Abstract:Terrains are visually important and commonly used in computer graphics. While many algorithms for their generation exist, it is difficult to assess the realism of a generated terrain. This paper presents a first step in the direction of perceptual evaluation of terrain models. We gathered and categorized several classes of real terrains and we generated synthetic terrains by using methods from computer graphics. We then conducted two large studies ranking the terrains perceptually and showing that the synthetic terrains are perceived as lacking realism as compared to the real ones. Then we provide insight into the features that affect the perceived realism by a quantitative evaluation based on localized geomorphology-based landform features (geomorphons) that categorize terrain structures such as valleys, ridges, hollows, etc. We show that the presence or absence of certain features have a significant perceptual effect. We then introduce Perceived Terrain Realism Metrics (PTRM); a perceptual metrics that estimates perceived realism of a terrain represented as a digital elevation map by relating distribution of terrain features with their perceived realism. We validated PTRM on real and synthetic data and compared it to the perceptual studies. To confirm the importance of the presence of these features, we used a generative deep neural network to transfer them between real terrains and synthetic ones and we performed another perceptual experiment that further confirmed their importance for perceived realism.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to evaluate and improve the perceived realism of terrain generation in computer graphics. Specifically, the author focuses on the following two core issues: 1. **What visual features make the terrain look more realistic?** 2. **What level of perceived realism has been achieved in synthetic terrains generated by existing computer graphics techniques?** To solve these problems, the author has carried out research in the following aspects: ### 1. **Perceived Realism Evaluation** The author introduced a new metric - **Perceived Terrain Realism Metrics (PTRM)** - to quantify the realism of the terrain. PTRM is based on geomorphons extracted from Digital Elevation Models (DEM), and these features include valleys, ridges, depressions, etc. ### 2. **User Studies** The author conducted two large - scale user studies: - **Experiment 1**: Compare the perceived realism of real and synthetic terrains. Pairs of terrain images were presented through the Mechanical Turk platform, and participants were asked to choose which one looked more realistic. - **Experiment 2**: Evaluate the perceived realism of terrains after feature transfer. CycleGAN was used to transfer the features of high - ranked terrains to low - ranked terrains, and user tests were conducted again. ### 3. **Feature Transfer** To verify the influence of specific geological features on perceived realism, the author used the CycleGAN deep neural network to transfer features (such as valleys, ridges, etc.) from high - ranked terrains to low - ranked terrains and vice versa. The results show that this feature transfer significantly affects the perceived realism of the terrain. ### 4. **Dataset and Method** The author collected real - terrain data from the Shuttle Radar Topography Mission and used multiple computer graphics algorithms to generate a synthetic - terrain dataset. All terrains were rendered and the corresponding geomorphons were generated. ### Main Contributions - **Introduction of PTRM**: A new metric for quantifying the perceived realism of terrains. - **User Studies**: Verified and measured the perceived realism of real and synthetic terrains through large - scale user studies. - **Identification of Geological Features Affecting Perceived Realism**: Discovered that certain geological features (such as valleys, ridges, etc.) have a significant impact on perceived realism. - **Public Dataset**: Provided a dataset containing real and procedural terrains, with perceptual evaluations and calculated geomorphons. Through these studies, the author hopes to provide a solid foundation for the evaluation of terrain - generation realism in computer graphics and to provide guidance for further improving the realism of synthetic terrains.

PTRM: Perceived Terrain Realism Metrics

A Review of Real-Time Terrain Rendering Techniques

An Automatic Digital Terrain Generation Technique for Terrestrial Sensing and Virtual Reality Applications

Digitally derived trail terrain metrics

A terrain and cloud computer image generation model

GeoMeter: Probing Depth and Height Perception of Large Visual-Language Models

Comparing metrics for evaluating 3D map quality in natural environments

Perceptual Point Cloud Qality Assessment for Immersive Metaverse Experience

Visual Perception of Real World Depth Map Resolution for Mixed Reality Rendering

PMT: New analytical framework for automated evaluation of geo-environmental modelling approaches

PatchRefiner: Leveraging Synthetic Data for Real-Domain High-Resolution Monocular Metric Depth Estimation

Realistic Terrain Visualization Based on 3D Virtual World Technology

Perception-Guided Quality Metric of 3D Point Clouds Using Hybrid Strategy

Effect of interpolation methods on quantifying terrain surface roughness under different data densities

ImplicitTerrain: a Continuous Surface Model for Terrain Data Analysis

Empir3D : A Framework for Multi-Dimensional Point Cloud Assessment

Algorithm Certainty Analysis of Spatial Data for Terrain Model

Deep Generative Framework for Interactive 3D Terrain Authoring and Manipulation

A Complex Terrain Simulation Approach Using Ensemble Learning of Random Forest Regression

GeodesicPSIM: Predicting the Quality of Static Mesh with Texture Map via Geodesic Patch Similarity