古遗址区形迹特征的影像融合信息识别分析——以陕西省五陵原遗址为例

Extraction of Structural Line Features of Tomb Sites Based on Remote Sensing Data:A Case Study in Wuling Hills of the Western Han Dynasty,Shaanxi Province

本研究以陕西省汉五陵原为研究区域,基于高分辨率影像(DOM航片)、多光谱影像(ALOS多光谱影像)和Quickbird融合影像数据,对五陵原遗址区结构性线状特征信息进行了提取。结果表明:融合后使影像隐含的空间纹理得到明显增强。用Gram-Schmidt方法的DOM和ALOS多光谱+NDVI融合影像,清晰地反映了遗址区的结构性线状特征;而Canny算法的边缘检测,能很好地提取遗址区显性和大量隐性的线状特征,对古遗迹的解译具有重要指示意义。

At present,it is widely believed that remote sensing technology is an important means for modern archaeological research.When remote sensing technology is used to assist in archaeological research,it can not only give us more effective means of scientific analysis,but also opportunities to discover or recover the historical monuments that traditional ways can not achieve.In contemporary remote sensing archaeology single source of image data are always used,which recognize the ground or underground ancient ruins with visual interpretation according to expert experience by remote sensing data analysis on the basis of landscape morphology,surface features shadow,frost or snow,land temperature,vegetation and various signs.In this study,high-resolution image DOM,ALOS multi-spectral data and QuickBird fused data are used in detecting tomb archaeological remains in Wuling Hills,Shaanxi Province.Structural line features are extracted based on image fusion algorithms and edge detection algorithms.The results show that： firstly,compared to single data source,the image fusion based on multi-source data can not only maintain the spatial detail of high-resolution images but also the spectral signature of multi-spectral images.Hidden line features can be enhanced by image fusion obviously.It can be a valuable approach for extraction of structural line features of the site.Secondly,the amount of spatial information and spatial distribution patterns will be different based on different spatial resolution and image fusion methods.In term of line feature extraction of the remains,higher resolution does not mean a more suitable data resource to the archaeological research.The fusion of DOM,ALOS＋NDVI images based on Gram-Schmidt algorithm is the most effective method for structural line features retrieval.Finally,the edge detection method will enhance the linear characteristic of the remains.The explicit and a large number of implicit linear characteristics can be obtained by edge detection method using Canny algorithm.It is an important indicator for inspection and interpretation of the remains.