Specifications

MCU

STMicro STM32N657X0H3Q — Arm Cortex-M55 up to 800 MHz with hardware ISP

NPU

Neural-ART Accelerator, 600 GOPS

System RAM

256 Mbit (32 MB) @ 200 MHz, 16-bit

Storage

512 Mbit (64 MB) NOR flash

Thermal sensor

Lynred ATI320 LWIR microbolometer

Thermal resolution

320 x 240 pixels

Frame rate

60 FPS

Pixel pitch

12 µm

Spectral range

LWIR 8-14 µm

Thermal sensitivity (NETD)

< 50 mK

AI inference latency

< 100 ms (model dependent)

Image processing latency

< 50 ms

USB interface

USB 2.0 High-Speed via JST — UVC (MJPEG, H.264, YUV422) and RAW8/RAW16

Ethernet interface

100 Mb/s PHY via JST — MJPEG and H.264 over RTP/RTSP/RTCP

Onboard AI functions

Object detection, classification, AI-assisted upscaling, temperature-triggered events

Power

5 V via USB JST connector

Dimensions

28.5 x 30.5 x 29.5 mm

Weight

38 g

Mounting

4x M2 mounting holes on four sides

Firmware

Proprietary firmware with failsafe recovery bootloader

Compliance

NDAA and TAA compliant

Overview

The 8devices 8Sight T100 is a compact long-wave infrared (LWIR) thermal camera built for unmanned systems, robotics and industrial machine vision. It pairs a Lynred ATI320 microbolometer with an STMicro STM32N657X0H3Q Cortex-M55 microcontroller that carries a 600 GOPS Neural-ART NPU and a hardware ISP, so thermal image processing and neural-network inference both run on the camera itself.

Because inference is executed on-board, the T100 operates completely offline. Customers can deploy custom models for object detection, anomaly detection or temperature-triggered events without any cloud dependency. The camera streams 320 x 240 thermal video at 60 FPS with a sensitivity below 50 mK and image-processing latency under 50 ms.

The 28.5 x 30.5 x 29.5 mm housing, 38 g weight and four M2 mounting holes are sized for UAS and other space-constrained mobile platforms. Host integration is through USB 2.0 (UVC MJPEG/H.264/YUV422 plus RAW8/RAW16) or a 100 Mb/s Ethernet PHY supporting MJPEG and H.264 over RTP/RTSP/RTCP, and the unit is NDAA and TAA compliant.

Key Benefits

Offline edge AI. A 600 GOPS Neural-ART NPU runs detection and classification on-camera with no cloud link. Real thermal fidelity. Lynred ATI320 microbolometer delivers 320 x 240 LWIR at 60 FPS with sub-50 mK NETD. UAS-ready. 38 g, 28.5 x 30.5 x 29.5 mm and four M2 mount points. Dual connectivity. UVC USB 2.0 and 100 Mb/s Ethernet with RTP/RTSP/RTCP. Compliance-friendly. NDAA and TAA compliant for public-sector deployments.

Applications

Unmanned aerial systems and UAS video transmission, security and surveillance, robotics and autonomous platforms, industrial thermal inspection, predictive maintenance, anomaly and temperature-triggered event detection, border and perimeter monitoring.

Request a Quote — 8DEVICES 8SIGHT T100 — LWIR THERMAL EDGE AI CAMERA WITH STM32N6 NEURAL-ART NPU

QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.

Get Your Quote →

Related Products

Cvitek Cv1842H Edge Ai Camera Module Neoeyes Ne503 Hailo 15H Openmv N6 Ae3 Cameras Otobrite Gmsl2 Camera Series Basler Ace 2 Gmsl