高级检索

基于归一化互相关的超声全矩阵图像拼接方法

Image stitching method for ultrasonic full matrix images based on normalized cross-correlation technology

  • 摘要: 由于超声相控阵探头尺寸限制,单次检测难以完成大型构件的整体成像,影响构件整体性能评估的准确性。为了突破传统图像拼接方法对机械运动装置和编码器的依赖,文章提出了基于归一化互相关模板匹配技术的全矩阵相位迁移超声图像拼接方法,通过融合不同位置的相位迁移图像实现全景成像。仿真和实验结果表明,该方法能够实现低分辨率相位迁移图像的准确拼接,同时优化图像边缘区域的成像效果。通过对比噪声比量化分析可知,拼接后的图像整体质量得到提升。

     

    Abstract: Due to the limited size of phased array transducers, it is difficult to obtain a full view of large components, which affects the ability to accurately evaluate these components. A full matrix phase shift migration ultrasound image stitching method based on normalized cross-correlation (NCC) template matching technology is proposed to overcome the dependence of traditional image stitching methods on mechanical motion devices and encoders. A full-view image is obtained by fusing ultrasound images from different locations. Simulation and experimental results show that this method ensures the stitching accuracy of low-resolution phase-shift migration ultrasound images and improves the image quality at the edges. Quantitative analysis of the Contrast-to-Noise Ratio (CNR) shows that the overall quality of the stitched images has been effectively improved.

     

/

返回文章
返回