ML Anthology

Improving Object Detection in VHR Aerial Orthomosaics

Tanguy Ophoff, Kristof Van Beeck, Toon Goedemé
ECCVW 2022 pp. 268-282
doi:10.1007/978-3-031-25082-8_18 /eccvw/2022/ophoff2022eccvw-improving/

Abstract

In this paper we investigate how to improve object detection on very high resolution orthomosaics. For this, we present a new detection model ResnetYolo, with a Resnet50 backbone and selectable detection heads. Furthermore, we propose two novel techniques to post-process the object detection results: a neighbour based patch NMS algorithm and an IoA based filtering technique. Finally, we fuse color and depth data in order to further increase the results of our deep learning model. We test these improvements on two distinct, challenging use cases: solar panel and swimming pool detection. The images are very high resolution color and elevation orthomosaics, taken from plane photography. Our final models reach an average precision of 78.5% and 44.4% respectively, outperforming the baseline models by over 15% AP.

PDF ECCVW Semantic Scholar

Cite

Text

Ophoff et al. "Improving Object Detection in VHR Aerial Orthomosaics." European Conference on Computer Vision Workshops, 2022. doi:10.1007/978-3-031-25082-8_18

Markdown

[Ophoff et al. "Improving Object Detection in VHR Aerial Orthomosaics." European Conference on Computer Vision Workshops, 2022.](https://mlanthology.org/eccvw/2022/ophoff2022eccvw-improving/) doi:10.1007/978-3-031-25082-8_18

BibTeX

@inproceedings{ophoff2022eccvw-improving,
  title     = {{Improving Object Detection in VHR Aerial Orthomosaics}},
  author    = {Ophoff, Tanguy and Van Beeck, Kristof and Goedemé, Toon},
  booktitle = {European Conference on Computer Vision Workshops},
  year      = {2022},
  pages     = {268-282},
  doi       = {10.1007/978-3-031-25082-8_18},
  url       = {https://mlanthology.org/eccvw/2022/ophoff2022eccvw-improving/}
}