Simulation-based analysis of range and cross-range resolution requirements for the identification of vehicles in ladar imagery

We analyze the number and range accuracy of pixels required to distinguish vehicle types in ladar imagery. A simulation is used to generate large numbers of synthetic ladar images of a set of vehicles, and an automated geometry-matching algorithm used to identify vehicles is presented with each of these images. By systematically varying the range resolution, cross-range resolution, and other parameters, we can estimate joint ladar-algorithm system performance as a function of the sensor parameters. The results of these studies will provide guidance to ladar researchers and designers in setting fundamental system parameters for specific applications or operational scenarios. We describe the simulation, identification algorithm, and experimental results.