摘要
对分辨率较高的图像来说,调焦评价函数的计算耗时较多,会影响调焦的实时性。选用整幅图像中细节最丰富的子图像作为调焦对象,在保证调焦精度的同时,可改善调焦效率。为快速找到此区域,提出了基于自适应遗传算法(AGA)的寻优算法。在此算法中,用离焦子图像的加权熵作为适应度函数,自适应的改变遗传算法的交叉概率和变异概率,以避免结果陷入局部极值。实验结果表明,此算法具有运算复杂度低,稳定性好等优点,所选定的子图像具有良好的调焦特性曲线。
Since it will be taken so much time to compute the focusing evaluation function for high resolution images, the real-time character of focusing system is sharply influenced. This can be improved by selecting the sub-image with the most detail to focus. Adaptive genetic algorithm （AGA） which can search this region rapidly is presented. It uses weighing entropy as the fitness function and changes the probabilities of crossover and mutation self-adapting to search the global convergence. The experiment proves that this algorithm is efficient and stable. At the same time, the focus characteristic curve of the selected region is fine.
出处
《光学技术》
EI
CAS
CSCD
北大核心
2006年第6期851-854,共4页
Optical Technique
基金
黑龙江省科技攻关基金资助项目(GB02A402)
关键词
自动调焦
遗传算法
熵
auto-focus
genetic algorithm
entropy
