Holtrics

Xiao Tian Quan AI Box

Xiao Tian Quan AI Box is a product integrating Internet of Things (IoT) and Artificial Intelligence (AI) technologies. It is built on deep learning-based object recognition algorithms and adopts a dedicated neural network model. Combined with an intelligent analysis and early-warning platform, it can achieve a wide range of AI vision recognition functions, including smoking, mobile phone use, safety helmet detection, smoke, flame, reflective clothing, sleeping on duty, intrusion, vehicle features and license plates, as well as open flame and smoke detection, and unauthorized personnel entry. The algorithms are highly accurate, provide second-level alerts, and support cost reduction through equipment reuse.

Application Scenario One (Group-Level Application):

A certain China Railway Bureau currently has more than 10 shield tunneling projects distributed nationwide, with each project equipped with over 20 cameras, totaling more than 200 video streams. All streams are unified and connected to the group company’s integrated video management system. At present, the group company intends to upgrade its video management to enable automated AI early-warning functions. We have analyzed two possible implementation methods:

1.Replace the cameras to enable AI early-warning functions.

2.Deploy the Xiao Tian Quan AI server at the group company’s integrated video management system site.

Application Scenario Two (Group-Level Streaming Media Server):

This solution enables the access, aggregation, computation, storage, and processing of massive resources. It supports multi-protocol integration, including mainstream standard protocols such as , RTSP/Onvif, and RTMP, as well as vendor-specific private protocols and SDK integration. It also supports multi-protocol video stream output, allowing centralized and unified management and monitoring of all video streams accessed from cameras, DVRs, and other devices.

Application Scenario Three (Project-Level Application):

A typical smart construction site requires approximately 35 cameras, of which around 10 algorithm capabilities. The unit price of smart cameras generally ranges b C$1,500–C$5,700. Assuming Model A smart cameras meet the project requirements, t price of Model A smart cameras is C$1,500, while standard cameras are priced at C$95. T comparison with the Xiao Tian Quan solution is as follows: need AI etween he unit he cost

Smart Camera Solution:

Model A smart camera: C$1,500 × 10 units = C$15,000

Xiao Tian Quan Solution:

Standard camera: C$95 × 10 units = C$950

Xiao Tian Quan AI Box: C$5,700

Total cost = C$6,650

TBM d-Twin

The platform system integrates DiTing V3.0 IoT Platform, the Digital Twin Platform, and the TBM d-Twin Application Platform, providing a comprehensive solution that combines IoT, BIM, GIS, and AIoT technologies for industrial applications. DiTing V3.0 is an IoT platform that features high integration and flexible scalability, designed to address complex and ever-changing IoT requirements. It adopts XTQservice to support high-frequency data acquisition and supports multiple transmission protocols such as Socket, WebSocket, and MQTT, with powerful concurrent data processing capabilities of up to 50,000 messages per second per server. Data storage is managed by Hadoop, InfluxDB, and other databases. The acquisition service allocates resources according to collection needs to ensure efficient reception, storage, and processing, while the acquisition client provides breakpoint resume and other stability features. It supports more than one hundred types of AIoT models and edge computing technologies, covering various industrial equipment and systems, enabling efficient interconnection across brands and protocols. The Digital Twin Platform applies virtual reality technology integrated with IoT + BIM + GIS to construct digital twins of target objects. Through comprehensive perception of domains, processes, and elements (people, machines, materials, methods, and environment), it achieves precise mapping from the physical world to digital space. Real-time data synchronization and dynamic updates are supported, while early warnings and alerts are pushed in parallel. Multi-dimensional visualization provides decision-makers with global auxiliary decision-making. TBM d-Twin is a shield tunneling TBM application platform built for customers on the basis of the CBIM digital twin platform and the DiTing V3.0 IoT platform. After eight years of development, it has become stable, reliable, and efficient, with a cumulative access of 2,001 TBM units. It supports access for millions of devices and has PB-level data storage and processing capabilities.

Platform Features:

High integration and flexible scalability.

Support for high-frequency data acquisition and powerful concurrent processing.

Multi-protocol compatibility including Socket, WebSocket, and MQTT.

Efficient and stable data reception, storage, and processing with Hadoop and InfluxDB.

Support for over one hundred AIoT models and edge computing technologies.

Real-time synchronization, dynamic updates, and comprehensive early-warning functions.

Multi-dimensional visualization and decision-making support.

Proven reliability with large-scale device access and PB-level data handling capacity.

Athena V1.0

Athena V1.0 is a large-scale smart construction site platform built for customers based on the CBIM digital twin platform and the IoT platform (DiTing V3.0). The platform takes the elements of construction projects—people, machines, materials, methods, and environment—as its foundation, with data-driven control at its core. It enables intelligent management and control of construction sites in real-time, dynamically, and in three dimensions, ensuring that the entire process and all elements are visible, controllable, and analyzable.

Core Functions:

Three Controls: Quality Control, Progress Control, Cost Control

Three Managements: Safety Management, Equipment Management, Document Management

One Coordination: Integrated Coordination

Business Scope – Progress Control

Production Plan Compilation

Visual Progress Confirmation

Planned vs. Actual Progress Analysis

Performance Evaluation and Assessment

Automatic Matrix Early Warning

Progress Milestone Early Warning

Quality Control

Technical Handover

Process Monitoring

Stage Acceptance

Cost Control

Material Procurement Planning

Material Procurement

Material Consumption

Settlement

Equipment Management

Equipment Ledger

Equipment Entry Acceptance

Equipment Operation Monitoring

Equipment Inspection

Equipment Fault Diagnosis

Safety Management

Risk Identification

Risk Inspection

Risk Warning

Hazard Elimination

Hidden Danger Investigation

Document Management

Engineering Drawings

Construction Plans

Process Acceptance

Image Archiving

Beaver V1.0

Beaver V1.0 is a digital-twin-based intelligent construction management platform built for customers on the CBIM digital twin platform and the IoT platform (DiTing V3.0). The platform is founded on automated equipment and monitoring, with a data-driven engine at its core, enabling full-domain, full-process, and all-element management of deep foundation pit construction. It transforms “passive monitoring” into “data-driven proactive control.”

Core Functions:

1.Digital Twin

– Situational awareness of key project elements (safety, progress, quality, environment, personnel, equipment).

2.Intelligent Monitoring

– Automated acquisition, transmission, storage, and analysis of monitoring data. Data-driven alerts notify engineering personnel of risks and validated results, realizing self-driven management of safety monitoring.

3.Smart Dewatering

– Integrated system linking dewatering operation monitoring, control systems, and alarm systems.

4.Smart Support

– Utilizes axial force servo systems for fully automated monitoring and control, enabling more precise, faster, and safer management.

5.Smart Reinforcement

– Real-time collection of parameters such as slurry flow, rotation speed, pressure, depth, and density from mixing piles and other field equipment. Data-driven alerts provide automatic warnings for minimum slurry content, notify relevant personnel, and generate regular quality evaluation reports.

Data-Driven Engine：

Data Collection：

Digital Twin

Smart Reinforcement

Document Management

Collaborative Management

Data Driven

Smart Support

Smart Dewatering

Intelligent Monitoring

Deep Foundation Pit Intelligent Construction Management Platform (Beaver V1.0)

Based on a data-driven engine and built around seven key links, the platform realizes “smart construction of deep foundation pits.”

Digital Twin (Situational Awareness)

Smart Monitoring (Automated Monitoring System)

Smart Dewatering (Automated Dewatering Monitoring and Control System)

Smart Support (Support Axial Force Servo System)

Smart Reinforcement

Document Management (Full-Process Management)

Collaborative Management (Data Collaboration)

Dragon Turtle V3.0

Dragon Turtle V3.0 is a TBM (Tunnel Boring Machine) digital twin training platform built for customers based on the CBIM digital twin platform and the IoT platform (DiTing V3.0). It adopts virtual simulation, immersive VR, and real operation consoles in combination to train TBM operators and tunnel construction personnel.

Core Functions:

1.3D simulation of TBM physical components

2.TBM assembly

3.Full-process demonstration of TBM construction operations

4.TBM construction site simulation

5.Realistic TBM construction process display (data-model driven)

6.TBM operation simulation training

7.Simulation training for TBM structure, electrical, and hydraulic system recognition and maintenance

8.3D simulation handling of TBM faults

9.TBM knowledge mapping, 4D micro-learning, 3D task-based assessment, and practical evaluation