近年來迅速增長的農地內建物面積已造成耕地的流失，並對我國之糧食生產與安全存糧造成威脅。現行的農地使用樣態清查方式是透過空拍航照圖或衛星影像進行人工判釋。鑒於此類方法需要較多的人力資源，難以達成每年清查的目標，本研究的主要目的在於探討如何以高度自動化的方式進行農地內建物清查。發展迅速的深層類神經網路已被廣泛用於遙測影像之地表覆蓋分類，本研究以遷移學習方法結合預訓練網路與FCN網路架構進行像元式的建物偵測。然而受限於預訓練網路所使用的自然色影像資料集，此類網路僅適用於三個波段的影像資料。為使用多光譜影像中的所有波段進行建物偵測，本研究採用FCN-8S-Conv-TL之模型架構，於遷移學習之目標網路前新增一個接收所有波段並產製深度為3的特徵圖的卷積層。本研究使用全色態銳化之Pléiades衛星影像進行實驗，成果達到94.2%的F1 score精度指標，並以高雄沿岸地區及桃園蘆竹區之影像進行可轉移性試驗。

The rapid growth of built-up surfaces in agricultural land has caused the loss of arable land and pose a threat to food crops production in recent years. Common practices of tallying buildings in agricultural land are based on manual digitization using aerial photos and satellite images. Such practices are tedious and consuming, making it difficult to carry out the investigation annually. As a result, it is on an urgent need to develop a highly automated method to fulfill the task. Emerging deep convolutional neural network (CNN)-based methods have been widely used to solve land cover classification problems. We adapted off-the-shelf pre-trained classification network into fully convolutional networks (FCNs) using transfer learning technique for pixel-wise building detection. However, the pre-trained network restricts the number of input bands to 3 since it was trained on a natural scene image dataset. FCN-8s-Conv-TL is proposed to make full use of all spectral bands by adding a convolutional layer which generates 3-channel feature maps before the pre-trained network. Experiments on a pansharpened Pléiades satellite image dataset casting VGG19, a pre-trained network for large-scale image classification, into FCN-8s-Conv-TL were conducted. The classification accuracy F1 score of 94.2% was achieved. Furthermore, transferability of the network was validated using image subsets in different areas.

Badrinarayanan, V., Kendall, A., &Cipolla, R. (2015). SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation. Retrieved from http://arxiv.org/abs/1511.00561
Chen, L.-C., Zhu, Y., Papandreou, G., Schroff, F., &Adam, H. (2018). Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation. Retrieved from https://arxiv.org/abs/1802.02611
Feng, Y., Lu, D., Moran, E., Dutra, L., Calvi, M., deOliveira, M., …DeOliveira, M. A. F. (2017). Examining Spatial Distribution and Dynamic Change of Urban Land Covers in the Brazilian Amazon Using Multitemporal Multisensor High Spatial Resolution Satellite Imagery. Remote Sensing, 9(4), 381. https://doi.org/10.3390/rs9040381
Food and Agriculture Organization of the United Nations. (2002). The State of Food Insecurity in the World 2002.
Fu, G., Liu, C., Zhou, R., Sun, T., &Zhang, Q. (2017). Classification for High Resolution Remote Sensing Imagery Using a Fully Convolutional Network. Remote Sensing, 9(5). https://doi.org/10.3390/rs9050498
Garcia-Garcia, A., Orts-Escolano, S., Oprea, S., Villena-Martinez, V., &Garcia-Rodriguez, J. (2017). A Review on Deep Learning Techniques Applied to Semantic Segmentation. Retrieved from http://arxiv.org/abs/1704.06857
Jiao, L., Liang, M., Chen, H., Yang, S., Liu, H., &Cao, X. (2017). Deep Fully Convolutional Network-Based Spatial Distribution Prediction for Hyperspectral Image Classification. IEEE Transactions on Geoscience and Remote Sensing, 55(10), 5585–5599. https://doi.org/10.1109/TGRS.2017.2710079
Kastens, J. H., Kastens, T. L., Kastens, D. L. A., Price, K. P., Martinko, E. A., &Lee, R. Y. (2005). Image masking for crop yield forecasting using AVHRR NDVI time series imagery. Remote Sensing of Environment. https://doi.org/10.1016/j.rse.2005.09.010
Krizhevsky, A., Sutskever, I., &Hinton, G. E. (2012). ImageNet Classification with Deep Convolutional Neural Networks. Proceedings of the 25th International Conference on Neural Information Processing Systems - Volume 1, 1097–1105. Retrieved from http://dl.acm.org/citation.cfm?id=2999134.2999257
Kumar, B., Pandey, G., Lohani, B., &Misra, S. C. (2019). A multi-faceted CNN architecture for automatic classification of mobile LiDAR data and an algorithm to reproduce point cloud samples for enhanced training. ISPRS Journal of Photogrammetry and Remote Sensing, 147, 80–89. https://doi.org/10.1016/J.ISPRSJPRS.2018.11.006
Long, J., Shelhamer, E., &Darrell, T. (2015). Fully convolutional networks for semantic segmentation. 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 3431–3440. https://doi.org/10.1109/CVPR.2015.7298965
Maggiori, E., Charpiat, G., Tarabalka, Y., &Alliez, P. (2016). Recurrent Neural Networks to Correct Satellite Image Classification Maps. https://doi.org/10.1109/TGRS.2017.2697453
Marcos, D., Hamid, R., &Tuia, D. (2016). Geospatial Correspondences for Multimodal Registration. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 5091–5100. https://doi.org/10.1109/CVPR.2016.550
Pan, B., Shi, Z., Xu, X., Shi, T., Zhang, N., &Zhu, X. (2019). CoinNet: Copy Initialization Network for Multispectral Imagery Semantic Segmentation. IEEE Geoscience and Remote Sensing Letters, 16(5), 816–820. https://doi.org/10.1109/LGRS.2018.2880756
Pan, S. J., &Yang, Q. (2010). A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering, 22(10), 1345–1359. https://doi.org/10.1109/TKDE.2009.191
Paszke, A., Chaurasia, A., Kim, S., &Culurciello, E. (2016). ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation. Retrieved from https://arxiv.org/abs/1606.02147
Piramanayagam, S., Saber, E., Schwartzkopf, W., &Koehler, F. W. (2018). Supervised classification of multisensor remotely sensed images using a deep learning framework. Remote Sensing, 10(9), 1–25. https://doi.org/10.3390/rs10091429
Ronneberger, O., Fischer, P., &Brox, T. (2015). U-Net: Convolutional Networks for Biomedical Image Segmentation. Retrieved from http://arxiv.org/abs/1505.04597
Shin, H.-C., Roth, H. R., Gao, M., Lu, L., Xu, Z., Nogues, I., …Summers, R. M. (2016). Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning. IEEE Transactions on Medical Imaging, 35(5), 1285–1298. https://doi.org/10.1109/TMI.2016.2528162
Simonyan, K., &Zisserman, A. (2014). Very Deep Convolutional Networks for Large-Scale Image Recognition. Retrieved from http://arxiv.org/abs/1409.1556
Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., &Salakhutdinov, R. (2014). Dropout: A Simple Way to Prevent Neural Networks from Overfitting. Journal of Machine Learning Research, 15, 1929–1958. Retrieved from http://jmlr.org/papers/v15/srivastava14a.html
Sublime, J., Troya-Galvis, A., Puissant, A., Sublime, J., Troya-Galvis, A., &Puissant, A. (2017). Multi-Scale Analysis of Very High Resolution Satellite Images Using Unsupervised Techniques. Remote Sensing, 9(5), 495. https://doi.org/10.3390/rs9050495
Wang, D., Wan, B., Qiu, P., Su, Y., Guo, Q., Wang, R., …Wu, X. (2018). Evaluating the Performance of Sentinel-2, Landsat 8 and Pléiades-1 in Mapping Mangrove Extent and Species. Remote Sensing, 10(9). https://doi.org/10.3390/rs10091468
Wang, J., Shen, L., Qiao, W., Dai, Y., &Li, Z. (2019). Deep Feature Fusion with Integration of Residual Connection and Attention Model for Classification of VHR Remote Sensing Images. Remote Sensing, 11(13). https://doi.org/10.3390/rs11131617
Wei, Y., Wang, Z., &Xu, M. (2017). Road Structure Refined CNN for Road Extraction in Aerial Image. IEEE Geoscience and Remote Sensing Letters, 14(5), 709–713. https://doi.org/10.1109/LGRS.2017.2672734
Wu, G., Guo, Y., Song, X., Guo, Z., Zhang, H., Shi, X., …Shao, X. (2019). A Stacked Fully Convolutional Networks with Feature Alignment Framework for Multi-Label Land-cover Segmentation. Remote Sensing, 11(9). https://doi.org/10.3390/rs11091051
Zeiler, M. D., &Fergus, R. (2013). Visualizing and Understanding Convolutional Networks. Retrieved from https://arxiv.org/abs/1311.2901
內政部. (2017). 修正全國區域計畫.