Research on the Application of Multimodal Perception and Fusion Interaction Technology in Smart Parks

To tackle the core challenges in the intelligent management of people, vehicles, facilities, objects, and operations within full-scenario smart parks, this paper develops a multimodal perception system integrating visual, acoustic, biometric, and environmental sensing technologies, drawing on multiple practical park projects. Leveraging the Lingyun Middle Platform of Flying Connected, this system enables unified access to various terminal devices, and accomplishes intelligent management of personnel, vehicles, environments, and emergency responses through three core processes: data denoising, cross-modal feature alignment, and fusion decision-making. Moreover, by virtue of the large-model adaptation capability of Lingyun Agent Studio, a full-link intelligent architecture is constructed. This research proposes a multi-modal fusion framework tailored to smart park scenarios and establishes a lightweight edge-cloud collaborative deployment solution. The industrial value of the proposed technology is verified through practical applications, which provides a replicable technical paradigm for the vertical application of multimodal perception and fusion interaction technologies as well as the digital transformation of smart parks.