基于二维HHT的隧道超前探测图像识别与检测

Advanced Detection Images in Tunnel Based on Bidimensional HHT

为有效分析隧道超前探测图像以避免灾害的发生,研究应用二维希尔伯特—黄变换(二维HHT)进行隧道超前探测图像的识别与检测。应用二维经验模态分解(BEMD)方法将隧道超前探测图像分解为不同频率的本征模态函数图(IMF)分量,去除含噪声的高频分量,得到重构后的新图像;再对新图像进行Hilbert变换,进而进行复信号分析,并求取图像的瞬时参数,突出图像特征。研究表明:二维HHT方法能较好地去除图像的噪声部分,并结合复信号分析所到的瞬时参数图,突出超前探测图像的异常体特征。

For the purpose of avoiding the disaster happening through effective analysis of advanced detection images in tunnel,a method for tunnel advanced detection images identification and detection is sudied based on 2D HHT（Hilbert-Huang Transform）method. BEMD method is applied to divide the advanced detecting images into IMF（Intrinsie Mode Function） component with different frequency and eliminate the component with high frequency,so as to obtain the new image after refactoring. Then,the hilbert transform was transformed to calculate the complex signal of the new image and parameters, draw up the instantaneous parameters and stress the characteristics of the image. The results show that 2D HHT method can eliminate the noise of the imgaes, and stress the unusual parts of advanced detection images by combine with the instantaneous parameters obtained by complex signal analysis.