Synthetic Brick Dataset for Object Detection and 6-DoF Pose Estimation

The synthetic data generator leverages open-source tools, including Blender-based Python libraries such as bpy, geometry_script, and blenderproc, to create large datasets of rendered images with randomised parameters. In this work, the generator is configured to model the specific brick type used in the real-world experiments, with physical dimensions of 240 × 115 × 50 mm and a light reddish surface texture, alongside synthetic mortar using randomised noise textures, with controlled variation in coverage and coarseness.     «

The synthetic data generator leverages open-source tools, including Blender-based Python libraries such as bpy, geometry_script, and blenderproc, to create large datasets of rendered images with randomised parameters. In this work, the generator is configured to model the specific brick type used in the real-world experiments, with physical dimensions of 240 × 115 × 50 mm and a light reddish surface texture, alongside synthetic mortar using randomised noise textures, with controlled variation in...     »