A resection method based on enhanced continuous taboo search

Resection is one of important issues in machine vision. Although L2 norm based le ast square method is reasonably fast, the globally optimal solution cannot be obt ained theoretically due to its non-convexity of the objective function. Optimiza tion using the L∞ norm has been becoming an effective way to solve parameter es timation problems in multiview geometry. But the computational cost increases rap idly with the size of measurement data. In the paper, we propose a novel approach under the framework of enhanced continuous taboo search (ECTS) for resection in multiview geometry. ECTS is an optimization method in the domain of artificial in telligence, which has an interesting ability of covering a wide solution space by promoting the search far away from current solution and consecutively decreasin g the possibility of trapping in the local minima. We propose the corresponding w ays in the key steps of ECTS, diversification and intensification. We also present theoretical proof to guarantee the global convergence of search with probabilit y one. Experimental results validate that the ECTS can obtain the global optimum effectively and efficiently. Potentially, the novel ECTS framework can be employed in many applications of multi-view geometry.