Multi-Temporal Land Cover Classification with Sequential Recurrent Encoders
Earth observation (EO) sensors deliver data at daily or weekly intervals. Most land use and land cover classification (LULC) approaches, however, are designed for cloud-free and mono-temporal observations. The increasing temporal capabilities of today’s sensors enable the use of temporal, along with...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2018-03-01
|
| Series: | ISPRS International Journal of Geo-Information |
| Subjects: | |
| Online Access: | http://www.mdpi.com/2220-9964/7/4/129 |
| _version_ | 1818903155217268736 |
|---|---|
| author | Marc Rußwurm Marco Körner |
| author_facet | Marc Rußwurm Marco Körner |
| author_sort | Marc Rußwurm |
| collection | DOAJ |
| description | Earth observation (EO) sensors deliver data at daily or weekly intervals. Most land use and land cover classification (LULC) approaches, however, are designed for cloud-free and mono-temporal observations. The increasing temporal capabilities of today’s sensors enable the use of temporal, along with spectral and spatial features.Domains such as speech recognition or neural machine translation, work with inherently temporal data and, today, achieve impressive results by using sequential encoder-decoder structures. Inspired by these sequence-to-sequence models, we adapt an encoder structure with convolutional recurrent layers in order to approximate a phenological model for vegetation classes based on a temporal sequence of Sentinel 2 (S2) images. In our experiments, we visualize internal activations over a sequence of cloudy and non-cloudy images and find several recurrent cells that reduce the input activity for cloudy observations. Hence, we assume that our network has learned cloud-filtering schemes solely from input data, which could alleviate the need for tedious cloud-filtering as a preprocessing step for many EO approaches. Moreover, using unfiltered temporal series of top-of-atmosphere (TOA) reflectance data, our experiments achieved state-of-the-art classification accuracies on a large number of crop classes with minimal preprocessing, compared to other classification approaches. |
| first_indexed | 2024-12-19T20:47:03Z |
| format | Article |
| id | doaj.art-6c788f96b65b4288aaf42678e429815c |
| institution | Directory Open Access Journal |
| issn | 2220-9964 |
| language | English |
| last_indexed | 2024-12-19T20:47:03Z |
| publishDate | 2018-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | ISPRS International Journal of Geo-Information |
| spelling | doaj.art-6c788f96b65b4288aaf42678e429815c2022-12-21T20:06:12ZengMDPI AGISPRS International Journal of Geo-Information2220-99642018-03-017412910.3390/ijgi7040129ijgi7040129Multi-Temporal Land Cover Classification with Sequential Recurrent EncodersMarc Rußwurm0Marco Körner1Technical University of Munich, Arcisstraße 21, 80333 Munich, GermanyTechnical University of Munich, Arcisstraße 21, 80333 Munich, GermanyEarth observation (EO) sensors deliver data at daily or weekly intervals. Most land use and land cover classification (LULC) approaches, however, are designed for cloud-free and mono-temporal observations. The increasing temporal capabilities of today’s sensors enable the use of temporal, along with spectral and spatial features.Domains such as speech recognition or neural machine translation, work with inherently temporal data and, today, achieve impressive results by using sequential encoder-decoder structures. Inspired by these sequence-to-sequence models, we adapt an encoder structure with convolutional recurrent layers in order to approximate a phenological model for vegetation classes based on a temporal sequence of Sentinel 2 (S2) images. In our experiments, we visualize internal activations over a sequence of cloudy and non-cloudy images and find several recurrent cells that reduce the input activity for cloudy observations. Hence, we assume that our network has learned cloud-filtering schemes solely from input data, which could alleviate the need for tedious cloud-filtering as a preprocessing step for many EO approaches. Moreover, using unfiltered temporal series of top-of-atmosphere (TOA) reflectance data, our experiments achieved state-of-the-art classification accuracies on a large number of crop classes with minimal preprocessing, compared to other classification approaches.http://www.mdpi.com/2220-9964/7/4/129deep learningmulti-temporal classificationland use and land cover classificationrecurrent networkssequence encodercrop classificationsequence-to-sequenceSentinel 2 |
| spellingShingle | Marc Rußwurm Marco Körner Multi-Temporal Land Cover Classification with Sequential Recurrent Encoders ISPRS International Journal of Geo-Information deep learning multi-temporal classification land use and land cover classification recurrent networks sequence encoder crop classification sequence-to-sequence Sentinel 2 |
| title | Multi-Temporal Land Cover Classification with Sequential Recurrent Encoders |
| title_full | Multi-Temporal Land Cover Classification with Sequential Recurrent Encoders |
| title_fullStr | Multi-Temporal Land Cover Classification with Sequential Recurrent Encoders |
| title_full_unstemmed | Multi-Temporal Land Cover Classification with Sequential Recurrent Encoders |
| title_short | Multi-Temporal Land Cover Classification with Sequential Recurrent Encoders |
| title_sort | multi temporal land cover classification with sequential recurrent encoders |
| topic | deep learning multi-temporal classification land use and land cover classification recurrent networks sequence encoder crop classification sequence-to-sequence Sentinel 2 |
| url | http://www.mdpi.com/2220-9964/7/4/129 |
| work_keys_str_mv | AT marcrußwurm multitemporallandcoverclassificationwithsequentialrecurrentencoders AT marcokorner multitemporallandcoverclassificationwithsequentialrecurrentencoders |