Crop classification in South Korea for multitemporal PlanetScope imagery using SFC-DenseNet-AM
In this manuscript, a new methodology based on a deep learning model using a Siamese network and attention module was proposed to classify crop cultivation areas, such as onion and garlic, from multitemporal PlanetScope images in South Korea. To consider the seasonal characteristics of crops in the...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-02-01
|
| Series: | International Journal of Applied Earth Observations and Geoinformation |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843223004430 |
| _version_ | 1797357811811745792 |
|---|---|
| author | Seonkyeong Seong Anjin Chang Junsang Mo Sangil Na Hoyong Ahn Jaehong Oh Jaewan Choi |
| author_facet | Seonkyeong Seong Anjin Chang Junsang Mo Sangil Na Hoyong Ahn Jaehong Oh Jaewan Choi |
| author_sort | Seonkyeong Seong |
| collection | DOAJ |
| description | In this manuscript, a new methodology based on a deep learning model using a Siamese network and attention module was proposed to classify crop cultivation areas, such as onion and garlic, from multitemporal PlanetScope images in South Korea. To consider the seasonal characteristics of crops in the model, training data were constructed from multitemporal satellite images. It was generated using PlanetScope satellite imagery from January and April, corresponding to the seasonal growth period of onion and garlic, in South Korea. Image patches were generated by considering the ratio of crops to minimize the influence of imbalanced data in the training process. Siamese FC-DenseNet with an attention module model (SFC-DenseNet-AM) is proposed, and the attention module is used to classify cultivated crop areas. Based on the proposed network, we extract cultivated crop areas using preliminary cultivation information. The results of the experiment using PlanetScope images indicate that image classification for cultivated areas was effectively performed using the proposed deep learning model. The model's performance, with F1-scores of 0.823 (garlic) and 0.774 (onion), was verified through an ablation study. |
| first_indexed | 2024-03-08T14:50:42Z |
| format | Article |
| id | doaj.art-e1b33aa9f5274130a221ac874f2f9d3c |
| institution | Directory Open Access Journal |
| issn | 1569-8432 |
| language | English |
| last_indexed | 2024-03-08T14:50:42Z |
| publishDate | 2024-02-01 |
| publisher | Elsevier |
| record_format | Article |
| series | International Journal of Applied Earth Observations and Geoinformation |
| spelling | doaj.art-e1b33aa9f5274130a221ac874f2f9d3c2024-01-11T04:30:29ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322024-02-01126103619Crop classification in South Korea for multitemporal PlanetScope imagery using SFC-DenseNet-AMSeonkyeong Seong0Anjin Chang1Junsang Mo2Sangil Na3Hoyong Ahn4Jaehong Oh5Jaewan Choi6Satellite Planning Division, National Meteorological Satellite Center, Korea Meteorological Administration, Jincheon, Republic of KoreaDepartment of Agricultural and Environmental Sciences, Tennessee State University, Nashville, TN, USANational Land Satellite Center, National Geographic Information Institute, Suwon, Republic of KoreaClimate Change Assessment Division, National Institute of Agricultural Sciences, Rural Development Administration, Jeollabuk-do, Republic of KoreaClimate Change Assessment Division, National Institute of Agricultural Sciences, Rural Development Administration, Jeollabuk-do, Republic of KoreaDepartment of Civil Engineering, Interdisciplinary Major of Ocean Renewable Energy Engineering, Korea Maritime and Ocean University, Busan, Republic of KoreaDepartment of Civil Engineering, Chungbuk National University, Cheongju, Republic of Korea; Corresponding author.In this manuscript, a new methodology based on a deep learning model using a Siamese network and attention module was proposed to classify crop cultivation areas, such as onion and garlic, from multitemporal PlanetScope images in South Korea. To consider the seasonal characteristics of crops in the model, training data were constructed from multitemporal satellite images. It was generated using PlanetScope satellite imagery from January and April, corresponding to the seasonal growth period of onion and garlic, in South Korea. Image patches were generated by considering the ratio of crops to minimize the influence of imbalanced data in the training process. Siamese FC-DenseNet with an attention module model (SFC-DenseNet-AM) is proposed, and the attention module is used to classify cultivated crop areas. Based on the proposed network, we extract cultivated crop areas using preliminary cultivation information. The results of the experiment using PlanetScope images indicate that image classification for cultivated areas was effectively performed using the proposed deep learning model. The model's performance, with F1-scores of 0.823 (garlic) and 0.774 (onion), was verified through an ablation study.http://www.sciencedirect.com/science/article/pii/S1569843223004430Attention moduleCultivated areaDeep learningSFC-DenseNet-AMmultitemporal PlanetScope imagery |
| spellingShingle | Seonkyeong Seong Anjin Chang Junsang Mo Sangil Na Hoyong Ahn Jaehong Oh Jaewan Choi Crop classification in South Korea for multitemporal PlanetScope imagery using SFC-DenseNet-AM International Journal of Applied Earth Observations and Geoinformation Attention module Cultivated area Deep learning SFC-DenseNet-AM multitemporal PlanetScope imagery |
| title | Crop classification in South Korea for multitemporal PlanetScope imagery using SFC-DenseNet-AM |
| title_full | Crop classification in South Korea for multitemporal PlanetScope imagery using SFC-DenseNet-AM |
| title_fullStr | Crop classification in South Korea for multitemporal PlanetScope imagery using SFC-DenseNet-AM |
| title_full_unstemmed | Crop classification in South Korea for multitemporal PlanetScope imagery using SFC-DenseNet-AM |
| title_short | Crop classification in South Korea for multitemporal PlanetScope imagery using SFC-DenseNet-AM |
| title_sort | crop classification in south korea for multitemporal planetscope imagery using sfc densenet am |
| topic | Attention module Cultivated area Deep learning SFC-DenseNet-AM multitemporal PlanetScope imagery |
| url | http://www.sciencedirect.com/science/article/pii/S1569843223004430 |
| work_keys_str_mv | AT seonkyeongseong cropclassificationinsouthkoreaformultitemporalplanetscopeimageryusingsfcdensenetam AT anjinchang cropclassificationinsouthkoreaformultitemporalplanetscopeimageryusingsfcdensenetam AT junsangmo cropclassificationinsouthkoreaformultitemporalplanetscopeimageryusingsfcdensenetam AT sangilna cropclassificationinsouthkoreaformultitemporalplanetscopeimageryusingsfcdensenetam AT hoyongahn cropclassificationinsouthkoreaformultitemporalplanetscopeimageryusingsfcdensenetam AT jaehongoh cropclassificationinsouthkoreaformultitemporalplanetscopeimageryusingsfcdensenetam AT jaewanchoi cropclassificationinsouthkoreaformultitemporalplanetscopeimageryusingsfcdensenetam |