Intelligent Computing Unit for Unmanned Duty surveillance Turntables
Category：
Main Products
Keywords：
WhatsApp：
Product Overview
Product Overview
Designed for intelligent control of electro-optical turntables, this compact unit enables fully automated target search, detection, tracking, and one-click target acquisition—without human intervention. Built on advanced digital vision network technology, it drives the turntable to perform autonomous patrols along preconfigured routes, delivering real-time visible light and IR video analytics, human and vehicle target recognition, and multi-level intrusion detection and alerting.
The unit supports video structuring, high-definition video storage, over-the-air algorithm updates, and seamless integration with third-party platforms via standard network protocols including RTSP, RTP, and GB28181.
Choose Your Model
High-performance model: NVIDIA computing board, power supply module, switching module, control module.
Standard-grade model: Huawei computing board, power supply module, switching module, control module.
Intelligent Computing Board: Available in high-performance and domestic versions, designed for integration into the turntable’s internal structure.
Product Highlights
* Turntable intelligent control: Supports full control over turntable patrol patterns (routes and speeds) and visible/IR camera coordination.
* AI-powered image analysis: Visible light and IR target (people, vehicles) detection , tracking, matching, fusion detection, fixed target change detection.
* Tracking Processing: Visible light tracking, IR tracking, human/vehicle target tracking.
* Automatic Alarm: Real-time intrusion alerts for personnel and vehicles, with alarm reporting and retrieval.
* High-performance video processing: Supports multiple video streams, multi-channel real-time H.264/H.265 decoding, and streaming output.
* Capable of storing 24 hours of HD footage on a minimum of two video streams.
* Online Algorithm Upgrade
* Video Structuring Processing, Transmission, Storage, and Invocation
* Device Status Online Monitoring
* Configuration Management: Enables configuration of full-scene capture, boundary lines, fixed target change detection, turntable scan patterns, camera parameters (IP, port, frame rate), algorithm settings, and alarm levels.
Specification | High-performance model | Standard-grade model | |
Model | HT-AIPT-DC004 | HT-AIPT-DD004 | |
System Parameters | Processor | NVIDIA | Huawei Ascend |
Calculation Power | >100Tops INT8 | ||
Memory Capacity | DDR 32GB@1800MHz | ||
Storage Specification | SSD 4T (expandable), Minimum 90-day 2K video storage supported per channel | ||
Input | Input Video | Supports RTSP, RTMP, GB28181 transmission protocol video stream, not less than 8 channels of 2K clear smart processing | |
Output | Condensed Video Stream | Configurable frame sampling, compression, and dynamic target video clip recording. Encoding format: H.264 or H.265 (configurable) | |
Summary Image Stream | Images with targets (background images and target images can be configured). Encoding format: JPEG. Network protocol: HTTP (GB1400 image library standard) | ||
Non-Structured Feature Stream | Feature vectors for search. Encoding format: customizable base64 vector array. Network protocol: HTTP | ||
Structured Target Feature Stream | Target attributes generated based on the configured algorithm. Encoding format: text (JSON). Network protocol: HTTP | ||
Alarm Information | Event description, alarm details, proactive alarm push. Encoding format: text. Network protocol: HTTP. | ||
Note | Output configuration can be set as needed via SDK or interface | ||
Video coding and decoding | H.265/H.264 coding and decoding | 16 channels of 4K@25fps, or 8 channels of 2K@25fps | |
Resolution | 256x128 (H264), 128x128 (H265) to 3840x2160 | ||
ROI Encoding | Supports up to 32 regions encoding | ||
| Storage Capability | Visible light and thermal imaging video streams can be independently stored after splitting, supporting two storage modes: 12 hours × 90 days of direct storage and independent storage of condensed video streams. | |
Video Processing | Framework | Caffe/Tenorflow | |
Configuration Method | Supports online update configuration for algorithms | ||
AI Processing | I. Capable of intelligent recognition and fusion processing of visible light and infrared thermal images, enabling detection, classification, and intelligent analysis of personnel, vehicles, and non-motorized vehicles. | ||
General Specifications | External Interface | Gigabit Ethernet Port *2, 485 *2 | |
Rated Power Consumption | 150W | 250W | |
Environmental Adaptability | Operating temperature: -40°C ~ 55°C | Storage temperature: -40°C ~ 60°C | |
MTBF | 3000h | ||
Softerware Interface
For any cooperation or support please contact:
Bella Yang
E-mail: yangqing@htmw.cn / xinqing_2011@foxmail.com
Tel/Whatsapp: +86 188 1011 7965
Follow the official account