基于地理模型和多源遥感的林草生态综合监测方法——以日照市莒县为例

Integrated Monitoring Method for Forest and Grass EcologyBased on Geographic Models and Multi source Remote Sensing——Taking Juxian County in Rizhao City as an Example

随着生态系统保护工作的不断推进,林草生态系统成为人们关注的重点问题。分析林草生态监测现状可确定目前林草生态监测存在的问题。针对当前林草生态综合监测过程,由于遥感图像分析能力不足,造成监测结果精度不足等问题,提出基于地理模型和多源遥感的林草生态综合监测方法。采集遥感图像,完成3D标定及预处理工作。通过纹理特征提取与支持向量机算法,对林草生态环境遥感影像进行判读。应用多源遥感技术构建地理加权回归模型,实现林草地图斑监测。将图斑带入马尔科夫矩阵模型中,获取林草生态综合变化监测结果。至此,基于地理模型和多源遥感的林草生态综合监测方法设计完成。通过实验证实,此方法的遥感图像分类精度以及监测精度均优于当前方法。

Accompanying with continuous development of ecosystem protection,forest grass ecosystems have become a key issue that should be paid more attention.Analyzing present condition of forest and grass ecological monitoring can determine the problems that currently exist in forest and grass ecological monitoring.A comprehensive monitoring method for forest and grass ecology based on geographic models and multi source remote sensing has been proposed to address issues,such as insufficient remote sensing image analysis capabilities and insufficient accuracy of monitoring results in current process of comprehensive monitoring of forest and grass ecology.Remote sensing images have been collected,and 3D calibration and preprocessing work has been finished.By using texture feature extraction and support vector machine algorithm,remote sensing images for forest and grass ecological environment has been studied.By using multi source remote sensing technology,a geographically weighted regression model for monitoring forest and grass map patches has been constructed.Importing the patches into the Markov matrix model,the monitoring results of comprehensive changes in forest and grassland ecology have been obtaiined.Thus,the design of a comprehensive monitoring method for forest and grass ecology based on geographic models and multi source remote sensing has been completed.Through experiments,it has been confirmed that the remote sensing image classification accuracy and monitoring accuracy of this method are superior to the current method.