Crop Rows Identification and Segmentation (Sugar Cane – Brazil)

This page contains our Orthomosaic Dataset of RGB aerial images for Crop Rows Identification and Segmentation and its Ground Truth (GT). It depicts a sugar sane plantation in Northern Brazil, captured as a single UAV-acquired sugarcane field orthomosaic. The image was captured from an Horus Aeronaves (https://horusaeronaves.com/) fixed-wing UAV employing a visible light RGB Canon G9X camera.  The UAV captured the data following a flight altitude of 125 to 200 meters, resulting in an approximate  spatial resolution of 5cm/pixel.

The images can be downloaded at full resolution by either accessing the direct download links under the illustrative low-resolution images below or by clicking at the images themselves.

Original Image

[Link for direct download]

Ground Truth

From the dataset an expert biologist produced a human-made ground truth (GT). The expert classified the whole image manually, using the GNU image manipulation program (GIMP). All pixels of the image were classified into two classes: crop row (green) and background (red).  Out-of-boundaries pixels are black.  This GT representation allows the direct usage of the GT in machine learning environments that employ the Pascal VOC standard.

[Link for direct download (ground truth image)]

Acknowledgements

We wish to thank Horus Aeronaves (https://horusaeronaves.com/) for acquiring and offering the dataset.

This work was the result of a collaborative effort of a team of engaged researchers:

• Alexandre Monteiro <xandemont@gmail.com>
• Paulo Cesar Pereira Junior <pcpereiraj@gmail.com>
• Antonio Carlos Sobieranski <asobieranski@gmail.com>
• Rafael da Luz Ribeiro <rl.ribeiro@outlook.com>

Citing this Dataset

@misc{CropRowsDataset2019,
  author = {Pereira Junior, P.C. and von Wangenheim, A.},
  title = {Orthomosaic Dataset of RGB aerial Images for Crop Rows Detection},
  year = {2019},
  publisher = {LAPIX/UFSC},
  journal = {INCoD Datasets Repository},
  howpublished = {\url{http://www.lapix.ufsc.br/weed-mapping-sugar-cane}}
}