A forest interannual phenology monitoring method based on multi-source remote sensing

Abstract

The invention discloses a forest interannual phenology monitoring method based on multi-source remote sensing. Firstly, all satellite remote sensing images with cloud cover lower than 80% are collected, and then the integration method of different satellite remote sensing images is corrected to improve the accuracy of different sensors. Spatial and spectral matching; Next, a modified continuous change detection and classification model is used to generate daily vegetation index curves; Finally, enhanced vegetation index, normalized difference vegetation index and surface water index are tested using logistic regression models based on daily synthetic imagery Extract optimal forest interannual SOS. The invention improves the integration method of different satellite data, increases the observation frequency; proposes the MCCDC model, takes the radiation difference into consideration and optimizes the model algorithm, shortens the calculation time while ensuring the accuracy, and finally generates daily clear cloud-free remote sensing images; introduces 3 The vegetation index estimates forest interannual SOS, and evaluates the difference of different indices in evaluating forest SOS.

Description

technical field [0001] The invention belongs to the technical field of forest phenology monitoring, and in particular relates to an interannual forest phenology monitoring method based on multi-source remote sensing. Background technique [0002] In recent decades, remote sensing has gradually become an effective means of monitoring forest phenology dynamics due to the limitations of large-scale monitoring and mapping of ground phenology observation networks. Among them, some low spatial resolution sensors are widely used in phenological information extraction due to their high temporal resolution, such as Advanced Very High Resolution Radiometer (AVHRR) and Moderate Resolution Imaging Spectrometer (MODIS) with high temporal resolution and The ability to obtain remote sensing data with a spatial resolution of 500 meters to 1100 meters has the advantage of large-scale observation. On this basis, domestic and foreign scholars have proposed many models and algorithms to estimat...

AI Technical Summary

Problems solved by technology

[0009] (2) The premise of accurate estimation by the CCDC model is the need for a large number of clear Landsat observations. If there is continuous rainy and cloudy weather, the problem of model overfitting or underfitting may occur, which is not universal, and the CCDC model Differences in radiation between different sensors are not taken into account

Therefore, the CCDC model judges each observation value pixel by pixel, and iteratively updates the model parameters, which requires a large amount of calculation and low work efficiency. For land plots with constant land cover types in the Landsat scene, the actual observation value and the model parameters are compared pixel by pixel. Model predictions are time consuming

Images