Keywords
Building rooftops
HSRRS
Mask R-CNN
Instance segmentation
Submitted : 2024-04-22
Accepted : 2024-05-07
Published : 2024-05-22
Abstract
Traditional Chinese villages serve as crucial repositories of traditional culture. However, In China, the urgent task of preserving information about traditional village architecture has arisen due to the degradation of these villages’ appearance caused by rapid urbanization in recent years. This paper proposes a method for preserving information about traditional village rooftops based on high spatial resolution remote sensing imagery. Leveraging an improved Mask R-CNN model, the method conducts target recognition on the rooftops of traditional village buildings and generates vectorized representations of these rooftops. The precision rate, recall rate, and F1-score achieved in the experimental results are 93.26%, 86.33%, and 92.02%, respectively. These findings indicate the effectiveness of the proposed method in preserving information about traditional village architecture and providing a viable approach to support the sustainable development of traditional villages in China.
References
Xiong Y, Chen Q, Zhu M, et al., 2020, Accurate Detection of Historical Buildings Using Aerial Photographs and Deep Transfer Learning, IEEE International Symposium on Geoscience and Remote Sensing (IGARSS), 1592–1595.
Liu J, Wang S, Hou X, et al., 2020, A Deep Residual Learning Serial Segmentation Network for Extracting Buildings From Remote Sensing Imagery. International Journal of Remote Sensing, 41(14): 5573–5587.
He K, Gkioxari G, Dollár P, et al., 2017, Mask R-CNN. Proceedings of the IEEE International Conference on Computer Vision, 2961–2969.
Li Y, Xu W, Chen H, et al., 2021, A Novel Framework Based on Mask R-CNN and Histogram Thresholding for Scalable Segmentation of New and Old Rural Buildings. Remote Sensing, 13(6): 1070.
Zhan Y, Liu W, Maruyama Y, 2022, Damaged Building Extraction Using Modified Mask R-CNN Model Using Post-Event Aerial Images of the 2016 Kumamoto Earthquake. Remote Sensing, 14(4): 1002.
Tejeswari B, Sharma SK, Kumar M, et al., 2022, Building Footprint Extraction from Space-Borne Imagery using Deep Neural Networks. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 43: 641–647.
Wang W, Shi Y, Zhang J, et al., 2023, Traditional Village Building Extraction Based on Improved Mask R-CNN: A Case Study of Beijing, China. Remote Sensing, 15(10): 2616.
Wang M, Shen L, 2024, High-Resolution Remote Sensing Imagery for the Recognition of Traditional Villages. Journal of Architectural Research and Development, 8(1): 75–83.
Radford A, Kim J W, Hallacy C, et al., 2021, Learning Transferable Visual Models from Natural Language Supervision. International Conference on Machine Learning. PMLR, 8748–8763.
Zhan Y, Liu W, Maruyama Y, 2022, Damaged Building Extraction Using Modified Mask R-CNN Model Using Post-Event Aerial Images of the 2016 Kumamoto Earthquake. Remote Sensing, 14(4): 1002.