Smart Construction Site Attendance System: AI‑Powered Contractual Attendance Management

The talk introduces a smart‑construction‑site contract attendance management system, aiming to strengthen the supervision of personnel who are responsible for contract fulfillment (履约人员). The system leverages Tencent Cloud AI technologies to achieve intelligent attendance, extending traditional attendance methods with site‑specific requirements.

What is contract attendance? It is the attendance tracking of personnel involved in contract execution on construction projects, such as project managers, chief engineers, safety officers, design representatives, and testing engineers. Accurate attendance helps prevent impersonation and ensures compliance with national highway and water‑transport engineering regulations.

Why is contract attendance needed? It provides a critical control point for contract performance, improves project quality, and supports regulatory oversight. Managing the authenticity of on‑site personnel is essential, especially in harsh construction environments.

Technology analysis

The system is built on six key pillars:

Cloud‑based deployment for low cost, easy maintenance, and data authenticity.

Artificial intelligence, including facial detection, landmark positioning, face comparison, and liveness detection.

Cloud computing platform to handle device data processing and resource sharing.

Big‑data integration for collection, storage, and analysis of attendance records, enabling decision‑making for managers and regulators.

Mobile applications (smartphone, tablet, mobile app) for on‑site, location‑aware check‑ins.

Support for smart devices such as facial‑recognition terminals and front‑end punch‑in hardware.

Main technical components

1. Face comparison and verification – Tencent Cloud’s facial similarity algorithm evaluates whether two face images belong to the same person, with a similarity threshold (e.g., 75‑80%).

2. Identity verification (人证核身) – Users submit a selfie or video that is compared with a pre‑registered photo to confirm identity, reducing manual verification effort.

3. Liveness detection – 3D facial reconstruction and feature‑point tracking ensure the subject is a live person, preventing spoofing.

4. Location‑Based Service (LBS) – Tencent Cloud’s LBS obtains GPS or network‑based coordinates, defines fixed attendance zones, and triggers alerts when a check‑in occurs outside the permitted area.

5. Additional technologies include cloud servers, SMS services, and OpenCV for preliminary mobile‑side validation.

System architecture

The solution follows a “cloud + edge” model: a PC‑based management console, Tencent Cloud services, and mobile/terminal attendance devices. It supports multiple projects, fixed‑point and mobile check‑ins, and real‑time data synchronization.

Functional overview

The platform provides facial‑recognition terminals, a mobile app for face check‑in and identity verification, a cloud management portal (attendance, approval, statistics, settings), and APIs for integration. It records attendance, validates identity, flags abnormal locations or missing key personnel, and offers real‑time alerts.

Application outlook

Facial‑recognition technology can be extended to equipment operation (e.g., tower cranes), safety supervision, and remote monitoring. Combining AI verification with LBS helps prevent impersonation and improves overall site safety. The speaker notes that the system has been deployed in Chengdu since May 2008 and continues to evolve.

Q&A highlights

• Concerns about GPS spoofing are acknowledged; no solution is foolproof, but the system mitigates risk through multi‑factor verification. • The solution is already in use on construction sites, with rapid deployment (two months for initial rollout). • Cost and worker acceptance are addressed by offering both fixed‑point terminals and mobile app check‑ins, reducing the need for extensive hardware.