Automated Inspection Region Generation via Pixel Identifiability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing methods for setting inspection regions in manufacturing lines are inefficient and prone to errors, especially when dealing with varying shapes, positions, or orientations of inspection targets, and fail to accurately extract inspection regions in cases with low contrast between foreground and background.
Innovation Solution
An image processing method that calculates an identifiability value for each pixel using feature values from multiple images of known inspection results, allowing for the automatic generation of an inspection region by setting pixels with high identifiability values as the target region, thereby improving precision and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If binarization or color gamut extraction is used to automatically determine the inspection region, then the setting time is reduced, but the inspection precision deteriorates when the contrast between foreground and background is low
Solution Approach 1:
The patent performs preliminary actions by collecting multiple images of the inspection target from different positions or conditions before determining the inspection region. These pre-collected images are stored and later used to calculate identifiability values, allowing the system to automatically determine accurate inspection regions without manual setting for each new inspection case, thus reducing setting time while maintaining precision.
Solution Approach 2:
The patent introduces an intermediary element - the identifiability value calculation based on multiple pre-collected images. This intermediary mechanism serves as a bridge between the low-contrast image data and the precise inspection region determination. By using identifiability values derived from multiple images as an intermediary indicator, the system can accurately identify inspection regions even when direct contrast is low, resolving the contradiction between speed and precision.
2Measurement precision
If the inspection region is precisely set for only the target portion to prevent erroneous detection, then the inspection precision is improved, but the processing time increases due to manual setting requirements
Solution Approach 1:
The patent implements self-service by enabling the inspection region determination to be performed automatically without manual intervention. The system uses pre-collected images and identifiability value calculations to autonomously determine the inspection region, eliminating the need for manual setting while maintaining precise inspection. This self-service mechanism resolves the contradiction by achieving both precision and time efficiency.
Solution Approach 2:
The patent performs preliminary actions by collecting and storing multiple images in advance. These pre-collected images serve as a foundation for automatic inspection region determination, eliminating the need for manual setting during actual inspection. This preliminary preparation enables the system to quickly and accurately determine inspection regions automatically, resolving the time-consuming manual setting issue while maintaining precision.
3Measurement precision
If multiple images are processed to determine the inspection region, then the inspection precision is improved, but the processing throughput requirement increases and takes longer time in low-capability systems
Solution Approach 1:
The patent performs the computationally intensive processing of multiple images in advance, before actual inspection occurs. By collecting and processing multiple images to establish identifiability values during a preliminary phase, the system reduces the processing burden during high-speed inspection. This preliminary action resolves the contradiction by shifting processing demands to a lower-priority time frame, enabling both high precision and maintained throughput during production.
Data Source
AI summary
An image processing apparatus includes an identifiability value obtaining portion, an inspection region generation portion, and an image inspection portion. The identifiability value obtaining portion is configured to obtain, for each pixel address constituting an image plane, an identifiability value for identifying which of a first inspection result and a second inspection result the pixel address corresponds to. The inspection region generation portion is configured to generate an inspection region serving as a target of image processing by setting a portion of the image plane including the pixel address where the obtained identifiability value satisfies a specific condition as the inspection region. The image inspection portion is configured to perform image processing for inspection on a partial image corresponding to the inspection region among a third image obtained by imaging a third target object.