Terminology

Before using the software, please understand the following concepts to better use the software.

  • Project solution

    Project solution, referred to as Solution, is a comprehensive solution for implementing 3D vision-guided projects and can handle one or more grasping scenarios. A "project solution" consists of one or more "workspaces".

  • Workspace

    Workspace, referred to as Space, is a specific configuration for a specific grasping scenario, including scene configurations such as robots and cameras, as well as grasping configurations such as vision algorithms and motion planning.

  • ROI

    ROI (Region of Interest) is a specific area in the data. Setting ROI can improve processing efficiency and accuracy, save resources, etc.

  • Hand-eye calibration

    Hand-eye calibration refers to calculating the conversion relationship between the robot coordinate system and the camera coordinate system by collecting the calibration object pose and robot pose captured by the camera. Using the calculated transformation matrix, the target point in the camera coordinate system can be converted to the robot coordinate system, thereby realizing the operation of the robot under 3D vision guidance. The accuracy of hand-eye calibration is crucial and will directly affect the accuracy and stability of the system.

  • Tools

    The end-effector tool is a device specially designed and installed at the mechanical interface to enable the robot to complete its task, such as a gripper or suction cup.

  • Scene objects

    Scene objects refer to various objects in the real robot working scene, generally including pallets, material bins, brackets, etc.

  • Path planning

    Path planning refers to planning the robot path between the "start planning point" and the "grasp point".

  • Octree resolution

    Octree resolution refers to the side length of a small cube (voxel) centered at the midpoint of the point cloud. In the collision detection process, the point cloud itself is not directly involved, but is achieved by converting the point cloud into an octree. If these voxels collide with other objects, it is considered that the point cloud has collided. By adjusting the octree resolution parameter, the side length of the voxel can be changed, thereby affecting the accuracy and efficiency of collision detection.