基于二次移动平均法估计背景光照的二值化方法

An Binarization Method by Background Estimation Based on Double Moving Average

针对不均匀光照图像的阈值分割问题,提出一种基于二次移动平均法估计背景光照的改进二值化方法。首先,利用二次移动平均预测法对一维图像空间的灰度序列进行趋势预测,根据预测趋势寻找前景与背景的分界点以此进行背景估计,然后利用背景差法分离出目标图像,最后采用最大类间方差法进行全局分割获得分割结果。为验证算法的有效性,实验采用50幅非均匀光照条件下的图像作为测试样本,并与几种局部阈值分割算法进行了对比。实验结果表明,与传统阈值分割算法相比,该算法能够在减少光照影响获取较为完整的图像信息的同时,在处理速度上也具有明显的提升。

An improved binarization method based on moving average method to estimate background illumination is proposed for threshold segmentation problem of images with uneven illumination.First,the quadratic moving average prediction method is used to predict the trend of the grayscale sequence in the one-dimensional image space.Then the boundary point between the fore-ground and the background is found according to the predicted trend to estimate the background,and the background difference method is used to separate the target foreground image.Finally,the maximum inter-class variance method is used to perform global segmentation to obtain segmentation results.To verify the effectiveness of the algorithm,50 images under non-uniform lighting con-ditions are used as test samples in the experiment,and compared with several local threshold segmentation algorithms.Compared with the traditional threshold segmentation algorithm,experimental results show that the algorithm has better performance in reduc-ing the impaction caused by non-uniform illumination and enhance the quality of images.At the same time,it also has a significant improvement in processing speed.