Orthorectified images of Poaceae species, specifically Lolium perenne L. (perennial ryegrass) and Phleum pratense L. (timothy grass), are presented in this dataset. The original images were captured using high-resolution RGB cameras, yielding an average size of 2736 x 2736 pixels per image. To correct for the “fisheye” distortion introduced by the APEMAN 60 camera and to ensure accurate spatial representation, the images underwent an orthorectification process, resulting in a standardized size of 912 x 912 pixels. This dataset is designed for use in deep learning applications, such as the classification and detection of weeds within Poaceae grass species.

[Methods for processing the data] The images are ortho-rectified using a vertical plane through calculation of a homography matrix, supported with the RANSAC algorithm (Vincent and Laganiére, 2001). Vincent, E. and R. Laganiére. 2001. Detecting planar homographies in an image pair. In Image and Signal Processing and Analysis ISPA 2001. Proceedings of the 2nd International Symposium on, pp. 182–187. IEEE.

This research was funded by Grant PID2020-113229RBC43/AEI/https://doi.org/10.13039/501100011033.

[Description of methods used for collection/generation of data] Handheld camera (APEMAN A60)

Ortho_rye, Ortho_timothy

Peer reviewed