ImplicitTerrain: a Continuous Surface Model for Terrain Data Analysis

TL;DR

ImplicitTerrain introduces an implicit neural representation for continuous, differentiable terrain modeling, enabling accurate surface fitting, topological feature retrieval, and topographical analysis, advancing terrain data analysis methods.

Contribution

The paper presents the first feasible continuous terrain surface modeling pipeline using implicit neural representations, improving accuracy and topological analysis over traditional discrete methods.

Findings

Superior surface fitting accuracy demonstrated

Effective topological feature retrieval achieved

Supports various topographical feature extractions

Abstract

Digital terrain models (DTMs) are pivotal in remote sensing, cartography, and landscape management, requiring accurate surface representation and topological information restoration. While topology analysis traditionally relies on smooth manifolds, the absence of an easy-to-use continuous surface model for a large terrain results in a preference for discrete meshes. Structural representation based on topology provides a succinct surface description, laying the foundation for many terrain analysis applications. However, on discrete meshes, numerical issues emerge, and complex algorithms are designed to handle them. This paper brings the context of terrain data analysis back to the continuous world and introduces ImplicitTerrain (Project homepage available at https://fengyee.github.io/implicit-terrain/), an implicit neural representation (INR) approach for modeling high-resolution terrain continuously and differentiably. Our comprehensive experiments demonstrate superior surface fitting accuracy, effective topological feature retrieval, and various topographical feature extraction that are implemented over this compact representation in parallel. To our knowledge, ImplicitTerrain pioneers a feasible continuous terrain surface modeling pipeline that provides a new research avenue for our community.