提要:图片城市主义范式下,街景照片成为游客感知城市环境的一种新型大数据。然而,由于技术限制,基于照片大数据的旅游情感研究仍处于初步探索阶段。本文从情感地理学视角出发,引入计算机视觉领域的深度学习方法,实证测度了西安市典型旅游街区的情感空间特征。研究发现:基于街景照片的深度学习模型能够比较准确地探测城市旅游情感空间特征;西安市典型旅游街区旅游积极和消极情感在空间上呈现不同类型的集聚格局;西安市典型旅游街区内部呈现多种旅游情感复杂交错的特征;问卷数据一定程度上验证了深度学习模型预测结果的合理性。研究结果为游客与城市环境互动研究提供了新的方法论视角,也为城市旅游环境的改造和规划提供了理论依据和实践指导。关键词关键词:街景照片;深度学习;旅游情感;空间探测中图分类号图分类号:F590.1文献标识码文献标识码:A基于街景照片的城市旅游情感空间探测研究卢欢1,2,杨淼甜1,2,张妍妍1,2,李君轶1,2(1.陕西师范大学地理科学与旅游学院,西安710119;2.陕西省旅游信息科学重点实验室,西安710119)URBANTOURISMSENTIMENTSPATIALDETECTIONURBANTOURISMSENTIMENTSPATIALDETECTIONBASEDONSTREETVIEWPHOTOSBASEDONSTREETVIEWPHOTOSLUHuan1,2,YANGMiao-tian1,2,ZHANGYan-yan1,2,LIJun-yi1,2(1.SchoolofGeographyandTourism,ShaanxiNormalUniversity,Xi'an710119,China;2.ShaanxiKeyLaboratoryofTourismInformatics,Xi'an710119,China)AbstractAbstract:Undertheparadigmofpictorialurbanism,streetviewphotoshavebecomeanewtypeofbigdatatechniquefortouriststoperceivetheurbanenvironment.However,duetotechnicallimitations,tourismsenti‐mentresearchesbasedonphotosarestillatthepreliminaryexplorationstage.Thispaperintroducesadeeplearningmodelinordertomeasurethetourismemotionsofstreetviewphotosbythecomputervision.Item‐piricallyanalysestheemotionalspatialcharacteristicsoftypicaltouristblocksinXi'an.Itshowsthedeeplearningmodelcanaccuratelydetectthespatialcharacteristicsofurbantourismemotionsusingstreetviewphotosasadatasource;thepositiveemotionsandnegativeemotionsoftypicaltouristblocksinXi'anshowdifferenttypesofagglomerationpatterns;thetourismemotionsinsidetypicaltouristblocksinXi'anshowcomplexanddiversecharacteristics.ItisfoundthatalthoughtheBella...