ST x SingTown Technology | Jointly Launching OpenMV N6, Defining a New Benchmark for Embedded Machine Vision

As the official founding team of OpenMV, SingTown Technology and ST, leveraging over a decade of deep collaboration, continue to drive innovation in embedded machine vision technology. Recently, SingTown Technology launched the OpenMV N6 machine vision smart camera module based on the STM32N6 chip. With an exceptional balance of AI computing power and low power consumption, it provides high-performance and user-friendly solutions for fields such as robotics, industrial inspection, and smart surveillance.

SingTown's collaboration with ST commenced in 2015. Together, based on the STM32F4, they developed the world's first embedded machine vision product, the OpenMV Cam, pioneering the field of embedded image recognition products. Subsequent iterations have progressively adopted STM32F7/H7 series chips, continuously pushing the boundaries of performance.

The OpenMV N6, launched in 2025, is equipped with the latest STM32N6 chip. It not only fully leverages the powerful performance of this chip and the mature STM32 ecosystem, but also provides developers with a high-performance, low-power, and easy-to-develop machine vision solution. This helps customers achieve innovative applications in fields such as robotics, AI machine vision, and industrial inspection.

Analysis of Core Technical Highlights of OpenMV N6

Featuring ultimate AI performance with underlying optimisation

The OpenMV N6 is the first machine vision module powered by MicroPython, featuring the STM32N6 as its main controller. With its Cortex-M55 core running at 800MHz and a Neural-ART Accelerator (NPU) delivering 600 GOPS, AI processing performance is significantly enhanced. The underlying algorithms are deeply optimised for the STM32N6's NPU, enabling the YOLOV8N model to run at over 30 FPS at a resolution of 256×256. It supports AI tasks such as gesture recognition, image segmentation, and object detection.

Furthermore, the STM32N6's built-in 4.2MB RAM, dual 64-bit AXI memory buses, combined with the H.264 encoder and JPEG codec engine, significantly enhance video processing and data transfer efficiency. This hardware architecture enables it to efficiently handle the massive data within neural networks when running machine learning algorithms, delivering a 600-fold improvement in AI performance compared to STM32 chips without an NPU.

It is noteworthy that the OpenMV N6, built on the MicroPython framework, significantly lowers the development barrier. Engineers can quickly debug applications without frequently compiling firmware. Coupled with its on-device visual processing capabilities, it avoids the privacy risks associated with uploading data to the cloud, thereby extending embedded vision technology to a broader range of application scenarios.

Diverse Visual Capabilities and Hardware Expansion

The OpenMV N6 module supports a range of high-end camera modules, including:

480FPS High Frame Rate 1-Megapixel Colour Global Shutter Camera

5-megapixel HDR camera

Lepton Infrared Thermal Imaging and RGB Colour Image Fusion Mode

Capable of meeting the requirements of professional scenarios such as industrial temperature measurement, dynamic target tracking, and robotics.

In terms of interface configuration, it integrates

2.4GHz Wi-Fi b/g/n

Bluetooth LE 5.1 Module

Gigabit Ethernet Interface (PoE Power Supply Board also available)

Equipped with a USB-C high-speed port (480Mb/s)

32MB Flash (400MB/s)

64MB PSRAM (800MB/s)

Supports JTAG and STLINK programming

Fully compatible with embedded development requirements.

Designed for low-power applications

The OpenMV N6 excels in power consumption control, with a minimum power draw of just 1mA. In deep sleep mode, battery life can extend up to one year, and no additional cooling devices are required. Its hardware interfaces support Ethernet PoE power boards and 3.7V lithium battery charge-discharge management. It has already achieved large-scale application in scenarios such as logistics robots and intelligent security terminals. Its fanless design not only simplifies the development of 3D-printed enclosures but also aligns with the integration needs of industrial equipment for miniaturisation and low power consumption.

In the future, SingTown Technology will also launch OpenMV products based on the STM32MP2 processor platform.OpenMV MP2Further expanding the application boundaries of high-performance MPU scenarios. The new solution will continue the technical roadmap of "high performance + low power consumption + easy development," leveraging the continuous iteration of the STM32 platform to provide more flexible machine vision solutions for fields such as industrial automation and smart transportation, driving embedded vision technology toward broader application scenarios.