Fruit tree leaf nitrogen content estimation method based on Stacking ensemble learning

Abstract

The invention relates to a fruit tree leaf nitrogen content estimation method based on Stacking ensemble learning. The method comprises the following steps: adopting a brand new logic processing process design; on the basis of remote sensing image spectra of sampling positions in the period that fruit tree leaves grow mature and biochemical components are stable, combining data values of various corresponding preset spectral vegetation indexes, and nitrogen content corresponding to the sampling positions; obtaining target spectrum vegetation indexes for the analysis of preset type correlation coefficients between the indexes and the nitrogen content; accordingly combining a model training process under a preset logic, and constructing a fruit tree leaf nitrogen content estimator corresponding to a target area; in practical application, applying the fruit tree leaf nitrogen content estimator to estimate fruit tree leaf nitrogen content at target positions in a target area. According to the method, the whole design scheme realizes effective integration of remote sensing image spectral features, the limitation of single model data feature extraction is overcome, the generalization ability of the model is enhanced, and long-period and large-range observation can be formed for orchards, so that the consumption of manpower and time of a traditional chemical detection method in the prior art is reduced, and the monitoring cost is greatly saved.

Description

technical field [0001] The invention relates to a method for estimating the nitrogen content of fruit tree leaves based on Stacking integrated learning, and belongs to the technical field of monitoring biochemical components of crops by using satellite remote sensing images. Background technique [0002] Nitrogen is an essential nutrient element in the growth process of fruit trees, plays an important role in regulating the physiological and biochemical processes of fruit trees, and also affects the final quality and yield of fruits. In actual agricultural production, the application of nitrogen fertilizer often relies on manual experience. Nitrogen deficiency will affect the growth and development of fruit trees, while excessive nitrogen fertilizer will lead to fruit production reduction and environmental pollution. Quickly and accurately obtaining the Leaf Nitrogen Content (LNC) of fruit trees is of great significance for the scientific and rational application of nitrogen...

AI Technical Summary

Problems solved by technology

[0004] The effective extraction of spectral features of satellite remote sensing images is an important factor affecting the estimation model. Traditional machine learning models have limited feature extraction capabilities, while deep learning models require the support of massive samples, and their internal mechanisms are often difficult to express clearly. Therefore, there is an urgent need for a Integrate the results of multiple models and have an integrated learning model with higher estimation accuracy and stronger generalization ability

Images