Classification method of tropical natural forest vegetation type groups in Hainan

Abstract

The invention discloses the classification method of tropical natural forest vegetation type groups in Hainan. In the invention, multi-temporal SAR data and mono-temporal less-cloud-covered optical remote sensing data are used to analyze a typical tropical forest land spectrum and scattering characteristic difference and dry and rainy season change information. A Hainan tropical natural forest vegetation type group type and a Hainan tropical natural forest growth distribution rule are combined and a remote sensing vegetation type group classification system suitable for a Hainan tropical natural forest is provided. Based on that, a Hainan tropical natural forest classification method based on machine learning is established. The method provides a solution for the remote sensing classification problem of a natural forest vegetation type group (three-level class) in the Hainan Island, and classification precision is higher than 80%.

Description

technical field [0001] The invention relates to the technical field of remote sensing, in particular to a classification method for vegetation type groups of tropical natural forests in Hainan based on multi-source and multi-temporal remote sensing data. Background technique [0002] Tropical natural forest is the forest ecosystem with the most abundant species in the world. It is mainly distributed in tropical climate regions near the equator, and plays an important role in global climate change, ecological cycle, water and soil conservation, and air purification. Therefore, it is of great significance to carry out resource investigation and monitoring of tropical natural forests. [0003] The natural forests of Hainan Island in my country are the most perfect and most abundant tropical natural forests in China. They are the core of the Hainan Island ecosystem and have high value in environmental protection. They have brought great benefits to the development of local touri...

AI Technical Summary

Problems solved by technology

However, the existing methods and technologies mostly focus on the identification of coniferous forests, broad-leaved forests, and mixed coniferous and broad-leaved forests in large-scale areas, and lack methods for remote sensing classification and identification of finer internal forest types in tropical natural forests. This is mainly due to the relatively hot and humid conditions of forests in tropical areas. The vegetation in the forest grows tall and dense, and the trees are divided into multiple layers from the canopy to the understory. The layers overlap each other, forming a very complex spatial structure and intricate. The distribution of tree species, and the fact that Hainan Island has more layers of natural forests, is located in complex mountainous areas and is often covered by rain and fog, all these pose a huge challenge to remote sensing tropical natural forest monitoring and investigation

Images