Purpose Visualization of bridge construction scenes often suffers from low efficiency in presenting multidimensional information, slow update rates, and unsatisfactory cognitive effectiveness. To address these issues, this study adopts bridge construction as a representative case and proposes a digital twin–based scene visualization method that supports multi-scale information. The proposed method aims to enable efficient, clear, and accurate representation of information within digital twin scenes. Design/methodology/approach In this study, a large suspension bridge currently under construction across a valley is used as a case study. Remote sensing data, environmental monitoring data, and bridge design data are collected. A visual encoding method is then developed to provide an integrated representation of multidimensional information within the bridge construction scene. This includes geometric characteristics, physical attributes, behavioral states, and rule-based constraints, which are mapped onto appropriate visual variables. Furthermore, a multiscale hierarchical visualization scheme is established to accommodate requirements across different spatiotemporal scales, ensuring consistent representation from the macroscopic to the microscopic level and throughout all stages of construction. Finally, an eye-tracking experiment is conducted to analyze and evaluate the effectiveness of the proposed visualization method in conveying information. Findings Compared with traditional approaches, the proposed method enables cross-spatiotemporal visual representation of multidimensional information, including geometric characteristics, physical attributes, and behavioral states. In practical applications, it achieves an average frame-rate improvement of at least 51.12% and demonstrates enhanced dynamic updating capability. Based on the results of the eye-tracking experiment, the proposed method yield the highest fixation duration ratio (36.92%) and the shortest time to first fixation (441.48 ms), together with the highest level of cognitive accuracy (53.34%). Originality/value The originality of this work lies in redefining digital twin visualization as an information communication problem rather than a purely visual rendering task. By introducing rule-constrained encoding and hierarchical spatiotemporal organization, the proposed framework establishes a systematic mechanism for delivering critical scene information efficiently and in an interpretable manner. The incorporation of quantitative cognitive validation further bridges the gap between visualization design and human perception, providing both methodological support for communication-oriented digital twin systems and practical insights for the development of information-centric visualization applications in infrastructure construction.
Guo et al. (Fri,) studied this question.