Method for predicting soil moisture by utilizing surface reflection signals and random forest regression algorithm

Abstract

The invention belongs to the technical field of soil water content prediction and relates to a method for predicting soil moisture by utilizing surface reflection signals and a random forest regression algorithm. The method comprises the following steps: training and establishing a random forest algorithm model according to obtained surface reflectance and satellite elevation angles of different soil sample areas and soil humidity data of the sample areas; obtaining the soil type, the surface reflectance and the satellite elevation angle of the to-be-detected area, inputting the surface reflectance and the satellite elevation angle of the to-be-detected area into the random forest algorithm model, and obtaining the predicted soil humidity of the to-be-detected area. According to the method, through an all-weather, wide-coverage, continuous and stable signal source provided by a satellite, and in combination with a random forest algorithm in machine learning, relatively accurate soil humidity prediction can be carried out on a large area and an unknown area lacking remote sensing data.

Description

technical field [0001] The invention belongs to the technical field of soil water content prediction, and in particular relates to a method for predicting soil water content by using surface reflection signals and a random forest regression algorithm. Background technique [0002] Soil moisture is an important basic parameter in the study of climate, hydrology, ecology and agriculture. It directly controls the transport and balance of water and heat between the land and the atmosphere. At present, remote sensing technology can obtain regional large-scale land soil moisture change information, and apply it to various fields such as land hydrology research, waterlogging and drought detection, crop growth situation assessment, and natural and ecological environment research. However, due to the need for multiple sites and high-cost monitoring equipment, it is usually impractical to continuously monitor the soil moisture content of large areas through on-site observations, and d...

AI Technical Summary

Problems solved by technology

However, due to the need for multiple sites and high-cost monitoring equipment, it is usually impractical to continuously monitor the soil moisture content of large areas through on-site observations, and due to the limitations of the surface and geographical environment, there are still a large number of soils that cannot be obtained in the field. In the unknown region of humidity, it is still a major challenge to continuously measure it on a global scale or a large scale with a certain temporal and spatial resolution.

Although traditional measurement methods (weight-moisture method, electrical resistance method, soil moisture meter method, etc.) are very accurate, there are problems such as damage to the observed object, difficulty in data assimilation between instrument types, limited temporal and spatial resolution, and high cost.

In the field of microwave remote sensing, various ground-based, airborne, space-borne, active and passive remote sensing systems such as synthetic aperture radar and microwave radiometers (such as SMOS, SMAP and AQUARIUS projects, etc.) have been applied to monitor soil moisture, but these methods There are still some limitations, such as high cost, low time resolution, etc.

Images