Glossary

The terms a new user meets in these docs, in alphabetical order. The terms apply to both Bin Picking Studio (BPS) and Locator Studio (LS) unless marked otherwise. Each term links to the page that explains it.

Action request: A request that the robot controller sends to the Studio through the Action Request Communication Channel, for example a scan or trajectory request. The Studio executes the request and sends a response back to the robot controller. See Action requests and Communication channels.
Action Request Server port (ARS port): The port of the TCP server that the vision controller creates and to which the robot controller connects to send action requests. The server runs on a default port; leave the Action Request Server port (ARS port) field on the Network page empty unless you need a custom port. A custom port must also be set on the robot side, in the Robot module. See Robot interface.
Bin picking routine (BPS): The sequence of trajectories and operations executed by the robot when the bin picking procedure is called in the robot program. It begins in the Start pose and ends in the End pose with the picked object attached to the gripper. See Grasping methods.
Binpicking memory (BPS): The bin picking memory holds data about the localized objects, including IDs, pick-failed counters and path plans. The Binpicking memory settings section defines the path planning memory, and an initialization request can clear it. See Initialization request actions.
Bounding box: A region of interest for the localization. It is used to localize objects only in defined areas of the scene, or to eliminate false positives outside this region. Using a bounding box is optional; without one, localization searches the whole point cloud. See Bounding boxes.
Calibration: The procedure that finds the transformation between camera space (the origin of the 3D sensor’s point cloud) and the target space, usually robot space. The result is the calibration matrix, which is applied to the coordinates of each localized object. Calibration is specific for each vision system. See What is calibration.
Calibration space: The coordinate space a vision system is calibrated to, chosen together with the sensor position in the Calibration space and sensor position field of the vision system. The options are robot space with an external fixed sensor (extrinsic calibration), robot space with the sensor on the robotic flange (hand-eye calibration), and marker space with a single sensor position (LS only). See Vision system configuration.
Deployment: Starting a fully configured solution, so that your application runs. A solution can be deployed in Production or Simulation mode. See Deployment page.
Environment (BPS): The setup step in which you build a 3D model of the robotic cell, define the robot’s working space and test its movement. The model can be checked against a scan from a calibrated sensor. See Environment and Scene.
Extrinsic calibration: Robot-camera calibration for a 3D sensor mounted in a fixed position, usually above the bin. The calibration matrix transforms directly to the robot base coordinate system. The sensor must not move relative to the robot after calibration. See Extrinsic calibration.
Fast scan: A sensor mode that captures only the texture frame. The positions of objects are recalled from memory and only the presence of each object is checked in the texture frame, so full scans still need to be triggered regularly. Available for CAD Picking and CAD Localization. See Fast scan.
Global Vision system settings: Settings on the solution’s Settings page that apply to all vision systems unless overridden. Every vision system can override a whole settings section with its own values. See Settings.
Grasping method (BPS): Defines the bin picking routine: the waypoints, the trajectories between them and the actions taken at each waypoint. A solution can have up to 5 grasping methods, and each vision system uses one of them. See Grasping methods.
Gripper (BPS): The tool on the robot that grasps the picked object. Its CAD model is needed to calculate valid paths with real collision avoidance, and the tool point is defined on it. See Gripper.
Gripping point (BPS): A point on the picked object where the gripper can grasp it, defined with respect to the origin of the object’s CAD model. During grasping, the tool point and a gripping point are aligned. Gripping points are only defined in CAD Picking solutions; in AI Picking, BPS finds them automatically. See Gripping point.
Hand-eye calibration: Robot-camera calibration for a 3D sensor mounted on the robotic arm, behind the last joint. The calibration matrix transforms to the robot’s TCP. The robot must not move during the scan acquisition. See Hand-eye calibration.
Invariance (BPS): A rotation of the picked object (for a gripping point) or of the gripper (for the tool point) around an axis that gives the same grasp. Invariances let one gripping point or tool point describe several grasping configurations. Redundant invariances reduce bin picking performance. See Invariance.
Localization engine: The part of the Studio that finds objects in the point cloud. There are three engines, each enabled by a software module: the CAD-based, the neural network-based (AI) and the layer-based localization engine. See Localization engines and Software modules.
Marker space calibration (LS): Marker-camera calibration that uses a marker pattern to define a common coordinate space for the camera and the robot. The vision system is calibrated to marker space with a single calibration point, and a custom coordinate space with its origin in the marker origin is defined on the robot controller. See Marker space.
MultiView: A module for CAD localization that captures scans of the scene from several viewpoints, merges them and runs the localization on the merged point cloud. It works either with several statically mounted sensors (Static MultiView) or with one sensor on the robotic arm (Hand-eye MultiView). See MultiView Localization.
Picked object: The object that the robot picks, given as a 3D CAD model (STL or PLY) on the Picked object page of the solution. Localization uses this model to find the objects in the point cloud. Only CAD-based solution types have this step. See Mesh tools and Reported pose.
Picking path stage (BPS): One row of the grasping method table. It combines the settings of a waypoint and of the trajectory leading to it. The mandatory stages are Start, Approach, Grasp, Deapproach and End. See Configuration overview.
Robot interface: The part of the vision controller that communicates with the Robot module on the robot controller. It is configured on the Network page. See Robot interface and Robot module.
Robot module: The software installed on the robot controller to communicate with the Studio. It contains the Photoneo robotic API, example programs and an integration guide for the robot brand. See Robot communication overview.
Robot State Server port (RSS port) (BPS): The port of the TCP server that the Robot module creates and to which the vision controller connects to read the robot state data. Leave the Robot State Server port (RSS port) field empty to use the default port. A custom port must also be set in the Robot module. See Robot interface.
Scanning profile: A set of scanning parameters of the 3D sensor. Each vision system can use a different scanning profile, so that the scan quality suits the objects it localizes. See Scanning profiles.
Scene (BPS): The 3D model of the real robotic cell, built in the Scene tab of the Environment page. It must match the real world at least within the reach of the robot, because the sensor never sees every object in the cell. See Scene.
Scene state (BPS): One form of the robotic cell. If the cell changes during bin picking, you can define up to four scene states and assign objects to them. The robot switches between scene states with the Change scene state request. See Scene states.
Solution: A configuration that you create and manage on the Solutions page. Each solution has a solution type, a name and a Unique ID, which the robot uses in its requests. See Solutions page.
Solution type: Chosen when you create a solution. It decides the product (BPS or LS), the localization method and what is reported to the robot. There are five types: CAD Localization, AI Localization, Layer Localization, CAD Picking and AI Picking. See Solution types.
Tool point (BPS): The point on the gripper that BPS uses instead of the TCP defined on the robot controller. It is defined with respect to the origin of the gripper’s CAD model, which must be where the gripper is mounted on the robot flange. See Tool point.
Vision system: A unit of a solution that uses one 3D sensor (or several, with Static MultiView) with its own calibration, scanning profile and settings. Most action requests contain the ID of the vision system they apply to. See 3D sensors and scanning profiles.
Waypoint (BPS): A point in space that the robot’s tool point moves through during the bin picking routine. You can define its position and a procedure to run when the robot reaches it. See Grasping methods.