S2 is an embedded artificial intelligence super calculator, which is equipped with a mobile processor with the top computing power — NVIDIA Jetson TX2, and provides the acceleration SDK for mainstream artificial intelligence algorithm models, it is the most suitable high-performance platform for edge computing.
TA adopts highly integrated design, which has superior physical form, saves users’ tedious debugging steps, and is easy to install and operate, greatly lowering the development threshold of terminal intelligence.
It can be widely used in various terminal equipments such as unmanned aerial vehicles, unmanned vehicles, fixed camera lens and so on, and can be applied to automatic obstacle avoidance of unmanned aerial vehicles, automatic identification and tracking detection of people and vehicles, fire prevention and control, logistics transportation of unmanned vehicles, unmanned scale automatic weighing objects, environmental pollution management, high voltage line fault detection, pavement pit monitoring and other intelligent identification application fields.
The unique advantages of S2 enable enterprises, research institutes, government agencies or individuals to develop, research, learn, and apply AI.
S2 can be used in the research and development of intelligent applications of various terminal equipment, while S2 + unmanned aerial vehicle provides users with a choice of large area and efficient operation mode.
Unmanned aerial vehicle has a God’s perspective and can be observed in a wide range. It works flexibly and responds quickly. When it is installed with AI intelligent brain, it can bring users more efficient working experience.
At present, we provide customers with two solutions of “S2 + unmanned aerial vehicle” target recognition and tracking. To this end, our research and development personnel have been preparing for more than a year, testing and upgrading repeatedly, only to bring mature solutions to users and ensure safety and reliability.
First, look at the target recognition SDK:
The research and development team of the company independently develops the map marking software, and confirms and locates the target object based on visual recognition by connecting S2 and cooperating with the optical sensor of the unmanned aerial vehicle.
1. The UAV is equipped with S2 module and the cloud platform of the network interface, which can be identified with high definition (1080P) in real time and can reach more than 15 fps. The SDK adopts the latest acceleration framework and algorithm optimization model, which can be used for environmental detection, forest fire detection, road inspection, earthwork damage, vehicle detection, etc.
2. Record the location of the target in real time, support suspicious alarm, support viewing the detection result record.
3. Support data backup, and automatically back up the recognition result images and data to USB drive.
4. Support model re-training, re-cutting and optimization. When customers have targeted requirements, they can reduce the unused scale and run faster.
5. Cloud platform and picture transmission with network port can be matched, such as Huixing sea EH314, EH640 and other optical zoom cameras and VDC, Xlink30 picture transmission and so on.
The solution we recommend to users:
Second, look at the target tracking SDK:
S2 collects video information and flight attitude, communicates with flight control at the same time, and the flight control adjusts the position of the aircraft to track.
1. The software code is open-source, including ros/mavros interface, gazebo simulation interface, kcf target tracking algorithm, capsule control, capsule attitude data reading, PX4 target tracking code, supporting the secondary development and upgrade of tracking algorithm
2. Direct adaptation of PX4 unmanned system
3. Support ROS/Gazeb tracking algorithm 3D simulation
4. Position tracking target object
5. Three-axis capsule, high resolution, ROS/serial interface module, support the three-axis angle control of the capsule, speed control mode, pwm mode control support 10 times zoom control.
At present, the whole scheme of S2 and UAV is being tested and will meet you soon.
At present, most of the embedded development versions of intelligent applications of terminal devices on the market stay in the bulk state of modules and boards, which not only require a lot of time for debugging, but also have poor security, moreover, the compatibility with various terminal devices is also poor. The Birth of S2 directly saves the above trouble, and conducts AI development, research, learning, applications and so on save a lot of labor and time costs.
The combination of S2 and UAV directly opens up a broader market for UAV application, which also indicates that huixinghai has officially stepped into the AI intelligent application field. At present, Huixing Sea’s unmanned aerial vehicle detection scheme based on intelligent identification of people and vehicles has been officially listed, and it also accepts fire prevention and control, personnel search and rescue, environmental pollution management, engineering vehicle management, road pit monitoring and other intelligent identification UAV industry application scheme customization. Welcome to consult!
Supercombustion! S2 intelligent module goes online, endowing terminal devices with the ability of artificial intelligence