Monitoring Of Straw Burning And Its Application Of Results Based On Remote Sensing Technology

This study selects Jiutai District,Changchun City,Jilin Province as the research area.It aims to rely on data sources such as multi-spectral remote sensing images,high temperature fire data and land classification vector data,and use remote sensing to realize the location and area measurement of farmland fire areas.The person responsible for confirming the ownership of the combined land.Based on the above objectives,this study uses software such as ENVI5.3 and Arc GIS 10.2.2 to analyze the spectral characteristics of the farmland’s fired area in the original band of sentinel data and the index band,and use the separation index to screen out the dominant bands that separate the farmland’s fired area,and use principal component analysis to further Mining the information that characterizes the fired area of farmland,and finally establishes the recognition rule classification through the CART classification tree algorithm to obtain the distribution of the fired area of the farmland,and based on this,conducts an example analysis of the estimation of straw burning products;uses the maximum between-class variance method to process the MODIS remote sensing image data,supplemented by the fired area of the farmland,to obtain the fire point of straw burning Identify the data;based on the results of many years of spatial analysis,explore the temporal and spatial distribution of straw burning,and finally analyze the temporal and spatial evolution characteristics of the intensity of straw burning based on the fire point density.The relevant research results of this article can provide evidence for the measures taken by relevant government departments to address the current situation of straw burning.The main conclusions of this paper are as follows:(1)The separation index is used to evaluate the potential of the original waveband and the index waveband to extract the farmland fired area.The results show that the BAI index and the NARROW-NIR,NIR,RED-EDGE3 and other indicators in the original waveband have the highest potential value for extracting the farmland fired area.(2)Principal component analysis of the above four indicators is carried out to further dig out the characteristics of the overfire area of farmland.The experimental results show that most of the information in the original band is concentrated on PC1 and PC2.The CART classification tree algorithm is further used to generate ground object distinguishing rules,and the overfire area of farmland is extracted.The Kappa coefficients of the two phases of data reached 0.853 and 0.843 respectively,and the overall accuracy reached 89.47% and 88.22% respectively,which is a certain improvement compared with the same type of research.(3)Using the standard deviation ellipse analysis method,the fire point results extracted by the maximum between-cluster variance method were compared with the official data and found that the two have a high degree of similarity,which verified the applicability of the maximum between-cluster variance method to extract the MODIS data for straw burning fire points.(4)The results of the analysis of the temporal and spatial changes of the annual fire point show that the comprehensive Moran index at the county level is 0.03,and the local spatial autocorrelation results are not significant,indicating that most of the straw burning phenomenon in Jiutai District is due to the spontaneous behavior of farmers and is not influenced by others.Obvious aggregation characteristics;the focus of the straw burning fire point generally shows a development trend of moving to the northeast;the areas where the intensity of straw burning is increased are mainly distributed in the north of the central area,and the straw burning intensity of most administrative village units in the southern region has not changed significantly.