Optimal Camera Parameter Selection for State Estimation with Applications in Object Recognition

摘要: In this paper we introduce a formalism for optimal camera parameter selection iterative state estimation. We consider framework based on Shannon's information theory and select the parameters that maximize mutual information, i.e. captured image conveys about true of system. The technique explicitly takes into account priori probability governing computation information. Thus, sequential decision process can be formed by treating posteriori at current time step in as next step. convergence proven. We demonstrate benefits our approach using an active object recognition scenario. results show outperforms random strategy, both sense rate number views necessary to return decision.