改进的快速Otsu自适应分割算法及其应用*

对Otsu算法因灰度直方图峰值不明显导致分割效果差提出了改进,根据分割目标背景信息的先验值对类内灰度值进行调整,通过迭代计算,使类间方差最大化,从而自动确定阈值,并应用于虹膜图像分割。实验结果表明,该算法对虹膜图像分割效果好,运算速度快,具有一定的鲁棒性和自适应性,可用于虹膜图像的实时分割。

Improved fast adaptive Otsu segmentation algorithm & its application

This paper introduced an improved Otsu algorithm for getting a good segmentation effect even if the histogram’s peak was not obvious for image segmenting.The algorithm took advantage of the prior knowledge about target and background to decrease classification difficult by clustering known background gray value,then achieved a satisfied threshold automatically through iterating procedure based on Otsu method.Finally,the target could be extracted by comparing every pixel’s gray value with the achieved threshold,and used it in iris image segmentation.From the segmentation results in eye image database,the algorithm’s performance on real-time and robustness is satisfied.The algorithm can be used in the real-time system of iris image segmentation.