University campuses provide vital ecosystem services, including carbon storage, urban cooling, and biodiversity conservation. Effective green infrastructure management requires comprehensive tree inventories that integrate species identification with 3D structural attributes however, conventional manual field surveys are labor-intensive. While UAV-LiDAR efficiently captures tree height and crown geometry over wide areas, ground-level imagery offers essential visual characteristics for species identification. Standard spatial datasets like Google Street View often lack coverage in pedestrian zones and off-road green spaces. Linking independently acquired ground imagery with aerial measurements offers a practical solution to bridge this gap. This study presents a multimodal framework integrating 360-degree ground-level imagery, Structure-from-Motion (SfM), UAV-LiDAR, and 3D Gaussian Splatting (3DGS) at the Asian Institute of Technology campus in Thailand. Ground-level 360-degree imagery along walking/cycling routes and UAV-LiDAR data were collected concurrently in December 2025. Image-detected tree stems are localized using SfM camera poses and dense point clouds. Meanwhile, semantically classified LiDAR point clouds are converted into a Canopy Height Model (CHM) for treetop detection and marker-controlled watershed segmentation to delineate individual crowns. The framework aims to link ground-observed stems with aerial crown candidates within an interactive, georeferenced 3DGS and CesiumJS web platform. Preliminary segmentation results are promising for isolated trees, but overlapping crowns remain challenging. In study area, ground detections were manually linked to 41 of 50 manually labeled trees. CHM segmentation (h = 0.6 m; Gaussian smoothing = 0.5 pixels) matched 41 trees using one-to-one matching with IoU ≥ 0.5 against crown-top reference regions derived from manually labeled point clouds. Seven lacked ground matches, giving 48 trees represented by at least one source. This measures sample coverage, not crown delineation accuracy, and suggests complementary aerial and ground observations. Ongoing work focuses on automated stem-to-crown matching to support species-specific biomass estimation and campus green infrastructure management.