Inspection and structural health monitoring (SHM) techniques are generally applied to reduce uncertainties associated with predicting damage occurrence, propagation, and overall bridge performance in life-cycle management. These techniques are crucial for ensuring the safety and longevity of bridge structures, particularly in harsh environments. This paper proposes a novel approach for reconstructing realistic three-dimensional (3D) models and detecting concrete surface spalling for cable-stayed bridges in cold regions. Specifically, this innovative method employs a multi-source registration method that utilizes point cloud data obtained from both unmanned aerial vehicles (UAVs) and terrestrial laser scanners (TLSs). First, point cloud data are collected from UAVs and TLSs according to flight path and scan planning. Next, a filter combining Radius Outlier Removal (ROR) and Statistical Outlier Removal (SOR) is applied to preprocess the data for noise reduction. Then, the Iterative Closest Point (ICP) algorithm is used to register the multi-source point cloud data. Finally, the realistic 3D model of the bridge is reconstructed based on the registered point cloud data. This is demonstrated through a field test of an extra-large cable-stayed prestressed concrete bridge with a length of 1020m in Northeast China. The results indicate that, as a non-destructive inspection method, the realistic 3D bridge model reconstructed from point cloud data is both effective and efficient in reproducing the bridge’s details regarding geometric shape, completeness, and texture. This model accurately identifies various bridge defects and deformations, such as concrete surface spalling. These capabilities are essential for precise condition assessment, prediction, and maintenance planning, while this advanced approach significantly enhances damage detection, providing a more reliable basis for structural evaluations. Integrating state-of-the-art inspection and SHM technologies with an intelligent fusion mechanism for multi-source bridge data creates a digitally connected, real-time interactive system for bridge inspection and health monitoring, enabling continuous monitoring and proactive maintenance strategies through updatable digital twins.