3D Fully Convolutional Neural Networks with Intersection Over Union Loss for Crop Mapping from Multi-Temporal Satellite Images

TL;DR

This paper presents a 3D Fully Convolutional Neural Network with an Intersection Over Union loss function for improved crop mapping from multi-temporal satellite images, demonstrating superior accuracy over existing methods.

Contribution

It introduces a novel 3D FCN architecture combined with an IOU loss function specifically designed for crop classification from multi-temporal remote sensing data.

Findings

Achieved a Kappa coefficient of 91.8% in crop classification.

Outperformed existing methods in crop type mapping.

Demonstrated the effectiveness of IOU loss in remote sensing applications.

Abstract

Information on cultivated crops is relevant for a large number of food security studies. Different scientific efforts are dedicated to generating this information from remote sensing images by means of machine learning methods. Unfortunately, these methods do not take account of the spatial-temporal relationships inherent in remote sensing images. In our paper, we explore the capability of a 3D Fully Convolutional Neural Network (FCN) to map crop types from multi-temporal images. In addition, we propose the Intersection Over Union (IOU) loss function for increasing the overlap between the predicted classes and ground reference data. The proposed method was applied to identify soybean and corn from a study area situated in the US corn belt using multi-temporal Landsat images. The study shows that our method outperforms related methods, obtaining a Kappa coefficient of 91.8%. We conclude that using the IOU loss function provides a superior choice to learn individual crop types.