The digital protection of Hakka cultural heritage in Henan Province has the technical foundation of multi-source perception, 3D reconstruction, knowledge mapping and collaborative operation. Focusing on the surrounding buildings, traditional villages, folk objects, oral data and activity scenes, this paper constructs a computing framework that integrates image acquisition, point cloud registration, semantic coding, terminal collaboration and feedback update. A total of 4680 samples of indoor and outdoor images, UAV aerial survey data, laser point clouds and text records were collected, including 3020 groups of architectural images and point clouds, and 1660 groups of text and audio samples. The system achieves 93.6% accuracy of geometric reconstruction, 91.9% of texture fitting and 92.8% of semantic mapping completeness, 93.2% of knowledge retrieval success rate, 91.7% of cross-end call consistency rate, and 90.9% of scene update completion rate. This method realizes the structured storage, scene reconstruction, knowledge association and terminal call of heritage information, and can support the display guide, environmental monitoring, content push and service adjustment synchronously, so that the protection, dissemination and sharing run in the unified computing link. The empirical results show that the framework can enhance the fineness, cross-end collaboration and service ability of Hakka cultural heritage digital expression in Henan Province, and provide technical support for regional cultural resource management.
