Visual Image Interpretation For Kailash Sacred Landscape Through Social Media

With the rapid development of various social media and photo sharing websites,it seems that Internet has entered a “graph-reading era”.Social media photos not only record people’s aesthetic taste and expression of the scenery,but also in turn influence their understanding on the landscape.In this context,this paper discusses the visual image of Kailash Sacred Landscape based on social media photos,hoping to innovate the research approach of landscape visual image from the perspective of pictorial big data.Taking “vision” as a starting point,this paper reviewed the relevant research on landscape from two perspectives: “visual environment” and “visual media”.Then,the convergence and distinction of landscape image research in various research fields were distinguished.Finally,we focused on visual landscape research in the era of big data,so as to build a theoretical basis for the construction of a cognitive system of landscape image.After a brief introduction of the historical background of Kailash Sacred Landscape,this paper established a database for landscape image research with geospatial data and social media data,and also the research areas were delimited.Considering the characteristics of big data and artificial intelligence technology,a smart landscape image cognitive system was built,which includes four parts: image environment,image structure,image element and image feeling.Also,the generation mode of visual image was analyzed from three dimensions: the dynamic experience,the mechanism of host-guest interactions and characteristics of visual paradigms.In the research on the cognition of landscape visual image,the analysis on image environment revealed the high visual prominence of Mt.Kailash among the region and the existence of natural visual corridors based on the surrounding valleys.The analysis on image structure revealed the clear and complete image spatial structure formed along the outer circumambulation route.The analysis on image element extracted the visual items from social media photos through image semantic segmentation,and then summarized them into 8 types of image elements,whose frequency and diversity were related to the visibility of Mt.Kailash.The analysis on image feeling distilled six types of feeling from the related text of photos,whose distribution showed significant spatial aggregation and certain similarity.In the research on the generation mode of landscape visual image,the analysis on the dynamic experience of landscape image revealed the consistency among landscape imageability,image element and image feeling which all influenced by the landscape visibility and image structure.The analysis on the mechanism of host-guest interactions adopted the geographical weighted regression model to analyze the mechanism of “lyricism through things”.The result reflected that theme elements of photos showed the greatest impact on the feeling expression of image.The analysis on visual paradigms of image extracted 9 kinds of paradigms from social media photos based on hierarchical clustering method.The transformation between paradigms partly reflected people’s spiritual sublimation during the process of pilgrimage.This paper verified the effectiveness of the constructed landscape image cognitive system for analyzing the visual image of landscape in social media photos.Finally,targeted strategies were given for the planning and development of Kailash Sacred Landscape according to the research findings.