Defect Detection Accuracy in Concrete Crack Width Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for determining the width of cracks in concrete surfaces using image data often inaccurately assess the maximum crack width due to factors like chipping or bubbles, leading to overestimation.
Innovation Solution
An information processing device and method that detect defects and determine attribute information by extracting feature amounts from a broader image region, using techniques like HOG features and SVMs, to accurately calculate the maximum crack width.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If the width of a crack is determined using luminance distribution data of a local region centered on pixels representing the crack, then the crack width can be determined automatically from image data, but the maximum crack width may be overestimated due to surface irregularities such as chipping or bubbles
Solution Approach 1:
The patent extracts only the necessary luminance distribution information from the local region around the crack, separating the relevant crack signal from irrelevant surface irregularity signals. By focusing extraction on the crack region itself rather than the entire local area, the method eliminates interference from chipping and bubbles while maintaining automated determination capability.
Solution Approach 2:
The patent applies different processing approaches to different regions: for pixels representing the crack, it uses luminance distribution verification against a database to determine crack width, while for other pixels in the local region, it uses different criteria. This localized quality approach ensures that surface irregularities in non-crack regions do not affect the crack width measurement accuracy.
2Loss of information
If the maximum crack width is calculated based on crack width determination for each pixel in the crack, then comprehensive crack information is obtained, but errors from surface irregularities propagate to give an inaccurate maximum width
Solution Approach 1:
The patent performs preliminary verification of luminance distribution characteristics before determining crack width for each pixel. By pre-establishing a database of correct luminance distribution patterns corresponding to actual crack widths, the system can pre-filter out pixels where surface irregularities have distorted the luminance pattern, preventing error propagation to the maximum width calculation.
Solution Approach 2:
The patent uses feedback from the luminance distribution verification process to identify and exclude pixels with measurement errors caused by surface irregularities. The system compares actual luminance distribution against expected patterns in the database, and when mismatches indicate chipping or bubbles rather than true crack features, those pixels are excluded from the maximum width calculation, ensuring accurate results while maintaining comprehensive crack information.
Data Source
Figure 1
Figure 2
Figure 3A~3D
AI summary
Making it possible to improve the accuracy of determinations in a technique that determines the attributes of a defect of an object appearing in an image. Detecting a defect of an object in an input image. Extracting means for extracting a feature amount pertaining to a partial image of the defect from the input image, on the basis of a result of detecting the defect. Determining an attribute of the defect using the feature amount pertaining to the partial image of the defect.