3D POSITIONING SYSTEM
Product Center
The 3D point cloud trajectory guidance and positioning system locates the workpiece through industrial camera and visual software, realizes the trajectory guidance based on point cloud, and guides the manipulator to realize a certain process flow according to the required trajectory.
Product details
3D POSITIONING SYSTEM
System specification
-
High accuracy
-
High frame clearing rate
-
Strong flexibility
The 3D point cloud trajectory guidance and positioning system locates the workpiece through industrial camera and vision software to realize the trajectory guidance based on point cloud, that is, no matter what pose the object to be guided is placed (the pose is required to be within the travel range of the manipulator), the vision system can locate the object and guide the manipulator to realize a certain process flow according to the required trajectory.
System composition
-
3D industrial camera
Select the type and model of camera with corresponding field of view and accuracy according to the situation.
-
Roboeye.AI
The component of vision scheme automatically generates AI model based on deep learning algorithm for recognition and positioning, and plans the motion trajectory of the robot at the same time.
-
Controller
Visual scheme components.
-
3D industrial robot
Select different types of robots according to different processes.
-
Robot fixture
Design robot fixture according to workpiece shape and process.
-
Automatic station integration
In order to realize the process, the station is integrated with software and hardware.
Application scenarion
-
01.
Automobile glass gluing.
-
02.
Car front instrument panel guide cutting.
-
03.
Sole gluing.
-
04.
Mobile phone adhesive.
-
05.
Grinding process of castings.
-
06.
Drilling trajectory guidance.
-
07.
Weld trajectory tracking.
Core technology
| Object detection | Point cloud matching |
|---|---|
| AI model using simulated data (patent) | Original AI model (patent) |
| It does not need or only needs a small amount of annotation data, and the degree of automation of model making is high. | “ZERO” sample |
| Without the serious problem of over fitting of traditional deep learning algorithm, the anti-interference ability is improved. | It can effectively deal with and improve the recognition rate of cloud missing and thin parts. |
| High precision edge location through dynamic domain and shape prior knowledge modeling (patent) | The model can be generalized to any part without further training for different parts. |
| Through the automatic sample enhancement technology, the model can resist the unknown ambient light interference other than the situation contained in the training data. | |
| Through in-depth learning, model simplification and refining technology, model efficiency can be improved to a greater extent. | |
product display
Related Products
