An Integrated Method Using a Convolutional Autoencoder, Thresholding Techniques, and a Residual Network for Anomaly Detection on Heritage Roof Surfaces DOI Creative Commons
Yongcheng Zhang, Liulin Kong, Maxwell Fordjour Antwi‐Afari

et al.

Buildings, Journal Year: 2024, Volume and Issue: 14(9), P. 2828 - 2828

Published: Sept. 8, 2024

The roofs of heritage buildings are subject to long-term degradation, resulting in poor heat insulation, regulation, and water leakage prevention. Researchers have predominantly employed feature-based traditional machine learning methods or individual deep techniques for the detection natural deterioration human-made damage on surfaces building preservation. Despite their success, balancing accuracy, efficiency, timeliness, cost remains a challenge, hindering practical application. paper proposes an integrated method that employs convolutional autoencoder, thresholding techniques, residual network automatically detect anomalies roof surfaces. Firstly, unmanned aerial vehicles (UAVs) were collect image data roofs. Subsequently, artificial intelligence (AI)-based system was developed detect, extract, classify by integrating threshold networks (ResNets). A project selected as case study. experiments demonstrate proposed approach improved accuracy efficiency when compared with single method. addresses certain limitations existing approaches, especially reliance extensive labeling. It is anticipated this will provide basis formulation repair schemes timely maintenance preventive conservation, enhancing actual benefits restoration.

Language: Английский

Eco-efficient industrial processes: Leveraging ai-powered management for reduced environmental footprint DOI Creative Commons

Sandugash Mombekova,

Sabira Akhmetova,

G.S. Shaimerdenova

et al.

E3S Web of Conferences, Journal Year: 2025, Volume and Issue: 614, P. 05005 - 05005

Published: Jan. 1, 2025

Artificial intelligence is the ability of artificial to perform actions that were previously only accessible human brain. Its algorithms work with data on basis tools tasks and solve named tasks. Each time, pioneering systems become more efficient by analyzing parameters. The creation development (AI) a complex multi-sided process based several factors. has great potential problems needs. This could include lot data, making best use manufacturing processes, weather forecasting, developing new drugs, much more. AI can significantly improve efficiency productivity various fields activity, such as industry, transportation, health care, education, industry. It also lead lower costs, fewer errors, better quality goods services. Today, scientists study interest in creating intelligence. According researchers, allows person focus creative aspects life helps free from everyday monotonous activities. On other hand, there no concern about possible negative effects loss jobs, control over technology, serious issues related autonomous systems.

Language: Английский

Citations

0
Visual defect detection for historical building preservation DOI
Mengqi Cheng, Xiaoling Zhang,

Leihua Xia

et al.

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 128376 - 128376

Published: June 1, 2025

Language: Английский

Citations

0
An Integrated Method Using a Convolutional Autoencoder, Thresholding Techniques, and a Residual Network for Anomaly Detection on Heritage Roof Surfaces DOI Creative Commons
Yongcheng Zhang, Liulin Kong, Maxwell Fordjour Antwi‐Afari

et al.

Buildings, Journal Year: 2024, Volume and Issue: 14(9), P. 2828 - 2828

Published: Sept. 8, 2024

The roofs of heritage buildings are subject to long-term degradation, resulting in poor heat insulation, regulation, and water leakage prevention. Researchers have predominantly employed feature-based traditional machine learning methods or individual deep techniques for the detection natural deterioration human-made damage on surfaces building preservation. Despite their success, balancing accuracy, efficiency, timeliness, cost remains a challenge, hindering practical application. paper proposes an integrated method that employs convolutional autoencoder, thresholding techniques, residual network automatically detect anomalies roof surfaces. Firstly, unmanned aerial vehicles (UAVs) were collect image data roofs. Subsequently, artificial intelligence (AI)-based system was developed detect, extract, classify by integrating threshold networks (ResNets). A project selected as case study. experiments demonstrate proposed approach improved accuracy efficiency when compared with single method. addresses certain limitations existing approaches, especially reliance extensive labeling. It is anticipated this will provide basis formulation repair schemes timely maintenance preventive conservation, enhancing actual benefits restoration.

Language: Английский

Citations

0