Estimating sensor performance and target population size with multiple sensors

A Bayesian inference approach to exploit multisensor target detections for the estimation of sensor performance is presented. Additionally, it is shown that a closed form approximate Bayesian estimator for the target population can be derived. This Bayesian method is applied to real data in the case of satellite detection of Automatic Identification System messages resulting in a global map of sensor performance. The resulting sensor detection performance estimates are also suitable for the tuning of tracking and fusion algorithms which make use of sensor models.