invention relates to a fire monitoring method, in particular to a visible light-thermal infrared based multispectral multi-scale forest fire monitoring method which utilizes the advantages of a large-scale satellite remote sensing monitoring means and a small-scale near ground monitoring means to invent a novel monitoring method through reasonable configuration and mutual coordination of the two means. In the invention, the spectral features as well as picture pattern and textural features of forest fire points are researched, the forest fire recognition algorithms of visible light images and thermal infrared images are organically combined, and the interference and the like on forest combustion smoke and open fire recognition, which are caused by the cloud, fog, lamplight, red substances and the like of a forest district, are eliminated, so that the visible light-thermal infrared based multispectral multi-scale forest fire monitoring method by which various factor interferences can be eliminated is invented, thus a fire alarm automatic recognition function can be achieved, the manpower cost for 24-hour manual monitoring can be greatly lowered, and the accuracy of forest fire monitoring can be effectively improved.

Description

technical field [0001] The invention relates to a fire monitoring method, in particular to a multi-spectral and multi-scale forest fire monitoring method based on visible light-thermal infrared, which combines two forest fire monitoring images of visible light and infrared to automatically identify fire and give an alarm. Background technique [0002] Forest fire is a frequent natural disaster, which causes huge loss and serious harm to forest resources, forest ecosystem and human life and property. In all aspects of forest fire disaster prevention, mitigation and relief, the most critical and core link is forest fire monitoring and alarm. Only by timely detection, alarm and appropriate fire suppression measures at the initial stage of forest fires can the fire be effectively controlled and the loss of forest fires reduced to a minimum. [0003] There are mainly four methods of forest fire monitoring at home and abroad: ground inspection, fixed-point observation of cloud pl...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is aimed at the deficiencies of the artificial 24-hour monitoring method adopted in the existing forest fire monitoring and the monitoring method using visible light monitoring with a lower false alarm rate. Spectral features, image patterns, and texture features of open flames, organically synthesize the forest fire recognition algorithms of visible light and thermal infrared images, and eliminate the interference of clouds, fog, lights, red objects, etc. in forest areas on the recognition of forest burning smoke and open flames Invented a multi-spectral and multi-scale forest fire monitoring method based on visible light-thermal infrared, which can effectively improve the accuracy of forest fire monitoring

Images