| Forest fires are a major threat to the sustainable development of China’s forestry.In order to fundamentally reduce or even eliminate the occurrence of forest fires,the investigation of forest fires is an effective measure to prevent forest fires.In view of the large amount of work and the difficulty of forest fire investigation,and the incomplete and inaccurate data acquisition,the Laodutian in Huizhou,Guangdong Province and the Heshan Fendong in Guangdong Province are used as research areas,and the fire-burning grounds are taken as research objects,and the construction of the drone-based platform is adopted.Multi-spectral load data acquisition system,obtain high-resolution multi-spectral images of forest fires,preprocess the image data with Agisoft Photoscan software,and extract and interpret remote sensing information by ERDAS remote sensing image processing software,and finally propose a set of drone forest The fire detection system integrates and applies the technical solutions and work processes to obtain the damage area of the forest fire area,the damage of the damaged tree species and the forest.Combined with the image processing results of the drone,the artificial fire survey technology was used to conduct forest fire survey,and the results of the drone survey were compared to evaluate the applicability of the drone in the forest fire survey.The research results are as follows:(1)The UAV survey technology determined that the damage area of Laodutian in Huizhou,Guangdong was 10.84hm~2,and the damage area of Guangdong Heshan Fendong was 75.70hm~2.The standard area investigation method determined that the damage area of Laodutian in Huizhou,Guangdong was 11.40 hm~2,and the damage area of Guangdong Heshan Fendong was 79.60 hm~2.The victim area obtained by the drone survey technology meets the accuracy requirements.(2)The UAV survey technology determined that the damaged tree species in Laodutian,Huizhou,Guangdong Province were eucalyptus and Hemu,and the tree species affected by the Heshan Fendong in Guangdong was Pinus massoniana.The standard survey method was used to treat the damaged tree species in Laodutian,Huizhou,Guangdong,and the tree species in Heshan,Guangdong.It is a masson pine.The victim species identified by the drone survey are consistent with the victim species identified by the standard survey.(3)Through the standard survey,the comprehensive damage index of Laodutian overfire forest in Huizhou,Guangdong Province was 0.8585.The forest was divided into severely affected forests.The comprehensive damage index of the overfired forest in Heshan Fendong in Guangdong was determined to be 0.7313.The forest was divided into severely affected forests.The results of the drone survey are consistent with the standard survey results.(4)The unit area of the UAV survey in Laodutian,Huizhou,Guangdong,and Heshan,Guangdong,was 0.88 h/hm~2and 0.17 h/hm~2,respectively.The unit area of the standard survey was 2.63 h/hm~2and 0.69 h/hm~2,respectively.Using the calculation formula of efficiency,the following conclusions are drawn:When the area of the test area is the same,the efficiency of using the drone survey is higher than that of the standard survey method;with the increase of the test area,the efficiency of the drone survey and the standard survey are improved.And the efficiency of UAV surveys has increased more than the standard survey.(5)The unit cost of drone survey in Laodutian,Huizhou,Guangdong,and Heshan,Guangdong,was 3083.03 yuan/hm~2,589.56 yuan/hm~2,respectively;the unit cost of standard survey was 612.28 yuan/hm~2,603.08 yuan/hm~2.Through cost comparison,the following conclusions are drawn:the unit area cost of the standard survey is relatively stable;the unit area cost of the drone survey decreases greatly with the increase of the test area;it is preliminarily believed that when the survey area reaches 80hm~2or more,the application is not The unit area cost of the man-machine survey is lower.Through research,it is confirmed that the drone forest fire investigation system and remote sensing image digital processing technology can collect,process and interpret the relevant information of the burnt land accurately,timely and conveniently,and obtain the accuracy requirements of the forest damage area and estimate the damaged tree species.Consistent with the survey of forest damage and the standard survey,the goal of rapid and accurate access to forest damage,damaged tree species and forest damage was achieved,reducing the time of forest fire investigation and improving the efficiency and safety of forest fire investigation.New methods and ideas for forest fire disaster assessment and investigation work will be helpful to expand the applicability of UAV system platforms in forest fire investigations and other aspects,and further promote the scientific and informational forest fire investigation. |
