Universal Picking Solutions for Industry 4.0
InPicker is especially suitable for all types of picking solutions for complex elements, with irregular shapes and multiple structures.
A single object, which can be in contact with other objects, but without being stacked.
3D training is required.
Position and orientation are approximately known.
It is the simplest version of the Picking problem.
Objects always come in similar positions with little variation
Ideal for blisters, pallets, etc.
RANDOM BIN PICKING
There are no restrictions when randomly stacking the pieces inside the bin
Risk of collision with the bin, the environment and other parts
Extraction of objects from a belt, either static or in motion
Objects are usually stacked on the belt
The environment is usually not complex
SEMI-STRUCTURED BIN PICKING
Parts are positioned in the bin with some degree of organization and predictability, although exact location may change
Simplifies the vision process to detect and inspect parts
Supported 3D Vision Technologies
It is a flexible solution to determine static 3D surfaces:
2 cameras are used in a controlled light environment.
Objects must have well defined textures.
Simplifies the mounting of cameras on the robot claw.
Process speed is very fast and allows to define the 3D map in total or in sections
Segments object structures with unknown shapes.
Generates a dense 3D image with two cameras and a specific light pattern projection (2D).
Allows to create definitions of point maps in motion. However, if high resolutions in Z are required, it is recommended that the scenes be static.
There are different versions of projections that condition the resolution in X, Y, Z.
Similar to texture projection, but the projected pattern is known.
Possible to project multiple different patterns
3D is calculated according to the deformation of this pattern.
Requires only one camera and one projector.
A system using a line laser as projection element and a monochrome camera as capture element.
Each captured frame generates a contour profile.
The movement of the camera or the object allows to accumulate consecutive profiles, generating a 3D image of the scanned object.
Allows to capture real time 3D images without movement.
A modulated light is sent by a transmitter that illuminates the scene on which it is intended to extract the 3D information.
The reflected light is detected by the smart sensor, which determines the time-of-flight for each pixel and the distance between pixel and object. In this way a 3D image is generated from the information of each camera pixel.