View in EDS

Automated Identification and Numbering of Brick in Orthoimages of Ancient City Walls with Settlement.

Saved in:
Bibliographic Details
Title: Automated Identification and Numbering of Brick in Orthoimages of Ancient City Walls with Settlement.
Authors: Zou, Zheng, Zhao, Xuefeng, Zhao, Peng, Qi, Fei, Li, Jiaqi, Jie, Yaqi
Source: Journal of Computing in Civil Engineering; Sep2025, Vol. 39 Issue 5, p1-16, 16p
Subject Terms: MACHINE learning, OBJECT recognition (Computer vision), DETECTION algorithms, ANCIENT cities & towns, BUILDING information modeling
Abstract: In the maintenance of ancient city walls, segmenting bricks and identifying their row and column positions can help record and track the development trends of defects, while also providing significant potential support for updating data in building information modeling (BIM)-based digital management systems. Traditional image processing faces challenges with misaligned bricks and complex surface textures. Deep learning algorithms can segment each brick accurately, but the results require further sorting due to their unordered nature. Additionally, settlement causes changes in the positions of bricks, making it challenging to automatically determine the brick layer. This paper proposes a pipeline that combines the YOLOv8 object detection algorithm with the unsupervised clustering technique DBSCAN to achieve continuous brick numbering in city wall orthoimages. Initially, a large orthoimage is divided into smaller sections using a sliding window. In each section, YOLOv8 accurately detects brick positions, while DBSCAN efficiently handles the disorder in the detection results. The sliding window technique helps overcome clustering failures caused by settlement, and the results are effectively filtered and integrated. The proposed method was tested on a section of the Forbidden City's city wall with uneven settlement, which is over 3 m high and more than 30 m long. The results demonstrate that the method can automatically recognize and number the bricks correctly. Furthermore, experimental validation on a section of the Xi'an Hanguang Gate, with no additional training, confirmed the model's generalization ability. The method automatically calculates the clustering radius, enabling accurate clustering of rows and columns of bricks, followed by correct integration of the numbering, further validating the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]
Copyright of Journal of Computing in Civil Engineering is the property of American Society of Civil Engineers and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Complementary Index
Description
Abstract:In the maintenance of ancient city walls, segmenting bricks and identifying their row and column positions can help record and track the development trends of defects, while also providing significant potential support for updating data in building information modeling (BIM)-based digital management systems. Traditional image processing faces challenges with misaligned bricks and complex surface textures. Deep learning algorithms can segment each brick accurately, but the results require further sorting due to their unordered nature. Additionally, settlement causes changes in the positions of bricks, making it challenging to automatically determine the brick layer. This paper proposes a pipeline that combines the YOLOv8 object detection algorithm with the unsupervised clustering technique DBSCAN to achieve continuous brick numbering in city wall orthoimages. Initially, a large orthoimage is divided into smaller sections using a sliding window. In each section, YOLOv8 accurately detects brick positions, while DBSCAN efficiently handles the disorder in the detection results. The sliding window technique helps overcome clustering failures caused by settlement, and the results are effectively filtered and integrated. The proposed method was tested on a section of the Forbidden City's city wall with uneven settlement, which is over 3 m high and more than 30 m long. The results demonstrate that the method can automatically recognize and number the bricks correctly. Furthermore, experimental validation on a section of the Xi'an Hanguang Gate, with no additional training, confirmed the model's generalization ability. The method automatically calculates the clustering radius, enabling accurate clustering of rows and columns of bricks, followed by correct integration of the numbering, further validating the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]
ISSN:08873801
DOI:10.1061/JCCEE5.CPENG-6672