Image Processing for Foreign Matter Detection in Rotating Transparent Containers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The detection of small detection targets, such as foreign matter, in transparent containers is challenging due to light reflection caused by the curvature of the container's bottom surfaces and the shapes of the medium surfaces inside the containers.
Innovation Solution
An image processing device that compares multiple images of a detection target inside a transparent container captured while rotating the container. It determines candidate regions that move in accordance with the rotation and uses both a first learning model for image information and a second learning model for chronological changes to accurately identify the presence or absence of detection targets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple images are captured while rotating the transparent container, then the detection accuracy of small detection targets is improved, but the detection complexity and processing time increase
Solution Approach 1:
The patent divides the detection process into multiple discrete image capture steps during container rotation, with each image representing a specific angular position. This segmentation allows the system to track movement of detection targets across multiple discrete frames, improving detection accuracy while managing complexity through structured processing of individual image segments
Solution Approach 2:
The patent performs preliminary image capture during container rotation before final detection analysis. By capturing multiple images in advance at different rotation positions and identifying candidate regions that move in accordance with rotation, the system prepares processed data that simplifies the final detection decision, reducing the complexity of the main detection algorithm
2Reliability
If multiple learning models are used for determination, then the detection reliability is improved, but the processing time and computational load increase
Solution Approach 1:
The patent employs two different learning models for determination: a first learning model for image information and a second learning model for chronological change information. This partial application of multiple models allows the system to leverage different computational approaches for different types of analysis, improving reliability by cross-validation while avoiding the excessive computational burden of using all possible models for every aspect of detection
Solution Approach 2:
The patent creates separate determination pathways using different learning models - one processing image information and another processing chronological change information. This copying of the determination function into parallel models allows independent optimization of each model for its specific task, improving overall reliability while managing computational load through specialized rather than universal processing
3Measurement precision
If candidate regions are tracked during rotation, then the distinction between detection targets and background is improved, but the computational complexity increases
Solution Approach 1:
The patent utilizes the dynamic rotation of the container to create movement patterns that distinguish detection targets from background elements. By tracking candidate regions across multiple images captured during rotation and identifying regions that move in accordance with the rotation, the system exploits the dynamic state of the system to simplify target-background distinction, converting a static detection problem into a dynamic tracking problem that is more easily solvable
Data Source
AI summary
An image processing device compares multiple images capturing a detection target accumulated at a bottom surface of a transparent container or a detection target accumulated at a surface at which a medium enclosed inside a transparent container contacts another medium inside the transparent container, the images being captured while rotating the transparent container, to determine candidate regions that move in a movement direction in accordance with the rotation, the image processing device determines the presence or absence of the detection target by using first determination results obtained by using a first learning model and image information for the candidate regions to determine whether or not the candidate regions are the detection target, and second determination results obtained by using a second learning model and information indicating a chronological change in the candidate regions to determine whether or not the candidate regions are the detection target.