figure6

Using Remote Sensing to detect bunkers in case of Dezi Freeman

Update: Desi Freeman Located Following Extensive Search

Update (April 2026):
Authorities have confirmed that fugitive Dezi Freeman has been located following one of the largest manhunts in Victoria’s history. Freeman was found on a remote rural property near Thologolong, close to the Murray River in north-east Victoria, after more than 200 days on the run.

Police operations culminated in a confrontation at a makeshift hideout consisting of shipping containers and scattered debris, where Freeman had been living off-grid. During the encounter, he emerged from the structure armed and was fatally shot by police.

The location highlights just how difficult detection can be in rugged, remote terrain. The property was isolated, surrounded by dense bushland and mountainous country, with limited visibility from roads or nearby access points.

This outcome reinforces several key points discussed in this article:

  • Remote and off-grid structures can remain undetected for extended periods, even during intensive search operations.
  • Terrain, vegetation, and improvised shelters significantly reduce visibility from conventional surveillance methods.
  • Advanced techniques such as remote sensing, aerial imaging, and thermal detection remain critical tools—but are not foolproof in complex environments.

Despite the scale of the search effort—including thousands of public tip-offs—Freeman was able to evade authorities for months, raising ongoing questions about how he avoided detection and whether assistance played a role.

The case ultimately underscores both the potential and the limitations of modern detection technologies when applied to real-world scenarios involving remote bunkers or concealed living spaces.

Remote sensing and GIS (Geographical Information Systems) can aid in the prevention of criminal activity in the field of forensic geoscience by detecting secret underground infrastructure, such as bunkers. The benefits of using remote sensing and GIS over more traditional field-based methods include the ability to investigate dangerous areas from a distance and the capacity to investigate much wider study areas. The manhunt for Dezi Freeman is such a case which could benefit from desk-based investigations. On 26th August 2025, Dezi Freeman fled after killing two police officers at an estate on Rayner Track, Porepunka, Victoria, Australia (Turnbull and Mitchell, 2025). This estate is situated near Mount Buffalo National Park (Turnbull and Mitchell, 2025). After months with no sign of Freeman, some people believe that Freeman is living in a bunker or abandoned mine within the National Park. 

Remote sensing can use proxy methods to identify possible locations of bunkers. Vegetation anomalies (Pringle et al., 2024) are one of the key proxies that can be used to identify bunkers. This is because changes to soil moisture content and soil composition lead to deterioration in vegetation health (Morley et al., 2024). This can lead to changes to vegetation indices (namely NDVI, SR, and EVI), which may become lower than those of the surrounding area. For this investigation, we chose to focus on NDVI (Normalised Difference Vegetation Index), a metric which determines the health of vegetation using satellite imagery by dividing the difference between the near-infrared (NIR) and red reflectance by their sum (NASA Earthdata, 2024). NDVI values range from -1 to 1, with higher values indicating greener vegetation and lower values associated with barren land or snow (NASA Earthdata, 2024). Landsat imagery was obtained to use for calculating NDVI. The average NDVI was calculated for the spring months (September-November) in 2023, 2024, and 2025, using NDVI values derived from a single image for each month. Imagery was selected based on the amount of cloud cover, which obscured the imagery, with the clearest being chosen. 

The locations of historical mining infrastructure were obtained for Mount Buffalo National Park, as shown in Figure 1. These areas are mostly densely clustered around the Northern boundary of the National Park; however, it is interesting to note that there are some sparsely distributed abandoned mines West of the estate from which Freeman fled. Interestingly, no significant difference in NDVI was observed in the areas where historical mineshafts are located (see Figures 2, 3, and 4). However, areas of difference were identified between 2025 and the previous two years (see Figures 5 and 6). It must be noted, however, that the months used for 2025 were August-October rather than September-November, as November had not been entered yet. This may have caused a reduction in the average NDVI, and it would be interesting to recalculate this average once we have reached November. Furthermore, the results could be further improved by calculating NDVI for more than three years to get a better understanding of NDVI changes in the area over the last decade or so. Additionally, the NDVI averages could be improved further by using more than three NDVI values for each year and by selecting multiple satellite images for each month. Alternatively, EVI (Enhanced Vegetation Index) could be investigated; this is a similar metric to NDVI, but it has a greater sensitivity in areas of dense vegetation (USGS, 2024). A further area of interest which could be investigated is the evidence of spoil heaps or new tracks (Morley et al., 2024). Additionally, changes to soil moisture content and drainage rate (Melillos et al., 2018) could be further examined by looking for any signs of excessive drainage (Morley et al., 2024). 

Figures –

Figure 1- A map showing the location of the estate where Dezi Freeman fled in relation to the locations of historical mining activity within the Mount Buffalo National Park 
Figure 2- A map showing the average NDVI for Spring 2023 calculated by finding the mean NDVI value for 4th September 2023, 31st October 2023 and 16th November 2023. 
figure2
figure3
Figure 3- A map showing the average NDVI for Spring 2024 calculated by finding the mean NDVI value for 15th September 2024, 1st October 2024 and 2nd November 2024. 
Figure 4- A map showing the average NDVI for Spring 2025 calculated by finding the mean NDVI value for 24th August 2025, 9th September 2025 and 4th October 2025.
figure4
figure5
Figure 5- Map showing the difference between 2025 and 2023 average NDVI values, with the lower values highlighting areas where the NDVI was lower for 2025. 
figure6
Figure 6- Map showing the difference between 2025 and 2024 average NDVI values, with the lower values highlighting areas where the NDVI was lower for 2025. 

Bibliography- 

Melillos, G., Kyriacos Themistocleous and Hadjimitsis, D.G. (2020). Detecting underground structures in vegetation indices: MSR, RDVI, OSAVI, IRG, time series using histograms. pp.5–5. doi:https://doi.org/10.1117/12.2569930.

Melillos, G., Kyriacos Themistocleous, Papadavid, G. and Hadjimitsis, D.G. (2018). Detection of Military Underground Structures through the Remote Sensing Investigation of Phenological Cycle of Crops. 

Advances in Remote Sensing, 07(03), pp.235–244. doi:https://doi.org/10.4236/ars.2018.73016.

Morley, A.M., Mather, T.A., Pyle, D.M. and Kendall, J-Michael. (2024). Detecting shallow subsurface anomalies with airborne and spaceborne remote sensing: A review. Science of Remote Sensing, [online] 11, p.100187. doi:https://doi.org/10.1016/j.srs.2024.100187.

NASA Earthdata. (2024). Normalized Difference Vegetation Index (NDVI) | NASA Earthdata. [online] Available at: https://www.earthdata.nasa.gov/topics/land-surface/normalized-difference-vegetation-index-ndvi.

Pringle, J.K., Ruffell, A., Styles, P., Stringfellow, M., Stimpson, I.G., Banham, S.G., Wisniewski, K.D., Owen, S., Hobson, L. and Thompson, J. (2024). Forensic geoscience non-invasive detection and characterisation of underground clandestine complexes, bunkers, tunnels and firing ranges. Forensic Science International, 359, pp.112033–112033. doi:https://doi.org/10.1016/j.forsciint.2024.112033.

Shepherd, T. (2025). Dezi Freeman may be hiding in any one of hundreds of mineshafts near Porepunkah, local guide says. [online] the Guardian. Available at: https://www.theguardian.com/australia-news/2025/sep/03/dezi-freeman-may-be-hiding-in-any-one-of-hundreds-of-mine-shafts-near-porepunkah-expert-says-ntwnfb.

Turnbull, T. and Mitchell, O. (2025). Dezi Freeman: Porepunkah shooting ignites conspiracy theorists fears. BBC NEWS. [online] 29 Aug. Available at: https://www.bbc.com/news/articles/ckgyk7ry8rdo.

USGS (2024). Landsat Enhanced Vegetation Index | U.S. Geological Survey. [online] www.usgs.gov. Available at: https://www.usgs.gov/landsat-missions/landsat-enhanced-vegetation-index.