基于小波分析的显微视觉景深扩展

摘要

光学显微视觉系统的主要特点在于景深短,难以获取反映显微场景的全面信息.为解决该问题,提出基于小波分析的光学显微视觉系统景深扩展策略,分为局部清晰图像获取和多聚焦图像融合两个阶段:首先,以定义的小波系数活性水平为依据,构造了新型清晰度评价函数和聚焦曲线全局极值搜索策略来实现快速自动聚焦、获取局部显微场景信息的目的;然后,为实现多局部聚焦显微图像的融合,设计了小波系数活性水平选择型融合规则来融合获取的多个局部显微场景信息.实验表明,提出方法可有效扩展光学显微镜的景深.%The main characteristic of optical microscopic visual is a short depth of focus, so it is difficult to obtain overall microscopic scene information.To overcome the problem, a wavelet analysis based strategy was proposed.It consists of two parts: the local information acquisition and the global information fusion.Firstly, based on the designed wavelet activity level measurement, a robust auto-focusing algorithm with new image sharpness function and fast global maximum searching strategy were designed to acquire local microscopic visual information.Secondly, an image wavelet activity level measurement based multi-focus images fusion scheme was presented.At last, the experimental results verify the validity of the proposed strategy.