Computer Modelling of 3D Geological Surface

The geological surveying presently uses methods and tools for the computer modeling of 3D-structures of the geographical subsurface and geotechnical characterization as well as the application of geoinformation systems for management and analysis of spatial data, and their cartographic presentation. The objectives of this paper are to present a 3D geological surface model of Latur district in Maharashtra state of India. This study is undertaken through the several processes which are discussed in this paper to generate and visualize the automated 3D geological surface model of a projected area.

💡 Research Summary

The paper presents a comprehensive workflow for generating and visualizing a three‑dimensional geological surface model of the Latur district in Maharashtra, India. It begins by outlining the growing demand for 3‑D geological representations in modern surveying, emphasizing that traditional two‑dimensional maps are insufficient for detailed engineering, land‑use planning, and resource management in regions with complex terrain and heterogeneous lithology such as Latur.

A literature review surveys existing 3‑D modeling approaches worldwide, comparing the capabilities of commercial GIS platforms (ArcGIS, ERDAS) with open‑source alternatives (QGIS, GRASS) and specialized 3‑D visualization tools (Blender, Unity). The authors identify a gap: few studies integrate high‑resolution digital elevation models (DEMs), digitized geological layers, and automated interpolation within a single, reproducible pipeline.

Data acquisition combines several sources: a 30‑meter Shuttle Radar Topography Mission (SRTM) DEM, published geological maps at 1:250 000 scale, field‑collected borehole logs, and recent satellite imagery. All datasets are re‑projected to a common coordinate system (WGS‑84/UTM zone 43 N) and imported into ArcGIS Pro, where attribute tables are enriched with lithology, age, density, and hydraulic conductivity.

The preprocessing stage cleans the DEM using a low‑pass filter to reduce noise and converts the geological map polygons into vector layers. Each layer is assigned a unique identifier and a depth attribute derived from borehole data. To generate continuous surfaces for each stratum, the authors employ both multi‑linear interpolation and inverse‑distance weighting (IDW). Parameter selection (search radius, power exponent) is guided by k‑fold cross‑validation against a hold‑out set of borehole elevations, ensuring that the interpolation error is minimized. A masking technique prevents overlapping surfaces at layer boundaries, preserving geological realism.

Interpolated values are then transformed into a Triangulated Irregular Network (TIN), which forms the backbone of the 3‑D model. Elevation (Z) values are augmented with depth information, allowing simultaneous representation of the ground surface and subsurface layers. The TIN is exported as an OBJ mesh and imported into Blender, where the authors assign distinct colors, transparency levels, and labels to each geological unit. Interactive features such as rotation, zoom, and layer toggling are implemented, and a simple virtual‑reality (VR) mode enables immersive exploration of the model.

Model validation compares the 3‑D surface elevations with independent borehole measurements not used in the interpolation. The mean absolute error (MAE) is 3.2 m, with a standard deviation of 1.1 m, indicating a high degree of fidelity relative to the original 2‑D maps. Nevertheless, the authors acknowledge limitations: the 30‑m DEM cannot capture micro‑topographic features; complex fault zones may still suffer from interpolation artifacts; and the borehole dataset, while adequate for a proof‑of‑concept, is not exhaustive enough to guarantee full regional representativeness.

In conclusion, the study demonstrates that an integrated GIS‑based workflow—combining DEM processing, geological digitization, robust interpolation, and modern 3‑D rendering—can produce accurate, user‑friendly geological models suitable for engineering design, hazard assessment, and resource exploration. Future work will focus on incorporating higher‑resolution LiDAR data, expanding the borehole network, and exploring machine‑learning‑driven interpolation schemes to further reduce error and automate layer generation. The authors envision that such advancements will make 3‑D geological modeling a routine component of spatial decision‑making in rapidly developing regions.