A pre-configured Blender scene designed to generate synthetic training datasets for machine learning object detection tasks. The setup uses Blender's built-in particle system to scatter a target object alongside distractor items, introducing lighting variations and natural occlusions to improve model robustness.

The package includes the Bl-Vision add-on, which automatically generates bounding box labels compatible with YOLO models during the rendering process. Users can replace the default object with their own model to quickly generate hundreds of annotated images for fine-tuning computer vision algorithms without writing custom code.