robokudo.annotators.simple_yolo_annotator

Classes

SimpleYoloAnnotator

Base class for all RoboKudo annotators.

Module Contents

class robokudo.annotators.simple_yolo_annotator.SimpleYoloAnnotator(name: str = 'SimpleYoloAnnotator')

Bases: robokudo.annotators.core.BaseAnnotator

Base class for all RoboKudo annotators.

This class provides core functionality for CAS access, tree structure navigation, and GUI result handling. It serves as the foundation for implementing annotators in the RoboKudo framework.

model

The YOLO model instance.

id2name

The YOLO models id to name map.

roi_comparator

A comparator for ROIs used to associate YOLO and RoboKudo ROIs.

update() → py_trees.common.Status

Run YOLO inference on the cas color image and combine classifications with existing object hypotheses.

Runs objects detection on the image and uses the RoiComparator find the closest bounding box created for an ObjectHypothesis by previous annotator. If there is a similar ROI present in an ObjectHypothesis, a Classification is attached to it.

static add_to_image(obj: robokudo.types.scene.ObjectHypothesis, classification: robokudo.types.annotation.Classification, image: numpy.typing.NDArray[numpy.uint8]) → numpy.typing.NDArray[numpy.uint8]

Add the object hypothesis along with the classification name and confidence to the visualization image.