SYSTEM BASED ON A LIDAR WITH A COMPUTER CONTROL FOR IDENTIFICATION OF SMOKE, IN PARTICULAR FOR IDENTIFICATION OF FOREST FIRE AT EARLY STAGE

摘要

This invention relates to a method and an active system for detection and localization of early stage forest fires using lidar. In the simplest configuration the system includes a lidar and a control computer that operates the beam-scanning system and performs automatic recognition of the smoke signature in the lidar signal on the basis of a neural-network algorithm. The scanning procedure is optimized for the given topography and other characteristics of the area under surveillance. The neural network is simulated or implemented as a co-processor. To cover wider areas, several lidar stations may be linked together in a network, which allows simultaneous scanning of the suspicious areas by several neighboring lidars in order to guarantee maximum efficiency and false alarm reduction. The system allows the detection and localization of fires earlier and farther than passive systems, whose sensitivity is lower.