Cropland Mapping in Tropical Smallholder Systems with Seasonally Stratified Sentinel-1 and Sentinel-2 Spectral and Textural Features
Mapping arable field areas is crucial for assessing agricultural productivity but poses challenges in sub-Saharan agroecosystems because of diverse crop calendars, small and irregularly shaped fields, persistent cloud cover, and lack of high-quality model training data. This study proposes several m...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-06-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/12/3014 |
| _version_ | 1797592770740748288 |
|---|---|
| author | Manushi B. Trivedi Michael Marshall Lyndon Estes C.A.J.M. de Bie Ling Chang Andrew Nelson |
| author_facet | Manushi B. Trivedi Michael Marshall Lyndon Estes C.A.J.M. de Bie Ling Chang Andrew Nelson |
| author_sort | Manushi B. Trivedi |
| collection | DOAJ |
| description | Mapping arable field areas is crucial for assessing agricultural productivity but poses challenges in sub-Saharan agroecosystems because of diverse crop calendars, small and irregularly shaped fields, persistent cloud cover, and lack of high-quality model training data. This study proposes several methodological improvements to overcome these challenges. Specifically, it utilizes long-term MODIS data to stratify finer Sentinel-2 reflectance and Sentinel-1 backscatter image features on a per-pixel basis. It also incorporates texture features and employs a machine learning approach with over 300,000 samples. The eastern region of Ghana was stratified into seven seasonal strata exhibiting distinct vegetation seasonality, capturing diversity in crop calendars, using long-term MODIS (2001–2009) normalized difference vegetation index phenology. Three years (2017–2019) of Sentinel-1 and Sentinel-2 original bands at 20 m were composited into dry and wet seasonal features according to the strata, from which spectral, polarimetric, and texture features were extracted. The field boundaries were digitized using PlanetScope images (2018–2019). Random Forest classifier with 10-fold cross-validation and recursive feature elimination was used for feature selection and model building. Including topographic variables, out of 137 image features, only 11 features were found important. Sentinel-2 SWIR-based spectral features were most important, followed by Sentinel-1 polarimetric (VV) and elevation features. Half of the 11 features were variance texture features, followed by spectral features. The Random Forest classifier produced a 0.78 AUC score with overall precision, recall, and F1-score of 0.96, 0.78, and 0.85, respectively. While the precision for both classes was >0.90, the recall rate for arable areas was half that of non-arable areas. Future studies could improve the technical workflow with reliable balanced sampling, narrowband hyperspectral images, and fully polarized SAR images. |
| first_indexed | 2024-03-11T01:58:50Z |
| format | Article |
| id | doaj.art-583b3f27320b4e2db7f1c3fac7edb88e |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-11T01:58:50Z |
| publishDate | 2023-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-583b3f27320b4e2db7f1c3fac7edb88e2023-11-18T12:25:08ZengMDPI AGRemote Sensing2072-42922023-06-011512301410.3390/rs15123014Cropland Mapping in Tropical Smallholder Systems with Seasonally Stratified Sentinel-1 and Sentinel-2 Spectral and Textural FeaturesManushi B. Trivedi0Michael Marshall1Lyndon Estes2C.A.J.M. de Bie3Ling Chang4Andrew Nelson5School of Integrative Plant Science, Cornell University, Tower Rd, Ithaca, NY 14850, USAFaculty of Geo-Information Science and Earth Observation, University of Twente, Hengelosestraat 99, 7514 AE Enschede, The NetherlandsGraduate School of Geography, Clark University, 950 Main Street, Worcester, MA 01610, USAFaculty of Geo-Information Science and Earth Observation, University of Twente, Hengelosestraat 99, 7514 AE Enschede, The NetherlandsFaculty of Geo-Information Science and Earth Observation, University of Twente, Hengelosestraat 99, 7514 AE Enschede, The NetherlandsFaculty of Geo-Information Science and Earth Observation, University of Twente, Hengelosestraat 99, 7514 AE Enschede, The NetherlandsMapping arable field areas is crucial for assessing agricultural productivity but poses challenges in sub-Saharan agroecosystems because of diverse crop calendars, small and irregularly shaped fields, persistent cloud cover, and lack of high-quality model training data. This study proposes several methodological improvements to overcome these challenges. Specifically, it utilizes long-term MODIS data to stratify finer Sentinel-2 reflectance and Sentinel-1 backscatter image features on a per-pixel basis. It also incorporates texture features and employs a machine learning approach with over 300,000 samples. The eastern region of Ghana was stratified into seven seasonal strata exhibiting distinct vegetation seasonality, capturing diversity in crop calendars, using long-term MODIS (2001–2009) normalized difference vegetation index phenology. Three years (2017–2019) of Sentinel-1 and Sentinel-2 original bands at 20 m were composited into dry and wet seasonal features according to the strata, from which spectral, polarimetric, and texture features were extracted. The field boundaries were digitized using PlanetScope images (2018–2019). Random Forest classifier with 10-fold cross-validation and recursive feature elimination was used for feature selection and model building. Including topographic variables, out of 137 image features, only 11 features were found important. Sentinel-2 SWIR-based spectral features were most important, followed by Sentinel-1 polarimetric (VV) and elevation features. Half of the 11 features were variance texture features, followed by spectral features. The Random Forest classifier produced a 0.78 AUC score with overall precision, recall, and F1-score of 0.96, 0.78, and 0.85, respectively. While the precision for both classes was >0.90, the recall rate for arable areas was half that of non-arable areas. Future studies could improve the technical workflow with reliable balanced sampling, narrowband hyperspectral images, and fully polarized SAR images.https://www.mdpi.com/2072-4292/15/12/3014arable areamachine learningSentinelMODISelevationSAR |
| spellingShingle | Manushi B. Trivedi Michael Marshall Lyndon Estes C.A.J.M. de Bie Ling Chang Andrew Nelson Cropland Mapping in Tropical Smallholder Systems with Seasonally Stratified Sentinel-1 and Sentinel-2 Spectral and Textural Features Remote Sensing arable area machine learning Sentinel MODIS elevation SAR |
| title | Cropland Mapping in Tropical Smallholder Systems with Seasonally Stratified Sentinel-1 and Sentinel-2 Spectral and Textural Features |
| title_full | Cropland Mapping in Tropical Smallholder Systems with Seasonally Stratified Sentinel-1 and Sentinel-2 Spectral and Textural Features |
| title_fullStr | Cropland Mapping in Tropical Smallholder Systems with Seasonally Stratified Sentinel-1 and Sentinel-2 Spectral and Textural Features |
| title_full_unstemmed | Cropland Mapping in Tropical Smallholder Systems with Seasonally Stratified Sentinel-1 and Sentinel-2 Spectral and Textural Features |
| title_short | Cropland Mapping in Tropical Smallholder Systems with Seasonally Stratified Sentinel-1 and Sentinel-2 Spectral and Textural Features |
| title_sort | cropland mapping in tropical smallholder systems with seasonally stratified sentinel 1 and sentinel 2 spectral and textural features |
| topic | arable area machine learning Sentinel MODIS elevation SAR |
| url | https://www.mdpi.com/2072-4292/15/12/3014 |
| work_keys_str_mv | AT manushibtrivedi croplandmappingintropicalsmallholdersystemswithseasonallystratifiedsentinel1andsentinel2spectralandtexturalfeatures AT michaelmarshall croplandmappingintropicalsmallholdersystemswithseasonallystratifiedsentinel1andsentinel2spectralandtexturalfeatures AT lyndonestes croplandmappingintropicalsmallholdersystemswithseasonallystratifiedsentinel1andsentinel2spectralandtexturalfeatures AT cajmdebie croplandmappingintropicalsmallholdersystemswithseasonallystratifiedsentinel1andsentinel2spectralandtexturalfeatures AT lingchang croplandmappingintropicalsmallholdersystemswithseasonallystratifiedsentinel1andsentinel2spectralandtexturalfeatures AT andrewnelson croplandmappingintropicalsmallholdersystemswithseasonallystratifiedsentinel1andsentinel2spectralandtexturalfeatures |