Forest Frre Monitoring Method And Application Based On Remote Sensing Data

When the deep forest fire is out of control,it spreads around the fire.In the process,a fire is formed and it also causes great damage.Forest fires destroy the natural and social values of forests and affect natural organisms and their environment.At the same time,it also brings greater losses to the social economy.Therefore,monitoring research on forest fires is of great significance.Since the development of remote sensing technology,satellite imagery has played an important role in fire monitoring.In this paper,the Bilahe forest farm in Daxing'anling is selected as the research area.And select the appropriate band parameters of the MODIS data to monitor the fire point.The background information of the bright temperature anomaly signal is constructed by the longterm sequence of MODIS data,which can remove the influence of most terrain and meteorological factors,and use the local environmental change index to confirm the non-fire point information and eliminate it.Then,the extracted single-phase data background information is added to identify the abnormal information in the spatial domain relative to the surrounding area and confirm the fire point,and combine the advantages of the two methods to improve the accuracy of the fire point algorithm.Finally,based on the fire point information,combined with the ALICE(Absolutely Llocal Index of Change of the Environment)index and the GEMI-B(Global Environment Monitoring Index-B)index,the information of the burnt area is quickly extracted.The Landsat data was monitored by a variety of spectral index indicators.The separation threshold is set by the distribution characteristics of the eight spectral indices on typical features in the study area.The algorithm of multi-index burnt area extraction is constructed to improve the recognition accuracy of burnt area.A multi-index based on extraction algorithm is constructed to improve the recognition accuracy of burnt area.The main conclusions are as follows:(1)On the basis of analyzing the advantages and disadvantages of common monitoring algorithms,combined with the advantages of RST(Robust Satellite Techniques)and fire product fire point algorithm based on background field information,the monitoring algorithm is improved and the accuracy of the algorithm is improved.At the same time,multiple threshold experiments were carried out on the Bilahe forest farm.When the ALICE value was chosen to be 2.6,the fire point monitoring results were better.(2)By changing the core parameters of the RST algorithm,three different RST algorithms are derived to obtain the ALICE value of the fire point information.The threshold is selected by the method of accuracy evaluation,and the three methods are compared with the MODIS fire product fire point algorithm.It is found that the monitoring accuracy is that: continuous date mean RST> space domain mean RST> MODIS fire product fire point algorithm> monthly mean RST.The continuous date mean RST algorithm has the highest precision and has better adaptability to the research area of this paper.(3)The paper constructs a double threshold by combining fire point information with ALICE index and GEMI-B index,and then extract the information of the burnt area through the threshold.The GEMI-B threshold is 0.24 and the ALICE threshold is 3.The method can quickly detect the location of the burnt area and most of the burnt area pixels.It is similar to the visual interpretation of MODIS image data and higher resolution Landsat image data.It shows that this method is feasible and has certain accuracy.(4)The paper analyzes the spectral characteristics of four typical features in the study area by Landsat data.The sensitivity and separation ability of each spectral index to the ground object are studied,and the threshold of each index is determined.It can separate information on features other than the burnt area,and accurate identification of the burnt area.