Grain Cracking Analysis via Adaptive Neural Network Axes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image analysis methods for determining the degree of grain cracking in forage harvesters are inaccurate and complex, failing to account for fluctuations in grain size due to external influences during plant growth, which affects the corn silage processing score (CSPS) and milk yield.
Innovation Solution
A computer-implemented image analysis method using neural networks to classify image pixels into grain and non-grain components, determining the long and short main axes of grain components, and calculating the degree of grain cracking with an adaptive limit value that adjusts dynamically based on actual grain size, enabling accurate determination of the CSPS and optimizing grain processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image analysis methods are used to determine grain cracking degree, then the analysis can be performed, but the accuracy is insufficient and the system becomes complex
Solution Approach 1:
The image analysis is divided into two distinct stages: first stage for classifying image pixels into grain components and non-grain components, and second stage for determining the long and short main axes of grain components. This segmentation allows each stage to be optimized independently, improving accuracy while managing complexity through modular processing
Solution Approach 2:
The patent introduces an adaptive limit value for the short main axis length that dynamically adjusts based on actual grain size measurements. Instead of using a fixed threshold, the system learns and adapts the limit value from the data, which improves measurement precision by accounting for natural variations in grain size while maintaining a relatively simple analysis framework
2Adaptability or versatility
If fixed limit values are used for grain size classification, then the analysis is simple, but external influences on grain size are not accounted for
Solution Approach 1:
The patent changes the classification parameter from a fixed limit value to an adaptive limit value that is determined based on actual grain size measurements. The system calculates the long and short main axes of grain components and uses these measurements to dynamically set the classification threshold, allowing the system to adapt to external influences on grain size while maintaining measurement precision
Solution Approach 2:
The system performs self-calibration by using the measured grain components themselves to determine the appropriate limit value for classification. The grain size measurements feed back into the classification process, allowing the system to automatically adjust to variations in grain size without requiring external calibration data or manual intervention
Data Source
AI summary
An image analysis method and system for the computer-implemented determination of the degree of grain cracking of grains within a flow of harvested material processed by working units of a forage harvester. The flow of harvested material comprises whole grains and crushed grains as grain components and non-grain components. The at least one working unit is automatically controlled depending on the degree of grain cracking. Images of the flow of harvested material are cyclically recorded using an optical recording device and transmitted to an image analysis apparatus for evaluation. The image analysis apparatus, in a first stage classifies image pixels contained in the images into grain components and non-grain components, and in a second stage, determines a length of a long main axis and a short main axis of each classified grain component via a length-width comparison, with the first stage and the second stage performed by a neural network.