The global distribution of karst landscapes is extensive, and in regions with suitable climates, a large number of natural caves have developed. However, these caves are often concealed by vegetation, making them difficult to detect. To address the challenges of locating and detecting caves in karst regions under vegetative cover, this study integrates Unmanned Aerial Vehicle (UAV) remote sensing and Light Detection and Ranging (LiDAR) technology to explore a method for identifying hidden caves in vegetated areas using both aerial and ground-based approaches. The findings demonstrate that the penetration capabilities of LiDAR dense point clouds in vegetated areas can effectively identify depression structures in the terrain, which are indicative of concealed caves that are not easily detectable by the naked eye. Additionally, as natural underground spaces, caves can be readily converted for purposes such as civil defense or storage, significantly reducing labor and material costs. To this end, a centimeter-level precision three-dimensional (3D) model of the caves was developed, and the concept of “Cubic Space Volume” was proposed to simulate the effectively usable space within the caves, providing valuable practical reference. This study not only offers a new approach for cave detection and exploration but also enriches the application of remote sensing technology in cave exploration research.