中文摘要：

研究文献时发现虽然采用固定步长调焦具有调焦过程稳定的特点，但由于调焦时间较长、效率较低等问题，很难应用于实际要求实现快速调焦的系统中。现有的变步长调焦点搜索算法存在计算公式推导过程复杂、算法计算量较大等问题，在实际要求实现快速调焦的系统中，也很少使用。为了解决此问题，本文提出一种新的基于图像处理自适应变步长调焦点搜索算法。该算法的具体调焦策略：当远离最佳调焦位置时，任意清晰度评价函数上两点组成的直线与水平直线构成的夹角较小，增加调焦步长，实现粗调焦；当处于最佳调焦位置附近时，减小调焦步长，实现细调焦。最后分别使用变步长、固定步长和本文提出的基于图像处理的变步长搜索算法对不同的清晰度评价函数进行自动调焦，实验结果表明：本文提出的变步长搜索算法在保证调焦精度的前提下，可以对基于灰度时域的清晰度评价函数实现快速调焦。通过对比调焦时间发现，本文提出的基于图像处理的自适应变步长搜索算法具有调焦时间较短、计算量较小等优点。

英文摘要：

Abstract： References found that the fixed step size is used to stabilize the focusing process. However, it is difficult to be used for the practical requirements of the system due to the longer focusing time and low efficiency. In addition, amongthe existing variable step size focus search algorithms, the existing variable step size adjustment focus search algorithm has quite complex calculation formula derivation process and the computation of the algorithm is very difficult. Therefore, in the actual applications which require rapid focus, the existing algorithm is rarely used. In order to realize the fast focusing, the process of the derivation of the focusing algorithm needs to be simpleand the algorithm calculation should be simple so that the algorithm can be applied to the system of fast focusing. In this paper, the variable step size focus search algorithm is studied, and a new adaptive variable step size algorithm based on image processing is proposed. The strategy of the new algorithm is： the coarse focusing is obtained by increasing the focus step when the the system consists of a straight line of the clarity evaluation function of any two points far away from the best focusing position; a fine focusing is realized by decreasing the focus step by nearby the peak of focusing curve; finally, the variable step size, fixed step size and variable step size search algorithm based on image processing are used to automatically focus the different definition evaluation functions. Experimental results show that, in this paper, the proposed variable step size search algorithm can ensure the accuracy of the focus, can achieve rapid focus for domain definition of the evaluation function, and the adaptive variable step size search algorithm based on image processin~ has the advantages of shorter focusing time, less computation and so on.