摘要
针对面状焊接缺陷超声B扫描检测图像受噪声污染造成的质量下降问题,利用多尺度小波分析方法,对超声B扫描图像进行了增强处理。分析了缺陷信号和噪声的小波变换特性,分别在软、硬阈值函数条件下,采用固定阈值、Stein无偏似然估计阈值、混合型阈值、极大极小准则阈值等方法进行了信号增强处理,比较了不同阈值方法小波变换的处理效果。结果表明:小波变换方法能有效地提高超声B扫描图像质量,硬阈值条件下固定阈值小波变换方法的图像增强处理效果较好。
Aimed at that ultrasonic B scan image of planar defect was degraded because of contaminated by noise,the degraded image was en- hancement processed by using wavelet analysis.The wavelet transform characters of defect signal and noise were analyzed by means of Sqt- wolog,Stein Unbiased Risk Estimate,Heursure and Minimaxi methods in defect signal enhancement processing by using hard-thresholding and soft-thresholding function respectively.The results show that wavelet transform can effectively enhance the degraded image by using Sqtwolog under the condition of hard-thresholding function,and image can be obtained with better resolution.
出处
《焊接》
北大核心
2007年第1期21-24,共4页
Welding & Joining
关键词
面状缺陷
B扫描图像
小波变换
图像增强
planar defect
B scan image
wavelet transform
image enhancement
