Vision intelligent humanoid robot machine vision platform based on Huawei's ascend310  Atlas200 QUANAI ROBOT

In recent years, edge computing is gradually moving towards intelligent edge. All love technology continues to keep pace with the times. It is an innovative machine vision solution jointly built by Huawei atlas artificial intelligence computing platform

Quanai robot kit product positioning:

The quantum robot machine vision development kit is a developer board form product with the core of the atlas200 AI acceleration module. Its main function is to open the interface of atlas200 to the outside world, so that users can use atlas200 quickly and simply. It can be used in the pre research and development of many fields such as security, UAV, robot, optical industry detection, intelligent medicine, video server, etc. Atlas 200 AI acceleration module (atlas 200 for short) is a high-performance AI intelligent computing module, which integrates the Hisilicon ascend 310 AI processor, and can realize image, video and other data analysis and reasoning calculation. It is a high-performance and low-power AI chip specially designed by Huawei for image recognition, video processing, reasoning calculation, machine learning and other fields. Two AI cores are built in the chip, which can support 128 bit wide lpddr4x, and can realize the maximum computing power of 16tops (int8).

Software environment: independent, open source, open, modular and platform

Huawei mindstudio: mindstudio provides a user-friendly programming interface and graphic based debugging function, allowing the offline model to be automatically managed using the simulation environment

Support python, C + +, opencv

Support open-source platforms such as cafe and tensorflow

Hardware features: super computing power, ultra-low power consumption, ultra small volume:

•16 TOPS INT8 @ 30 W

•1 USB type-AB | 2 camera MIPI|

• 1 GE network port | 2 SD card slot

•8 GB memory

•6.5V-17VDC Wide voltage range power supply

•Operating temperature: 0°C to 75°C

•Dimensions: 100 x 75 x 24 mm

 demo, convenient for developers to learn and deploy training

A. Real time object recognition, target tracking and obstacle avoidance

B. Human motion recognition of humanoid robot in agent sense interaction, Lora wireless networking, cluster control

C. Machine vision direction recognition and path planning of maywell wheel car