Inspection Image Region Matching for Defect Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for comparing current and past inspection results of infrastructure are inefficient, particularly when the captured images have different ranges, making it difficult to focus on regions of interest and perform effective confirmation work.
Innovation Solution
An information processing apparatus and method that searches for and extracts regions of interest from past inspection images to be compared with current images, enabling efficient display and confirmation of inspection results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If an image captured in the widest possible range is used to improve inspection efficiency, then the inspection coverage is improved, but it becomes difficult to focus on regions of interest when comparing with past inspection results
Solution Approach 1:
The patent divides the comparison task into two levels: first, automatically extracts and compares defect regions using image processing to identify changes; second, allows manual selection and detailed comparison of specific regions of interest. This segmentation resolves the contradiction by handling broad coverage automatically while enabling focused examination when needed.
Solution Approach 2:
The patent introduces an automatic image processing system as an intermediary between the wide-range inspection images and the user. This intermediary automatically extracts defect regions, performs initial comparisons, and presents results in a structured format, enabling users to efficiently navigate and focus on specific regions of interest without manually analyzing entire wide-range images.
2Device complexity
If simple superposition of current and past inspection results is performed, then the display process is simple, but effective confirmation work cannot be performed due to different image capturing ranges
Solution Approach 1:
The patent changes the parameters of image comparison by performing automatic extraction and alignment of defect regions before display. Instead of directly superimposing entire images with different capturing ranges, the system extracts relevant defect regions, adjusts their positioning and scaling parameters, and then presents them for comparison. This maintains relatively simple display processes while significantly improving comparison accuracy.
3Measurement precision
If high-definition images are used for defect detection, then the detection precision is improved, but the time required for image processing and comparison increases
Solution Approach 1:
The patent extracts only the essential defect regions from high-definition images using automatic image processing, rather than processing and comparing entire high-definition images. By taking out only the relevant defect portions for detailed comparison while using automated analysis, the system maintains high detection precision while significantly reducing the time required for processing and comparison.
Data Source
AI summary
An information processing apparatus executes instructions to search for inspection information in which a first image in an inspection target is captured and a defect region of the first image are associated, the inspection information corresponding to a second image in which the inspection target is captured, to extract a region corresponding to the second image from the searched information corresponding to the second image, to generate information for presenting the extracted region corresponding to the second image so that the region is comparable against the second image, and to associate and to save, as inspection information, the first image and the defect region of the first image. The saved inspection information corresponding to the second image is searched for, and a partial region of the first image corresponding to the second image and a partial defect region is included in the partial region are extracted from the searched inspection information.