基于Hausdorff距离与模拟退火算法的图像匹配技术

Image matching technique based on Hausdorff distance and simulated annealing algorithm

在图像匹配过程中,经常会有目标图像被遮掩,匹配效率不高等问题,使得应用较为困难。对图像和模板之间的Hausdorff距离进行改进,将其作为检测模板和图像中轮廓相似性的标准,同时引入模拟退火算法,由于该算法具有鲁棒性和易于全局最优的特点,可以较快地完成全局搜索。利用人物图像和涂鸦图像进行仿真实验,实验结果表明,文章算法有着较好的图像匹配结果,且运行效率高于传统Hausdorff距离法。

In the image matching process,there are some problems existing usually,obscuring target image,not high matching efficiency,which make the application more difficult.By improving the Hausdorff distance between the image and the template and making it as a template and the contour detection similarity criteria, while the introduction of simulated annealing algorithm, due to the robustness and ease of global optimum characteristics of the algorithm,the global search can be completed quickly.By using images of people and graffiti image to do simulation experiment, and the results showed that the algorithm has a better image matching results, and ran more efficient than general Hausdorff distance method.