Underground Personnel Safety Monitoring Solution_Smart Energy Solution_Solution_Xian Xitu Zhiguang Intelligent Technology Co., Lt

Smart Energy Solution

Smart Manufacturing Solution

Smart Education Solution

Solution Overview

The underground personnel safety monitoring plan takes 3D facial recognition technology as its

core and integrates a personnel positioning system. It realizes accurate identity verification of

underground staff, standardized duty management, and closed-loop management of hidden

dangers.

Aiming at common coal mine pain points such as complex underground environments, fraudulent

shift duty and difficult traceability of hidden dangers, the solution applies contactless biometric

technology. It ensures real on-duty authenticity, access verification in high-risk areas and

real-time data synchronization, greatly improving the safety production level of mines.

The plan promotes the management transformation from post-incident accountability to

proactive prevention, and fully complies with the policy requirements of the National Mine

Safety Supervision Bureau for the leading personnel underground duty system.

Underground Personnel Safety Monitoring Solution

Solution Architecture

The solution adopts a terminal-edge-cloud collaborative architecture to realize full-process

closed-loop management for underground mines.

Terminal Devices: Intrinsically safe and explosion-proof 3D facial recognition devices are deployed

in key underground areas (such as wellheads, working faces and high-risk locations) for personnel

identity verification and attendance recording. The equipment supports hybrid deployment to

cover underground supervision blind spots.

Edge Processing: The devices are integrated with an offline resume transmission function and

interfaced with the personnel positioning platform, ensuring automatic data synchronization

under unstable mine network conditions.

Cloud Platform Management: The software platform provides centralized management of

equipment, personnel permissions, regional configuration and data analysis. It supports API

integration with provincial regulatory platforms and enables intelligent early warning and report

export.

The solution architecture highlights human-machine-environment linkage. It verifies personnel

identity via 3D facial recognition and links turnstiles as well as sound-light alarm systems to issue

real-time alerts for abnormal conditions including missed check-in and overtime attendance. A

full closed-loop mechanism is formed, covering access verification, dynamic monitoring and

hidden danger rectification.

Core Strengths

Technological Leadership

3D structured light facial recognition technology features high environmental adaptability. It can
identify dark-complexioned faces without being affected by coal dust stains. With a recognition
speed of less than 0.5 seconds, the built-in liveness detection effectively defends against photo
and mask spoofing attacks.

Efficient Management

Contactless recognition reduces manual intervention, with individual passage time of less than 1
second. Automated data recording lowers management costs and improves the accuracy of
quantitative statistics for shift management.

Compliance Assurance

Compliant with policies such as the Regulations on Coal Mine Leaders’ Underground Duty and
Safety Supervision Inspection, the system provides complete logs and reports to meet regulatory
and audit requirements.

System Integration

Supports seamless integration with existing personnel positioning and early warning platforms,
breaks data silos, and realizes closed-loop management of hidden dangers (hidden danger
registration — rectification assignment — on-site review — closure).

Related Products

2D Face SDK Andorid

The 2D Face SDK is a facial algorithm development toolkit designed for the Android operating System（mobile phones, tablets, Android industrial computers, etc.）;. Built around 2D RGB images captured by mobile phone cameras, it encapsulates a full suite of algorithm capabilities including face detection, recognition,matching, liveness detection, and attribute analysis, and provides API interfaces that comply with Android development standards.

2D Face SDK for Linux

The 2D Face SDK (Software Development Kit) is a set of face-related algorithms and interfaces developed for the Linux operating system. Built around 2D images (RGB/grayscale images), it fulfills functions such as face detection, recognition and analysis. Developers can quickly integrate face capabilities into products including Linux servers, embedded devices (such as Linux industrial PCs, edge boxes), Linux desktop applications, and other products by calling the APIs provided by the SDK.

2D Face SDK for Windows

The 2D facial algorithm SDK is a standardized software development kit developed for the Windows operating system（supporting and Windows 10/11 and Windows Server）. It encapsulates core face-related algorithmic capabilities based on 2D images. Developers do not need to develop underlying algorithms from scratch; by simply calling the API interfaces provided by the SDK, they can quickly integrate functons such as face detection, recognition and analysis into Windows applications, including desktop software, client programs and industrial PC applications.

3D Face SDK for Andorid

The 3D Face SDK (Android Version) is a professional software development toolkit built for Android mobile devices, purpose-built for mobile application scenarios involving 3D facial recognition. It encapsulates core functions including face detection, 3D modeling, facial keypoint localization, liveness detection and facial feature extraction. Powered by computer vision, deep learning and 3D reconstruction technologies, the SDK is compatible with Android hardware configurations (depth cameras, RGB cameras included) and mainstream software environments running Android 8.0 and higher. Developers can rapidly embed high-precision, highly stable 3D facial processing capabilities into Android apps with simple API invocations. There is no requirement to build low-level algorithms from the ground up, which drastically lowers the technical barrier and shortens development lead times for mobile 3D face applications.

3D Face SDK for LINUX

The 3D Face SDK for Linux Platform is a software development toolkit designed for Linux operating systems (Ubuntu, CentOS, Debian), compatible with architectures including x86_64 and ARM64. It integrates core algorithms: 3D face detection, modeling, recognition, liveness detection, and point cloud analysis. The SDK is delivered as static libraries (.a), dynamic libraries (.so), or source code (for partially open-source SDKs). It supports major Linux programming languages: native C/C++, Python, and Go. Developers can rapidly deploy 3D facial capabilities onto Linux servers, embedded hardware, industrial PCs and other end devices. The SDK is optimized to match Linux’s attributes: multithreading, low resource footprint, and open-source ecosystem.

3D Face SDK for Windows

The 3D Face SDK is a software development toolkit tailored for Windows operating systems (Windows 10/11, Windows Server). It integrates core algorithms and interfaces for 3D face capture, modeling, detection, recognition, and analysis.

Object Detection and Classification

Locates and categorizes objects in images and video streams, including pedestrians, motor vehicles, non-motor vehicles, animals, and miscellaneous goods. Typical algorithms and models: YOLO, Faster R-CNN, SSD, RetinaNet, EfficientDet, etc. Core Capabilities： Performs simultaneous detection for multi-class, multi-scale, and multiple objects. Enables detection of tiny objects, occluded objects, and densely clustered objects. Deployment scenarios: security surveillance, traffic management, retail analytics, industrial inspection, robot perception.

Multi-Target Tracking (MOT)

It continuously localizes identical objects and sustains consistent ID association across consecutive video frames. Typical algorithms: DeepSORT, ByteTrack, StrongSORT, FairMOT, etc. Core Capabilities: Persistent ID retention across frames. Re-identification of targets after occlusion. Application Scenarios: Passenger flow counting, vehicle flow statistics, behavior analysis, security monitoring, robot following.

Image Segmentation

Semantic Segmentation: Assigns category labels to every pixel in an image (roads, buildings, sky, vegetation, human figures, etc.). Typical models: U-Net, DeepLab, SegFormer, etc. Application scenarios: Autonomous driving perception, medical image segmentation, remote sensing analysis, industrial defect region localization. Instance Segmentation: It executes category classification while differentiating distinct individual objects belonging to the same class (e.g., multiple pedestrians, multiple vehicles). Typical models: Mask R-CNN, SOLO, YOLACT, etc. Application scenarios: Robotic grasping, industrial sorting, security traffic/people counting, retail merchandise recognition. Panoptic Segmentation: Integrates semantic and instance segmentation branches: it generates instance masks for foreground objects and semantic labels for background zones. Application scenarios: Autonomous driving, urban environment perception, high-accuracy scene comprehension.

Keypoint & Pose Estimation

Human Pose Estimation: Detects human body keypoints (head, shoulders, elbows, wrists, hips, knees, ankles, etc.) and outputs human skeletal structures. Typical models: OpenPose, HRNet, YOLO-Pose, etc. Applications: Behavior analysis, fall detection, motion recognition, fitness and sports posture correction, human-computer interaction. Hand & Limb Keypoint Estimation: Supports 21 hand keypoints, finger pose tracking and gesture recognition. Applications: VR/AR interaction, sign language recognition, dexterous robotic manipulation, contactless gesture control.

Application Scenarios

Underground Shift Duty Management

Verify the attendance of leaders on underground shift duty, prevent proxy check-in and shift replacement, and realize genuine duty performance and strict assessment.

High-Risk Area Control

Key locations such as mining working faces and electromechanical chambers mandate on-duty identity verification, reducing supervision blind spots.

Hidden Danger Closed-Loop Management

Used for the hidden danger rectification process, the system links responsible persons with 3D facial data to ensure traceable rectification deadlines and acceptance results.

Emergency Response

Quickly verify the identity and location of underground personnel in the event of disasters to support emergency rescue decision-making.

Peak Shift Handover Management

Manage centralized personnel entry and exit at the mine with intelligent scheduling and identification to prevent congestion.

Implementation Case

Guizhou Tianlun Energy Shuicheng Jiyuan Coal Mine

Deploy the underground personnel safety supervision system to realize precise control of shift duty performance and significantly reduce the accident rate.

Shandong Energy Group Beidou Tiandi

Covering multiple projects such as Shilaosu Coal Mine, the system achieves the integration of personnel positioning and hidden danger management through 3D facial recognition technology.

Zhangji Coal Mine, Huainan, Anhui

After applying the system, personnel management efficiency has been improved by 50%, and job substitution has been basically eliminated.

Smart Energy Solutions

Smart Energy Solution

Solution Overview

Solution Architecture