Testing of Pattern Recognition Processes in Visual and Thermal PIV

The ability to estimate water velocity and bathymetry in the riverine environment from imagery data is an important development with applications in scientific, industrial, and military sectors. Forward deployed units operating in unfamiliar environments would be able to better navigate rivers for transit and crossing if conditions such as water velocity and depth could be easily acquired through imagery from a UAS or satellite. Some other potential applications of this capability are safe navigation, river discharge estimation, and dredging. Toolboxes currently under development by the Coastal Imaging Research Networkprovide the ability to determine the surface velocity of rivers by tracking thermal or visual features on the river’s surface. The tracking of thermal features is a more robust method than visual, since these are always present in some capacity on the water surface.

The goals of this study are to gather infrared imagery data on a river system in Annapolis, MD from onboard a UAV, in order to test methods currently under development for UAV image rectification and determination of water surface velocity in different environmental conditions. This infrared data will be utilized to test the accuracy of methods developed to estimate river bathymetry. We will develop a solution to correct the deficiencies of determining surface velocities during hours of low thermal contrast through a pattern recognition imagery rectification method.