Project

Real-time Smart Crop Management System Designing

A Wageningen crop model GECROS, after calibrating and validating against data from field experiments, will be employed to simulate the production functions considering the temporal and spatial heterogeneity of a crop, by making use of real-time sensor data. It is expected that such an integrated approach will generate a powerful tool to support farmers in making real-time decisions towards smart crop management.

Background

Farmers always need to make field management decisions for meeting multiple objectives: increasing productivity, enhancing resource use efficiency, and improving profitability by optimizing cost to benefit ratios. To determine an optimal fertilization needed to maximize yield, economic profit, or fertilizer use efficiency, production functions, e.g. the relationship between yield and fertilizer application, are established usually from field experimentation. However, experimentation for diverse environments is always laborious. Also, production functions based on field experiments are static, while field crop growth is dynamic, constantly responding to growth environmental variables and varying from year to year. Dynamic crop growth simulation models can help to simulate crop yields if they are well calibrated and validated using experimental data. However, even with the best available crop models, the prediction of crop yields under diverse environmental conditions is full of uncertainties, largely due to difficulties in accurately simulating many intermediate crop characteristics and soil water and nutrient availabilities for crop uptake. Nowadays, owing to the availability of information-based technological facilities, instant and accurate in situ data could make farming processes be data-driven.

Project description

In the context of developing smart crop management tools, this study aims to develop a methodology, in which a state-of-the-art generic crop model, GECROS, will be used as an engine. Real-time sensor technologies will be explored to complement GECROS in the system. This integrated system tool can be ultimately used to generate real-time crop production functions, upon which farmers can determine the optimum resource management strategies.

Results

    1. The tight associations of photosynthesis parameters captured in GECROS with leaf nitrogen content were established for four major crops;

    1. Various machine learning algorithms for predicting crop nitrogen parameters in GECROS in dependence on real-time hyperspectral sensor data have been established.

    Publication