Submit

Login Login

Register Register

Home / Journals / SDHM / Online First / doi:10.32604/sdhm.2025.073124
Journal Menu
Special Issues
Table of Content

All issues

Online First

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Open Access

ARTICLE

Vision-Based Crack Detection for Wall-Climbing Robot on Building Surface

Xianghui Li1,2, Xin Fu3, Libo Pan2, Fancong Zeng1,2,*, Zhijiang Zuo1,2
1 State Key Laboratory of Precision Blasting, Jianghan University, Wuhan, 430056, China
2 School of Intelligent Manufacturing, Jianghan University, Wuhan, 430056, China
3 Dongfeng Special Component Co., Ltd., Shiyan, 442000, China
* Corresponding Author: Fancong Zeng. Email: email
(This article belongs to the Special Issue: Non-contact Sensing in Infrastructure Health Monitoring)

Structural Durability & Health Monitoring https://doi.org/10.32604/sdhm.2025.073124

Received 11 September 2025; Accepted 10 November 2025; Published online 08 December 2025

Abstract

The present study proposes an autonomous visual inspection system based on Wall-Climbing Robot (WCR), with a view to addressing the shortcomings of traditional building crack detection methods, namely their low measurement accuracy, high manual dependence and insufficient environmental adaptability. The system has been developed to construct a crack recognition model with robust illumination adaptation by fusing the improved YOLOv5s target detection algorithm with the Canny edge enhancement algorithm. The system has been realized as a lightweight deployment on an embedded device (MaixCAM). The robot platform employs a design scheme integrating a dual-chamber negative pressure adsorption mechanism and a differential drive system, which effectively addresses the key technical challenges of stable motion and real-time image acquisition on the vertical wall. Concurrently, the embedded vision processing module accomplishes efficient data parsing within dynamic environments. The experimental findings demonstrate that the system’s detection accuracy consistently maintains a range of 88.3% to 95.6% under conditions of 1000-50 lux illumination. In comparison with conventional detection methods, the recognition accuracy of various types of building cracks is enhanced by 17.3%. This study proposes a pioneering technical solution for the intelligent detection of complex building surface defects, which holds significant engineering application value.

Keywords

Crack detection; wall-climbing robot; autonomous visual inspection system; building surface; YOLOv5s; Canny edge enhancement

  • 668

    View

  • 376

    Download

  • 1

    Like

Related articles
Copyright© 2026 Tech Science Press
© 1997-2026 TSP (Henderson, USA) unless otherwise stated

Share Link