This paper reports on the characterisation of water sprays for the purpose of modelling the effectiveness of such sprays in mitigating the impacts of wildfires on residential buildings. Four types of water spray nozzles were experimentally tested and analysed using a single high-speed camera and image analysis software. The sprays were generated by a misting nozzle, a hollow-cone nozzle, a rotating 'butterfly' sprinkler and a deflector plate sprinkler. Back-illumination was used to produce silhouette images of droplets within each spray. The point spread function half-width of each droplet image was determined, to give an indication of the distance of the droplet from the focal plane. In this way a control volume was defined for the measurements without introducing a bias towards larger or smaller droplets. The high-speed videos were analysed using a custom-built script in the software Matlab. Overlapping droplet images were identified through the high spatial rate of change of grey-level gradient around their perimeters. These images were separated, reconstructed and corrected for any resulting change in local contrast. The size of each droplet was measured and the velocities of individual droplets were determined by tracking the droplets between video frames. Results are presented on the spatial distribution of droplet sizes and velocities within each spray in terms of probability density functions. Commentary is also provided on how this data will be used in future computational fluid dynamics analyses of sprays implemented on the exterior of residential buildings and the effects of wind thereon.