Thundercomm TurboX™ AI KIT
Thundercomm AI Kit 是一个面向开发者,拥有最强性能的智能视觉的嵌入式开发设备。 Thundercomm AI Kit基于高通SDA845平台,提供可进行开发的设备、算法SDK、教程、代码,帮助开发人员和终端厂商进行最终产品的原型设计。 点此查看更多炫酷应用






Product Show

11161146b661.jpg        

Product interface diagram

Key Advantage

预定制的算法SDK

提供高清人脸识别,物体识别等算法SDK,让开发者及厂商提供方便快速的集成AI算法,开发AI应用,完成产品解决方案。


强大的运算力

幸运赛车首页 集成了全新的Qualcomm Hexagon 685 DSP,新的向量引擎和异构(CPU,DSP和GPU)计算能力,强大的运算能力和视觉处理能力,可以实时完成高清AI视觉处理。


多样灵活的视觉方案

幸运赛车首页 提供多种camera方案支持,包括内置Ultra HD camera,外置USB camera及IP camera等,可以让开发者及厂商灵活选择自己需要的方案。


Product Hardware SPEC

模块 条目 规格
Computing CPU Qualcomm SDA845 processor, 8x Qualcomm® Kryo™ 385 @ 2.8GHz
GPU Adreno 630 GPU
DSP Hexgan 685 DSP
ISP Qualcomm Spectra™ 280 image sensor processor
Memory RAM 8GB LPDDR4
ROM 64GB UFS2.1
SD Card Support MicroSD Card
Camera Built in Camera 8MP, FOV 120 degree;
External Camera USB3.0 camera (accessory, not included in release)                    
Display HDMI 1x Micro HDMI, 4k Ultra HD
Type C 4k Ultra HD
Video VPU Support for HDR10, HLG, and H.265 (HEVC) 4k@60fps video capture
Audio Speaker 1x speaker
Audio out 1x Headsetout
Microphone 1x microphone
Wireless Connectivity WIFI 2.4G/5G, 802.11 a/b/g/n/ac 2x2 MIMO
Wired Connectivity USB 3x USB3.0 host, 1x 3.0 Type C
Ethernet 1x GbE.
Debug Port Micro USB 1x Micro USB (connecting to serial port for debug purpose)
Indicator LED RGB Led: Power status, WIFI status, Ethernet status;
Input Buttons Power key/ Volume up/down key/Camera snapshot key
Sensors 9-axis sensor Gyroscope & accelerometer & geomagnetic sensor
Power Power charging 12v DC

Product Software SPEC

模块 特性
OS Android O
AI Framework Tensorflow, Tensorflow Lite, Caffe/Caffe2, ONNX,SNPE, Android NN
AI Algorithm SDK Face Detection & Face Recognition & Emotion Detection & Age Detection & Gender Detection algorithm & Eye blink
Object Detection algorithm
AI Sample Application Face Recognition & Emotion Detection & Age Detection & Gender Detection algorithm demo
Objection Detection algorithm demo
Camera Support USB3.0 Camera, 1080p@30fps
Ultra HD camera preview, recording and snapshot
RTSP Client
Graphic OpenGLES3.2, OpenCL2.0 full