Unsupervised ground filtering of airborne-based 3D meshes using a robust cloth simulation

• DOI: 10.1016/j.jag.2022.102830
• Publication Date: 2022-06-09T22:28:36Z
• Authors: Dayu Yu; Lianlian He; Fan Ye; Liangcun Jiang; Chenxiao Zhang; Zhe Fang; Zheheng Liang
• Keywords: Bounding volume; Bounding overwatch

Airborne platforms have been improved in the past decade to provide geographic information systems (GISs) with large-scale 3D geographical information. Objectification of such organized meshes is a significant challenge for GISs. The ground filtering key step meeting this challenge, however, its accuracy highly affected by negative blunders and unbalanced vertex density. This paper proposes novel method differentiating geometric primitives from realistic based on cloth simulation filter. Within method, fall piece simulated flipped mesh, stationary shape considered be fitted ground. Utilizing spatial continuity meshes, collision detection bounding volume hierarchy introduced, making results independent Further, correction scan line ray casting proposed make it applicable data blunders. assessed quantitatively visually over several datasets different densities, scenes, noise distributions. Results demonstrate that robust suitable landscapes not impacted density noise.