关键词: 多传感器方法, 计算机视觉, 四旋翼无人机, 林火监测, 定位

Abstract: In In order to improve the effectiveness of forest fire prevention and control strategy and solve the problems of forest fire monitoring, rapid fire point identification and effective location, binocular fire point positioning technology of quadrotor UAV equipped with computer vision system was proposed. Firstly, it is equipped with infrared thermal imager, camera, laser radar and other equipment to detect forest smoke and hidden fire points, and use image intelligent recognition technology for data processing to achieve accurate fire point identification; Secondly, the binocular plane image algorithm combined with the UAV's own attitude and position information is used to propose a mathematical model of fire location, and the fire location coordinates are obtained. Finally, 8 sets of experiments were designed to verify the accuracy and reliability of the UAV identification and positioning system. The results show that the maximum error of longitude is （6.20×10-6 ）o, the maximum error of latitude is （ 3.21×10-6）o, and the maximum error of elevation is 2.73×10-3 m, which can achieve accurate positioning. Therefore, it provides new methods and ideas for forest fire prevention and control.

Key words: multi-sensor method, computer vision, quadrotor unmanned aerial vehicle, forest fire monitoring, location