A minimum size AI depth learning board based on the NVIDIA TX2 module. The interface board is connected to the TX2 core module using a connector. The compact design and small form factor bring deep learning to more and more applications. This interface board is compatible with the TX1 and TX2 core modules.
Can be equipped with 4G communication module
Can be equipped with wife and bluetooth modules
Support 4k × 2k 30Hz encoding (HEVC), 4k × 2k 60Hz decoding (10-bit)
Support linux operating system
Support TCP, RTSP network protocol
Board configuration
NVIDIA Pascal GUP, 256 NVIDIA CUDA cores
Dual-core Denver 2, quad-core ARM Cortex-A57 processor
8G Bytes 128bit LPDDR4
32G Bytes eMMC
Peripheral interface
1 way Micro HDMI output
2-way USB 3.0 interface (Type-A)
1-way Micro USB 2.0 interface
1 way 10/100/1000MBPS Ethernet interface
1 channel mini PCIe interface, can be connected to 4G communication module
1 channel UART debug serial port
1 way JTAG debug interface
Physical characteristics
Size: with core plate 87mm*53mm*22mm, with core plate + fan 85mm*50mm*40mm
Weight: 40g, with fan 125g, TX2 core board 90g
Environmental adaptability
Working temperature: -20 ° C ~ +60 ° C
Working humidity: 10% to 80%
Electrical characteristics
DC power supply, voltage +12V@3A
Power consumption: ≤20W
Embedded industrial machine vision system
Embedded medical image recognition system
Analysis of traffic statistics, face recognition and abnormal behavior in shopping malls
Customs intelligent inspection system: HD face capture
Intelligent parking management system: intelligent image recognition, accurate license plate recognition
Urban Road Monitoring System: Road illegal driving behavior
ADAS system: lane departure, obstacle detection, vehicle detection, and subsequent development of panoramic stereo imagery surveillance system
Forest fire prevention: dynamic monitoring, fire warning