Optical Crop Assessment Using Siamese Image Similarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for optimizing combine harvester settings rely heavily on manual adjustment or supervised learning, which is time-consuming and requires extensive adaptation for different crop types, lacking efficiency and flexibility.
Innovation Solution
A system utilizing a Siamese neural network for optical assessment of crop, comparing real-time images with a reference image to determine similarity, allowing for automatic adjustment of harvester parameters without the need for extensive training on various crop types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If supervised learning methods are used for crop assessment, then measurement precision can be improved, but loss of time increases due to extensive training requirements
Solution Approach 1:
The system performs preliminary action by capturing reference images of crops under ideal harvesting conditions before actual harvesting begins. These reference images are stored and used for comparison during harvesting, eliminating the need for time-consuming supervised training while enabling accurate crop assessment through immediate image comparison
Solution Approach 2:
The system creates copies of ideal crop images (reference images) that represent optimal harvesting conditions. These copied reference images are then compared against real-time crop images using image processing algorithms, providing rapid assessment without requiring extensive training data collection and labeling
2Adaptability or versatility
If manual adjustment of harvester settings is used, then adaptability to different crop types can be improved, but productivity decreases due to operator workload
Solution Approach 1:
The system enables self-service by automatically comparing real-time crop images against stored reference images and generating real-time feedback on crop conditions. This automated image assessment allows the harvester to adapt to different crop types without manual intervention, maintaining high productivity while achieving versatile adaptability
Solution Approach 2:
The system implements continuous feedback by comparing current crop images with reference images and providing real-time information about crop properties and harvesting quality. This feedback loop enables automatic adjustment of harvesting parameters based on actual crop conditions, improving both adaptability and productivity
3Measurement precision
If extensive training data is collected for different crop types, then measurement precision improves, but device complexity increases
Solution Approach 1:
The system extracts only the essential features needed for crop assessment by comparing images against pre-stored reference images that encapsulate ideal crop characteristics. This extraction approach avoids the need for complex supervised learning models while maintaining measurement precision through targeted feature comparison
Data Source
AI summary
A system for optical assessment of crop in a harvesting machine comprising: a camera configured to record an image of the crop; and an image processing system configured to receive the supplied image from the camera and a reference image and generate an output value based on a degree of similarity between the reference image and the supplied image.