Image Region Extraction for Faster Conveyor Article Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing systems for automated recognition of articles on conveyor belts face challenges in efficiently recognizing the position and quality of multiple articles due to high computational requirements, especially when using convolutional neural networks (CNNs), particularly when CPU or GPU performance is low.
Innovation Solution
An image processing apparatus that extracts specific areas from images based on distance and height information, using a combination of image pickup apparatuses and recognition units to narrow down the area for recognition processing, thereby reducing computational load and enhancing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If recognition processing is performed on the entire image using CNN, then recognition accuracy is improved, but processing time increases and productivity decreases
Solution Approach 1:
The patent divides the image into multiple regions based on distance information from the image pickup apparatus. By segmenting the image into near, middle, and far regions, the system performs recognition processing only on relevant regions rather than the entire image, thereby reducing computational load while maintaining recognition accuracy for articles at different positions on the conveyor belt
Solution Approach 2:
The patent extracts specific regions of interest from the full image based on distance thresholds. By taking out only the portions of the image where articles are located at optimal recognition distances, the system eliminates unnecessary processing of empty or irrelevant areas, improving processing speed without sacrificing recognition accuracy
2Reliability
If recognition processing is performed on all articles conveyed by the belt conveyor, then comprehensive quality control is achieved, but processing time becomes too long for successive articles
Solution Approach 1:
The patent performs preliminary filtering by calculating distance information for all articles and identifying which ones fall within the optimal recognition distance range before performing full recognition processing. This preliminary action allows the system to pre-determine which articles require detailed analysis, reducing the number of articles that undergo computationally intensive recognition processing while ensuring all articles within the critical distance range are thoroughly inspected
Solution Approach 2:
The patent applies partial action by performing comprehensive recognition processing only on articles that meet specific distance criteria, while using simplified or no processing for articles outside this range. This selective approach ensures that quality control is maintained for all articles that could potentially be processed, while articles too far away or already clearly visible require less intensive analysis
Data Source
AI summary
An image processing apparatus that processes an image of a recognition target on a mounting surface captured by an image pickup apparatus, the image processing apparatus includes an extraction unit configured to extract an area from the image based on a first distance from the recognition target to the image pickup apparatus, a height of a designated recognition target, and a second distance from the image pickup apparatus to the mounting surface, and a recognition unit configured to perform recognition processing on the area.