Optimization of Edge Detectors for Topographic Maps of Cave Inscriptions

We propose an optimization-based framework for designing linear edge detectors for topographic maps of cave inscriptions. This optimization framework requires an edge model, which is constructed from sample images. The optimization criteria are described and genetic algorithms are employed in an iterative fashion for finding optimal edge detectors. Thus, the optimization framework consists roughly of three steps: modeling a representative edge, forming the objective function for optimization, and choosing the optimization method. An additional preprocessing step is also described for removing special types of noise before applying the edge detectors. Simulation results are presented to illustrate the performance of our method on sample topographic maps.