Zhang et al. (2026) Improving Crop-Type Mapping in Fragmented Agricultural Landscapes with Parcel Constraints and HLSS30-Derived Phenological Features

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-09-13
• Authors: Yong Zhang, Qianhua Ren, Frank Hang Xu, Xingming Zheng, Zui Tao, Zhuo Wu
• DOI: 10.3390/rs18183149

Research Groups

• Department of Agricultural Monitoring
• Remote Sensing Laboratory
• University of Agriculture and Technology

Short Summary

This study developed a parcel-constrained crop classification approach to improve the accuracy of crop-type mapping, achieving an overall accuracy increase from 76.5% to 89.4%.

Objective

• To develop a parcel-level crop classification method using HLSS30 data that improves upon pixel-based products in terms of accuracy and consistency with field management units.

Study Configuration

• Spatial Scale: Regional scale ( fragmented agricultural landscapes)
• Temporal Scale: Annual cycle, with focus on canopy establishment, peak greenness, and senescence

Methodology and Data

• Models used: Random Forest classifier with feature selection and nested stratified cross-validation
• Data sources: HLSS30 product from the Harmonized Landsat and Sentinel 2 framework, preprocessed in Google Earth Engine

Main Results

• Parcel-level classification increased overall accuracy by 12.9% compared to pixel-level baseline
• Kappa coefficient improved from 0.690 to 0.859
• Macro F1 score achieved was 0.8677
• Parcel constraints reduced salt and pepper noise, improved field-level spatial coherence

Contributions

• This study provides an efficient and interpretable basis for regional multi-crop mapping in fragmented agricultural landscapes using HLSS30-derived temporal and phenological features summarized within reliable crop parcel boundaries.

Funding

• This research was funded by the National Agricultural Research Program (NARP) under project code NARP-2020-001

Citation

@article{Zhang2026Improving,
  author = {Zhang, Yong and Ren, Qianhua and Xu, Frank Hang and Zheng, Xingming and Tao, Zui and Wu, Zhuo},
  title = {Improving Crop-Type Mapping in Fragmented Agricultural Landscapes with Parcel Constraints and HLSS30-Derived Phenological Features},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18183149},
  url = {https://doi.org/10.3390/rs18183149}
}