Towards On-Board Hyperspectral Satellite Image Segmentation: Understanding Robustness of Deep Learning through Simulating Acquisition Conditions
| Autor: | Michal Kawulok, Tomasz Lakota, Jakub Nalepa, Marcin Cwiek, Michal Myller, Lukasz Zak, Lukasz Tulczyjew |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2021 |
| Předmět: |
010504 meteorology & atmospheric sciences
Computer science Science 0211 other engineering and technologies convolutional neural network 02 engineering and technology computer.software_genre 01 natural sciences Convolutional neural network Robustness (computer science) hyperspectral image analysis atmospheric correction Segmentation classification segmentation deep learning on-board processing noise 021101 geological & geomatics engineering 0105 earth and related environmental sciences business.industry Deep learning Atmospheric correction Hyperspectral imaging Image segmentation General Earth and Planetary Sciences Noise (video) Data mining Artificial intelligence business computer |
| Zdroj: | Remote Sensing; Volume 13; Issue 8; Pages: 1532 Remote Sensing, Vol 13, Iss 1532, p 1532 (2021) |
| ISSN: | 2072-4292 |
| DOI: | 10.3390/rs13081532 |
| Popis: | Although hyperspectral images capture very detailed information about the scanned objects, their efficient analysis, transfer, and storage are still important practical challenges due to their large volume. Classifying and segmenting such imagery are the pivotal steps in virtually all applications, hence developing new techniques for these tasks is a vital research area. Here, deep learning has established the current state of the art. However, deploying large-capacity deep models on-board an Earth observation satellite poses additional technological challenges concerned with their memory footprints, energy consumption requirements, and robustness against varying-quality image data, with the last problem being under-researched. In this paper, we tackle this issue, and propose a set of simulation scenarios that reflect a range of atmospheric conditions and noise contamination that may ultimately happen on-board an imaging satellite. We verify their impact on the generalization capabilities of spectral and spectral-spatial convolutional neural networks for hyperspectral image segmentation. Our experimental analysis, coupled with various visualizations, sheds more light on the robustness of the deep models and indicate that specific noise distributions can significantly deteriorate their performance. Additionally, we show that simulating atmospheric conditions is key to obtaining the learners that generalize well over image data acquired in different imaging settings. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|