【摘要】 超声图像的分割问题是超声图像分析和理解中的一大难题.由于超声图像背景较暗,光照不均的特点,传统的分割方法在超声图像中不能得到很好的分割效果,本文在频域范围基于相位的相关性讨论图像的最佳阈值,通过选取适合超声图像的参数获得最佳阈值并对其进行自动分割,实验证明,针对超声图像,本文的方法优于传统的在灰度域讨论最佳阈值的方法.更多还原

【Abstract】 Segmentation of ultrasonic image is a difficult problem in the analysis and comprehension.For the dark background and uneven illumination,perfect segmentation result can′t obtain through traditional segmentation method.In this article,we discuss optimal threshold base on phase correlation in frequency field.The ultrasonic image was segmented automatically by the selection of coefficient that fit for ultrasonic image.The experiment result shows that for ultrasonic image,this method is better than traditional method that discuss optimum threshold in gray level.更多还原