Study On Remote Sensing Assessment Model And Method Of Land Ecological Quality

Based on the study of land use classification methods based on machine learning,the thesis established an evaluation index system based on land ecology and other theories,built a land ecological quality evaluation model based on the ideal point method,and conducted remote sensing evaluation models and methods of land ecological quality in Guang'an.The research provided a decision basis and technical support for the land ecological quality supervision work of the land remediation department,and the topic selection has important application value.The main achievements and innovations of the paper are as follows:1.In order to reflect the land ecological background information better,according to sustainable development,landscape ecology,land ecology and other related theories,this thesis chose 14 land ecological indicators from four aspects of ecological background,ecological stress,ecological structure and ecological benefits,which have established an indicator system for land ecological quality evaluation.2.By adopting a combination of subjective and objective evaluation index weighting method,this thesis established an improved ideal point model,and this thesis comprehensively evaluated the land ecological quality of Guang'an City in 2000,2005,2010,and 2015.3.Based on the hotspot model,spatial differentiation analysis was carried out on the land ecological quality of Guang'an City in 2000,2005,2010,and 2015.The ecological hotspots were mainly distributed in the areas along the Qujiang,Jialing River,and mountainous areas in the east.The principal component analysis method was used to extract the main controlling factors of land ecological quality in 2000,2005,2010 and 2015,and the proportion of forest land area and temperature were the most important controlling factors.4.This thesis established a land use classification method based on support vector machine model and multi-feature vectors which combined with fractal texture,color,vegetation index,water body index,bare land index,building index,etc.This thesis also carried out the transfer learning method and convolutional Neural Network on land use classification method study.In a word,the land use classification method based on machine learning is helpful to improve the efficiency of feature extraction and the accuracy of land use classification.