Student Research Projects offered by Bernie Jenny, Associate Professor, Faculty of Information Technology, Monash University, Melbourne
Last update: 15 October 2026
Most research topics listed below are suitable for a PhD Thesis, an Honours Thesis, a Masters by Coursework Project, a final year project or other types of student projects. Topics are available to Monash students and visiting students.
Please contact Bernie if you have questions about any topic, or would like to discuss your own project ideas with a focus on cartographic maps and globes, AR/VR, or machine learning/AI for visualising spatial data.
Creating a map consists of a series of steps, such as researching and compiling data, selecting a geographic area and a map projection, integrating and harmonising vector and raster data, styling map layers, annotating and labelling objects, adding interactive elements, and creating a UI and layout. You develop a harness for AI agents that integrate a variety of existing tools for automating the map creation process and assessing the resulting maps.
We are developing an app for raster-to-raster transformation with machine learning (ML) for cartographers. The app can train ML models for creating shaded relief from elevation models, classifying images, beautifying aerial imagery, etc. The app is developed in Swift for macOS and currently supports different types of U-Nets. You add diffusion models for difficult problems, such as creating terrain sketches and rock drawings from elevation models. National Geographic Society is a partner of this project. The goal is to integrate your work into Eduard (https://eduard.earth), our specialised tool for terrain mapping.
Three-dimensional maps rendered from digital elevation models lack detail in steep areas, and often do not have a natural look. Image diffusion AI can add these details, as shown by a recent map of Yosemite National Park. The goal of this project is to implement a rendering pipeline for creating large and highly-detailed maps enhanced by image diffusion, using local AI, preferably for macOS. You could use mflux or another framework. The goal is to integrate your work into Eduard (https://eduard.earth), our specialised tool for terrain mapping.
Guides for rock climbers and alpinists often feature route maps that are created with ink drawing. They show the shape of rock walls and indicate sections traversable by alpinists. The goal of this project is to develop a reproducible workflow for creating such representations. A machine learning model could be trained from manually created drawings that are aligned with computer-generated 3D views.
When a digital map is zoomed out, it needs to reduce the number of details shown to remain readable. The geometric simplification of lines for representing roads, rivers, political boundaries, etc. is a core problem where only limited machine learning methods exist. You design and implement a line simplification method that is trained with detailed lines and lines simplified by humans. You can make this tool available to cartographers, develop an evaluation metric, or evaluate results with a user study.
You explore how virtual three-dimensional maps, globes, and tools for spatial navigation can be embedded in the physical environment. You ideate, develop, implement, and evaluate new immersive visualisation types to better understand spatial data and/or navigation tools. You invent dedicated interactions for manipulating maps, globes, and navigation tools in mixed reality, using mid-air gestures and physical devices, such as digital pencils. Your research targets state-of-the-art mixed-reality headsets and smart glasses. It aims at developing methods that intuitively visualise complex geospatial phenomena or enable quick and unambiguous understanding of critical information. Applications include teaching and education in multi-user setups, the exploration of complex datasets in earth sciences, or navigation and way-finding.
The Logitech Muse digital pencil allows for drawing in three-dimensional space with the Apple Vision Pro headset. You explore how it can be used to annotate globes in a teaching context: a presenter explains a geographic phenomenon on a physical globe augmented with virtual visualisations and uses the digital pencil to write and highlight elements on the globe. The annotations are shared with the audience, who also wears Vision Pro headsets, or follows the presentation on devices with 2D displays. Your contribution is integrated into Globus, our application for Apple platforms, developed in collaboration with the David Rumsey Map Library at Stanford University.
Viewers can be positioned inside a virtual globe and experience a 360° immersive globe. 360° globes are highly engaging, but currently there are no dedicated tools for augmenting them with information visualisations. You explore the design space for visualising geographic data (e.g., population of the biggest cities, trade flows between countries, etc.) with different techniques. You select and implement visualisation techniques, and evaluate them in a user study. Your contribution is integrated into Globus, our application for Apple platforms, developed in collaboration with the David Rumsey Map Library at Stanford University.
Virtual three-dimensional maps can be placed on physical tables or walls, and globes can be embedded in portals placed on a physical wall. A portal is a window into a virtual world that can be created for the Apple Vision Pro: https://youtu.be/yhgQbYFY48c?si=wohFZ9cKiaLvulGF&t=657. You create interactive 3D terrain maps for the Vision Pro that can be placed on physical surfaces, or you embed virtual globes into portals placed on physical walls. Alternatively, you add terrain elevation to printed maps to make them pop out into the third dimension, or you overlay maps with geospatial visualisation, such as value-proportional bars, circles, etc. Your contribution is integrated into Globus, our application for Apple platforms, developed in collaboration with the David Rumsey Map Library at Stanford University.
Virtual globes are an engaging medium for understanding geographic space, but they can only show half of Earth’s surface. You explore combinations of a globe and diverse types of virtual maps that show the currently hidden face of Earth. The globe and the map are coordinated, such that reorienting the globe changes the area shown by the map, and vice versa. The map is projected on a curved surface placed behind the globe, or it forms a halo around the globe. Other geometries are to be explored, for example, a continuous blend of the globe and the map, or a virtual globe sticking out of a map placed on a physical wall.
Ocean current models are vector fields with a third dimension. You create particle system visualisation to show near-real-time ocean current patterns on virtual globes. You use the Apple Metal API for the Apple Vision Pro. Alternatively, you visualise wind, temperature and pressure and other atmospheric attributes. Your contribution is integrated into Globus, our application for Apple platforms, developed in collaboration with the David Rumsey Map Library at Stanford University.
Explore the design space of pop-out paper maps and data visualisations, and identify techniques for showing data on interactive paper maps and charts. This results in a catalogue of ideas; you create three-dimensional pop-up maps and charts with coloured paper, cardboard or plywood using an in-house laser cutter; you evaluate design ideas in a user study.
](Popup.jpg)
In September 2026, the United Nations General Assembly approved a resolution to replace the distorting Mercator map projection with equal-area projections, such as the Equal Earth projection (which was co-invented by Bernie Jenny). The next step is now to create a collection of maps using the Equal Earth projection that educators worldwide can download and print for free. This Equal Earth Atlas should be available in multiple languages, and cover a wide variety of global topics, such as weather and climate, human population, economy, technology, conflicts, etc. You can contribute to a variety of areas: develop a thematic structure of the atlas, create maps, develop software, plan outreach campaigns, etc.
When zooming out of Google Maps and other digital map platforms, the world is shown immense areal distortion. The goal of this project is to develop an adaptive projection algorithm that adjusts the geometry of the map in a smooth transitions between the Mercator projection for large map scales, and an area-preserving projection for world maps. This research will serve as a demonstrations for authors of digital map platforms how such a system can be implemented, and is highly relevant and timely in consideration of the resolution United Nations resolution encouraging map authors to replace the distorting Mercator map projection with equal-area projections.
Graduated dot maps are an advanced cartographic visualization technique that uses variably sized dots to represent different quantitative values across a geographic area. There is currently no out-of-the-box software that generates graduate dot maps. You explore algorithms for creating graduated dot maps, and implement an editor for creating graduated dot maps.
Merging and branching flow maps are difficult to create. You create a tool for cartographers that enables cartographers to quickly create flow maps with merging and branching flows.
Relative elevation models visualise a river valley's topography by removing the overall downstream slope of the river, so the subtle landforms around the channel become visible. Current methods for creating relative elevation models are limited. The goal of this project is to develop a method combining machine learning for identifying river plains and an algorithm for removing the downstream slope using the multigrid method or the successive over-relaxation method. Your contribution will be integrated into Eduard (https://eduard.earth), a specialised application for cartographers for creating terrain images for maps.
Plan oblique relief is a cartographic technique that displays 3D terrain and mountains in a partial side-view profile on top of a flat, undistorted two-dimensional map. Combining map data (such as roads, rivers, etc.) with plan oblique relief requires specialised algorithms for transforming vector data. You review and implement algorithms for plan oblique transformation and clipping of lines and polygons. Your contribution will be integrated into Eduard (https://eduard.earth), a specialised application for cartographers for creating terrain images for maps.
Terrain bending adds depth and a horizon to three-dimensional terrain maps. Terrain Bender is an early application for bending digital elevation models: https://terraincartography.com/terrainbender/. The goal of this project is to develop algorithms for real-time rendering with the Apple Metal API and make a tool available to professional cartographers.
Creating maps with ray-traced terrain shading requires specialised expertise in Blender or other software. The goal of this project is to develop hardware-accelerated shading using the Apple Metal API. Your contribution will be integrated into Eduard (https://eduard.earth), a specialised application for cartographers for creating terrain images for maps.
