Recovering Scale in Relative Pose and Target Model Estimation Using Monocular Vision