Leveraging Artificial Intelligence to Assist Human in Monitoring Construction Site Safety
Project Description
The construction industry is one of the most hazardous industries, reporting the highest fatality rates consecutively from 2013 to 2019 in Hong Kong. Site accidents, such as collision with moving objects and falling from height, are mainly caused by the workers’ unawareness of the dynamic operations of heavy machines, and non-compliance to safety guidelines like not wearing proper personal protective equipment (PPE). Current practices mostly involve on-site safety officers to manually observe site operations, while human beings may easily overlook any potential hazards when monitoring a very large site, failing to alert and prevent accidents proactively. Exploring advanced technologies is essential to facilitate site safety monitoring. In particular, artificial intelligence (AI) has automated different applications, including surveillance video analytics using computer vision and deep learning algorithms. AI models can potentially self-learn how to analyze construction site videos, e.g. localize different workers and machines, and classify their types and activities. While human observers may lose attention over time, AI models can simultaneously monitor different site areas non-stop, providing 24/7 unbiased surveillance. Therefore, this project aims to explore different advanced technologies and develop solutions for more effective site safety monitoring.
Supervisor
CHENG Jack Chin Pang
Quota
10
Course type
UROP1000
UROP1100
UROP2100
UROP3100
UROP4100
Applicant's Roles
- Review the latest literature and solutions for construction site safety monitoring (e.g. computer vision, deep learning, Internet-of-Things (IoT) sensing)
- Develop new functionalities and ideas for site safety monitoring (to be matched with applicants' strengths and interests), including but not limited to:
o AI-driven video analytics, e.g. worker tracking, PPE detection, unauthorized entry to danger zones (Python programming needed)
o Edge computing for more efficient data processing
o Motion and proximity tracking among moving objects with motion sensors
o Unmanned aerial vehicles (UAV) for mobile analytics
- Validate the ideas on actual construction sites with a working prototype (e.g. From an idea to an executable program with a user-friendly user interface)
- Possibly join some meetings with clients or partnering companies (online/face-to-face) to understand the project requirements and devise suitable solutions
- Develop new functionalities and ideas for site safety monitoring (to be matched with applicants' strengths and interests), including but not limited to:
o AI-driven video analytics, e.g. worker tracking, PPE detection, unauthorized entry to danger zones (Python programming needed)
o Edge computing for more efficient data processing
o Motion and proximity tracking among moving objects with motion sensors
o Unmanned aerial vehicles (UAV) for mobile analytics
- Validate the ideas on actual construction sites with a working prototype (e.g. From an idea to an executable program with a user-friendly user interface)
- Possibly join some meetings with clients or partnering companies (online/face-to-face) to understand the project requirements and devise suitable solutions
Applicant's Learning Objectives
- Enhance knowledge of the latest applications of AI-related technologies in civil and construction engineering
- Obtain hands-on experience in implementing AI-related programming techniques (e.g. From data collection to algorithm design, model optimization and deployment)
- Strengthen skillset to communicate with industry clients/partners and manage their expectations
- Understand how products are conceptualized and actualized, like product design, engineering and testing.
- Experience technology transfer into commercial products
- Obtain hands-on experience in implementing AI-related programming techniques (e.g. From data collection to algorithm design, model optimization and deployment)
- Strengthen skillset to communicate with industry clients/partners and manage their expectations
- Understand how products are conceptualized and actualized, like product design, engineering and testing.
- Experience technology transfer into commercial products
Complexity of the project
Moderate