Sentinel-2 cropland mapping using pixel-based and object-based time-weighted dynamic time warping analysis

TLDR

Efficient cropland mapping is essential for sustainable agriculture and monitoring, and the high spatial and temporal resolution of Sentinel‑2 offers new opportunities, yet current methods largely rely on pixel‑based time‑series analysis. This study evaluates the performance of a time‑weighted dynamic time‑warping (TWDTW) method on Sentinel‑2 time series for pixel‑based and object‑based crop classification across Romania, Italy, and the USA. The authors compared TWDTW outputs to Random Forest for both pixel and object units and assessed each method’s sensitivity to training samples. Object‑based TWDTW achieved higher accuracies (78–96 %) and faster computation than pixel‑based TWDTW, matched Random Forest in Romania and Italy but lagged in the USA, and proved less sensitive to training samples, making it valuable where training data are scarce.

Abstract

Efficient methodologies for mapping croplands are an essential condition for the implementation of sustainable agricultural practices and for monitoring crops periodically. The increasing spatial and temporal resolution of globally available satellite images, such as those provided by Sentinel-2, creates new possibilities for generating accurate datasets on available crop types, in ready-to-use vector data format. Existing solutions dedicated to cropland mapping, based on high resolution remote sensing data, are mainly focused on pixel-based analysis of time series data. This paper evaluates how a time-weighted dynamic time warping (TWDTW) method that uses Sentinel-2 time series performs when applied to pixel-based and object-based classifications of various crop types in three different study areas (in Romania, Italy and the USA). The classification outputs were compared to those produced by Random Forest (RF) for both pixel- and object-based image analysis units. The sensitivity of these two methods to the training samples was also evaluated. Object-based TWDTW outperformed pixel-based TWDTW in all three study areas, with overall accuracies ranging between 78.05% and 96.19%; it also proved to be more efficient in terms of computational time. TWDTW achieved comparable classification results to RF in Romania and Italy, but RF achieved better results in the USA, where the classified crops present high intra-class spectral variability. Additionally, TWDTW proved to be less sensitive in relation to the training samples. This is an important asset in areas where inputs for training samples are limited.