Plant Leaf Damage Quantification Using Dual CNN Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for quantifying biotic damage on plant leaves are subjective and inefficient, particularly in real-world conditions with non-ideal imaging, leading to inaccurate and time-consuming damage assessment.

Innovation Solution

A computer-implemented method using two separately trained convolutional neural networks (CNNs) for leaf segmentation and damage quantification, employing machine learning to process leaf images and provide objective damage-per-leaf indicators, even under suboptimal imaging conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If farmers manually inspect and quantify damage on plant leaves, then they can assess damage using visual inspection, but the process is subjective and inconsistent across different farmers

Engineering Contradiction:

Improvedamage quantification accuracyVSAvoidmanual inspection dependency

Core Design Contradiction:

Measurement precisionVSExtent of automation

Solution Approach 1:

The patent replaces manual visual inspection with an automated computer-based image processing system. Farmers take photos of leaves using mobile devices, and a computer automatically analyzes the images to quantify damage, substituting human subjective assessment with objective computational analysis.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables farmers to independently perform damage assessment without requiring expert pathology knowledge. The computer automatically processes images and provides damage quantification, allowing farmers to self-assess their crop conditions accurately.

Inventive Principle:

Principle #25Self-service

2Ease of operation

If farmers take leaf images under real-world conditions, then they can capture damage information in the field, but the images are of poor quality due to non-focused smartphones and poor illumination

Engineering Contradiction:

Improvefield imaging convenienceVSAvoidimage quality

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The patent converts the limitation of using non-professional mobile device cameras into an advantage by developing image processing algorithms that are specifically optimized for poor-quality, non-controlled images. The system learns to extract meaningful damage information even from blurry or poorly lit photos taken in field conditions.

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The system performs preliminary image processing and enhancement before damage analysis. The computer applies various image processing techniques to improve image quality and extract relevant features, preparing the images for accurate damage quantification despite their initial poor quality.

Inventive Principle:

Principle #10Preliminary action

3Productivity

If classical image processing techniques are used to estimate damage, then the process can be automated, but the methods are time-consuming and less accurate

Engineering Contradiction:

Improvedamage assessment speedVSAvoiddamage quantification accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent replaces classical image processing techniques with a neural network-based deep learning system. The neural network automatically learns to identify and quantify damage patterns, providing both faster processing and higher accuracy compared to traditional computer vision methods.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system changes the approach from rule-based image processing to learning-based parameter extraction. The neural network automatically determines the most relevant image parameters and features for damage assessment, adapting to different damage types and conditions without manual programming.

Inventive Principle:

Principle #35Parameter changes

4Reliability

If farmers apply pest control measures based on subjective damage assessment, then they can respond to damage, but the efficiency is reduced due to inaccurate quantification

Engineering Contradiction:

Improvepest control effectivenessVSAvoiddamage quantity accuracy

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The system provides accurate, objective feedback on damage levels to farmers, enabling them to make informed decisions about pest control measures. The precise damage quantification allows farmers to apply control measures only when necessary and at the appropriate intensity, improving the overall effectiveness of pest management.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS12450862B2Quantifying biotic damage on plant leaves, by convolutional neural networks

Publication Date: 2025.10.21 BASF SE

AI summary

To quantify biotic damage in leaves of crop plants, a computer receives (701A) a leaf-image taken from a particular crop plant. The leaf-image shows at least one of the leaves of the particular crop plant. Using a first convolutional neural network (CNN, 262), the computer processes the leaf-image to derive a segmented leaf-image (422) being a contiguous set of pixels that show a main leaf of the particular plant completely. The first CNN has been trained by a plurality of leaf-annotated leaf-images (601A), wherein the leaf-images are annotated to identify main leaves (461). Using a second CNN (272), the computer processes the single-leaf-image by regression to obtain a damage degree (432).