Journal
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
Volume 59, Issue 6, Pages 4972-4983Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TGRS.2020.3015272
Keywords
Clouds; Image segmentation; Cloud computing; Semantics; Remote sensing; Image resolution; Annotations; Cloud detection; machine learning (ML); multispectral; neural networks; remote sensing
Categories
Funding
- Academy of Finland Flagship Program: Finnish Center for Artificial Intelligence, FCAI
Ask authors/readers for more resources
A new CNN model architecture is proposed for pixel-level classification of remote sensing images, particularly suitable for cloud detection tasks. By cascading model components, it supports a wider range of spatial features and achieves significant improvement in pixel accuracy.
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 a 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 clouds boundary region. For the cloud-ambiguous edge patches requiring further segmentation, the framework then delegates computation to a fine-grained model component. We apply the architecture to a cloud detection data set 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.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available