Regional crop yield assessment method based on a multi-source information step-by-step correction crop model

Abstract

The invention discloses a regional crop yield assessment method based on a multi-source information step-by-step correction crop model. The method comprises the steps of decomposing the crop model, obtaining multi-source information, correcting the crop model step by step, and correcting a growth period module, an LAI module and a yield module in sequence; parameters of each functional module aresubjected to the following parameter optimization process: parameter optimization, wherein the initial parameter space consists of dozens of groups of parameters; operating the crop model to obtain asimulation result corresponding to each group of parameters, establishing a cost function between the result of the current optimization function module and the correction data; searching a global optimal solution of a cost function in a result of the current parameter space, if the optimal solution does not meet an optimization termination condition, updating a parameter value of a current function module in the parameter space towards a direction jointly decided by the global optimal solution and the local optimal solution, re-calculating a simulation result, and performing judgment; and ifthe optimal solution meets the optimization termination condition, ending the correction process, and outputting the corrected optimal parameter and the optimized simulation result of the corresponding functional module.

Description

technical field [0001] The invention relates to the technical field of agricultural remote sensing, and more particularly relates to a regional crop yield estimation method based on multi-source information step-by-step correction of crop models. Background technique [0002] Since the 1960s, the technology of using crop growth models and remote sensing observation data to monitor crop growth status and estimate yield has achieved rapid development and established a relatively complete theoretical system and rich application practices. By simulating the photosynthesis, respiration, transpiration and nutrient accumulation of crops, the crop growth model can continuously simulate the crop growth state with the time step of "day", but the regional yield simulation process is often affected by the non-uniform surface environment. However, it is difficult to obtain large-scale initial data of the model, such as soil parameters and local meteorological conditions; remote sensing o...

AI Technical Summary

Problems solved by technology

In the process of remote sensing assimilation of crop models, high temporal resolution remote sensing images that can effectively reflect crop growth and development changes and phenological information often have low spatial resolution, which limits the accuracy of assimilation simulations

However, the coverage of high-spatial-resolution remote sensing images is narrow, the product time series is short, and the calculation and processing pressure is high, making it difficult to play an effective role.

[0007] 2. The technical stability of crop yield estimation in existing regions is low

However, the existing regional yield estimation technology often only focuses on how remote sensing assimilation can improve the simulation of a single intermediate variable to improve the accuracy of the final yield, while ignoring the overall accuracy of the simulation of key crop growth processes, resulting in poor repeatability of regional yield estimation technology , inconsistent accuracy within the region, etc.

[0008] 3. The existing regional crop yield estimation technology is difficult to carry out commercial regional yield estimation application

The disadvantage of low accuracy and low stability is that the existing regional crop yield estimation technology is difficult to be replicated and transplanted to other regions or applied to other crop simulation processes.

Images