Critical Assessment of Cocoa Classification with Limited Reference Data: A Study in Côte d’Ivoire and Ghana Using Sentinel-2 and Random Forest Model
Cocoa is the economic backbone of Côte d’Ivoire and Ghana, making them the leading cocoa-producing countries in the world. However, cocoa farming has been a major driver of deforestation and landscape degradation in West Africa. Various stakeholders are striving for a zero-deforestation cocoa sector...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2024-02-01
|
Series: | Remote Sensing |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-4292/16/3/598 |
_version_ | 1797318259471548416 |
---|---|
author | Nikoletta Moraiti Adugna Mullissa Eric Rahn Marieke Sassen Johannes Reiche |
author_facet | Nikoletta Moraiti Adugna Mullissa Eric Rahn Marieke Sassen Johannes Reiche |
author_sort | Nikoletta Moraiti |
collection | DOAJ |
description | Cocoa is the economic backbone of Côte d’Ivoire and Ghana, making them the leading cocoa-producing countries in the world. However, cocoa farming has been a major driver of deforestation and landscape degradation in West Africa. Various stakeholders are striving for a zero-deforestation cocoa sector by implementing sustainable farming strategies and a more transparent supply chain. In the context of tracking cocoa sources and contributing to cocoa-driven deforestation monitoring, the demand for accurate and up-to-date maps of cocoa plantations is increasing. Yet, access to limited reference data and imperfect data quality can impose challenges in producing reliable maps. This study classified full-sun-cocoa-growing areas using limited reference data relative to the large and heterogeneous study areas in Côte d’Ivoire and Ghana. A Sentinel-2 composite image of 2021 was generated to train a random forest model. We undertook reference data refinement, selection of the most important handcrafted features and data sampling to ensure spatial independence. After refining the quality of the reference data and despite their size reduction, the random forest performance was improved, achieving an overall accuracy of 85.1 ± 2.0% and an F1 score of 84.6 ± 2.4% (mean ± one standard deviation from ten bootstrapping iterations). Emphasis was given to the qualitative visual assessment of the map using very high-resolution images, which revealed cases of strong and weak generalisation capacity of the random forest. Further insight was gained from the comparative analysis of our map with two previous cocoa classification studies. Implications of the use of cocoa maps for reporting were discussed. |
first_indexed | 2024-03-08T03:49:51Z |
format | Article |
id | doaj.art-71d1f567226347738b27414b670d720d |
institution | Directory Open Access Journal |
issn | 2072-4292 |
language | English |
last_indexed | 2024-03-08T03:49:51Z |
publishDate | 2024-02-01 |
publisher | MDPI AG |
record_format | Article |
series | Remote Sensing |
spelling | doaj.art-71d1f567226347738b27414b670d720d2024-02-09T15:21:36ZengMDPI AGRemote Sensing2072-42922024-02-0116359810.3390/rs16030598Critical Assessment of Cocoa Classification with Limited Reference Data: A Study in Côte d’Ivoire and Ghana Using Sentinel-2 and Random Forest ModelNikoletta Moraiti0Adugna Mullissa1Eric Rahn2Marieke Sassen3Johannes Reiche4Laboratory of Geo-Information Science and Remote Sensing, Wageningen University, Droevendaalsesteeg 4, 6708 PB Wageningen, The NetherlandsLaboratory of Geo-Information Science and Remote Sensing, Wageningen University, Droevendaalsesteeg 4, 6708 PB Wageningen, The NetherlandsInternational Center for Tropical Agriculture (CIAT), Km 17 Recta Cali-Palmira, Cali 763537, ColombiaPlant Production Systems Group, Wageningen University, Bornsesteeg 48, 6708 PE Wageningen, The NetherlandsLaboratory of Geo-Information Science and Remote Sensing, Wageningen University, Droevendaalsesteeg 4, 6708 PB Wageningen, The NetherlandsCocoa is the economic backbone of Côte d’Ivoire and Ghana, making them the leading cocoa-producing countries in the world. However, cocoa farming has been a major driver of deforestation and landscape degradation in West Africa. Various stakeholders are striving for a zero-deforestation cocoa sector by implementing sustainable farming strategies and a more transparent supply chain. In the context of tracking cocoa sources and contributing to cocoa-driven deforestation monitoring, the demand for accurate and up-to-date maps of cocoa plantations is increasing. Yet, access to limited reference data and imperfect data quality can impose challenges in producing reliable maps. This study classified full-sun-cocoa-growing areas using limited reference data relative to the large and heterogeneous study areas in Côte d’Ivoire and Ghana. A Sentinel-2 composite image of 2021 was generated to train a random forest model. We undertook reference data refinement, selection of the most important handcrafted features and data sampling to ensure spatial independence. After refining the quality of the reference data and despite their size reduction, the random forest performance was improved, achieving an overall accuracy of 85.1 ± 2.0% and an F1 score of 84.6 ± 2.4% (mean ± one standard deviation from ten bootstrapping iterations). Emphasis was given to the qualitative visual assessment of the map using very high-resolution images, which revealed cases of strong and weak generalisation capacity of the random forest. Further insight was gained from the comparative analysis of our map with two previous cocoa classification studies. Implications of the use of cocoa maps for reporting were discussed.https://www.mdpi.com/2072-4292/16/3/598cocoa classificationcrop mappinglimited reference dataqualitative assessmentrandom forestfeature engineering |
spellingShingle | Nikoletta Moraiti Adugna Mullissa Eric Rahn Marieke Sassen Johannes Reiche Critical Assessment of Cocoa Classification with Limited Reference Data: A Study in Côte d’Ivoire and Ghana Using Sentinel-2 and Random Forest Model Remote Sensing cocoa classification crop mapping limited reference data qualitative assessment random forest feature engineering |
title | Critical Assessment of Cocoa Classification with Limited Reference Data: A Study in Côte d’Ivoire and Ghana Using Sentinel-2 and Random Forest Model |
title_full | Critical Assessment of Cocoa Classification with Limited Reference Data: A Study in Côte d’Ivoire and Ghana Using Sentinel-2 and Random Forest Model |
title_fullStr | Critical Assessment of Cocoa Classification with Limited Reference Data: A Study in Côte d’Ivoire and Ghana Using Sentinel-2 and Random Forest Model |
title_full_unstemmed | Critical Assessment of Cocoa Classification with Limited Reference Data: A Study in Côte d’Ivoire and Ghana Using Sentinel-2 and Random Forest Model |
title_short | Critical Assessment of Cocoa Classification with Limited Reference Data: A Study in Côte d’Ivoire and Ghana Using Sentinel-2 and Random Forest Model |
title_sort | critical assessment of cocoa classification with limited reference data a study in cote d ivoire and ghana using sentinel 2 and random forest model |
topic | cocoa classification crop mapping limited reference data qualitative assessment random forest feature engineering |
url | https://www.mdpi.com/2072-4292/16/3/598 |
work_keys_str_mv | AT nikolettamoraiti criticalassessmentofcocoaclassificationwithlimitedreferencedataastudyincotedivoireandghanausingsentinel2andrandomforestmodel AT adugnamullissa criticalassessmentofcocoaclassificationwithlimitedreferencedataastudyincotedivoireandghanausingsentinel2andrandomforestmodel AT ericrahn criticalassessmentofcocoaclassificationwithlimitedreferencedataastudyincotedivoireandghanausingsentinel2andrandomforestmodel AT mariekesassen criticalassessmentofcocoaclassificationwithlimitedreferencedataastudyincotedivoireandghanausingsentinel2andrandomforestmodel AT johannesreiche criticalassessmentofcocoaclassificationwithlimitedreferencedataastudyincotedivoireandghanausingsentinel2andrandomforestmodel |