All Issue

2026 Vol.59, Issue 10 Preview Page
Land cover classification of the Miho River riparian zone using high-resolution planetscope imagery and deep learning semantic segmentation

PlanetScope 고해상도 위성영상과 딥러닝 기반 의미론적 분할을 이용한 미호강 수변환경 토지피복 분류

Mingeun Songa
,
Chanul Choib
,
Van Linh Nguyenc
,
Giha Leed*
송 민근a
,
최 찬울b
,
Nguyen Van Linhc
,
이 기하d*
aPh.D. Student, School of Advanced Science and Technology Convergence, Kyungpook National University, Sangju, Korea
bPh.D. Student, School of Advanced Science and Technology Convergence, Kyungpook National University, Sangju, Korea
cPost-doctoral Researcher, Climate Crisis Institute, Kyungpook National University, Sangju, Korea
dProfessor, School of Advanced Science and Technology Convergence, Kyungpook National University, Sangju, Korea
a경북대학교 미래과학기술융합학부 박사과정
b경북대학교 미래과학기술융합학부 박사과정
c경북대학교 기후위기재난대응연구소 박사후연구원
d경북대학교 미래과학기술융합학부 교수
*Corresponding Author
31 October 2026. pp. 1285-1297
PDF XML Citation
Abstract

Monitoring land cover changes in riparian zones is important for quantitatively assessing river environmental changes associated with rainfall and flood events and for supporting river management. This study applied deep learning-based semantic segmentation models to PlanetScope high-resolution multispectral imagery (3 m) to classify land cover in the riparian zone of the Miho River, a tributary of the Geum River, and analyzed changes before and after a rainfall-related water-level rise in 2025. Training data were constructed from summer imagery acquired over the upper and lower reaches of the Miho River in 2023 and 2024, yielding 2,507 patches. Labels for five classes—water, grassland, woody vegetation, bare ground, and road—were generated by combining Google Dynamic World v1 with the Ministry of Environment land cover map and refined through boundary-pixel exclusion and spectral-index-based masking. Ten-channel inputs consisting of eight PlanetScope reflectance bands plus NDVI and NDWI were used, and U-Net, Attention U-Net, DeepLabV3+, and SegFormer were trained and evaluated using a spatial block split (1,865/343/299 train/validation/test patches). SegFormer (MiT-B4 + FPN) showed the highest performance with OA 96.82%, mIoU 0.8275, and mF1 0.8954, with a road-class IoU of 0.735. Applying SegFormer to imagery acquired before (August 10) and after (September 5) the water-level rise showed that water area increased by 60.3% (+36.2 ha), whereas bare ground decreased by 96.5% (-24.7 ha). These changes were consistent with the expansion of the water-covered area and reduction of exposed sandbars and gravel bars following the observed water-level rise. The results indicate that PlanetScope imagery and deep learning-based semantic segmentation can quantitatively assess riparian land cover changes associated with rainfall-related water-level rises.

Keywords
PlanetScope
Land cover classification
Deep learning
Semantic segmentation
SegFormer
Riparian monitoring

하천 수변구역의 토지피복 변화를 모니터링하는 것은 강우·홍수 사상에 따른 하천환경 변화를 정량적으로 파악하고, 하천 관리를 위한 기초자료를 확보하는데 중요하다. 본 연구는 금강 지류인 미호강 수변구역을 대상으로 PlanetScope 고해상도 다중분광 위성영상(3 m)에 딥러닝 기반 의미론적 분할 모델을 적용하여 토지피복을 분류하고, 2025년 수위 상승 전·후의 토지피복 변화를 분석하였다. 2023~2024년 미호강 상·하류 여름철 영상으로부터 총 2,507개 패치를 추출하고, Google Dynamic World v1과 환경부 중분류 토지피복도를 결합한 라벨에 경계부 화소 제외 및 분광지수 기반 마스킹을 적용하여 수체, 초지, 목본, 나지, 도로의 5개 클래스 학습 데이터셋을 구축하였다. PlanetScope 8밴드 반사율에 NDVI·NDWI를 추가한 10채널 입력과 공간 블록 분할(훈련 1,865/검증 343/테스트 299 패치)을 적용하여 U-Net, Attention U-Net, DeepLabV3+, SegFormer 4개 모델을 동일 조건으로 학습·비교하였다. 테스트셋 평가 결과 SegFormer (MiT-B4+FPN)가 전체 정확도 96.82%, mIoU 0.8275, mF1 0.8954로 가장 높은 성능을 나타냈으며, Att-U-Net (mIoU 0.7965), U-Net (0.7691), DeepLabV3+(0.7251)에 비해 특히 도로 클래스에서 높은 IoU (0.735)를 보였다. 학습된 SegFormer 모델을 수위 상승 전(8월 10일)·후(9월 5일) 영상에 적용한 결과, 강우 후 수체 면적은 60.0 ha에서 96.2 ha로 60.3%(+36.2 ha) 증가한 반면, 나지는 25.6 ha에서 0.9 ha로 96.5% 감소하였다. 이는 수위 상승과 함께 하천 내 노출 나지(모래톱·자갈밭)가 감소하고 수체 면적이 확대된 양상을 정량적으로 보여준다. 본 연구는 PlanetScope 고해상도 위성영상과 딥러닝 기반 의미론적 분할을 활용하여 수위 상승 전·후의 하천 수변구역 토지피복 변화를 정량적으로 파악할 수 있음을 제시하였다.

키워드
PlanetScope
토지피복 분류
딥러닝
의미론적 분할
SegFormer
수변환경 모니터링
References
1

Allan, J.D. (2004). “Landscapes and riverscapes: The influence of land use on stream ecosystems.” Annual Review of Ecology, Evolution, and Systematics, Vol. 35, pp. 257-284.

10.1146/annurev.ecolsys.35.120202.110122
2

Bizzi, S., Demarchi, L., Grabowski, R.C., Weissteiner, C.J., and Van de Bund, W. (2016). “The use of remote sensing to characterise hydromorphological properties of European rivers.” Aquatic Sciences, Vol. 78, pp. 57-70.

10.1007/s00027-015-0430-7
3

Brown, C.F., Brumby, S.P., Guzder-Williams, B., Birch, T., Hyde, S.B., Mazzariello, J., and Tait, A.M. (2022). “Dynamic World, Near real-time global 10 m land use land cover mapping.” Scientific Data, Vol. 9, 251.

10.1038/s41597-022-01307-4PMC9184477
4

Chen, L.C., Zhu, Y., Papandreou, G., Schroff, F., and Adam, H. (2018). “Encoder-decoder with atrous separable convolution for semantic image segmentation.” Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany, pp. 833-851.

10.1007/978-3-030-01234-2_49
5

Cooley, S.W., Smith, L.C., Stepan, L., and Mascaro, J. (2017). “Tracking dynamic northern surface water changes with high-frequency planet CubeSat imagery.” Remote Sensing, Vol. 9, No. 12, 1306.

10.3390/rs9121306
6

Deng, J., Dong, W., Socher, R., Li, L. J., Li, K., and Fei-Fei, L. (2009). “Imagenet: A large-scale hierarchical image database.” 2009 IEEE Conference on Computer Vision and Pattern Re-cognition, IEEE, Miami, FL, U.S., pp. 248-255.

10.1109/CVPR.2009.5206848
7

Furuya, D.E.G., Aguiar, J.A.F., Estrabis, N.V., Pinheiro, M.M.F., Furuya, M.T.G., Pereira, D.R., Gonçalves, W.N., Liesenberg, V., Li, J., Marcato Junior, J., Osco, L.P., and Ramos, A.P.M. (2020). “A machine learning approach for mapping forest vegetation in riparian zones in an Atlantic biome environment using Sentinel-2 imagery.” Remote Sensing, Vol. 12, No. 24, 4086.

10.3390/rs12244086
8

Jeong, B., Lee, S., and Lee, M. (2025). “Comparative analysis of AI performance for riparian zone land use and land cover classification using loss functions adapted for data imbalance.” Korean Journal of Remote Sensing, Vol. 41, No. 2, pp. 489-499.

10.7780/kjrs.2025.41.2.2.8
9

Kim, J., Song, Y., and Lee, W.K. (2021). “Accuracy analysis of multi-series phenological landcover classification using U-Net-based deep learning model focusing on the Seoul, Republic of Korea. Korean Journal of Remote Sensing, Vol. 37, No. 3, pp. 409-418.

10.7780/KJRS.2021.37.3.4
10

Konapala, G., Kumar, S.V., and Ahmad, S.K. (2021). “Exploring Sentinel-1 and Sentinel-2 diversity for flood inundation mapping using deep learning.” ISPRS Journal of Photogrammetry and Remote Sensing, Vol. 180, pp. 163-173.

10.1016/j.isprsjprs.2021.08.016
11

Kutz, K., Cook, Z., and Linderman, M. (2022). “Object based classification of a riparian environment using ultra-high resolution imagery, hierarchical landcover structures, and image texture.” Scientific Reports, Vol. 12, 11291.

10.1038/s41598-022-14757-y35789170PMC9253135
12

Lee, I., Bak, S., Park, M., and Jang, S.W. (2026). “Drone imagery and artificial intelligence-based investigation and management of pollution source distribution in the Miho River mid-watershed.” Journal of Korea Water Resources Association, Vol. 59, No. 6, pp. 613-622. doi: 10.3741/JKWRA.2026.59.6.613.

10.3741/JKWRA.2026.59.6.613
13

Lee, S.Y., Jang, R.H., Han, Y.S., Jung, Y.H., Lee, S.I., Lee, E.P., and You, Y.H. (2018). “Health condition assessment using the riparian vegetation index and vegetation analysis of Geumgang mainstream and Mihocheon.” Korean Journal of Environment and Ecology, Vol. 32, No. 1, pp. 105-117. doi: 10.13047/KJEE.2018.32.1.105.

10.13047/KJEE.2018.32.1.105
14

Lin, T.Y., Dollár, P., Girshick, R., He, K., Hariharan, B., and Belongie, S. (2017). “Feature pyramid networks for object detection.” 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, Honolulu, HI, U.S., pp. 936-944.

10.1109/CVPR.2017.106
15

Oktay, O., Schlemper, J., Le Folgoc, L., Lee, M., Heinrich, M., Misawa, K., Mori, K., McDonagh, S., Hammerla, N.Y., Kainz, B., Glocker, B., and Rueckert, D. (2018). “Attention U-Net: Learning where to look for the pancreas.” arXiv preprint, arXiv:1804.03999.

16

Park, J., and Kang, G. (2022). “Water-related disaster monitoring using satellite big data based on Google Earth Engine cloud computing platform.” Korean Journal of Remote Sensing, Vol. 38, No. 6, pp. 1761-1775.

17

Rivas-Fandiño, P., Acuña-Alonso, C., Novo, A., Pacheco, F.A.L., and Álvarez, X. (2023). “Assessment of high spatial resolution satellite imagery for monitoring riparian vegetation: riverine management in the smallholding.” Environmental Monitoring and Assessment, Vol. 195, No. 1, 81.

10.1007/s10661-022-10667-836342553PMC9640423
18

Ronneberger, O., Fischer, P., and Brox, T. (2015). “U-Net: Convolutional networks for biomedical image segmentation.” International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), Munich, Germany, pp. 234-241.

10.1007/978-3-319-24574-4_28
19

Roy, A.G., Navab, N., and Wachinger, C. (2018). “Concurrent spatial and channel ‘squeeze & excitation’ in fully convolutional networks.” International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), Granada, Spain, pp. 421-429.

10.1007/978-3-030-00928-1_48
20

Shin, W.J. and Kim, J.Y. (2025). “A study on the effect of geomorphic characteristics in the Mihogang River channel on the 2023 flood event.” Journal of the Association of Korean Geographers, Vol. 14, No. 2, pp. 129-145. doi: 10.25202/JAKG.14.2.5.

10.25202/JAKG.14.2.5
21

Tan, M., and Le, Q. (2019). “Efficientnet: Rethinking model scaling for convolutional neural networks.” International Conference on Machine Learning, PmLR, Long Beach, CA, U.S., pp. 6105-6114.

22

Tarpanelli, A., Mondini, A.C., and Camici, S. (2022). “Effectiveness of Sentinel-1 and Sentinel-2 for flood detection assessment in Europe.” Natural Hazards and Earth System Sciences, Vol. 22, No. 8, pp. 2473-2489.

10.5194/nhess-22-2473-2022
23

Tockner, K., and Stanford, J.A. (2002). “Riverine flood plains: Present state and future trends.” Environmental Conservation, Vol. 29, No. 3, pp. 308-330.

10.1017/S037689290200022X
24

Xie, E., Wang, W., Yu, Z., Anandkumar, A., Alvarez, J.M., and Luo, P. (2021). “SegFormer: Simple and efficient design for semantic segmentation with transformers.” Advances in Neural Information Processing Systems, Vol. 34, pp. 12077-12090.

25

Zanaga, D., Van De Kerchove, R., De Keersmaecker, W., Souverijns, N., Brockmann, C., Quast, R., and Arino, O. (2022). ESA WorldCover 10 m 2021 v200, accessed 20 April 2026, <https://doi.org/10.5281/zenodo.7254221>.

10.5281/zenodo.7254221>
Information
  • Publisher :KOREA WATER RESOURECES ASSOCIATION
  • Publisher(Ko) :한국수자원학회
  • Journal Title :Journal of Korea Water Resources Association
  • Journal Title(Ko) :한국수자원학회 논문집
  • Volume : 59
  • No :10
  • Pages :1285-1297
  • Received Date : 2026-06-11
  • Revised Date : 2026-08-19
  • Accepted Date : 2026-08-23
  • DOI :https://doi.org/10.3741/JKWRA.2026.59.10.1285
Journal Informaiton Journal of Korea Water Resources Association Journal of Korea Water Resources Association
Online Submission
Archives
  • scopus
  • KCI-수자원
  • NRF
  • KOFST
  • KISTI Current Status
  • KISTI Cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close