Spatial Analysis And Remote-Sensing-Based Detection Of Neolithic Archaeological Sites In Lake Taihu Basin

Spatial information technology has been applied more widely and more deeply in the archaeological research than ever.There is five thousand years long history in China with rich and diverse cultural heritage,making it possible for archaeological remote sensing to play an important role in it.However,there are diverse and complex geomorphic types,which pose enormous challenges for archaeological research using remote sensing.The research objects are the neolithic archaeological sites in Lake Taihu Basin,whose excellent natural conditions gave birth to famous neolithic cultures such as Majiabang,Songze and Liangzhu.Though many representative archaeological sites were discovered,there are still more unknown sites which have changed their location,depth,shape and size during the long history of evolution.Some sites are buried in the farm land or modern building.This situation makes it difficult to detect these sites by usual archaeological methods,while the spatial information technology can overcome this difficulty to some extent.In order to deepen the application of spatial information technology in archaeological research,this paper fully discusses the application of spatial analysis methods and multi-source remote sensing data in archaeological application.First of all,this paper investigates the distribution characteristics of archaeological sites during the Neolithic Age,then quantitatively analyzes the relationship between sites and environmental variables.The correlation between environmental variables are calculated by Spearman correlation coefficient method in order to remove the variables with high correlation.After that,a maximum entropy model is employed to predict the site location probability.The impact of each variable are assessed and random samples are selected to test the effectiveness of this model.Moreover,the AUC value(the area under the ROC curve)are analyzed to assess the accuracy of model.Then the predictive results are classified by the threshold calculated from the model run.The areas of interest with high probability are extracted in order to detect the archaeological features using multi-source remote sensing images.The data fusion,spectral indices,orthogonal equations and spatial autocorrelation are performed to identify the weak archaeological information,which provides reference for the archaeological feature exploration in the subtropical monsoon climate zone.The result demonstrate that Maxent can be performed to predict the archaeological sites effectively.Moreover,the historical images,very high resolution images and Sentinel-2A can be processed to detect ancient ruins,ancient tombs and ancient river channels with unique advantages.Finally,with the integration of spatial predictive result,multi-source image and geological and geomorphological research results,this paper quantitatively analyzes the land use in the Liangzhu site group and the relationship between the Liangzhu cultural development and river change.This paper analyzes the spatial geographical features of neolithic archaeological site in Lake Taihu Basin,applying maximum entropy model to predict the unknown sites.Moreover,multi-source remote sensing data are processed and analyzed.Then the space detection system of archaeological features from macro to micro are proposed for the first time,which provides a new perspective for the archaeological site protection,city construction and social sustainable development.