PreprintArticleVersion 2Preserved in Portico This version is not peer-reviewed
Generative Adversarial Approach to Urban Areas NDVI Estimation Using Information Exclusively from Structural and Textural Analysis of Panchromatic Orthoimagery: A Case Study of Łódź, Poland
Version 1
: Received: 1 February 2022 / Approved: 3 February 2022 / Online: 3 February 2022 (15:11:58 CET)
Version 2
: Received: 19 April 2022 / Approved: 19 April 2022 / Online: 19 April 2022 (10:11:20 CEST)
How to cite:
Adamiak, M.; Będkowski, K.; Bielecki, A. Generative Adversarial Approach to Urban Areas NDVI Estimation Using Information Exclusively from Structural and Textural Analysis of Panchromatic Orthoimagery: A Case Study of Łódź, Poland. Preprints2022, 2022020054. https://doi.org/10.20944/preprints202202.0054.v2
Adamiak, M.; Będkowski, K.; Bielecki, A. Generative Adversarial Approach to Urban Areas NDVI Estimation Using Information Exclusively from Structural and Textural Analysis of Panchromatic Orthoimagery: A Case Study of Łódź, Poland. Preprints 2022, 2022020054. https://doi.org/10.20944/preprints202202.0054.v2
Adamiak, M.; Będkowski, K.; Bielecki, A. Generative Adversarial Approach to Urban Areas NDVI Estimation Using Information Exclusively from Structural and Textural Analysis of Panchromatic Orthoimagery: A Case Study of Łódź, Poland. Preprints2022, 2022020054. https://doi.org/10.20944/preprints202202.0054.v2
APA Style
Adamiak, M., Będkowski, K., & Bielecki, A. (2022). Generative Adversarial Approach to Urban Areas NDVI Estimation Using Information Exclusively from Structural and Textural Analysis of Panchromatic Orthoimagery: A Case Study of Łódź, Poland. Preprints. https://doi.org/10.20944/preprints202202.0054.v2
Chicago/Turabian Style
Adamiak, M., Krzysztof Będkowski and Adam Bielecki. 2022 "Generative Adversarial Approach to Urban Areas NDVI Estimation Using Information Exclusively from Structural and Textural Analysis of Panchromatic Orthoimagery: A Case Study of Łódź, Poland" Preprints. https://doi.org/10.20944/preprints202202.0054.v2
Abstract
Generative adversarial networks (GAN) opened new possibilities for image processing and analysis. Inpainting, dataset augmentation using artificial samples or increasing spatial resolution of aerial imagery are only a few notable examples of utilizing GANs in remote sensing. This is due to a unique construction and training process expressed as a duel between GAN components. The main objective of the research is to apply GAN to generate an artificial Normalized Difference Vegetation Index (NDVI) using panchromatic images. The NDVI ground-truth labels were prepared by combining RGB and NIR orthophoto. The dataset was then utilized as input for a conditional generative adversarial network (cGAN) to perform an image-to-image translation. The main goal of the neural network was to generate an artificial NDVI image for each processed 256px × 256px patch using only information available in the panchromatic input. The network achieved 0.7569 ± 0.1083 Structural Similarity Index Measure (SSIM), 26.6459 ± 3.6577 Peak Signal-to-Noise Ratio (PNSR) and 0.0504 ± 0.0193 Root-Mean-Square Error (RSME) on the test set. The perceptual evaluation was performed to verify the usability of the method when working with a real-life scenario. The research confirms that the structure and texture of the panchromatic aerial remote sensing image contains sufficient information for NDVI estimation for various objects of urban space. Even though these results can be used to highlight areas rich in vegetation and distinguish them from urban background, there is still room for improvement in terms of accuracy of estimated values. The purpose of the research is to explore the possibility of utilizing GAN to enhance panchromatic images (PAN) with information related to vegetation. This opens interesting possibilities in terms of historical remote sensing imagery processing and analysis. The panchromatic orthoimagery dataset was derived from RGB orthoimagery.
Keywords
generative adversarial networks; NDVI; green areas; orthophoto; artificial datasets.
Subject
Environmental and Earth Sciences, Remote Sensing
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Received:
19 April 2022
Commenter:
Adam Bielecki
Commenter's Conflict of Interests:
Author
Comment: - Evaluation metrics calculation was changed to match visualizations. - Ambiguous statements were fixed and extended. - Minor language corrections were introduced.
Commenter: Adam Bielecki
Commenter's Conflict of Interests: Author
- Ambiguous statements were fixed and extended.
- Minor language corrections were introduced.