# Evaluating Multi-Temporal Sentinel-1 and Sentinel-2 Imagery for Crop Classification: A Case Study in a Paddy Rice Growing Region of China

## TL;DR

The study shows that combining Sentinel-1 and Sentinel-2 satellite data improves crop classification accuracy, especially for rice and corn in a paddy rice region in China.

## Contribution

The paper introduces a practical classification strategy using combined SAR and optical data for improved crop mapping.

## Key findings

- Dual-polarimetric SAR decomposition parameters effectively distinguish different crop types.
- Multi-temporal optical data with low cloud cover supports accurate crop classification.
- Combining SAR and optical data significantly enhances classification accuracy for rice and corn.

## Abstract

What are the main findings?
The decomposition parameters mv derived from the dual-polarization model-based decomposition can effectively discriminate different crop types.Multi-temporal optical data with low cloud cover can effectively support crop classification. Incorporating dual-polarimetric SAR data further enhances the classification accuracy, particularly for rice and corn.

The decomposition parameters mv derived from the dual-polarization model-based decomposition can effectively discriminate different crop types.

Multi-temporal optical data with low cloud cover can effectively support crop classification. Incorporating dual-polarimetric SAR data further enhances the classification accuracy, particularly for rice and corn.

What is the implication of the main finding?
A practical classification strategy was proposed for crop type identification.A 10-m-resolution thematic map was produced to classify crops in the study area.

A practical classification strategy was proposed for crop type identification.

A 10-m-resolution thematic map was produced to classify crops in the study area.

Information on crop planting structure serves as a key reference for crop growth monitoring and agricultural structural adjustment. Mapping the spatial distribution of crops through feature-based classification serves as a fundamental component of sustainable agricultural development. However, current crop classification methods often face challenges such as the discontinuity of optical data due to cloud cover and the limited discriminative capability of traditional SAR backscatter intensity for spectrally similar crops. In this case study, we assess multi-temporal Sentinel-1 and Sentinel-2 Satellite images for crop classification in a paddy rice growing region in Helonghu Town, located in the central region of Xiangyin County, Yueyang City, Hunan Province, China (28.5° N–29.0° N, 112.8° E–113.2° E). We systematically investigate three key aspects: (1) the classification performance using optical time-series Sentinel-2 imagery; (2) the time-series classification performance utilizing polarimetric SAR decomposition features from Sentinel-1 dual-polarimetric SAR images; and (3) the classification performance based on a combination of Sentinel-1 and Sentinel-2 images. Optimal classification results, with the highest overall accuracy and Kappa coefficient, are achieved through the combination of Sentinel-1 (SAR) and Sentinel-2 (optical) data. This case study evaluates the time-series classification performance of Sentinel-1 and Sentinel-2 data to determine the optimal approach for crop classification in Helonghu Town.