| CT is the abbreviation of computer Tomography, which combined the X ray and the complicated computer signals processing method. CT has been applied in many fields of science, and in the late of 1980s, it is used to observe the inner structure of rock block. CT signals are usually expressed as CT number and displayed by graphic or image. However, because of the restrictions of computer gray image and the eyes of man's distinguish ability, we need to improve the image resolution of the CT, in order to make an accurate judgement when we have a direct analysis of the CT signals.In order to provide high-quality and clear computer gray images, on the basis of the principle of CT machine, this paper selected a group of the same samples of the first section of the measured CT scan data as the research object, according to the characteristics of these CT signals, the filtering function of Fourier Transform and wavelet transform were applied to divide the frequency, through the decomposition and reconstruction of CT signals, it is divided into low-frequency signals and high-frequency in two parts. The low-frequency signals reflect the main information, high-frequency signals reflect the details, and the overall goal is to deal with "general agreement, the details highlighted." "The overall agreement" reflects the low-frequency signals' comparability, "the details highlighted" reflects the clarity of high-frequency signal. Through the comparative analysis of low-frequency of the CT signals, we found that the consistency exists in low-frequency signals, after frequency demultiplication, the image resolution of high frequency signal increased more than before, and it was obviously matched with the cracks.Frequency separation shows that the low-frequency signal of CT is a certain signal, comparable between each other and it reflects the overall characteristic of the inner structure of the rock sample, the high-frequency signal is a random signal, it reflects when the rock samples on the load, the changes of density in rock and random interference caused by CT machines. Through illustration of a two-CT image showed that it can improve the image resolution significantly.In order to obtain internal structure changes of samples in the loading process, we employed concentric ring and fractured domain to make statistics of the high frequency signal, statistics show that the two methods have similar results of the law, and have close relations with the selection of regional, this indicates that it is feasible to use Fourier or Wavelet methods to separate frequency, in order to promote the resolution of the CT images. |
