A Crop Yield Estimation Method Based on Joint Learning of Deep Spatiotemporal Features

Abstract

The invention discloses a crop yield estimation method based on joint learning of deep spatio-temporal features. Obtain and preprocess the historical crop yield data and meteorological data of the region, preprocess the meteorological data to obtain meteorological parameters, and preprocess the yield data to obtain detrended yield, which are respectively used as the input and output of the deep learning model of the subsequent crop yield spatiotemporal characteristics Data; build a deep learning model of crop yield spatiotemporal characteristics, and optimize the hyperparameters of the model; take meteorological parameters as input, and take detrended yield as output to form a training set sample and then train the deep learning model of crop yield spatiotemporal characteristics to obtain model parameters , input the meteorological parameters of the crop yield to be measured into the trained model, output the estimation result, and obtain the crop yield estimation result. The invention combines time feature learning and space feature learning, and the crop yield estimation accuracy of the invention is higher and the stability is better in the research area with large spatial difference and complexity.

Description

technical field [0001] The invention relates to the field of agricultural meteorology, in particular to a crop yield estimation method based on joint learning of deep spatiotemporal features. Background technique [0002] The construction of crop yield estimation model is an important research method to quantitatively evaluate the response of crop growth to changes in meteorological resources. The impact of meteorological resources on crops has dynamic changes and cumulative effects in time series, and understanding these time series characteristics is conducive to optimizing crop production decisions; while the spatial distribution of meteorological resources has spatial heterogeneity, which leads to crop growth and yield distribution. Spatial distribution differences and affect the stability of the model. How to build a deep learning model to jointly learn the time series features and spatial features related to crop growth and meteorology is the current technical difficu...

AI Technical Summary

Problems solved by technology

The process mechanism model is difficult to apply in a large spatial scale because of its over-parameterization and data requirements. The statistical regression model is difficult to deal with the nonlinear relationship and collinearity in the data. The current machine learning method only uses the machine learning model to analyze the data. The crop yield estimation method using machine learning is lacking in the prior art, and there is even a lack of a method that can jointly use spatio-temporal feature learning and spatial feature learning for accurate estimation

Images