Damage Detection Apparatus Using Partial Image Omission Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately capture and record images of damage in large structures like bridges, as it is difficult to visually recognize small cracks and determine their position within a single image, leading to potential omissions in image capture.
Innovation Solution
A detection apparatus that includes a processor to determine the position of partial images relative to a wide-angle image, identify non-overlapping pairs of partial images with damaged portions, and detect image-capturing omission regions by estimating the continuation of damage outside the image frame, allowing for notification of missing image areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a wide-angle image capturing the entire subject is taken, then the position information of the damage is obtained, but the small cracks (0.1 mm width) become visually unrecognizable
Solution Approach 1:
The patent divides the image capture process into two segments: a wide-angle image for overall position coverage and multiple partial images for detailed damage inspection. The partial images are then positioned onto the wide-angle image, creating a composite view that provides both macroscopic location and microscopic damage details without requiring a single high-resolution image of the entire large area.
2Measurement precision
If a partial image with high resolution is captured to visually recognize the crack, then the damage detail is obtained, but the position information within the entire subject is lost
Solution Approach 1:
The patent merges the wide-angle image (providing position context) with multiple partial images (providing detailed damage views) into a single composite image. The partial images are positioned onto the wide-angle image at their corresponding locations, allowing the viewer to see both the overall structure and detailed damage features in one integrated visualization.
3Area of stationary object
If multiple partial images are captured to cover the entire subject, then the coverage area is improved, but image-capturing omission occurs at the boundaries between partial images
Solution Approach 1:
The patent performs preliminary overlap processing between adjacent partial images before final positioning. By detecting overlapping regions in advance and using them to verify continuous damage patterns, the system identifies and flags potential omission areas where damage might be missed at image boundaries, allowing for corrective action before final analysis.
Solution Approach 2:
The patent implements a feedback mechanism where the system detects whether damaged portions continuously extend across overlapping regions of adjacent partial images. When a discontinuity is detected in an overlapping region, the system generates an omission notification, providing feedback that triggers re-capturing of the affected area to ensure complete damage documentation.
4Loss of information
If separate recording of damage position is performed, then the position information is obtained, but the work efficiency is reduced
Solution Approach 1:
The patent automatically merges damage position information with damage detail information by positioning partial images onto the wide-angle image. This automated positioning and integration eliminates the need for separate manual recording of position data, as the position is automatically determined through image matching and coordinate transformation algorithms.
Data Source
AI summary
A detection apparatus includes a memory and a processor coupled to the memory. The processor configured to determine a position of each of a plurality of partial images relative to the wide-angle image, to extract, from among the plurality of partial images, a pair of partial images that are consecutive in an image-capturing order, that do not have an overlapping portion, and at least one of which includes an image of a damaged portion, to detect a region of the outside of the partial image to which a damaged portion is estimated to be continuous, as an image-capturing omission candidate region, to determine the image-capturing omission candidate region as an image-capturing omission region, in a case where the image-capturing omission candidate region is included in none of the other partial images, and to issue, based on the position, a notification of a position corresponding to the image-capturing omission region.