2025
1.
Amin Norouzi Kandelati; Amanda T. Stahl; Yan Yan; Siddharth Chaudhary; Steve Van Vleet; David I. Gustafson; Hans Kok; Kirti Rajagopalan
Regional-scale Field-level Estimation and Mapping of Tillage Practices in Areas with Crop Diversity Unpublished
2025.
Abstract | Links | BibTeX | Tags: Conservation Agriculture, Importance Weighting, Machine Learning, Satellite Imagery, Sustainable Farming Practices, Tillage Classification
@unpublished{norouzi_kandelati_regional-scale_2025,
title = {Regional-scale Field-level Estimation and Mapping of Tillage Practices in Areas with Crop Diversity},
author = {Amin Norouzi Kandelati and Amanda T. Stahl and Yan Yan and Siddharth Chaudhary and Steve Van Vleet and David I. Gustafson and Hans Kok and Kirti Rajagopalan},
url = {https://papers.ssrn.com/abstract=5035781},
doi = {10.2139/ssrn.5035781},
year = {2025},
date = {2025-09-01},
urldate = {2025-09-01},
publisher = {Social Science Research Network},
address = {Rochester, NY},
abstract = {The imperative for sustainable agriculture has led to the promotion of conservation tillage practices, which are critical for maintaining soil health and reducing negative environmental impacts. However, the ability to verify the adoption of these practices at a regional scale and with minimal costs remains a challenge, particularly in regions with diverse crop systems. This study presents a satellite-imagery-based random forest classifier that integrates crop class information and a customized importance weighting technique for field-level, regional-scale mapping of tillage practices in crop-diverse regions. We use eastern Washington State, U.S., with multiple grains, legumes, and canola as a case study. Using 577 field-level ground-truth data points across two counties, the tillage classifier achieved a median overall accuracy of 84%, representing a 15-percentage-point improvement over the baseline median overall accuracy of 69%, driven by the inclusion of crop class and the application of importance weighting. Specific crop classes, including grains, legumes, and canola, showed accuracy improvements of 6, 42, and 13 percentage points, respectively. While the model showed promising results, limitations such as overprediction of no-tillage in non-grain crop classes were identified, pointing to the need for further refinement with additional data. The model was applied to map tillage practices across dryland fields in eastern Washington State for the years 2012, 2017, and 2022, and the predictions were generally congruent with the USDA Census of Agriculture county-level statistics. This research offers a scalable tool for the verification of conservation tillage practice adoption, supporting both regional and global monitoring efforts for sustainable agricultural practices.},
keywords = {Conservation Agriculture, Importance Weighting, Machine Learning, Satellite Imagery, Sustainable Farming Practices, Tillage Classification},
pubstate = {published},
tppubtype = {unpublished}
}
The imperative for sustainable agriculture has led to the promotion of conservation tillage practices, which are critical for maintaining soil health and reducing negative environmental impacts. However, the ability to verify the adoption of these practices at a regional scale and with minimal costs remains a challenge, particularly in regions with diverse crop systems. This study presents a satellite-imagery-based random forest classifier that integrates crop class information and a customized importance weighting technique for field-level, regional-scale mapping of tillage practices in crop-diverse regions. We use eastern Washington State, U.S., with multiple grains, legumes, and canola as a case study. Using 577 field-level ground-truth data points across two counties, the tillage classifier achieved a median overall accuracy of 84%, representing a 15-percentage-point improvement over the baseline median overall accuracy of 69%, driven by the inclusion of crop class and the application of importance weighting. Specific crop classes, including grains, legumes, and canola, showed accuracy improvements of 6, 42, and 13 percentage points, respectively. While the model showed promising results, limitations such as overprediction of no-tillage in non-grain crop classes were identified, pointing to the need for further refinement with additional data. The model was applied to map tillage practices across dryland fields in eastern Washington State for the years 2012, 2017, and 2022, and the predictions were generally congruent with the USDA Census of Agriculture county-level statistics. This research offers a scalable tool for the verification of conservation tillage practice adoption, supporting both regional and global monitoring efforts for sustainable agricultural practices.
2.
Oishee Bintey Hoque, Abhijin Adiga, Aniruddha Adiga, Siddharth Chaudhary, Madhav V. Marathe, S. S. Ravi, Kirti Rajagopalan, Amanda Wilson,; Samarth Swarup
IGraSS: Learning to Identify Infrastructure Networks from Satellite Imagery by Iterative Graph-constrained Semantic Segmentation Proceedings Article
In: 2025 IEEE International Geoscience and Remote Sensing Symposium 2025.
Abstract | Links | BibTeX | Tags: Infrastructure Networks, Satellite Imagery, Semantic Segmentation
@inproceedings{nokey,
title = {IGraSS: Learning to Identify Infrastructure Networks from Satellite Imagery by Iterative Graph-constrained Semantic Segmentation},
author = {Oishee Bintey Hoque, Abhijin Adiga, Aniruddha Adiga, Siddharth Chaudhary, Madhav V. Marathe, S. S. Ravi, Kirti Rajagopalan, Amanda Wilson, and Samarth Swarup},
url = {https://arxiv.org/abs/2506.08137},
doi = { https://doi.org/10.48550/arXiv.2506.08137},
year = {2025},
date = {2025-08-08},
urldate = {2025-08-08},
organization = {2025 IEEE International Geoscience and Remote Sensing Symposium},
abstract = {Accurate canal network mapping is essential for water management, including irrigation planning and infrastructure maintenance. State-of-the-art semantic segmentation models for infrastructure mapping, such as roads, rely on large, wellannotated remote sensing datasets. However, incomplete or inadequate ground truth can hinder these learning approaches. Many infrastructure networks have graph-level properties such as reachability to a source (like canals) or connectivity (roads) that can be leveraged to improve these existing ground truth. This paper develops a novel iterative framework IGraSS, combining a semantic segmentation module\textemdashincorporating RGB and additional modalities (NDWI, DEM)\textemdashwith a graph-based ground-truth refinement module. The segmentation module processes satellite imagery patches, while the refinement module operates on the entire data viewing the infrastructure network as a graph. Experiments show that IGraSS reduces unreachable canal segments from 18% to 3%, and training with refined ground truth significantly improves canal identification. IGraSS serves as a robust framework for both refining noisy ground truth and mapping canal networks from remote sensing imagery. We also demonstrate the effectiveness and generalizability of IGraSS using road networks as an example, applying a different graph-theoretic constraint to complete road networks.},
keywords = {Infrastructure Networks, Satellite Imagery, Semantic Segmentation},
pubstate = {published},
tppubtype = {inproceedings}
}
Accurate canal network mapping is essential for water management, including irrigation planning and infrastructure maintenance. State-of-the-art semantic segmentation models for infrastructure mapping, such as roads, rely on large, wellannotated remote sensing datasets. However, incomplete or inadequate ground truth can hinder these learning approaches. Many infrastructure networks have graph-level properties such as reachability to a source (like canals) or connectivity (roads) that can be leveraged to improve these existing ground truth. This paper develops a novel iterative framework IGraSS, combining a semantic segmentation module—incorporating RGB and additional modalities (NDWI, DEM)—with a graph-based ground-truth refinement module. The segmentation module processes satellite imagery patches, while the refinement module operates on the entire data viewing the infrastructure network as a graph. Experiments show that IGraSS reduces unreachable canal segments from 18% to 3%, and training with refined ground truth significantly improves canal identification. IGraSS serves as a robust framework for both refining noisy ground truth and mapping canal networks from remote sensing imagery. We also demonstrate the effectiveness and generalizability of IGraSS using road networks as an example, applying a different graph-theoretic constraint to complete road networks.
