A high-performance concrete bridge crack identification method and system

By acquiring multi-exposure images from drones or track-based inspection robots and combining them with advanced image processing technology, aggregate texture interference is suppressed, and microcrack features of high-performance concrete bridges are accurately extracted. This solves the problems of recognition accuracy and real-time performance in existing technologies, and enables efficient bridge crack identification and safety assessment.

CN122368620APending Publication Date: 2026-07-10XINJIANG BEIXIN ROAD & BRIDGE GRP
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
XINJIANG BEIXIN ROAD & BRIDGE GRP
Filing Date
2026-04-17
Publication Date
2026-07-10

AI Technical Summary

Technical Problem

Existing technologies for identifying cracks in high-performance concrete bridges suffer from poor anti-interference capabilities, insufficient micro-crack extraction, low cross-scene recognition accuracy, high computational complexity, and inability to detect in real time, thus failing to meet the needs for accurate early micro-crack identification and safety assessment.

Method used

Multi-exposure image sequences are acquired by drones or track-based inspection robots equipped with fixed-focus industrial cameras. By combining homography matrix registration, Laplacian pyramid fusion, and adaptive grayscale stretching, aggregate texture interference is suppressed. Lightweight texture feature extraction and dual anti-interference algorithms are used to separate noise and crack features. A lightweight multi-scale feature segmentation network is constructed to extract micro-crack features and perform sub-pixel-level localization and quantization calculations to generate a standardized inspection report.

Benefits of technology

It achieves high-precision and low-cost microcrack identification, reduces false detection and false negative rates, supports real-time detection, has adaptive capabilities across bridges and environments, and can generate inspection reports that directly support the safety assessment of bridge structures.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122368620A_ABST
    Figure CN122368620A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of image processing technology and discloses a high-performance method and system for identifying cracks in concrete bridges. It involves acquiring 3-5 levels of multi-exposure image sequences using a drone or a track inspection robot. Preprocessing steps such as registration and fusion are performed to suppress aggregate texture interference and preserve weak edge details of microcracks. Then, HPC aggregate-specific features are extracted and matched with a texture prior library to distinguish the background from suspected cracks. Noise is separated using a dual anti-interference algorithm. A lightweight multi-scale network can accurately extract microcrack features at the 0.05mm level. Combined with 7×7 template Zernike matrix sub-pixel detection, the edge localization accuracy is improved to within 0.1 pixels, reducing false detection and false negative rates, and achieving accurate identification of early-stage microcracks in HPC bridges. The constructed lightweight multi-scale feature segmentation network has four parallel branches, a simplified structure, and high computational efficiency, enabling real-time detection directly on edge devices such as drones and track inspection robots.
Need to check novelty before this filing date? Find Prior Art