robokudo.descriptors.analysis_engines.tiago_demo_from_storage¶

Analysis engine for TIAGo robot perception using stored data.

This module provides an analysis engine that demonstrates TIAGo robot perception capabilities using stored camera data. It implements a pipeline that combines preprocessing, object detection, and parallel feature analysis.

The pipeline implements the following functionality:

Reading stored TIAGo camera data from MongoDB
Sequential preprocessing stage
Parallel object analysis stage
Color analysis
PCA-based pose estimation
Optional query handling (commented out)

Note

This engine processes stored data from the TIAGo robot’s camera system, making it useful for testing and developing perception algorithms without requiring direct robot access.

Classes¶

AnalysisEngine

Analysis engine for TIAGo robot perception with stored data.

Module Contents¶

class robokudo.descriptors.analysis_engines.tiago_demo_from_storage.AnalysisEngine¶

Bases: robokudo.analysis_engine.AnalysisEngineInterface

Analysis engine for TIAGo robot perception with stored data.

This class implements a pipeline that processes stored TIAGo camera data using a combination of sequential and parallel processing stages. It demonstrates advanced pipeline structuring with parallel feature analysis.

The pipeline includes:

Sequential preprocessing stage * Collection reader for stored data * Image preprocessing
Sequential object detection stage * Crop point cloud * Detect plane * Extract clusters
Parallel feature analysis stage * Color analysis * PCA-based pose estimation

Note

The pipeline includes commented-out query handling components that can be enabled for interactive operation.

name()¶

Get the name of the analysis engine.

Returns:: The name identifier of this analysis engine
Return type:: str

implementation()¶

Create a pipeline for TIAGo perception with stored data.

This method constructs a processing pipeline that combines sequential and parallel stages for processing stored TIAGo camera data. The pipeline is structured to optimize processing by running compatible analyses in parallel.

Pipeline structure:

Preprocessing sequence * Read stored data * Preprocess images
Object detection sequence * Crop point cloud * Detect plane * Extract clusters
Parallel analysis * Color analysis * PCA pose estimation

Returns:: The configured pipeline with parallel processing
Return type:: robokudo.pipeline.Pipeline