Skip to content

MultiSensor Step Scanning (LiDAR + Camera)

Namespace: AWSIM.Scanning
Script: MultiSensorStepScanning.cs

A unified, step-based scanning system for LiDAR and camera sensors in Unity.
Designed for AWSIM / RGLUnityPlugin, this tool enables synchronized LiDAR and camera captures on a per-step basis, useful for dataset generation, ground-truth logging, or robotic simulation. The most important aspect of this component is ensuring that all elements are captured simultaneously, preventing any blurring or misalignment in the camera image.

🧭 Overview

MultiSensorStepScanning coordinates LiDAR and camera captures frame-by-frame, ensuring precise, motion-free outputs.
Each “step” freezes the Unity simulation (Time.timeScale = 0), triggers all sensors, and saves outputs before resuming time.
The script also supports ground truth logging of detected objects via configured GroundTruthArea components.

Key Features

  • LiDAR Scanning
  • Builds RGL subgraphs per LiDAR sensor.
  • Supports local ROS axes or global ROS world transformations.
  • Optional voxel downsampling for .pcd output files.

  • Saves results under Assets/{outputDirectoryRoot}/Lidar.

  • Camera Scanning

  • Captures one image per configured camera each step.
  • Supports existing custom capture components or built-in fallback capture.
  • Configurable resolution, supersampling, and image format (.png / .jpg).

  • Saves images under Assets/{outputDirectoryRoot}/Camera.

  • Ground Truth CSV Logging

  • Aggregates visible objects from assigned GroundTruthArea volumes.
  • Logs object name, UUID, ROS-space position, yaw, and step index.

  • CSV outputs saved under Assets/{outputDirectoryRoot}/GroundTruthPos.

  • Triggering

  • Manually via Space key.
  • Or programmatically via TriggerAndSaveStep() / coroutine variant.

⚙️ How to Use

1. Add the Script

Attach MultiSensorStepScanning to an empty GameObject in your Unity scene.

2. Configure Sensors

In the Inspector panel:

LiDAR

  • Assign all GameObjects containing LidarSensor components.
  • Choose coordinate frame mode:
  • Local (default, ROS axes)
  • Global (transforms using worldOriginROS)
  • Enable/disable downsampling and adjust leaf size.

Camera

  • Assign all GameObjects with Camera components.
  • Optionally define:
  • Override resolution and supersampling.
  • Image format (PNG/JPG) and quality.
  • Custom capture component name (e.g., "MyCustomCameraFeature").

Ground Truth (optional)

  • Add one or more GroundTruthArea volumes.
  • Optionally specify a rosReference transform to define relative frame.
  • Enable CSV header writing if desired.

Output

  • Choose root folder name under Assets/.
  • Enable or disable per-sensor subfolders.
  • The step index will auto-increment after each capture.
alt text

⌨️ Triggering a Step

Manual: Press Space in Play Mode.

Programmatic: 

var scanner = FindObjectOfType<MultiSensorStepScanning>();
scanner.TriggerAndSaveStep();