建筑结构壁画形谱遥感特征提取研究进展  

Research progress on shape-spectrum feature extraction of ancient building structures and mural by remote sensing data

作　　者：郭明 唐星雨[1] 程鹏 孙梦溪 郭可才 GUO Ming;TANG Xingyu;CHENG Peng;SUN Mengai;GUO Kecai(School of Geomatics and Urban Spatial Informatics,Beijing University of Civil Engineering and Architecture,Beijing 102616,China;Engineering Research Centre of Representative Building and Architectural Heritage Database,Ministry of Education,Beijing 100044,China;Key Laboratory of Modern Urban Surveying and Mapping,National Administration of Surveying Mapping and Geoinformation,Beijing 100044,China;Beijing Key Laboratory for Architectural Heritage Fine Reconstruction&Health Monitoring,Beijing University of Civil Engineering and Architecture,Beijing 100044,China;Beijing Shenxindacheng Technology Co.Ltd.,Beijing 102444,China)

机构地区：[1]北京建筑大学测绘与城市空间信息学院,北京102616 [2]代表性建筑与古建筑数据库教育部工程研究中心,北京100044 [3]北京建筑大学建筑遗产精细重构与健康监测北京市重点实验室,北京100044 [4]现代城市测绘国家测绘地理信息局重点实验室,北京100044 [5]北京申信达成科技有限公司,北京102444

出　　处：《测绘科学》2023年第8期119-129,共11页Science of Surveying and Mapping

基　　金：国家自然科学基金项目(41971350,52130809);国家重点研究发展计划项目(2022YF0904400,2021YF0602005-03);分类发展定额项目一硕士研究生创新项目(2023年)(03081023002)。

摘　　要：针对建筑结构提取对于建筑遗产保护研究愈发重要,且传统建筑遗产测绘技术具有很大的局限性等问题,该文主要针对建筑结构壁画提取基于高光谱技术、深度学习、其他算法等方法进行综述,系统地分析了各方法的优缺点,进一步阐述了以壁画、古建筑结构为研究对象的提取结果,为建筑遗产保护提供有力的数据支持。研究得出基于高光谱的方法在壁画信息提取方面有较好的应用,但其提取对象较为单一,无法很好地提取较为复杂的结构特征等局限;深度学习的自主性可以避免人为因素的影响,大大提高工作效率;其他算法在提取方面有较为广泛的应用,需要针对具体的提取对象使用具体的提取方法,但总体上不太适用于复杂场景的提取。可以将多种方法进行融合来获取较高精度的结果,是提取建筑遗产特征信息研究方向的一个趋势。In view of the increasing importance of architectural structure extraction for architectural heritage protection research and the significant limitations of traditional architectural heritage mapping technology,the paper mainly focused on the extraction of architectural structure frescoes based on hyperspectral technology,deep learning,other algorithms and other methods for the review,systematically analyzed the advantages and disadvantages of each method,and further described the results of the extraction of frescoes,ancient architectural structures as the object of research to provide robust data support for the protection of architectural heritage.Protection to provided vital data support.The study concluded that the hyperspectral-based method had better application in the extraction of mural information,but its extraction object was relatively single,and it could not well extract the more complex structural features and other limitations;the autonomy of deep learning could avoid the influence of human factors,and significantly improved the efficiency of the work;other algorithms had a wide range of applications in extraction,and it was necessary to use specific extraction methods for the specific extraction object,but overall it was not very suitable for the extraction of complex scenes.The other algorithms had a broader application in extraction and needed to use specific extraction methods for specific extraction objects.However,in general,they were less suitable for extracting complex scenes.It was a trend in the research direction of extracting architectural heritage feature information that multiple methods could be fused to obtain higher precision results.

关 键 词：建筑遗产 建筑结构 壁画 古建筑 特征提取 深度学习 

分 类 号：P237[天文地球—摄影测量与遥感]