内蒙古视觉旅游形象的时空感知研究——基于DeepSentiBank的地理标记照片分析

Spatiotemporal analysis of visual tourism images in Inner Mongolia from the perspective of tourists

本研究以内蒙古为案例,运用DeepSentiBank解析了257633张地理标记照片的视觉内容,进而从游客感知视角分析了内蒙古旅游景观和视觉旅游形象的季节性差异及其空间分布特征。结果发现:1)内蒙古四季景观各具特色,春季沙漠、夏季草原、秋季林木、冬季冰雪是游客在各季节感知的突出景观;2)自然风光是内蒙古各季节最为突出的视觉旅游形象;3)内蒙古游客在照片中的情感表达存在显著的正面倾向,且正向评价的游客比例远高于负向评价的游客比例;4)内蒙古自然风光、人物、设施、动物四类旅游形象呈现"中东部集中连片,西部稀疏孤立"的空间分布特征,而建筑、植物旅游形象呈现"稀疏分散,极核显著"的空间分布特征。本研究对内蒙古旅游形象优化和旅游营销策略提升有一定的借鉴意义。

This research aims at finding seasonal visual tourism images in Inner Mongolia Autonomous Region.We collected 257,633 geo-tagged photos from Mafengwo app,and applied DeepSentiBank to analyze the visual content of these photos.The spatial distribution of tourism images in Inner Mongolia was also explored through kernel density estimation.The result reveals that different seasons have different landscape in Inner Mongolia.Tourists love grassland in summer and trees in autumn,while desert in spring and snow and ice in winter also impress tourists.Natural sceneries are the most prominent visual tourism images in all seasons of Inner Mongolia.We also found that,most tourists expressed positive emotions in their photos.In addition,the spatial distribution of tourism images natural scenery,people,facilities and animals in Inner Mongolia shows gathering and connecting in central and eastern areas,while sparse and isolated in west areas.However,architectural image and plant image present sparse decentralization but existing a few agglomeration centers in spatial distribution.Our research is helpful to optimize tourism image and improve tourism marketing strategy in Inner Mongolia.