Syringe Defect Detection via Trajectory Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for inspecting syringes for defects are expensive, error-prone, and time-consuming, and existing machine learning models are inadequate in distinguishing between contaminants and bubbles in syringes.
Innovation Solution
A system using machine learning techniques and heuristics to analyze images of syringes, which involves background subtraction, bounding box application, and trajectory evaluation to accurately identify defects such as contaminants, reducing false positives through agitation and tracking of object trajectories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If conventional machine learning models are used for syringe inspection, then automation is achieved, but accuracy in distinguishing contaminants from bubbles is insufficient
Solution Approach 1:
The inspection process is segmented into multiple specialized modules: background subtraction module, trajectory analysis module, and classification module. Each module handles a specific aspect of the inspection task, with the trajectory module specifically analyzing motion patterns to distinguish bubbles (which move) from contaminants (which remain stationary), thereby resolving the accuracy issue while maintaining automation.
Solution Approach 2:
The system introduces dynamic analysis by capturing multiple images over time and analyzing the trajectories of detected objects. Bubbles exhibit dynamic movement behavior while contaminants remain static, allowing the system to differentiate between them through temporal analysis rather than relying solely on static image classification.
2Measurement precision
If human inspection is used, then accuracy is maintained, but time consumption and manual effort increase significantly
Solution Approach 1:
The patent replaces the mechanical human inspection process with an automated computer vision system that captures images, processes them through algorithms, and makes classification decisions. This substitution eliminates manual effort and time consumption while maintaining high accuracy through sophisticated image processing and trajectory analysis capabilities.
Solution Approach 2:
The system performs self-inspection by automatically capturing images, processing them through the inspection pipeline, and generating defect classifications without requiring human intervention at any stage. The automated system serves itself by handling the entire inspection workflow from image acquisition to defect identification.
3Device complexity
If conventional inspection methods are used, then simplicity is maintained, but error rates increase
Solution Approach 1:
The system introduces trajectory analysis as an intermediary step between image capture and final classification. This intermediate analysis layer processes the motion patterns of detected objects to provide additional information that reduces classification errors, acting as a mediator that enhances reliability without making the overall system overly complex.
Data Source
AI summary
A processor receives an image of a syringe. After identifying a background and foreground of the image, where the foreground indicates pixels that may be associated with a defect, the processor subtracts the background to generate an updated image with an accentuated foreground. The processor applies a bounding box to a group of pixels in the foreground and inputs the bounding box into a classifier. The classifier outputs a label indicating whether the syringe is defective.