Turfgrass Divot Dataset (Synthetic ) for divot detection object detection system

The dataset provided below has been synthetically created using Blender. A fundamental analysis on this data was conducted utilizing the YOLO V3 object detection technique to identify divots or areas of damage.

Used for paper: 

Advancing Turfgrass Maintenance with Synthetic Data for Divot Detection

https://github.com/stevefoy/Turfgrass-Divot-Object-Detection

@inproceedings{IMVIP2024,
    author = {Stephen Foy and Simon McLoughlin},
    title = {Advancing Turfgrass Maintenance with Synthetic Data for Divot Detection},
    booktitle = {Irish Machine Vision and Image Processing Conference (IMVIP)},
    year = {2024}
}

Contents of the Zip File:

• synthDivot_416x416 Folder:

  • Train and validation subfolders
  • 1200 RGB PNG images
  • Corresponding masks for each image
  • Bounding box data in YOLO .txt format

• synthDivot_608x608 Folder:

  • Train and validation subfolders
  • 1200 RGB PNG images
  • Bounding box data in YOLO .txt format