Abstract

Satellite image time series (SITS) collected by modern Earth Observation (EO) systems represent a valuable source of information that supports several tasks related to the monitoring of the Earth surface dynamics over large areas. A main challenge is then to design methods able to leverage the complementarity between the temporal dynamics and the spatial patterns that characterize these data structures. Focusing on land cover classification (or mapping) tasks, the majority of approaches dealing with SITS data only considers the temporal dimension, while the integration of the spatial context is frequently neglected. In this work, we propose an attentive spatial temporal graph convolutional neural network that exploits both spatial and temporal dimensions in SITS. Despite the fact that this neural network model is well suited to deal with spatio-temporal information, this is the first work that considers it for the analysis of SITS data. Experiments are conducted on two study areas characterized by different land cover landscapes and real-world operational constraints (i.e., limited labeled data due to acquisition costs). The results show that our model consistently outperforms all the competing methods obtaining a performance gain, in terms of F-Measure, of at least 5 points with respect to the best competing approaches on both benchmarks.

Highlights

  • The Food and Agriculture Organization (FAO) of the United Nations predicts that in order to meet the needs of the expected 3 billion population growth by 2050, food production has to increase by 60% [1]

  • Satellite image time series (SITS) data constitutes a valuable source of information to assess the Earth surface dynamics

  • Applications range from food production estimation to natural resources mapping and biodiversity monitoring

Read more

Summary

INTRODUCTION

The Food and Agriculture Organization (FAO) of the United Nations predicts that in order to meet the needs of the expected 3 billion population growth by 2050, food production has to increase by 60% [1]. Despite STGCNNs being widely adopted to deal with spatio-temporal tasks, such as traffic forecasting [10], flood forecasting [11] and video activity recognition [12], surprisingly, to the best of our knowledge and according to a very recent literature survey [9], no study has been conducted yet to adopt such models in the context of satellite image time series data analysis, by leveraging RAGs in the classification process. The rest of the article is structured as follows: the literature related to our work is introduced in Section II; Section III introduces preliminary definitions about the land cover mapping task and the graph-based geographical area representation; Section IV describes the STEGON framework and Section V describes the data and the considered study areas.

METHOD OVERVIEW
ARCHITECTURE DETAILS OF THE ONE DIMENSIONAL CONVOLUTIONAL NEURAL NETWORK
SATELLITE IMAGE TIME SERIES DATA AND GROUND TRUTH
REGION ADJACENCY GRAPH STATISTICS
EXPERIMENTS
Findings
CONCLUSION
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call