Robotic Camera Inspection for Targeted Product Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current visual inspection techniques on production lines, including human inspections and existing automated systems, fail to effectively identify product defects, especially those that develop over time, leading to defective products entering the stream of commerce.
Innovation Solution
A robotic system equipped with a camera mounted on a robotic arm that visually inspects products before they enter the commerce stream, allowing for precise examination of surfaces and features, with the ability to reposition for closer inspection of abnormalities and create a defect database for future identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If human workers perform visual inspections, then the inspection process is simple and flexible, but defective products are missed and new product defects cannot be identified
Solution Approach 1:
The patent replaces human visual inspection with an automated robotic inspection system that uses cameras and image processing algorithms. The robot captures images of products, processes them through computer vision algorithms, and automatically identifies defects. This substitution eliminates human error and fatigue while maintaining operational simplicity through automated workflows.
Solution Approach 2:
The inspection system performs self-learning by automatically capturing images of defective products, processing them through machine learning algorithms, and updating its defect detection capabilities. The system improves its own performance over time by learning from newly identified defects without requiring external reprogramming or manual intervention.
2Measurement precision
If a robot performs detailed inspection of every surface and feature, then measurement precision improves, but inspection time increases
Solution Approach 1:
The patent divides the product inspection into multiple segments or stages. The robot first performs a preliminary inspection of entire product surfaces to identify potential defect areas. Then, it focuses detailed examination only on those specific regions where abnormalities are detected. This segmented approach maintains high measurement precision for critical areas while significantly reducing overall inspection time by avoiding exhaustive examination of all surfaces.
3Measurement precision
If the robot repositions itself for closer inspection of abnormalities, then defect detection precision improves, but productivity decreases
Solution Approach 1:
The robot performs partial repositioning only when abnormalities are detected during initial scanning. Instead of repositioning for every potential defect, the system uses image processing to preliminarily assess abnormalities and only triggers detailed close-up inspection when the abnormality threshold is exceeded. This partial action approach maintains high defect detection accuracy for genuine defects while avoiding unnecessary repositioning that would reduce productivity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of improved quality inspection includes (i) receiving a first command to capture a first image of a surface of a product, (ii) positioning, by actuating a plurality of rotatable joints, a camera at a first position that is substantially adjacent to the surface of the product, and capturing the first image of the surface of the product. The method further includes, after capturing the first image: (i) processing the first image to identify a defect in the first image and a relative location of the defect in the first image, and (ii) determining a second position of the camera in accordance with the first position of the camera and the relative location of the defect in the first image, and (iii) repositioning, by actuating the plurality of rotatable joints, the camera from the first position to the second position. The method further includes capturing the second image.