robokudo.types.annotation¶

Annotation types for robokudo.

This module provides various annotation types used for object classification, semantic information, location data, geometric shapes, and pose information. All annotation types inherit from the base Annotation class.

Classes¶

`Classification`	Classification annotation for objects.
`SemanticColor`	Semantic color annotation.
`LocationAnnotation`	Location annotation for objects.
`Plane`	Plane annotation for surface detection.
`Shape`	Base class for shape annotations.
`Cuboid`	Cuboid shape annotation.
`Sphere`	Sphere shape annotation.
`ColorHistogram`	Color histogram annotation.
`PoseAnnotation`	Pose annotation combining transform and annotation functionality.
`PositionAnnotation`	Position annotation combining position and annotation functionality.
`StampedPoseAnnotation`	Timestamped pose annotation.
`StampedPositionAnnotation`	Timestamped position annotation.
`StampedTransformAnnotation`	Timestamped transform annotation.
`Encoding`	An abstract Encoding Type.
`BoundingBox3DAnnotation`	3D bounding box annotation.
`CloudAnnotation`	Point cloud annotation.
`SpatiallyNearestAnnotation`	Annotation to describe the spatially nearest object to the camera.

Module Contents¶

class robokudo.types.annotation.Classification¶

Bases: robokudo.types.core.Annotation

Classification annotation for objects.

This class represents a classification result, including the type of classification (instance, class, or shape), the class name, confidence score, and class ID.

Variables:

classification_type – Type of classification (INSTANCE, CLASS, or SHAPE)
classname – Name of the classified class
confidence – Confidence score of the classification
class_id – Numeric identifier for the class

classification_type = ''¶

classname = ''¶

confidence = 0.0¶

class_id = 0¶

class robokudo.types.annotation.SemanticColor¶

Bases: robokudo.types.core.Annotation

Semantic color annotation.

This class represents semantic color information, including the color name and its ratio/proportion in the annotated object or region.

Variables:

color – Name of the color
ratio – Proportion/ratio of this color in the object/region

color = ''¶

ratio = 0.0¶

class robokudo.types.annotation.LocationAnnotation¶

Bases: robokudo.types.core.Annotation

Location annotation for objects.

This class represents a named location in the environment.

Variables:: name – Name of the location

name = ''¶

class robokudo.types.annotation.Plane¶

Bases: robokudo.types.core.Annotation

Plane annotation for surface detection.

This class represents a detected plane in 3D space, including its model parameters and inlier points.

Variables:

model – 4-dimensional plane equation parameters [a, b, c, d] for ax + by + cz + d = 0
inliers – List of point indices that belong to this plane

model = []¶

inliers = []¶

class robokudo.types.annotation.Shape¶

Bases: robokudo.types.core.Annotation

Base class for shape annotations.

This class serves as a base for specific shape types like Cuboid and Sphere.

Variables:

inliers – List of point indices that belong to this shape
type – Type of the shape (e.g., ‘Cuboid’, ‘Sphere’)

inliers = []¶

type = ''¶

class robokudo.types.annotation.Cuboid¶

Bases: Shape

Cuboid shape annotation.

This class represents a cuboid shape defined by three plane equations.

Variables:: model – Three plane equations defining the cuboid, shape (3,4)

model = []¶

class robokudo.types.annotation.Sphere¶

Bases: Shape

Sphere shape annotation.

This class represents a sphere shape defined by its center and radius.

Variables:

radius – Radius of the sphere
center – Center coordinates of the sphere

radius = 0.0¶

center¶

class robokudo.types.annotation.ColorHistogram¶

Bases: robokudo.types.core.Annotation

Color histogram annotation.

This class represents a 2D color histogram, typically containing hue and saturation information.

Variables:

hist – 2D histogram array
normalized – Whether the histogram is normalized

hist = None¶

normalized = False¶

class robokudo.types.annotation.PoseAnnotation¶

Bases: robokudo.types.tf.Pose, robokudo.types.core.Annotation

Pose annotation combining transform and annotation functionality.

This class inherits from both Pose and Annotation to provide pose information as an annotation type.

class robokudo.types.annotation.PositionAnnotation¶

Bases: robokudo.types.tf.Position, robokudo.types.core.Annotation

Position annotation combining position and annotation functionality.

This class inherits from both Position and Annotation to provide position information as an annotation type.

class robokudo.types.annotation.StampedPoseAnnotation¶

Bases: robokudo.types.tf.StampedPose, robokudo.types.core.Annotation

Timestamped pose annotation.

This class combines timestamped pose information with annotation functionality.

class robokudo.types.annotation.StampedPositionAnnotation¶

Bases: robokudo.types.tf.StampedPosition, robokudo.types.core.Annotation

Timestamped position annotation.

This class combines timestamped position information with annotation functionality.

class robokudo.types.annotation.StampedTransformAnnotation¶

Bases: robokudo.types.tf.StampedTransform, robokudo.types.core.Annotation

Timestamped transform annotation.

This class combines timestamped transform information with annotation functionality.

class robokudo.types.annotation.Encoding¶

Bases: robokudo.types.core.Annotation

An abstract Encoding Type.

This class represents various types of encodings such as feature vectors, latent space representations, or other variables.

Variables:: encoding – The encoded representation

encoding = None¶

class robokudo.types.annotation.BoundingBox3DAnnotation¶

Bases: robokudo.types.cv.BoundingBox3D, robokudo.types.core.Annotation

3D bounding box annotation.

This class combines 3D bounding box functionality with annotation capabilities.

class robokudo.types.annotation.CloudAnnotation¶

Bases: robokudo.types.cv.Points3D, robokudo.types.core.Annotation

Point cloud annotation.

This class combines 3D point cloud functionality with annotation capabilities.

class robokudo.types.annotation.SpatiallyNearestAnnotation¶

Bases: robokudo.types.core.Annotation

Annotation to describe the spatially nearest object to the camera. Supposed to be unique for an amount of objects or humans.