Environmental monitoring has always been rife with challenges. The most common include site accessibility, appropriate equipment, determining the right sampling design and amount of data required. However, some regions such as the wet-dry tropics of northern Australia, face additional challenges associated with the ‘wildness’ of the region. One of the primary challenges in monitoring freshwater ecosystems in this region is the ever present and very real threat of crocodile attack. Traditional monitoring methods for fish involve capture and count techniques, which require someone to be in or near the water, but this is not feasible in crocodile infested waters. We trialled the use of unbaited underwater video cameras (UVC) in wetlands of Kakadu National Park to determine if this technique can be readily applied in freshwater billabongs where the use of other sampling techniques are limited by the presence of crocodiles and extreme environmental conditions. The trial examined the utility of UVCs in shallow and channel billabongs with dense and sparse vegetation using both stationary and moving UVC deployments. UVCs were a feasible option for monitoring fish and with the appropriate design have the capacity and potential to capture more representative data than previously used traditional methods. We will discuss the design of the trial, the representativeness of the data collected, and comparisons with historical fish data for the sampled billabongs.