Computing robust viewpoints with multi-constraints using tree annealing

In order to compute camera viewpoints during sensor planning, Tarabanis et al. (1991) present a group of feature detectability constraints which include six nonlinear inequalities in an eight-dimensional real space. It is difficult to compute robust viewpoints which satisfy all feature detectability constraints. In this paper, the viewpoint setting is formulated as an unconstrained optimization problem. Then a tree annealing algorithm, which is a general-purpose technique for finding minima of functions of continuously-valued variables, is applied to solve this nonlinear multiconstraint optimization problem. Our results show that the technique is quite effective to get robust viewpoints even in the presence of considerable amounts of noise.