一种改进的Otsu图像阈值分割算法

Otsu法是图像阈值化分割法中的经典算法之一,在图像处理中有广泛的应用。本文针对红外吊舱图像的特点,在对Otsu法进行分析的基础上,提出了一种改进的Otsu图像阈值分割算法。首先,采用整幅图像的平均灰度值作为初始阈值对图像进行分割;然后,推导出了Otsu法四个参数的递推公式;最后,在缩小的灰度级范围内用递推的Otsu法来计算最终的分割阈值T。实验仿真结果表明,改进的Otsu算法在图像分割效果方面和时间效率方面明显优于Otsu法。

Improved image threshold segmentation algorithm based on OTSU method

As one of the classical methods for image threshold segmentation,OTSU algorithm has been widely applied in image processing.An improved method based on Otsu algorithm for IR images is presented.Firstly,the average level is used as the initial threshold to segment images.Secondly,the recurrence formulas of four parameters are deduced.Finally,the threshold value T is calculated using the four formulas within the reduced gray level range.Experiment results show that the improved method can get better segmentation effect and reduce the calculation time significantly.