| The inertia terrain-aided navigation system occupies an fundamental important place in the modern military technology and the image matching is the key technology for the navigation system based on scene matching. Although the matching accuracy is very high for the image matching system owing to its matching methods based on pixel traversal search, nevertheless, the computation is too large to meet the requirements of terrain matching in the aspect of high real-time. In order to improve the searching efficiency continuously, this paper studied the optimization of the terrain matching process.First of all, this paper studied in depth the processes and characteristic of terrain matching, and the component which affects the image matching algorithm most was analyzed. On the basis of the above research, a cross correlation matching algorithm using gray value was selected. Furthermore, by simplifying algorithm function expansion, the computational complexity of each match was greatly reduced.Meanwhile, the principles and characteristics of genetic algorithms and swarm algorithms were elaborated in detail. Considering the fact that genetic algorithm is easy to get a local optimum consequence, a Genetic Algorithm Based on Bee Colony Algorithm (BCGA) was proposed. Moreover, based on the parallel genetic algorithm, the multi-swarm evolution model was established. And various values of self-adaptive crossover probability were set according to different evolutionary stage. Finally, these algorithm were applied in the simulation of image matching, and the results indicted that the Genetic Algorithm Based on Bee Colony Algorithm presented in this article are applicable in the terrain matching, by mean of which, the matching time for real-time graph and base graph was abbreviated and consequently improve the matching efficiency. Worthy to note that, Genetic Algorithm Based on Bee Colony Algorithm proposed in this paper might be significant for other type of image matching. |
PDF Full Text Request