Plant Nutritional Diagnosis via Self-Calibrating Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for non-contact determination of a plant's nutritional status require a reference surface for radiometric calibration, which is cumbersome, costly, and prone to errors, limiting user accessibility and accuracy.
Innovation Solution
A method that calibrates digital images by determining relative color values for pixels, using HSB color space, histograms, and thresholding to separate background and leaves, allowing for fertilizer recommendations without special equipment or reference surfaces, utilizing commercially available smartphones and cameras.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a reference surface is attached to the image recording system for radiometric calibration, then measurement precision is improved, but device complexity and ease of operation deteriorate
Solution Approach 1:
The system uses the soil or background visible in the image itself as the reference surface for calibration, eliminating the need for external reference surfaces. The calibration is performed automatically through image analysis algorithms that identify and use the background proportion (at least 10-20% of total area) to establish radiometric relationships, making the system self-calibrating and removing complex manual calibration procedures
Solution Approach 2:
The reference surface (soil or background) is extracted from the image scene itself rather than being added as a separate component. The calibration process separates and utilizes the background portion of the image to perform radiometric calibration, removing the need for attached reference surfaces while maintaining measurement accuracy
2Measurement precision
If a reference surface is attached to the image recording system, then measurement precision is improved, but ease of operation worsens
Solution Approach 1:
The system automatically identifies and uses the soil or background in the image scene as the reference surface for calibration without requiring manual attachment or configuration by the user. The image analysis automatically performs radiometric calibration using the visible background proportion, eliminating complex manual calibration procedures and making the system easy to operate
Solution Approach 2:
The software algorithm acts as an intermediary that automatically handles the calibration process using the background visible in the image. Instead of requiring the user to manually attach and configure reference surfaces, the software mediates the calibration by automatically identifying the background portion and performing radiometric calibration, simplifying the user interface and operation
3Measurement precision
If special reference surfaces and equipment are used, then measurement precision is improved, but cost increases
Solution Approach 1:
The system replaces expensive, specialized reference surfaces and calibration equipment with the freely available soil or background already present in the field. By using the natural background as the reference surface, the system eliminates the need to purchase and maintain specialized calibration artifacts, significantly reducing costs while maintaining measurement precision through automatic image analysis
Solution Approach 2:
The image recording system is made universal by using the background/soil as the reference surface for calibration, which is available in all agricultural field conditions. This approach allows the same standard camera or smartphone to be used across different locations and conditions without requiring location-specific or condition-specific reference surfaces, reducing overall system costs while maintaining precision
Data Source
Figure 1a
Figure 1b
Figure 2
AI summary
The invention relates to a method for contactless determination of the current nutritional state of a plant population, in which a digital image composed of pixels is captured by means of an image acquisition system in at least two spectral channels, the current nutritional state is determined from the image by an image analysis, and the fertiliser recommendation and/or plant protection recommendation is derived therefrom.