Datatypes
Datatypes
Datatypes are the strongly typed contracts every Skill and Agent reads and writes - from scalars and vectors to geometry, images, annotations, and scene objects. Because one Skill's output is another's typed input, they compose without conversion glue.
Scalars & Color
Primitive values and packed color used throughout the SDK.
Bool
Single boolean value.
View →Int
32-bit signed integer.
View →Float
32-bit floating-point number.
View →String
UTF-8 encoded text string.
View →Rgba32
Packed RGBA color representation.
View →Images
Image data and its format metadata.
Image
Image data as NumPy array or raw bytes.
View →ImageFormat
Image format metadata and encoding information.
View →Vectors, Matrices & Positions
Linear-algebra primitives for transforms, poses, and coordinates.
Vector2D
2D vector with x and y components.
View →Vector3D
3D vector with x, y, and z components.
View →Vector4D
4D vector with x, y, z, and w components.
View →Mat3X3
3×3 matrix for 2D transformations.
View →Mat4X4
4×4 matrix for 3D transformations and poses.
View →Position2D
2D position with x and y coordinates.
View →Position3D
3D position with x, y, and z coordinates.
View →Geometry
Points, boxes, and shapes in 2D and 3D.
Points2D
2D point cloud or array of 2D points.
View →Points3D
3D point cloud or array of 3D points.
View →ListOfPoints3D
List containing multiple 3D point clouds.
View →Boxes2D
2D bounding boxes with multiple format support.
View →Boxes3D
3D bounding boxes with poses and rotations.
View →LineStrips2D
2D polylines and contours.
View →Circles
2D circles defined by centers and radii.
View →Ellipses
2D ellipses with centers, axes, and angles.
View →Lines
2D lines represented in polar form (rho, theta).
View →Mesh3D
3D triangle mesh with vertices and faces.
View →Annotations
Detection and segmentation annotation contracts.
Categories
Category definitions for annotations.
View →ObjectDetectionAnnotations
Object detection and instance segmentation annotations.
View →PanopticSegmentationAnnotation
Panoptic segmentation with masks and segment information.
View →GeometryDetectionAnnotations
Geometric shape detection annotations for circles, ellipses, and lines.
View →Scene Objects
Higher-level industrial objects for picking and packing workflows.
Part
Part with ID, 6-DOF pose, dimensions, and state.
View →Tray
Tray container with ID, pose, dimensions, and slot count.
View →Bin
Bin container with ID, pose, and parts.
View →Where to Go Next?
Continue to the next tutorial.
Inter-Device Communication
Pub/sub and client/server messaging across devices with BabyROS.
Next tutorial →