Evolutionary algorithms in airborne surveillance systems: image enhancement via optimal sightline control

Abstract

Airborne electro-optic systems require precise control of the camera line-of-sight vector to prevent unwanted and possibly mission-critical degradation of image quality. This is particularly true for long-range surveillance missions or operation in low signal-to-noise ratio environments. Recent work has shown that pixel smear during camera integration (jitter) can be reduced if adaptive friction compensation ‘shapes’ the jitter frequency content in addition to reducing the energy content. This article expands upon that work by automating the tuning process for the sightline control parameters using a genetic algorithm (GA) with a novel fitness function – the average modulation transfer function. The efficacy of this cost function is illustrated first through tuning of the traditional rate loop proportional + integral controller gains and second via tuning of a bespoke adaptive friction compensator. It is shown in both cases that controllers tuned using the GA with this fitness function are significantly better than the current root-mean-square figure of merit typically employed in stabilization loop design.