Harvester Crop-Flow Imaging for Real-Time Impurity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing harvesting machines struggle to accurately determine the degree of contamination in harvested crops, particularly inorganic impurities, which can lead to financial losses and health issues due to contaminants like sand and soil components, and existing methods like near-infrared spectroscopy provide imprecise estimates.
Innovation Solution
A self-propelled harvesting machine equipped with a camera system and an image evaluation device using machine learning algorithms, specifically convolutional neural networks like EfficientNet, to analyze crop flow images and determine the proportion of inorganic impurities in real-time, allowing for precise contamination classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If near-infrared spectroscopy is used to determine crude ash content, then measurement can be performed during harvesting, but the measurement precision is insufficient and estimates are imprecise
Solution Approach 1:
The patent replaces near-infrared spectroscopy (optical/chemical method) with a camera system capturing images in the visible spectrum (optical method). The camera system with machine learning algorithm provides both real-time measurement capability and high precision by analyzing visual characteristics of contaminants, achieving accurate differentiation between crop material and inorganic impurities through image processing rather than spectral analysis.
Solution Approach 2:
The patent changes the measurement parameter from near-infrared spectral properties to visible light image properties. By capturing images in the visible spectrum and analyzing visual features such as color, texture, and shape through machine learning, the system achieves both real-time operation and high measurement precision for contaminant detection.
2Measurement precision
If laboratory incineration is used to determine crude ash content, then measurement precision is high, but the method is not feasible for real-time harvesting operations
Solution Approach 1:
The patent replaces the mechanical/thermal process of laboratory incineration with an optical imaging system and machine learning algorithm. The camera system captures images of the crop flow, and the algorithm automatically identifies and quantifies contaminants based on visual characteristics, providing real-time results without requiring physical sample collection, transportation, or laboratory processing.
Solution Approach 2:
The patent creates a visual copy (image) of the crop flow and contaminants, then analyzes this copy using machine learning algorithms. Instead of physically processing the actual crop material through incineration, the system analyzes digital images that represent the crop flow, enabling rapid real-time assessment while maintaining measurement accuracy through sophisticated image processing techniques.
3Device complexity
If image analysis with traditional algorithms is used, then device complexity is low, but the ability to accurately differentiate contaminants from crop material is insufficient
Solution Approach 1:
The patent introduces a machine learning algorithm as an intermediary between the camera system and the contaminant detection process. The trained neural network model acts as a mediator that has learned to recognize the complex visual patterns distinguishing crop material from various contaminants, enabling accurate differentiation without requiring complex hardware modifications or manual analysis procedures.
Solution Approach 2:
The patent performs preliminary training of the machine learning algorithm using labeled image data before deployment. This preliminary action creates a trained model that already contains the knowledge needed to accurately differentiate contaminants from crop material, allowing the system to achieve high measurement precision during actual harvesting operations without requiring complex real-time processing or adjustment mechanisms.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to a self-propelled harvesting machine (1), wherein the harvesting machine (1) comprises a header (4) for receiving crop (2), working units (20) for processing the received crop (2), a transfer device (15) for discharging the processed crop (2), a camera system (16) for capturing images (41) of a crop stream (21) passing through the harvesting machine (1), and an image evaluation device (27) for evaluating the images (41), wherein the image evaluation device (27) is configured to analyze the images (41) of the crop stream (21) for a proportion of, in particular inorganic, impurities contained in the crop stream (21) by means of a machine learning algorithm.