基于动态超声图像的减压病气泡检测系统设计和实现

The Design and Implementation of Decompression Sickness Bubble Detection System Based on Dynamic Ultrasound Images

摘要: 本文详细介绍了基于动态超声图像的减压病气泡检测关键技术以及实现过程。重点介绍了检测过程中的两大关键技术：ROI轮廓提取和气泡检测。针对ROI轮廓提取，本文提出了同邻域距离判别法、同邻域轮廓曲率一致性判别法和同邻域梯度方向一致性判别法来精确定位ROI的轮廓。针对ROI区域中气泡的检测，本文提出了基于区域生长和闭合边缘路径的气泡检测方法，并分析了两种方法各自的适应情况。实际检测结果验证了本文算法的有效性。

Abstract: This paper introduces in detail the key technologies and the complete implementation procedure of decompression sickness bubble detection based on dynamic ultrasound image. It focuses on two key technologies: ROI contour extraction and bubble detection. With regard to ROI contour extraction, this paper proposes the neighborhood distance criterion, the neighborhood contour curvature consistency criterion and neighborhood gradient direction consistency criterion to precisely locate the outline of ROI. For bubble detection, this paper puts forward the bubble detection algorithms based on region growing and edge closed path, and analyzes the applicable situation of the two algorithms. The actual test results verify the validity of the algorithms.