OverflowError in Interactive Camera-Lidar Calibrator Causing Segmentation Fault

[jahanifaruga]

Checklist

• I've read the contribution guidelines.
• I've searched other issues and no duplicate issues were found.
• I'm convinced that this is not my fault but a bug.

Description

I’m using the interactive_camera_lidar_calibrator with a FLIR Blackfly S camera (monochrome) and a Velodyne VLP‑16. Whenever I select a point in RViz, the calibrator crashes with a “libshiboken: Overflow” warning followed by a segmentation fault.

Below is the console log:

ros2 run interactive_camera_lidar_calibrator interactive_calibrator
--ros-args -r pointcloud:=/velodyne_points
-r image:=/flir_camera/image_raw
-r camera_info:=/flir_camera/camera_info
-p camera_frame:=flir_camera
-p can_publish_tf:=false
-p use_calibration_api:=false
-p use_compressed:=false

[...]
Overflow: Value -2989252461 exceeds limits of type [signed] "i" (4bytes).
QPaintDevice: Cannot destroy paint device that is being painted
[ros2run]: Segmentation fault

Expected behavior

The calibrator should let me pick 3D↔2D correspondences without crashing, even if some lidar points are out of view.

Actual behavior

It throws “Overflow” errors in image_view.py:792 from large projected coordinates and segfaults. The calibrator window closes immediately.

Steps to reproduce

1. Run the camera (Spinnaker driver) and lidar (velodyne_all_nodes).
2. Launch interactive_camera_lidar_calibrator.
3. In RViz, pick “Publish Point” and click any point in the cloud.
4. Calibrator segfaults with the OverflowError.

Versions

OS: Ubuntu 22.04
ROS 2: Humble
Autoware: main branch
Camera: BFS-U3-200S6M-C (25 mm lens)

Possible causes

Not sure.

Additional context

No response