透過您的圖書館登入
IP:44.210.103.233
  • 期刊

A Deep Learning Approach for Building Segmentation in Taiwan Agricultural Area Using High Resolution Satellite Imagery

應用深度學習於高解析衛星影像臺灣農業區建物分塊

摘要


Understanding buildings in agricultural area is important because the arable land in Taiwan is limited. One of the practical ways is manual digitization from high resolution satellite imagery, which can acquire satisfying results without field investigation. However, such practice is tedious and labour intensive. Given these reasons, past research devoted to deep learning approaches have shown that convolutional neural networks are useful for building segmentation using satellite imagery. In this study, ENVINet5 model was trained and utilized from high resolution Pléiades pansharpened imagery. The training images (with the size of 2500 pixels × 2500 pixels) were randomly selected from 9 counties/cities to increase diversity because each county/city has different building patterns. The performance of ENVINet5 model was evaluated based on pixels and polygons, respectively. The pixel-based evaluation showed that the trained model can find 84% of building pixels. The polygon-based evaluation was carried out through calculating the number of building segments and comparing them with the reference data using IoU (Intersection of Union). The results showed that 92% of building segments were found, and the IoU of most building segments range between 0.6 and 0.9. The trained model was validated on the testing images for the transferability test. Moreover, an image tiling and stitching technique was proposed to deal with large satellite imagery. Finally, we compared the time costs of labelling with and without the aid of deep learning approach. The results showed that the time costs decreased by 7.3% with the help of deep learning approach.

並列摘要


臺灣的可耕地面積有限,清查建物的面積有助於了解土地利用的狀況。為了瞭解建物在臺灣農業區所佔的總面積,現有的做法之一是透過高解析衛星影像進行人工辨識,此法可以掌握建物的邊界、改善現地調查的不便。然而,卻需要大量人力資源的投入。過去的研究顯示,深度學習的方法可以有效地在高解析衛星影像進行建物分塊。因此,本研究使用ENVINet5深度學習模型及Pléiades彩色融合影像進行訓練,針對臺灣的農業區進行建物分塊。因為各地區的建物型態皆不相同,所以本研究使用九個不同的縣市的影像進行訓練,每張訓練影像的尺寸為2500像素×2500像素。模型的評估是透過驗證集中的像素以及分塊後的建物多邊形進行計算。前者結果顯示,經訓練的模型可以找出84%的建物像素;後者計算了建物多邊形的數量,並將其與參考建物以IoU(Intersection of Union)做比較。成果顯示,該模型可以在影像上偵測且分塊92%的建物,其IoU集中於0.6到0.9之間。該模型也以測試集做可轉移性試驗。另外,本研究提出了影像切圖與拼接的方法以處理大範圍的衛星影像。最後,我們將ENVINet5的成果輔助人工辨識建物,可以節省7.3%的時間成本。

參考文獻


Alshehhi, R., Marpu, P.R., Woon, W.L., and Dalla Mura, M., 2017. Simultaneous extraction of roads and buildings in remote sensing imagery with convolutional neural networks, ISPRS Journal of Photogrammetry and Remote Sensing, 130: 139-149.
Boonpook, W., Tan, Y., Ye, Y., Torteeka, P., Torsri, K., and Dong, S., 2018. A deep learning approach on building detection from unmanned aerial vehicle- based images in riverbank monitoring, Sensors, 18(11): 3921.
Chen, Y., Tang, L., Yang, X., Bilal, M., and Li, Q., 2020. Object-based multi-modal convolution neural networks for building extraction using panchromatic and multispectral imagery, Neurocomputing, 386: 136-146.
Esetlili, T.M., Balcik, F.B., Sanli, F.B., Ustuner, M., Kalkan, K., Goksel, C., Gazioğlu, C., and Kurucu, Y., 2018. Comparison of object and pixel-based classifications for mapping crops using rapideye imagery: A case study of menemen plain, Turkey, International Journal of Environment and Geoinformatics, 5(2): 231-243.
Gu, J., Wang, Z., Kuen, J., Ma, L., Shahroudy, A., Shuai, B., Liu, T., Wang, X., Wang, G., Cai, J., and Chen, T., 2017. Recent advances in convolutional neural networks, Pattern Recognition, 77: 354-377.

延伸閱讀