Semantic segmentation by convolutional neural networks (CNN) has advanced the state of the art in pixel-level classification of remote sensing images. However, processing large images typically requires analyzing the image in small patches, and hence features that have large spatial extent still cause challenges in tasks such as cloud masking. To support a wider scale of spatial features while simultaneously reducing computational requirements for large satellite images, we propose an architecture of two cascaded CNN model components successively processing undersampled and full resolution images. The first component distinguishes between patches in the inner cloud area from patches at the cloud's boundary region. For the cloudambiguous edge patches requiring further segmentation, the framework then delegates computation to a fine-grained model component. We apply the architecture to a cloud detection dataset of complete Sentinel-2 multispectral images, approximately annotated for minimal false negatives in a land use application. On this specific task and data, we achieve a 16% relative improvement in pixel accuracy over a CNN baseline based on patching.

One of the precepts of food security is the proper functioning of the global food markets. This calls for open and timely intelligence on crop production on an agroclimatically meaningful territorial scale. We propose an operationally suitable method for large-scale in-season crop yield estimations from a satellite image time series (SITS) for statistical production. As an object-based method, it is spatially scalable from parcel to regional scale, making it useful for prediction tasks in which the reference data are available only at a coarser level, such as counties. We show that deep learning-based temporal convolutional network (TCN) outperforms the classical machine learning method random forests and produces more accurate results overall than published national crop forecasts. Our novel contribution is to show that mean-aggregated regional predictions with histogram-based features calculated from farm-level observations perform better than other tested approaches. In addition, TCN is robust to the presence of cloudy pixels, suggesting TCN can learn cloud masking from the data. The temporal compositing of information do not improve prediction performance. This indicates that with end-to-end learning less preprocessing in SITS tasks seems viable.

We consider the use of remote sensing for large-scale monitoring of agricultural land use, focusing on classification of tillage and vegetation cover for individual field parcels across large spatial areas. From the perspective of remote sensing and modelling, field parcels are challenging as objects of interest due to highly varying shape and size but relatively uniform pixel content and texture. To model such areas we need representations that can be reliably estimated already for small parcels and that are invariant to the size of the parcel. We propose representing the parcels using density estimates of remote imaging pixels and provide a computational pipeline that combines the representation with arbitrary supervised learning algorithms, while allowing easy integration of multiple imaging sources. We demonstrate the method in the task of the automatic monitoring of autumn tillage method and vegetation cover of Finnish crop fields, based on the integrated analysis of intensity of Synthetic Aperture Radar (SAR) polarity bands of the Sentinel-1 satellite and spectral indices calculated from Sentinel-2 multispectral image data. We use a collection of 127,757 field parcels monitored in April 2018 and annotated to six tillage method and vegetation cover classes, reaching 70% classification accuracy for test parcels when using both SAR and multispectral data. Besides this task, the method could also directly be applied for other agricultural monitoring tasks, such as crop yield prediction.

No abstract