Camera kit with AI engine

Dual Core ARM Cortex-A35 processor with:

256MB DDR3 RAM (SIP package);

High performance Image Vision ISP (up to 5Mpix with HDR);

H.264/H.265/JPEG encoder;

Video scaling, OSD support;

OpenCV hardware accelerator;

Hardware NPU supports TensorFlow, Caffe and etc;

Hardware audio decoder (G.711, G.726, ADPCM).

Default sensor:
Option 1: Sony IMX462 (Full HD (1920x1080), Rolling shutter, ultra-low noise, HDR mode, 60fps);
(sensor + GMSL2 board adapter supplied with a camera kit);
Option 2: Sony IMX392, Global shutter, Full HD, 60fps;
Option 3: Sony IMX528, iTOF 3D sensor.

100Mbit LAN RJ45 port;

GMSL2 interface (coax cable with power,
15 meters cable length max);

Supports GMSL2 cameras also.

5” Color TFT LCD 720x1280 (MIPI-DSI interface);

microSD card slot (for video recording);

Possible extension for audio in/out;

CAN 2.0b bus (+ CAN FD) (based on STM32F7 controller);

Wi-Fi 802.11 b/g/n/ac (HostAP and STA, MIMO 2x2) + BLE 5.3;

12VDC input (low power consumption);

Temperature range: -20.. +70 ºC (possible extension).

Linux 5.15.78 (Fastboot) + Buildroot;

Demo application with RTSP server, H.265 encoder (streaming video from the camera over Wi-Fi network, LAN network);

Running BlueZ for future BLE applications;

Running WiFi HostAP (+ DHCP server) and WiFi STA – possible to get RTSP video stream using VLC application on mobile phone or tablet.

SDK for custom application development;

OpenCV HW accelerators support in SDK;

NN API in SDK (for example – face recognition).