This project makes robots understand humans, makes them smarter. It develops AI based software that makes robots aware, adaptive to the environment, simplifies the way we interact with them.
The first product of this company is a software that extracts position and orientation of any object in 3D using simple camera. It has numerous applications in Industry 4.0. For example, sensing and measurement, robotics pick and place, bin picking, indoor navigation. This project outperforms the competitive active sensing solutions in performance, price, size and other factors.
This project has developed an artificial intelligence-based visual recognition software that simplifies human-computer interaction through robot perception and adaptation to the environment. The core technology algorithm can use any vision sensor (RGB, DEPTH, LIDAR) to monitor the essential position of the object in real time, thereby outputting an object with 6DOF (ID, Position and Orientation).
The product can obtain the position and orientation of the 3D object through the integrated camera. The applicable scenarios are sensing and measurement of objects, 3D recognition and positioning of disorderly stacked workpieces, automatic grabbing and placement by robots, indoor navigation, etc.
The core technology developed in this project can convert any camera (industrial camera, professional camera, smart phone or even network camera) into a 3D measurement device. It can detect objects and extract their position and orientation from a single image, a single camera. Provide 6 data for any object captured:
Orientation angle: A，B，C
With this technology, data can be acquired in Indoor navigation, Industrial robotics manipulators, and Industrial AR / VR with faster speed, higher accuracy, and lower cost.
Detect and provide the position and orientation of any object in a single image frame without additional image processing or object detection.
Any vision sensor, network camera, industrial camera, professional camera, smartphone and so on can be used.
The recognition accuracy is 0.1mm, which is more than 10 times higher than similar technologies. Adjusting the optical performance of the camera can identify objects with a size of 0.1mm to 100m. Can recognize translucent and shiny objects under different lighting conditions.
Small size, light weight, low power consumption, can be used and installed anywhere, such as toys, robotic arms, drones and manufacturing.
Use GPU for fast detection more than 30 times per second without 3D point cloud processing. The data input device delay time is only 0.03 seconds, which is more than 2 times faster than similar competing technologies.
The project is applied to Industry 4.0 in the form of an application program. It has entered the industrialization stage and is looking for strategic partners in the Chinese market and the establishment of a joint venture.
If you want to cooperate further with the project, please contact email@example.com.