RVL Depth Decoder

This Foxglove extension registers a schema message converter that turns sensor_msgs/CompressedImage messages published with <encoding>; compressedDepth rvl payloads into standard sensor_msgs/Image frames. Only the RVL codec described in compressed_depth_image_transport is handled today.

Features

• Detects compressed depth messages that advertise the RVL codec.
• Mirrors the RVL bitstream decoder from the ROS encoder, supporting 16UC1 and 32FC1 output encodings (32F output is reconstructed from the quantized inverse-depth fields embedded in the compressed payload).
• Registers converters for both ROS 1 (sensor_msgs/…) and ROS 2 (sensor_msgs/msg/…) schema names so either dialect can be decompressed.

Usage

1. Install and build the extension:

   npm install
   npm run local-install

   Restart or reload Foxglove so it picks up the new extension bundle.

2. Load a bag or live connection that publishes RVL CompressedImage messages, e.g. /camera/aligned_depth_to_color/image_raw/compressedDepth.

3. When subscribing to the topic inside a panel, choose Convert to → sensor_msgs/Image (or sensor_msgs/msg/Image). Foxglove will invoke the converter and surface a regular depth image topic that downstream panels can consume.

The converter validates the RVL header (quantization parameters, rows, and columns) before decoding. If the payload is too small, malformed, or does not declare compressedDepth rvl, the converter returns undefined and Foxglove falls back to the original message.

Developing & Packaging

Standard extension workflows still apply:

• npm run build bundles the code.
• npm run local-install copies the build into ~/.foxglove-studio/extensions.
• npm run package creates a distributable .foxe archive (set the name, publisher, version, and description fields in package.json before shipping).