Vision Intelligence-based Construction Safety Management

한국연구재단(NRF), 기초연구실지원사업 2021.07 - 2023.02

---

This project aims to achieve safer construction sites through Vision Intelligence-based Construction Safety Management. It integrates site imagery with project information to enable proactive risk control. The core technology involves an integrated workflow utilizing deep learning for situational hazard identification and AR for safety facility installation verification. This system delivers automated safety monitoring, rapid hazard alerts, and supports data-driven risk analytics for continuous improvement.

---

[연구목표]

Enabling safer construction sites through Vision Intelligence-based Construction Safety Management that integrates site imagery with project information for proactive risk control

[연구목적]

Developing an integrated safety management workflow that combines (1) deep learning based situational hazard identification and (2) AR based verification of safety facility installation. The workflow supports automated safety monitoring, rapid hazard alerts, and data driven risk analytics and prediction

[연구 내용]

• Structuring safety planning and video acquisition logistics by defining vision intelligence targets, hazard rules, camera types, and measurement locations and timing

• Integrating project information such as daily work logs, risk assessment records, schedules, and BIM to provide context for risk identification and on site decision making

• Implementing deep learning based hazard identification functions including visual recognition and depth based assessment for dynamic risk scenarios

• Building an AR based safety facility installation verification method through 3D safety facility modeling and BIM based AR matching

• Establishing an operational loop for safety control actions including inspection and control, worker alerts, risk analytics and prediction, and continuous image collection for feedback and improvement

[연구지원 기관]

한국연구재단(NRF)

[연구기간]

Jul.2021 - Feb.2023