Pixel-wise Agricultural Image Time Series Classification: Comparisons and a Deformable Prototype-based Approach
This addresses crop segmentation for agricultural monitoring, offering a simple and interpretable method that is incremental but effective in unsupervised settings.
The paper tackles pixel-wise segmentation of satellite image time series for agricultural monitoring by introducing a deformable prototype-based method that adds invariance to spectral deformations and temporal shifts. It achieves the best performance in low-data regimes and significantly improves state-of-the-art unsupervised classification on four recent datasets.
Improvements in Earth observation by satellites allow for imagery of ever higher temporal and spatial resolution. Leveraging this data for agricultural monitoring is key for addressing environmental and economic challenges. Current methods for crop segmentation using temporal data either rely on annotated data or are heavily engineered to compensate the lack of supervision. In this paper, we present and compare datasets and methods for both supervised and unsupervised pixel-wise segmentation of satellite image time series (SITS). We also introduce an approach to add invariance to spectral deformations and temporal shifts to classical prototype-based methods such as K-means and Nearest Centroid Classifier (NCC). We study different levels of supervision and show this simple and highly interpretable method achieves the best performance in the low data regime and significantly improves the state of the art for unsupervised classification of agricultural time series on four recent SITS datasets.