solidity_rag_merge.py

What is this?

../../_images/solidity_rag_merge.png

(red region in top image is segmented using euclidean clustering, yellow, purple and red region is segmented by solidity_rag_merge.)

Segment image region which has high solidity, using RAG. It firstly segments image with SLIC Superpixels, generates RAG based on solidity, and then merges regions.

Subscribing Topic

  • ~input (sensor_msgs/Image)

    Raw image. In most cases, this is mask or depth image.

  • ~input/mask (sensor_msgs/Image)

    Region where the label should be ignored.

Publishing Topic

Default Topics

  • ~output (sensor_msgs/Image, encoding: 32SC1)

    Label image of region with high solidity.

Optional Topics

Following topics are published when ~debug is True, see Parameters for more detail.

  • ~debug/slic (sensor_msgs/Image, encoding: 32SC1)

    Label image segmented with SLIC Superpixels.

  • ~debug/rag (sensor_msgs/Image)

    RAG overlayed image.

  • ~debug/label_viz (sensor_msgs/Image)

    Label color overlayed image. This function can be replaced by label_image_decomposer.py but I added for easy debugging.

Parameters

  • ~approximate_sync (Bool, default: False)

    Whether to use approximate for input topics.

  • ~debug (Bool, default: True)

    Debug mode. In this mode, the node publishes ~debug/* topics.