robokudo.descriptors.analysis_engines.demo_from_storage

Analysis engine for demonstrating tabletop segmentation using stored data.

This module provides an analysis engine that processes stored camera data to perform tabletop segmentation. It reads data from MongoDB storage and applies a sequence of annotators for image preprocessing and object detection.

The pipeline consists of the following steps:

  1. Reading stored camera data

  2. Image preprocessing

  3. Point cloud cropping

  4. Plane detection

  5. Point cloud cluster extraction

Note

This demo uses the MongoDB storage interface and requires a properly configured MongoDB database with stored camera data.

Classes

AnalysisEngine

Analysis engine for demonstrating tabletop segmentation from stored data.

Module Contents

class robokudo.descriptors.analysis_engines.demo_from_storage.AnalysisEngine

Bases: robokudo.analysis_engine.AnalysisEngineInterface

Analysis engine for demonstrating tabletop segmentation from stored data.

This class implements a basic pipeline for tabletop segmentation using stored camera data. It reads data from MongoDB storage and processes it through a sequence of annotators to detect and segment objects on a table surface.

The pipeline includes:

  • Collection reader for accessing stored data

  • Image preprocessing for data preparation

  • Point cloud cropping to focus on relevant regions

  • Plane detection for table surface identification

  • Point cloud cluster extraction for object segmentation

Note

The pipeline is configured to use MongoDB storage by default and requires proper database configuration.

name()

Get the name of the analysis engine.

Returns:

The name identifier of this analysis engine

Return type:

str

implementation()

Create a basic pipeline for tabletop segmentation.

This method constructs the processing pipeline by configuring and connecting the necessary annotators in sequence.

Returns:

The configured pipeline for tabletop segmentation

Return type:

robokudo.pipeline.Pipeline