SEGMENTASI BANGUNAN PERKOTAAN PADA CITRA SATELIT BERESOLUSI TINGGI: CNN, U-NET (VGG16), DAN DEEPLABV3+ (RESNET-50)
DOI:
https://doi.org/10.55123/storage.v4i4.6552Keywords:
Segmentasi Bangunan, CNN, U-Net, DeepLabV3 , Citra Satelit, UrbanisasiAbstract
Seiring meningkatnya laju urbanisasi di Indonesia, kebutuhan pemetaan bangunan yang akurat menjadi semakin penting untuk mendukung perencanaan tata ruang, mitigasi bencana, dan pengelolaan infrastruktur perkotaan. Pendekatan konvensional berbasis survei manual dinilai kurang efisien, terutama di wilayah dengan pertumbuhan pesat. Oleh karena itu, pemanfaatan citra satelit dan Deep learning menjadi solusi potensial untuk identifikasi bangunan secara otomatis. Penelitian ini membandingkan performa tiga model segmentasi bangunan pada citra satelit resolusi tinggi: CNN konvensional (CNN-K), U-Net berbasis VGG16 (U-VGG), dan DeepLabV3+ dengan ResNet-50 (DL-ResNet). Dataset terdiri atas 1.216 patch citra dari kawasan Bali Selatan yang telah dilabeli dan diaugmentasi. Evaluasi dilakukan menggunakan metrik akurasi, IoU, dice coefficient, precision, recall, dan F1-score. Hasil menunjukkan U-VGG unggul (dice 89%, IoU 81%) dengan keseimbangan presisi dan efisiensi, sementara DL-ResNet mendekati hasilnya (dice 85%, IoU 80%) tetapi memerlukan sumber daya komputasi lebih besar. CNN-K mengalami overfitting dengan performa terendah.
Downloads
References
Alsabhan, W., Alotaiby, T. and Dudin, B., 2022. Detecting Buildings and Nonbuildings from Satellite Images Using U-Net. Computational Intelligence and Neuroscience, 2022. https://doi.org/10.1155/2022/4831223.
Altun, M. and Turker, M., 2025. Integration of convolutional neural networks with parcel-based image analysis for crop type mapping from time-series images. Earth Science Informatics, [online] 18(3). https://doi.org/10.1007/s12145-025-01819-8.
Ayala, C., Sesma, R., Aranda, C. and Galar, M., 2021. A deep learning approach to an enhanced building footprint and road detection in high-resolution satellite imagery. Remote Sensing, 13(16), pp.1–21. https://doi.org/10.3390/rs13163135.
Chang, F., Ma, T., Wang, D., Zhu, S., Li, D., Feng, S. and Fan, X., 2025. Method for building segmentation and extraction from high-resolution remote sensing images based on improved YOLOv5ds. PLoS ONE, 20(3 March). https://doi.org/10.1371/JOURNAL.PONE.0317106.
Gibril, M.B.A., Al-Ruzouq, R., Shanableh, A., Jena, R., Bolcek, J., Shafri, H.Z.M. and Ghorbanzadeh, O., 2024. Transformer-based semantic segmentation for large-scale building footprint extraction from very-high resolution satellite images. Advances in Space Research, [online] 73(10), pp.4937–4954. https://doi.org/10.1016/j.asr.2024.03.002.
Li, J., Cai, Y., Li, Q., Kou, M. and Zhang, T., 2024a. A review of remote sensing image segmentation by deep learning methods. International Journal of Digital Earth, [online] 17(1). https://doi.org/10.1080/17538947.2024.2328827.
Li, Z.H., Shi, A.C., Xiao, H.X., Niu, Z.H., Jiang, N., Li, H.B. and Hu, Y.X., 2024b. Robust Landslide Recognition Using UAV Datasets: A Case Study in Baihetan Reservoir. Remote Sensing, 16(14). https://doi.org/10.3390/rs16142558.
Mehta, Y., Baz, A. and Patel, S.K., 2024. Semantic segmentation of optical satellite images for the illegal construction detection using transfer learning. Results in Engineering, [online] 24, p.103383. https://doi.org/10.1016/J.RINENG.2024.103383.
Peng, F., Yao, S., Chen, Y. and Li, W., 2024. Unsupervised Domain Adaptive Transfer Learning for Urban Built-Up Area Extraction. p.10. https://doi.org/10.3390/proceedings2024110010.
Putu, N., Dewi, N.P., Budhi, P. and Purwanta, D., 2021. Big Data for Indonesian Marine Fisheries A Preliminary Research Plan.
Rahman, A.M., Zaber, M., Cheng, Q., Nayem, A.B.S., Sarker, A., Paul, O. and Shibasaki, R., 2021. Applying state-of-the-art deep-learning methods to classify urban cities of the developing world. Sensors, 21(22), pp.1–22. https://doi.org/10.3390/s21227469.
Ramalingam, A., Srivastava, V., George, S., Alagala, S. and Martin Leo Manickam, J., 2024. Building rooftop extraction from aerial imagery using low complexity UNet variant models. Journal of Spatial Science, 69, pp.1–28. https://doi.org/10.1080/14498596.2024.2302166.
Singla, J.G. and Vaghela, B., 2024. Semantic segmentation on multi-resolution optical and microwave data using deep learning. [online] pp.0–3. Available at: <https://arxiv.org/pdf/2411.07581>.
Sundaresan, A.A. and Solomon, A.A., 2025. Post-disaster flooded region segmentation using DeepLabv3+ and unmanned aerial system imagery. Natural Hazards Research, [online] 5(2), pp.363–371. https://doi.org/10.1016/j.nhres.2024.12.003
Suputra, P.H., Sensusiati, A.D., Artaria, M.D., Verkerke, G.J., Yuniarno, E.M. and Purnama, I.K.E., 2022. Automatic 3D Cranial Landmark Positioning based on Surface Curvature Feature using Machine Learning. Knowledge Engineering and Data Science, 5(1), p.27. https://doi.org/10.17977/um018v5i12022p27-40.
Tang, Y., Zhang, G., Liu, J.K. and Qin, R., 2025. Weakly supervised land-cover classification of high-resolution images with low-resolution labels through optimized label refinement. International Journal of Remote Sensing, [online] 46(5), pp.1913–1937. https://doi.org/10.1080/01431161.2024.2443612.
Vasavi, S., Sri Somagani, H. and Sai, Y., 2023. Classification of buildings from VHR satellite images using ensemble of U-Net and ResNet. Egyptian Journal of Remote Sensing and Space Science, [online] 26(4), pp.937–953. https://doi.org/10.1016/j.ejrs.2023.11.008.
Zeng, Y., Guo, Y. and Li, J., 2022. Recognition and extraction of high-resolution satellite remote sensing image buildings based on deep learning. Neural Computing and Applications, [online] 34(4), pp.2691–2706. https://doi.org/10.1007/s00521-021-06027-1.
Zhang, J., Li, Y., Yang, X., Jiang, R. and Zhang, L., 2025. RSAM-Seg: A SAM-Based Model with Prior Knowledge Integration for Remote Sensing Image Semantic Segmentation. Remote Sensing, 17(4), pp.1–26. https://doi.org/10.3390/rs17040590.
Zhang, T., Tang, H., Ding, Y., Li, P., Ji, C. and Xu, P., 2021. Fsrss-net: High-resolution mapping of buildings from middle-resolution satellite images using a super-resolution semantic segmentation network. Remote Sensing, 13(12). https://doi.org/10.3390/rs13122290.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Putu Haryaka Setadewa, Kadek Yota Ernanda Aryanto, Luh Joni Erawati Dewi
This work is licensed under a Creative Commons Attribution 4.0 International License.
