Abstract: In the context of national wisdom community construction, wisdom parking has become an important element of wisdom community construction. The current community parking management wisdom is generally low, the main root cause of which is the failure to form the dynamic integration and utilization of data resources such as IOT data and scheduling command data. In this paper, facing the demand of community parking management wisdom, for the problems of scattered parking functions, insufficient visualization ability and large differences in data resources, we propose the framework of wisdom parking multimodal data dynamic fusion by studying the wisdom parking scene and related data flow, and implement the application in the community, which solves the practical problems in the parking process and meets the expansion of multimodal data fusion technology in the community parking scene and the wisdom It meets the need for the expansion of multimodal data fusion technology in the community parking scene and the continuation of the construction of intelligent parking, and improves the scientific and intelligent level of urban parking management.

Key words: smart community, intelligent parking, multimodal data fusion framework, parking guidance, reverse car finding