Anonymous
Mar 30, 2026
Rating: 5/5
Impressive inference speed
The inference speed on this board exceeded our expectations for its price point. We successfully deployed YOLOv5 variants with good frame rates. Support team was also very responsive.
Description
Reviews
| Brand | - |
| Category | Robotics / STEM Teaching Tools / Development, Learning, Evaluation & Industrial Control Boards |
| Origin | - |
| Processor | HiSilicon Hi3516CV610 |
| AI Computing Power | 1 TOPS (INT8) |
| Memory | 512MB DDR4 |
| Storage | Supports up to 128GB microSD card |
| Video Input | 1x MIPI CSI interface |
| Video Output | HDMI 1.4, CVBS |
| Connectivity | 10/100Mbps Ethernet, Wi-Fi module ready |
| Operating System | Linux-based SDK provided |
| Power Supply | 5V DC via Type-C or GPIO |
| Dimensions | 60mm x 45mm x 12mm |
Anonymous
Mar 30, 2026
Rating: 5/5
Impressive inference speed
The inference speed on this board exceeded our expectations for its price point. We successfully deployed YOLOv5 variants with good frame rates. Support team was also very responsive.
Anonymous
Mar 27, 2026
Rating: 5/5
Excellent board for AI vision prototypes
This development board is perfect for getting started with edge AI. The 1 TOPS performance is sufficient for running multiple object detection models simultaneously. Documentation was clear and setup was straightforward.
Anonymous
Mar 25, 2026
Rating: 4/5
Great value for security applications
Used this for a custom smart camera project. The Hi3516CV610 chip handles video encoding very well. Only minor issue was the Wi-Fi module needed separate sourcing, but overall a solid choice.
Anonymous
Mar 13, 2026
Rating: 4/5
Solid performer with minor quirks
Overall a very capable board. The GPIO pins are versatile for connecting sensors. Had some initial trouble with the Linux image version, but a quick update resolved it. Would buy again.
Anonymous
Mar 12, 2026
Rating: 4/5
Good learning tool
As a student learning about computer vision, this board offered a great balance of price and performance. The SDK examples helped me understand neural network deployment quickly.
Anonymous
Mar 12, 2026
Rating: 5/5
Robust hardware for industrial use
We integrated this into an industrial control system. It runs stable under continuous operation and the thermal management is adequate without a large heatsink. Highly recommended for evaluation.
Q: Does this board come with a pre-installed OS?
A: The board typically comes with a bootloader. You will need to flash the Linux SDK image provided in the download section of our website to get the full operating system running.
Q: What is the power consumption during typical operation?
A: Under typical load with AI inference running, the board consumes approximately 2.5W to 3.5W. Idle consumption is significantly lower, around 1.2W.
Q: Is there a JTAG interface available for debugging?
A: Yes, the board includes a standard 20-pin JTAG header for low-level debugging and firmware development. Pinout details are available in the hardware reference manual.
Q: Can I run TensorFlow Lite models on this board?
A: Yes, the board supports TensorFlow Lite models through our NNIE (Neural Network Inference Engine) toolkit included in the SDK. Conversion tools are provided to optimize models for the NPU.
Q: What is the maximum resolution supported for video input?
A: The Hi3516CV610 supports up to 4MP (2560x1440) at 30fps via the MIPI CSI interface, making it suitable for high-definition surveillance applications.