Inspection Apparatus Dynamic Algorithm Switching for Unprocessable Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional inspection apparatuses face difficulties in generating a new optimized image processing algorithm when an image cannot be processed with the current algorithm, especially during operation on a manufacturing line, leading to erroneous processing results.
Innovation Solution
An inspection apparatus that generates a new image processing algorithm based on acquired images, updates the teacher data set, and automatically switches to the new algorithm to improve processing results, allowing continuous inspection even with images that cannot be properly processed by the current algorithm.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the current image processing algorithm is used for inspection, then the inspection process is simple and fast, but the processing results become erroneous when encountering unprocessable images
Solution Approach 1:
The inspection apparatus automatically generates new image processing algorithms using genetic algorithms when encountering unprocessable images, without requiring manual intervention. The system self-updates its processing capabilities by creating new algorithms from acquired images and switching to them, enabling autonomous improvement of processing accuracy.
Solution Approach 2:
The system changes the parameters of the image processing algorithm dynamically. When an unprocessable image is detected, the system generates a new algorithm with different processing parameters by optimizing based on the acquired image characteristics, thereby adapting to new inspection scenarios.
2Reliability
If a new image processing algorithm is generated when encountering unprocessable images, then the processing accuracy improves, but the inspection time increases due to algorithm generation and switching
Solution Approach 1:
The system performs preliminary actions by acquiring images and generating new algorithms in advance before actual inspection failures occur. The genetic algorithm optimization runs on acquired images to pre-generate suitable processing algorithms, which can then be quickly switched to when needed, reducing the impact of time loss.
Solution Approach 2:
The inspection apparatus dynamically switches between different image processing algorithms based on the processed images. The system monitors processing results and can switch to newly generated algorithms when current ones fail, creating a dynamic adaptive system that responds to real-time inspection conditions.
3Reliability
If the inspection apparatus waits for manual generation of teacher data sets, then the algorithm optimization is thorough, but the inspection line stops and time is lost
Solution Approach 1:
The inspection apparatus automatically generates teacher data sets and optimizes algorithms without requiring manual intervention. The system acquires images, generates target images and weighted images, and automatically runs genetic algorithm optimization to create new processing algorithms, maintaining continuous inspection operation.
Solution Approach 2:
The inspection line operates continuously while the system generates new algorithms in the background. The useful action of inspection continues uninterrupted as the apparatus acquires images and generates algorithms parallel to the inspection process, eliminating stops for manual data preparation.
Data Source
AI summary
An inspection apparatus and method are provided, wherein even when an image that cannot be processed by a current image processing algorithm is input to an image processing unit while a working line is in operation, the inspection can be continued by newly generating an image processing algorithm optimized in keeping with a particular image. The apparatus includes an erroneous recognition detector, a teacher data generator and a switching unit for switching the current image processing algorithm to a new image processing algorithm generated based on an updated teacher data group. As a result, the inspection can be continued without extremely decreasing the accuracy even when an unexpected image is input to the working line.