Logit-based Superpixel Semantic Segmentation of Images for Precision Agriculture

🧑 Corneliu Florea, Alexandru BURGHIU, Mihai Ivanovici
📅 2024

In this work we approach the problem of remote sensing image segmentation using a classical approach: the image is first segmented and, subsequently, each segment is labeled using a classifier. For segmentation, we rely on a superpixel framework and several methods are evaluated. For the classifier, again, several state-of-the-art algorithms are tested and performances are compared. The best performing method is obtained by a modified SEED superpixel algorithm with boosted trees for classification. The evaluation is carried out on the Agriculture-Vision database and the results are encouraging.

https://www.scientificbulletin.upb.ro/rev_docs_arhiva/fullf4b_945300.pdf

C. Florea, A. Burghiu, M. Ivanovici, Logit-based Superpixel Semantic Segmentation of Images for Precision Agriculture, Scientific Bulletin of Politehnica University, series C, vol. 86, issue 2, 2024,

@article{florealogit,
title={Logit-based Superpixel Semantic Segmentation of Images for Precision Agriculture},
author={FLOREA, Corneliu and BURGHIU, Alexandru and IVANOVICI, Mihai}
}