【摘要】 根据超声C扫描成像原理,分析了采样间隔对超声C扫描图像分辨率的影响.采用一种基于局部协方差特征的边缘自适应图像插值方法来提高C扫描图像的分辨率,同时将图像局部均方差作为边缘特征判据,使算法得到了简化.该方法根据低分辨率图像的局部协方差估计出高分辨率图像的插值加权系数,能够在放大图像的同时保留图像的边缘特征.试验结果表明,基于局部协方差特征的图像插值方法明显提高了C扫描图像分辨率,并且减小了边缘模糊效应,因此降低了采样间隔对图像分辨率的影响,提高了图像中缺陷分布评价结果的准确性.更多还原

【Abstract】 The effect of sampling step on the resolution of ultrasonic C-scan image was studied, and an edge adaptive image interpolation method based on the local covariance feature was presented to improve the C-scan image resolution. The method was also optimized using the local mean square deviation as a criterion of edge character. The basic idea of the method is to estimate the interpolation weight parameters of high-resolution image from the local covariance of low-resolution image, hence the edge character in high-resolution image was reserved after interpolation. The experimental results show that the interpolation method enables to improve the C-scan image resolution obviously and reduce the edge blur effect, which facilitates alleviating the effect of sampling step on the image resolution and improving the evaluation accuracy of the defect in the image.更多还原