#356:
H. AliAkbarpour,
K. Gao,
R. Aktar,
S. Suddarth, and
K. Palaniappan
High-Throughput Crop Phenotyping,
Springer International Publishing,
pgs. 55--69,
2021
Abstract,
Bibtex,
PlainText,
URL,
DOI,
Google Scholar
This work presents a 3D-enabled method to register aerial image sequences. Our approach is based on a novel Bootstrapped Structure-from-Motion (BSfM) followed by analytical homography reprojection or georegistration. BSfM is a fast and robust method to recover the 3D exterior orientation of the camera poses and the scene 3D structure from image sequences. The recovered 3D parameters are used in an analytical approach to estimate homography matrices that project the input images onto the dominant ground plane in the scene to produce a global mosaic of the field and plants. Preliminary experimental results validate the approach and show satisfactory results suitable for scaling up to support high-throughput field phenotyping (HTP) for agricultural crop field experiments.
@inbook{AliAkbarpour_BSfM_Bookchapter_2021,
author = "H. AliAkbarpour and K. Gao and R. Aktar and S. Suddarth and K. Palaniappan",
title = "Structure from motion and mosaicking for high-throughput field-scale phenotyping",
year = 2021,
booktitle = "High-Throughput Crop Phenotyping",
publisher = "Springer International Publishing",
pages = "55--69",
keywords = "structure-from-motion (sfm), video stabilization, image registration, orthorectification, mosaicking",
doi = "10.1007/978-3-030-73734-4_4",
url = "https://doi.org/10.1007/978-3-030-73734-4_4"
}
H. AliAkbarpour, K. Gao, R. Aktar, S. Suddarth, and K. Palaniappan. Structure from motion and mosaicking for high-throughput field-scale phenotyping. High-Throughput Crop Phenotyping, Springer International Publishing, pages 55--69, 2021.