Camouflage challenges in the age of drones and AI

Defence scientist Dr Bin Lee has completed a year-long posting to Quebec to conduct camouflage pattern research with Defence Research and Development Canada (DRDC) counterparts. The work was a contribution to our longstanding bilateral S&T collaboration in Camouflage Pattern Generation and Soldier Signature Modelling and Simulation.

'This collaboration allows us to share and leverage resources in the land electro-optical/infrared signatures and survivability areas with DRDC,' says Dr Lee, who has developed camouflage insights and innovations with Canada and other international partners for 26 years.

S&T activities under this collaboration have included the sharing of field trial techniques, data collection and analytic tools, methodologies and datasets, as well as soldier signature models. Dr Lee's work during his posting contributed to the ability of both nations to determine the effectiveness of current Camouflage, Concealment & Deception (CCD) techniques against current and specialised artificial intelligence (AI) targeting algorithms. Importantly, the research is providing techniques for countering AI-based detection and targeting in operational conditions.

'We've been collaborating through the NATO camouflage community for many years,' says Dr Lee. 'Back in 2012 I hosted the first ever NATO scientific trial in Australia which gathered 20 scientists from ten nations including Canadians and we continued on with the Canadian development bilateral agreement. We had further field trials in 2016 through to 2019, and my posting last year is a continuation of that work.'

New colour set for multi-terrain uniforms

One focus area for Dr Lee and the project team was the assessment of revisions to the Canadian multi-terrain camouflage uniform to make it suitable for different background environments.

Testing has expanded to cover an emerging threat - AI-enabled sensor systems. In the past, camouflage patterns and designs have been based on hiding from human vision. Part of the recent work involved collecting data sets and analysing the performance of an open source AI detection algorithm.

The project team defined and validated rigorous processes for assessing the performance of new camouflage patterns. Analysis of hyperspectral imagery data has enabled the computation and development of new pixelated Multi Terrain Canadian Disruptive Patterns (CADPAT (MT)) which were initially trialled in the Northern Territory. Following down-selection, prototype uniforms were manufactured in limited quantities for further testing.

Field-testing conducted in Canada demonstrated the improved performance of the new CADPAT (MT) pattern over the in-service variant. Further analysis demonstrated that some colour adjustments were needed to balance performance in more arid areas.

Mojave Desert trial

As a final check of the new CADPAT (MT) pattern Dr Lee and the project team took the new camouflage to the Mojave Desert for a field trial. The project team collected tactically relevant data to help quantify the detectability level of this new CADPAT variant.

Soldiers in tactical positions were imaged with ground and airborne images in realistic military scenarios. AI models were used to analyse the image sequence data. The assessment of AI-enabled sensor system technologies in the context of battlefield reconnaissance with conventional or autonomous surveillance systems against camouflaged assets is crucial to the design and development of future CCD techniques.

Results following analysis confirmed that the signature of the camouflage pattern was reduced over multiple deployment conditions and environments (in other words, objects were harder to detect). The CADPAT (MT) is entering service in the Canadian Armed Forces - an exemplar of the value and impact of Australia and Canada's fruitful Defence S&T collaboration delivering improved capability to the warfighter.

Watch a video produced by DRDC about the new CADPAT (MT) here.

DSTG enhances its science and technology impact by collaborating with research and industry partners, nationally and globally.

We access and leverage world-leading science, technology, knowledge and innovation, through collaboration with industry, academia and international agencies. Learn more about our partnerships.

Public link

Please use the above public link if you want to share this noodl on another website.