Precision Agriculture: Crop Mapping using Machine Learning and Sentinel-2 Satellite Imagery

TL;DR

This paper demonstrates the effectiveness of deep learning and pixel-based machine learning methods, especially a U-Net model, for accurate crop mapping using Sentinel-2 satellite imagery in precision agriculture.

Contribution

It introduces a novel application of U-Net for crop segmentation and reveals the surprising success of pixel-based methods with RGB bands in this context.

Findings

U-Net achieved a Dice coefficient of 0.8324.

Pixel-based methods with RGB bands were unexpectedly effective.

Satellite imagery can be effectively used for precise crop mapping.

Abstract

Food security has grown in significance due to the changing climate and its warming effects. To support the rising demand for agricultural products and to minimize the negative impact of climate change and mass cultivation, precision agriculture has become increasingly important for crop cultivation. This study employs deep learning and pixel-based machine learning methods to accurately segment lavender fields for precision agriculture, utilizing various spectral band combinations extracted from Sentinel-2 satellite imagery. Our fine-tuned final model, a U-Net architecture, can achieve a Dice coefficient of 0.8324. Additionally, our investigation highlights the unexpected efficacy of the pixel-based method and the RGB spectral band combination in this task.