森林类型遥感分类及变化监测研究进展

Research Progress of Remote Sensing Classification and Change Monitoring on Forest Types

森林是陆地生态系统最主要的植被类型,利用遥感技术对森林类型分类识别和动态监测对于全球碳循环研究和森林资源可持续发展具有重要意义。梳理了森林遥感分类的主要经典方法,从传统的基于像元的分类方法、面向对象方法再到新型基于红边波谱信息以及基于深度学习的分类方法,并详细介绍了现有的各种方法的应用案例及其优势。最后,提出了现阶段森林遥感分类和遥感变化监测研究中的局限性,为新形势下的森林资源动态监管提供借鉴。

Forest is one of the main vegetation type in the terrestrial ecosystem,and using remote sensing technology on discriminating and change monitoring forest types are of great significance importance for the global carbon cycle study and sustainable development of forest resources.This article reviewed the classical remotely sensed classification methods forest remote sensing classification methods,including pixel-based,object-oriented,red-edge spectral information based and deep learning methods,separately.We also introduced the details and individual advantages of these methods in the some specific applications.Finally,the limitations of the current study on forest remote sensing classification and change monitoring on forest types were indicated in order to provide reference for the dynamic supervision of forest resources under the new situation.