P2P-ST: A Pixel-to-Parcel Scale-Transform Method for Cropland Monitoring

This study proposes a pixel-to-parcel scale-transform framework, termed P2P-ST, based on pixel unmixing to enable the unified generation, representation, and management of multi-source remote sensing data at the parcel scale.

This repository contains the processed datasets and supporting code associated with the study “P2P-ST: A Pixel-to-Parcel Scale-Transform Method for Cropland Monitoring”. The study develops a pixel-to-parcel scale-transform framework that converts remote sensing signals from the pixel scale to the agricultural parcel scale, thereby reducing pixel-level heterogeneity, mixed-pixel effects, and salt-and-pepper noise in cropland monitoring applications.

The dataset includes parcel-level remote sensing time-series data derived from Sentinel-2 and MODIS imagery, evaluation results for multi-source data consistency, crop classification assessment outputs, valid-observation statistics during crop-sensitive growth stages, and data used to reproduce the main figures and tables in the manuscript. The accompanying code provides the main processing and analysis procedures for implementing and evaluating the P2P-ST framework.

The original satellite imagery was obtained from publicly available Sentinel-2 and MODIS products through Google Earth Engine and related NASA/ESA data portals. Because of the large volume of raw satellite imagery, the repository provides processed and analysis-ready data products rather than the full original image archive. These materials are intended to support transparency, reproducibility, and reuse of the results reported in the manuscript.