In order to support the collaborative brand building of universities and rural red tourism resources, this paper proposes an intelligent visual computing framework guided by cultural logic. A multi-modal data set for cross-scene brand collaborative analysis is constructed by focusing on the images of university exhibition halls, practical activity images, course texts, rural ruins images, village space images, route node texts and spatial attribute information. Then, the narrative semantics, visual symbols and rural context clues in heterogeneous scenes are aligned through the cultural logic constraint module. Finally, the brand consistency and cross-scene correlation strength are determined by combining image, text and spatial attributes. Experimental results show that the proposed method achieves 92.3% accuracy, 90.8% Macro-F1 value and 0.918 cross-scene consistency on the final brand collaborative discrimination task, which is superior to the comparison models in terms of representation stability and collaborative brand building effectiveness. This framework provides a computable path for data-driven red tourism brand construction in the university-rural integrated cultural communication scenario.
