robokudo.descriptors.analysis_engines.object_knowledge_from_storage

Analysis engine for object detection and knowledge visualization from stored data.

This module provides an analysis engine that demonstrates object detection and knowledge visualization using stored camera data. It combines static object detection with various visualization components to display object hypotheses, knowledge, and color information.

The pipeline implements the following functionality: - Reading stored camera data from MongoDB - Image preprocessing - Static object detection with predefined parameters - Visualization of object hypotheses - Visualization of object knowledge - Color analysis of detected objects

Note

This engine uses predefined object detection parameters optimized for a specific use case (mug detection). Adjust the parameters for other objects or scenarios.

Classes

AnalysisEngine

Analysis engine for object detection and knowledge visualization.

Module Contents

class robokudo.descriptors.analysis_engines.object_knowledge_from_storage.AnalysisEngine

Bases: robokudo.analysis_engine.AnalysisEngineInterface

Analysis engine for object detection and knowledge visualization.

This class implements a pipeline that combines static object detection with various visualization components. It processes stored camera data to detect objects and visualize their properties and associated knowledge.

The pipeline includes: - Collection reader for accessing stored data - Image preprocessing - Static object detection with predefined parameters - Object hypothesis visualization - Object knowledge visualization - Color analysis and visualization

Note

The static object detector is configured with specific parameters for mug detection. These parameters include bounding box dimensions and pose information that should be adjusted for different objects.

name()

Get the name of the analysis engine.

Returns:

The name identifier of this analysis engine

Return type:

str

implementation()

Create a pipeline for object detection and knowledge visualization.

This method constructs a processing pipeline that reads stored camera data, performs object detection with predefined parameters, and visualizes the results including object hypotheses, knowledge, and color information.

The static object detector is configured with specific parameters for a mug: - Bounding box: 397x126 pixels with size 49x106 - Position: (0.202, -0.109, 1.096) - Rotation: Quaternion (0.575, 0.666, -0.360, 0.310)

Returns:

The configured pipeline for object detection and visualization

Return type:

robokudo.pipeline.Pipeline