SKU: IOWC-1359
📋 GST Invoice Available

Realtek AMB82-Mini IoT AI Camera Dev Board w/ F37 Camera

₹3,779.00 (Inclusive of all taxes)
Out of Stock

The Realtek AMB82-Mini is an ultra-low-power AI camera development board built around the RTL8735BDM SoC, integrating dual-band Wi-Fi, BLE 5.1, an NPU, and a 1080p camera. It supports Arduino IDE, FreeRTOS, and real-time H.265/H.264 encoding for battery-powered computer vision and IoT endpoints.

Features

  • Realtek RTL8735BDM SoC integrates an ARMv8M MCU (500MHz), dual-band 802.11 a/b/g/n Wi-Fi, Bluetooth BLE 5.1, an ISP, and a 0.4 TOPS Neural Processing Unit (NPU) for on-device AI.
  • Includes a pre-attached Taiyee F37 camera module with a 1080p Full HD CMOS sensor and a 130° wide-angle lens, supporting real-time H.265/H.264/JPEG encoding up to 2K resolution.
  • Ultra-low-power architecture enables microampere-level sleep currents and millisecond boot times, ideal for battery-powered AI vision applications.
  • Comprehensive SDK support for Arduino IDE, FreeRTOS, IAR, and GCC, alongside extensive I/O including USB 2.0 host/device, microSD slot, SPI, I2C, UART, and up to 23 GPIOs.
  • Package includes the dev board, a dual-band (2.4GHz/5GHz) IPEX antenna, and the F37 camera module for immediate prototyping.

Out of stock

Description

The AMB82-Mini development board leverages the highly integrated Realtek RTL8735BDM SoC, an 8-in-1 solution designed for power-efficient AI-at-the-edge applications. It combines an ARMv8M microcontroller, dual-band Wi-Fi, Bluetooth 5.1, an audio codec, an Image Signal Processor (ISP), and a dedicated neural network engine capable of 0.4 TOPS. This convergence allows for sophisticated tasks like object detection and classification directly on the device without constant cloud dependency.

Central to its design is the support for real-time, multi-stream video encoding in H.265, H.264, and JPEG formats, handling up to 1080p@30fps. Paired with the included wide-FOV 1080p camera sensor, it forms a complete vision system. The board’s power management is engineered for always-on, battery-operated scenarios, featuring deep sleep modes and fast wake-up times measured in milliseconds.

For developers, it offers a rich set of physical interfaces—including USB host/device, a microSD card slot, multiple serial ports, and programmable GPIOs—facilitating integration with sensors and actuators. Support for Arduino IDE lowers the barrier to entry, while professional SDKs for FreeRTOS, IAR, and GCC cater to advanced firmware development for commercial IoT deployments.

Key Features

  • Ultra-Low-Power AI SoC: The RTL8735BDM integrates MCU, Wi-Fi, BLE, NPU, and ISP for efficient battery-powered vision AI.
  • On-Device Neural Processing: Dedicated 0.4 TOPS NPU accelerates machine learning inferences locally, reducing latency and bandwidth.
  • High-Definition Video Pipeline: ISP with HDR/3DNR/WDR feeds a hardware encoder supporting simultaneous H.265/H.264/JPEG streams.
  • Dual-Band Wi-Fi & BLE 5.1 Connectivity: Provides robust 2.4GHz/5GHz Wi-Fi for high-throughput video and Bluetooth for peripheral communication.
  • Extensive Development Framework Support: Compatible with Arduino IDE for quick prototyping and FreeRTOS/IAR/GCC for production-grade firmware.
  • Rich Peripheral Set: Features USB 2.0 host/device, microSD slot, I2S audio codec, multiple UARTs/SPIs/I2C, PWM, and up to 23 configurable GPIOs.
  • Integrated Camera System: Comes with a soldered 1080p F37 CMOS camera module featuring a 130° wide-angle lens, ready for vision applications.
  • Enhanced Security: Includes hardware cryptographic engine, secure boot (optional), TrustZone, and support for WPA3 Wi-Fi security.

Applications

  • Battery-powered smart security cameras and video doorbells with local person/object detection.
  • AIoT edge devices for retail analytics, occupancy sensing, and industrial machine vision.
  • Wireless baby monitors, pet cameras, and elderly care monitoring systems with low-power operation.
  • Prototyping for smart home hubs that require camera input, voice control (via mic), and wireless connectivity.
  • Educational and research platforms for embedded machine learning and computer vision.

Technical Specifications

Core SoCRealtek RTL8735BDM
MicrocontrollerARMv8M, up to 500MHz
Neural Processor (NPU)Intelligent Engine @ 0.4 TOPS
Memory128MB internal DDR2, 16MB external SPI NOR Flash
Wireless Protocols802.11 a/b/g/n Dual-Band (2.4GHz/5GHz) Wi-Fi, Bluetooth BLE 5.1
Camera SensorTaiyee F37, 1920x1080 Full HD CMOS, 130° FOV Lens
Video EncodingH.264/H.265/JPEG encoder, up to 1080p@30fps + 720p@30fps
Image Signal Processor (ISP)HDR, 3DNR, WDR
AudioIntegrated Audio Codec (ADC/DAC/I2S), 1x On-board Microphone
Interfaces2x Micro USB (Host/Device), MicroSD Card Slot, 3x UART, 2x SPI, 1x I2C, 8x PWM, up to 23x GPIO
Antenna ConnectorIPEX (for included dual-band antenna)
Development SupportArduino IDE, FreeRTOS, IAR, GCC
Security FeaturesHardware Crypto Engine, Secure Boot (Opt.), TrustZone, WEP/WPA/WPA2/WPA3/WPS

Frequently Asked Questions

How do I start programming the AMB82-Mini with the Arduino IDE?

First, install the Arduino IDE and add the Realtek AMB82 board support package via the Boards Manager. Connect the board to your computer via the Micro USB port labeled for programming. Select 'AMB82-Mini' as the board and the correct COM port. You can then open and upload example sketches for camera capture, Wi-Fi connectivity, or using the NPU for basic inference tasks, utilizing the extensive library support provided in the SDK.

What is the power requirement and typical operating current for battery use?

The board is designed for ultra-low-power operation and can run from a 5V supply via the Micro USB port or a 3.7V Li-Po battery connected to appropriate pins. During active sensing with Wi-Fi transmitting, current draw can be several hundred milliamps. However, leveraging its deep sleep modes, the system can achieve microampere-level currents, making it suitable for battery-powered applications where the device wakes periodically to capture and process an image.

Can I replace the attached F37 camera module with a different sensor?

The F37 camera module is directly soldered onto the AMB82-Mini development board. Replacing it is not intended for end-users, as it requires desoldering and precise matching of the MIPI CSI interface and power requirements. For prototyping with different sensors, you would need to design a custom carrier board that connects to the board's available GPIO and camera interface pins, which requires hardware design expertise.

How do I access the NPU for running custom AI models on the board?

To use the NPU, you must work within the professional SDK frameworks like IAR or the GCC-based FreeRTOS SDK provided by Realtek. The workflow involves converting your trained model (e.g., from TensorFlow Lite) into a format compatible with the Realtek toolchain using their conversion utilities. The compiled model is then loaded into the NPU memory, and your application code calls the NPU driver APIs to perform inferences on camera or sensor data.

Package Includes

  • 1 x Realtek AMB82-MINI Development Board
  • 1 x Dual-band Wi-Fi (2.4GHz/5GHz) IPEX Antenna
  • 1 x Taiyee F37 Camera Module (pre-attached)

Additional Information

Weight 0.037 kg
Dimensions 12 × 8 × 3 cm