The 20th and 21st joint COTA International Conference of Transportation Professionals (CICTP 2020-21)

Disaster along road identification is one of the key works of disaster monitoring. Remote sensing can make up for many shortcomings such as large consumption of ground observation resources, information lag, etc., and has gradually become an important research direction of road disaster identification in a large range. This paper proposes a road area hazards identification method based on remote sensing images, analyzes the characteristics of disasters from the images, and uses the maximum likelihood method combined with the fused ZY-3 images to achieve hazards classification and identification.. Taking the S306 highway in Qinghai Province as an example, GF-1C data is used to identify road area disasters, and GF-2 data with higher resolution is used to verify the accuracy. The verification results show that in the range of single image, the accuracy of GF-1C data identifying landslide by maximum likelihood method can reach 95%. The research results will greatly reduce the cost of disaster identification and have broad application prospects.

CICTP