RESUMEN
As forest terrain is complex, containing leaves and other obstacles, it is difficult to distinguish the signal of forest-fire smoke when using single-channel lidar. To address this difficulty, a scanning micropulse polarization lidar system is developed, and a new method to detect forest fires is proposed in this study. Based on the characteristics of the depolarization ratio of in-scene obstacles, a matrix is constructed to remove obstacle signals, which in turn reduces the misidentification rate. Artificial forest-fire tests are carried out to verify the correctness of the proposed method and the feasibility of early forest-fire detection using the scanning polarization lidar system. In the working mode, the developed polarizing lidar system can locate a forest fire within three minutes with the proposed method. The experimental results show that forest fires can be accurately detected in real time when using scanning polarization lidar.