Multi-Agent Pipeline for Crop-Type Classification and Label Refinement Using Sentinel-1 SAR Time Series and Field-Level Temporal Features in the Nakasatsunai Region, Hokkaido.
Saved in:
| Title: | Multi-Agent Pipeline for Crop-Type Classification and Label Refinement Using Sentinel-1 SAR Time Series and Field-Level Temporal Features in the Nakasatsunai Region, Hokkaido. |
|---|---|
| Authors: | Arai, Kohei1 (AUTHOR) arai@is.saga-u.ac.jp, Maruta, Ria2 (AUTHOR), Okumura, Hiroshi1 (AUTHOR) |
| Source: | Remote Sensing. Aug2026, Vol. 18 Issue 15, p2628. 29p. |
| Subjects: | Normalized difference vegetation index, Synthetic aperture radar, Random forest algorithms, Vegetation monitoring, Optical properties, Phenology |
| Geographic Terms: | Japan, Hokkaido (Japan) |
| Abstract: | Highlights: What are the main findings? A three-step, NDVI-based preprocessing strategy (noise removal, k-means subclass discovery, trajectory-based label correction) systematically surfaces mislabeled or non-growing samples within the coarse, nine-class eMAF crop schema, particularly for broad categories such as rice, beans, and millet. Field-level comparison of NDVI, RVI, and γVH/γVV time series shows that γVH/γVV correlates with NDVI more consistently than RVI does across crop types, supporting its use as a SAR-only proxy for optical vegetation dynamics when cloud cover or nighttime acquisition limits optical availability. What are the implications of the main findings? A modular pipeline architecture—separating acquisition, preprocessing, temporal analysis, and quality control into coordinated stages—offers a practical template for scaling label-refinement workflows to other regions or crop taxonomies without redesigning the full system. Combining SAR-derived indices (RVI, VV, VH, γVH/γVV) with NDVI-based temporal-shape features is likely to improve classifier robustness over single-sensor approaches, since the two data sources are sensitive to different aspects of crop phenology; this motivates the broader comparison with polarimetric indices (PRVI, DPSVI, DpRVI) discussed. Reference labels for crop-type mapping are frequently coarse, and administrative land-use registries such as Japan's eMAF (electronic Map of Agriculture and Forestry) database routinely group agronomically distinct crops under broad, ambiguous categories. This study addresses that problem for the Nakasatsunai region of Hokkaido, Japan, by combining Sentinel-1 synthetic aperture radar (SAR) time series with field-level optical vegetation-index analysis in a modular processing pipeline. The principal novelty of the work is not the pipeline architecture alone but a three-step, Normalized Difference Vegetation Index (NDVI)-driven label-refinement procedure—automatic removal of non-growing or low-amplitude field samples, Euclidean k-means subclass discovery within each coarse label, and trajectory-based label correction—that converts noisy nine-class eMAF labels into a more reliable training set prior to classifier training. The feature set combines the Radar Vegetation Index (RVI), VV and VH backscatter, the γVH/γVV polarization ratio, and NDVI, together with temporal-shape descriptors (phenological timing, peak magnitude, amplitude, maximum slope, and area under the curve) derived from monthly growth trajectories over the 2018 growing season. A Random Forest classifier, together with a gradient-boosting comparator, is evaluated before and after preprocessing under stratified k-fold cross-validation. Across n = 1208 field samples spanning the nine eMAF classes, classification accuracy improved from an overall accuracy of 71.8% on the raw labels to 82.6% after the three-step refinement; Cohen's kappa increased from 0.63 to 0.77. Correlation analysis indicates that γVH/γVV tracks field-level NDVI more consistently (mean Pearson r = 0.68) than RVI does (mean Pearson r = 0.43) across the eight classes with sufficient samples, motivating its use as a SAR-only phenological proxy; this comparison is extended to the polarimetric PRVI, DPSVI, and DpRVI indices in the discussion. The underlying 80–90% label-accuracy estimate is derived from NDVI trajectory inspection rather than independent, field-surveyed ground truth, and a factorial ablation is used to characterize, to the extent the cross-validated evidence allows, how much of the reported accuracy gain is attributable to label-error correction as opposed to NDVI–SAR feature fusion; both this attribution and the label-accuracy estimate itself are identified as priorities for field validation in future work. The proposed framework is intended to convert coarse, noisy crop labels into a structured and reliable dataset while producing interpretable, field-level phenological insight for agricultural monitoring. [ABSTRACT FROM AUTHOR] |
| Copyright of Remote Sensing is the property of MDPI and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Highlights: What are the main findings? A three-step, NDVI-based preprocessing strategy (noise removal, k-means subclass discovery, trajectory-based label correction) systematically surfaces mislabeled or non-growing samples within the coarse, nine-class eMAF crop schema, particularly for broad categories such as rice, beans, and millet. Field-level comparison of NDVI, RVI, and γVH/γVV time series shows that γVH/γVV correlates with NDVI more consistently than RVI does across crop types, supporting its use as a SAR-only proxy for optical vegetation dynamics when cloud cover or nighttime acquisition limits optical availability. What are the implications of the main findings? A modular pipeline architecture—separating acquisition, preprocessing, temporal analysis, and quality control into coordinated stages—offers a practical template for scaling label-refinement workflows to other regions or crop taxonomies without redesigning the full system. Combining SAR-derived indices (RVI, VV, VH, γVH/γVV) with NDVI-based temporal-shape features is likely to improve classifier robustness over single-sensor approaches, since the two data sources are sensitive to different aspects of crop phenology; this motivates the broader comparison with polarimetric indices (PRVI, DPSVI, DpRVI) discussed. Reference labels for crop-type mapping are frequently coarse, and administrative land-use registries such as Japan's eMAF (electronic Map of Agriculture and Forestry) database routinely group agronomically distinct crops under broad, ambiguous categories. This study addresses that problem for the Nakasatsunai region of Hokkaido, Japan, by combining Sentinel-1 synthetic aperture radar (SAR) time series with field-level optical vegetation-index analysis in a modular processing pipeline. The principal novelty of the work is not the pipeline architecture alone but a three-step, Normalized Difference Vegetation Index (NDVI)-driven label-refinement procedure—automatic removal of non-growing or low-amplitude field samples, Euclidean k-means subclass discovery within each coarse label, and trajectory-based label correction—that converts noisy nine-class eMAF labels into a more reliable training set prior to classifier training. The feature set combines the Radar Vegetation Index (RVI), VV and VH backscatter, the γVH/γVV polarization ratio, and NDVI, together with temporal-shape descriptors (phenological timing, peak magnitude, amplitude, maximum slope, and area under the curve) derived from monthly growth trajectories over the 2018 growing season. A Random Forest classifier, together with a gradient-boosting comparator, is evaluated before and after preprocessing under stratified k-fold cross-validation. Across n = 1208 field samples spanning the nine eMAF classes, classification accuracy improved from an overall accuracy of 71.8% on the raw labels to 82.6% after the three-step refinement; Cohen's kappa increased from 0.63 to 0.77. Correlation analysis indicates that γVH/γVV tracks field-level NDVI more consistently (mean Pearson r = 0.68) than RVI does (mean Pearson r = 0.43) across the eight classes with sufficient samples, motivating its use as a SAR-only phenological proxy; this comparison is extended to the polarimetric PRVI, DPSVI, and DpRVI indices in the discussion. The underlying 80–90% label-accuracy estimate is derived from NDVI trajectory inspection rather than independent, field-surveyed ground truth, and a factorial ablation is used to characterize, to the extent the cross-validated evidence allows, how much of the reported accuracy gain is attributable to label-error correction as opposed to NDVI–SAR feature fusion; both this attribution and the label-accuracy estimate itself are identified as priorities for field validation in future work. The proposed framework is intended to convert coarse, noisy crop labels into a structured and reliable dataset while producing interpretable, field-level phenological insight for agricultural monitoring. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 20724292 |
| DOI: | 10.3390/rs18152628 |
