Predictive Yield Map Correction for Real-Time Harvester Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current yield prediction systems for agricultural harvesters face limitations due to spectral response saturation and low signal-to-noise ratios in vegetation index values, leading to reduced sensitivity and accuracy in harvester control algorithms.
Innovation Solution
A system that selects aerial images based on sufficient vegetation index variability across the image, generating a predictive map using these images to control the harvester, and incorporates in-situ yield corrections for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If aerial images are used to predict yield, then yield prediction capability is provided, but sensitivity and accuracy are reduced due to spectral response saturation and low signal-to-noise ratios
Solution Approach 1:
The system changes the parameter being measured from saturated vegetation index values to in-situ yield sensor signals. By switching from optical spectral measurement to direct yield measurement, the system avoids spectral saturation while maintaining yield prediction capability. The in-situ sensor provides a different measurement parameter that does not suffer from the same saturation limitations.
2Loss of information
If vegetation index values are used for control, then yield information is obtained, but signal-to-noise ratio is low reducing control sensitivity
Solution Approach 1:
The system replaces the optical measurement system (vegetation index sensors) with a direct yield measurement system (in-situ yield sensor). This substitution transitions from indirect optical inference to direct physical measurement of yield, eliminating the signal-to-noise ratio problems associated with vegetation index interpretation while preserving yield information availability.
3Ease of operation
If current predictive maps are used for harvester control, then basic control capability is provided, but accuracy is insufficient for precise control
Solution Approach 1:
The system implements feedback by continuously comparing in-situ yield sensor measurements with predictive map values and using this information to correct and update the predictive map in real-time. This feedback loop transforms the static, inaccurate predictive map into a dynamic, accurate control guide, simultaneously maintaining ease of operation and improving control accuracy.
Data Source
AI summary
A predictive map including predictive values of an agricultural characteristic of a worksite is obtained. An in-situ value of the agricultural characteristic is identified based on a sensor signal provided by an in-situ sensor. A corrected predictive map, including corrected predictive values of the agricultural characteristic, is generated based on the in-situ value of the agricultural characteristic. A controllable subsystem of a work machine is controlled based on the location of the work machine and the corrected predictive map.