改进的Brenner图像清晰度评价算法

图像清晰度评价是基于数字图像的被动式自动调焦技术的基本问题之一.传统Brenner图像清晰度评价算法具有运算速度快特点,但是其评价准确性取决于阈值选取,且其灵敏度较低.针对上述问题,本文提出了一种改进算法.改进算法采用高通和带通两个滤波器对图像进行计算,克服阈值对传统Brenner算法评价结果的影响.为了衡量改进算法的性能,将其与传统的Brenner算法比较,并对评价算法的单峰性、无偏性、灵敏度、计算量等主要衡量标准逐一分析.实验结果表明:与传统的Brenner评价算法相比,改进算法在满足评价算法单峰性和无偏性前提下,提高了灵敏度,降低了计算次数.

An Improved Brenner Algorithm for Image Definition Criterion

The image definition criterion is one of basic problems in passive auto focusing technique based on digital image.The traditional Brenner Algorithm has a fast calculation speed,but the accuracy of evaluation result depends on the threshold value,and the Algorithm sensitivity is low. For this issue, an improved Algorithm is proposed. The improved function uses high pass filter and bandpass filter to evaluate the image, and overcomes the limitation of traditional Brenner Algorithm depending on the threshold value.In order to evaluate the improved Algorithm performance, the parameters of unimodality, accuracy, sensitivity, calculating cost are compared and analyzed.Compared with the traditional Brenner Algorithm, the experiments and analysis show that the improved Algorithm can meet unimodality, accuracy, and improve sensitivity and reduce calculating cost.