Products Description
The Intelligent Computing Module is a high-performance embedded computing unit purpose-built for edge AI applications, delivering powerful real-time inference and data processing capabilities in latency-sensitive or cloud-independent environments. Built on an advanced heterogeneo us computing architecture (CPU + NPU + GPU), it enables efficient neural network inference, algorithm deployment, and on-device AI acceleration with an optimal balance of high computing power and low energy consumption.
As part of the "Shanhai" Series, this China-manufactured module fully complies with ISO, CE, and FCC international standards, supports mainstream AI frameworks, and offers rich expansion interfaces for flexible system integration. It is widely used in smart security, industrial inspection, autonomous mobile robots (AMR), intelligent transportation, and other edge computing scenarios, making it a core component for enterprises and research institutions developing AI-powered devices and edge AI solutions.
Products Specification
|
Parameter Category |
Specifications |
|
Computing Architecture |
CPU + NPU + GPU Heterogeneous Architecture |
|
Supported AI Frameworks |
TensorFlow, PyTorch, ONNX, Caffe |
|
Operating Temperature |
-40℃ to +85℃ |
|
Storage Temperature |
-55℃ to +125℃ |
|
Expansion Interfaces |
USB 3.0 × 2, Ethernet (Gigabit) × 1, HDMI 2.0 × 1, MIPI-CSI 2.0 × 2, UART × 2 |
|
Inference Latency |
≤10ms (for ResNet-50 model, batch size = 1) |
|
Dimension (L×W×H) |
60mm × 40mm × 8mm |
Features
Broad AI Framework Compatibility
Native support for mainstream AI frameworks including TensorFlow, PyTorch, ONNX, and Caffe, eliminating the need for custom framework adaptation. Developers can directly deploy pre-trained models, reducing development cycles by an average of 30%.
Reliable Operation in Harsh Environments
Designed for industrial applications, the module operates reliably across an extended temperature range of –40°C to +85°C and supports wide input voltage fluctuations. It ensures stable performance in high-temperature workshops, outdoor security cabinets, and other demanding environments.
Rich Interfaces and Easy Integration
Equipped with dual MIPI-CSI camera interfaces, Gigabit Ethernet for high-speed data transmission, and HDMI 2.0 for video output. The compact 60 × 40 × 8 mm form factor enables seamless integration into space-constrained devices such as mini security cameras and compact AMRs.
Real-Time High-Speed AI Inference
Delivers ultra-low latency AI inference with ResNet-50 latency ≤ 10 ms (batch size = 1), meeting the stringent response requirements of time-critical applications such as industrial visual inspection with instant defect alerts and real-time obstacle avoidance for AMRs.
Advantages
Proven Radar Technology
Built on proprietary radar core technologies and supported by recognized R&D platforms, ensuring reliable performance and long-term stability.
10+ Years of Industry Expertise
Over a decade of experience across automotive, aerospace, and industrial applications, delivering solutions aligned with real-world system requirements.
Application
Service
Technical support (product structure, operation guide, common troubleshooting); Product documentation (user manual, operation videos, checklist)
Global Warranty Services
2-year worldwide warranty;Free repair or replacement of components due to product quality issues;Excludes damage caused by human error or improper use
FAQ
Q: What AI frameworks does the Intelligent Computing Module support?
A: It natively supports mainstream frameworks including TensorFlow, PyTorch, ONNX, and Caffe-no custom adaptation is required.
Q: Can the module operate in high-temperature environments?
A: Yes, it has an operating temperature range of -40℃ to +85℃, ensuring stable performance in high-temperature industrial workshops or outdoor scenarios.
Q: What expansion interfaces does the module provide?
A: It includes USB 3.0 × 2, Gigabit Ethernet × 1, HDMI 2.0 × 1, MIPI-CSI 2.0 × 2, and UART × 2-supporting camera connection, data transmission, and video output.
Q: Is a development kit provided for the module?
A:Yes, a complete development kit is included: development board, SDK with API docs, and 20+ pre-trained models, compatible with Windows/Linux systems.
Q: What is the inference latency of the module?
A: For the ResNet-50 model (batch size = 1), the inference latency is ≤10ms, meeting real-time response requirements.
Q: Does the product have international certifications?
A: Yes, it has obtained CE, FCC, and ISO 9001 certifications, complying with global market import standards.
Q: What is the warranty period of the module?
A: All models come with a 2-year global warranty, covering defects in materials and workmanship. Human error or improper use is not covered.
Hot Tags: intelligent computing module, China intelligent computing module manufacturers, suppliers, factory
