Semantic Segmentation Model for Crop Residue Estimation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing vision-based systems for estimating crop residue coverage in fields suffer from inaccuracies in crop residue estimates, which can impact soil productivity and erosion control.

Innovation Solution

A computing system utilizing a machine-learned semantic segmentation model processes imagery to semantically segment pixels as residue or non-residue pixels, determining a crop residue value based on the segmentation, and adjusts operational parameters of agricultural vehicles to maintain optimal residue levels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If computer-aided image processing techniques are used to estimate crop residue coverage, then the system can provide automated vision-based estimation, but the accuracy of crop residue estimates deteriorates

Engineering Contradiction:

Improveautomated vision-based estimationVSAvoidcrop residue estimate accuracy

Core Design Contradiction:

Extent of automationVSMeasurement precision

Solution Approach 1:

The patent replaces traditional computer-aided image processing techniques with a machine learning-based vision system. The machine learning model is trained on labeled training data to automatically distinguish crop residue from soil, replacing manual or rule-based image processing methods and achieving both automation and improved accuracy

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent transforms the image processing approach by changing from deterministic computer-aided processing to probabilistic machine learning classification. The system uses trained models that have learned optimal parameters for residue detection from training data, enabling accurate automated estimation

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS10769771B2Measuring crop residue from imagery using a machine-learned semantic segmentation model

Publication Date: 2020.09.08 CNH IND CANADA

AI summary

The present disclosure provides systems and methods that measure crop residue in a field from imagery of the field. In particular, the present subject matter is directed to systems and methods that include or otherwise leverage a machine-learned semantic segmentation model to determine a crop residue parameter value for a portion of a field based at least in part on imagery of such portion of the field captured by an imaging device. For example, the imaging device can be a camera positioned in a downward-facing direction and physically coupled to a work vehicle or an implement towed by the work vehicle through the field.