CAJ | 학술논문

针对实时性和聚焦精度两方面的要求,在分析对比了常用聚焦评价函数的基础上,提出了一种基于灰度差分的快速聚焦评价函数--利用相邻像素在水平和垂直方向的差分乘积作为核函数.该算法保持了较好的计算性能,但函数陡峭性提高,可加快搜索定焦的收敛速度并能提高聚焦精度,同时引入阈值滤除噪声,增强了算法的稳定性.与常用聚焦函数进行了实验对比,结果表明该算法既有较好的计算速度又有较高的灵敏度,适用于对实时性要求较高的场合,如显微镜自动聚焦系统等.
침대실시성화취초정도량방면적요구,재분석대비료상용취초평개함수적기출상,제출료일충기우회도차분적쾌속취초평개함수--이용상린상소재수평화수직방향적차분승적작위핵함수.해산법보지료교호적계산성능,단함수두초성제고,가가쾌수색정초적수렴속도병능제고취초정도,동시인입역치려제조성,증강료산법적은정성.여상용취초함수진행료실험대비,결과표명해산법기유교호적계산속도우유교고적령민도,괄용우대실시성요구교고적장합,여현미경자동취초계통등.
According to requirements of focusing accuracy and speed, by the analysis and comparison of the common auto-focusing evaluation function, this paper proposed a new evaluation function based on gray scale difference: using product of adjacent pixels difference in the horizontal and vertical direction as core function. The algorithm maintained a good computing performance, and its curve was more steep so that could speed focusing convergence rate and improve focusing accuracy. By using threshold value to filter noise, night enhance stability of the algorithm. The experiments show that the algorithm not only has good speed, but also has high sensitivity. It's suitable for system with high real-time demanding, such as microscope auto-focusing system.