A terahertz three-dimensional tomography method

Abstract

The present invention discloses a terahertz three-dimensional tomography method, wherein the sampling points are plotted according to the sampling order to draw a sampling point x-coordinate change graph; the redundant sampling points at the serpentine inflection points in the sampling point x-coordinate change graph are removed by first-order difference; the optimized sampling point x-coordinates are plotted according to the acquisition order to draw a new sampling point x-coordinate change graph, and a cutting algorithm is used to cut at the serpentine inflection points to divide the data into n groups; the x-coordinates of each group of sampling points and their corresponding terahertz time domain signals form a group of two-dimensional arrays, and the length of the jth group of two-dimensional arrays is n. j *t, after grouping, the length of the two-dimensional arrays in each group will be different. The nearest neighbor interpolation method is used to unify the length of the two-dimensional arrays in each group to obtain regularized terahertz imaging data; the regularized terahertz imaging data is input into the imaging module to generate a three-dimensional image or an image of a certain slice. The three-dimensional tomography displayed by this method is clearer and has universality. Any data can be displayed by this method.

Description

Technical field:

[0001] The invention belongs to the technical field of terahertz spectroscopy and imaging, and in particular relates to a terahertz three-dimensional tomography method. Background technology:

[0002] Terahertz waves are electromagnetic waves with a wavelength between infrared and microwaves. Because they are in the transition region from photonics to electronics, they have many unique properties, such as fingerprint spectroscopy, low energy, and special penetration. The interaction between terahertz waves and materials will produce a unique absorption spectrum - fingerprint spectrum. Polar molecules, such as water molecules, and non-polar molecules, such as carbon dioxide molecules, have very obvious differences in the absorption of terahertz waves, so terahertz absorption spectra are also of great value in testing molecular properties. Time-domain terahertz spectroscopy scanning technology uses high-precision delay lines to extend the sampling time of femtosecond pulses to tens of picoseconds, and reduces noise through hardware preprocessing. The time-domain signal generated by it is Fourier transformed to obtain the characteristic absorption spectrum of the substance being tested, and its spectral width can reach more than 5THz and the dynamic range can reach more than 70dB. This spectrum performance can meet the detection needs of most compounds, thus providing a large number of application scenarios for terahertz time-domain spectroscopy scanning.

[0003] At present, our company uses pulsed terahertz wave imaging system for imaging. The basic principle of pulsed terahertz wave imaging is that the intensity and phase of the THz electromagnetic wave passing through the imaging sample (or reflected from the sample) contain the spatial distribution of the complex dielectric function of the sample. The THz image of the sample can be obtained by recording the two-dimensional information of the intensity and phase of the transmitted THz electromagnetic wave and processing and analyzing it appropriately.

[0004] At present, terahertz three-dimensional tomography technology has a broad space for development. Improving the speed and resolution of three-dimensional tomography are both urgent problems to be solved. The existing three-dimensional tomography algorithms, such as terahertz peak intensity imaging, terahertz flight time imaging, and terahertz specific power spectrum imaging, can all be used for imaging display. These algorithms have their own advantages, but they cannot be displayed as a universal algorithm. This application provides a new terahertz three-dimensional tomography method. Summary of the invention:

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and seek to design a terahertz three-dimensional tomography method. The original data collected by the terahertz three-dimensional tomography system is used to process the original data and display it in three dimensions to perform a new terahertz three-dimensional tomography data processing and communication display method. After clever data processing and algorithm processing, the contextlib module in python is used for memory communication, the mayavi module is used for three-dimensional tomography visualization display, and the pyqt5 module is used for interface encapsulation and integration of the host computer. The three-dimensional tomography image displayed by this method is clearer and has universality. Any data can be displayed by this method.

[0006] In order to achieve the above-mentioned purpose, the present invention relates to a terahertz three-dimensional tomography method, which specifically comprises the following steps:

[0007] (1) Obtain the coordinates of N sampling points and the terahertz time-domain spectrum signals corresponding to the coordinates from the host computer. The total coordinate set P of the sampling points is {(x1, y1), (x2, y2), ..., (x N ,y N )}, set X: (x1, x2, ..., x N ) draws a graph of the x-coordinate changes of all sampling points in the sampling order;

[0008] (2) Using the first-order difference to remove the redundant sampling points at the serpentine inflection points in the sampling point x-coordinate change graph, the optimized set X′ is obtained;

[0009] (3) Draw a new x-coordinate change graph of all sampling points in the set X′ according to the order of collection, and use the cutting algorithm to cut at the inflection point of the snake to divide the data into n groups;

[0010] (4) After grouping, the x-coordinates of each group of sampling points and their corresponding terahertz time-domain signals form a two-dimensional array. The number of sampling points in the jth group is n. j , the length of the j-th two-dimensional array is n j *t, j∈(1, 2, ..., n), t is the sampling time. After grouping, the length of the two-dimensional array in each group will be different. The nearest neighbor interpolation method is further used to unify the length of the two-dimensional array in each group to obtain regularized terahertz imaging data;

[0011] (5) The regularized terahertz imaging data is input into an imaging module to generate a three-dimensional image or an image of a certain cut surface.

[0012] As an implementation method, the set X′ is optimized by the following steps:

[0013] (201) The x coordinates in the total set P are extracted in the order of collection to form a set X: (x1, x2, ..., x N ), subtract the x-coordinate value of the ith sampling point from the x-coordinate value of the ith sampling point according to the following formula, and perform the first-order difference method. If Δx<x value , then delete the x coordinate of the i+1 / ith sampling point in the set X to obtain the optimized set X′;

[0014] Δx=x i+1 -x i , i∈(1,2,...,N-1)

[0015] Among them, x i+1 is the x coordinate value of the i+1th sampling point, x i is the x coordinate value of the i-th sampling point, x value The x-direction threshold is set.

[0016] As another implementation, the set X′ is optimized by the following steps:

[0017] (201) The x coordinates in the total set P are extracted in the order of collection to form a set X: (x1, x2, ..., x N ), subtract the x-coordinate value of the ith sampling point from the x-coordinate value of the ith sampling point according to the following formula, and perform the first-order difference method. If Δx<x value , then delete the x coordinate of the i+1 / ith sampling point in the set X to obtain the optimized set X1;

[0018] Δx=x i+1 -x i , i∈(1,2,...,N-1)

[0019] Among them, x i+1 is the x coordinate value of the i+1th sampling point, x i is the x coordinate value of the i-th sampling point, x value is the set x-direction threshold;

[0020] (202) In order to further remove the data at the inflection point, all the y coordinates in the total set P are extracted in the order of collection to form a set Y: (y1, y2, ..., y N ), subtract the y coordinate value of the i+1th sampling point from the y coordinate value of the i-th sampling point according to the following formula, and perform the first-order difference method. If Δy < y value , then delete the x coordinate of the i+1 / ith sampling point in the set X1 to obtain the further optimized set X′;

[0021] Δy=y i+1 -y i, i∈(1,2,...,N-1)

[0022] Among them, y i+1 is the y coordinate value of the i+1th sampling point, y i is the x coordinate value of the i-th sampling point, y value The y-direction threshold is set.

[0023] Specifically, in step (3), a peak-finding algorithm is used to find all the peaks and troughs in the x-coordinate change graph of the new sampling point, and the x-coordinate change graph is divided into n segments by cutting at the peaks and troughs. Each segment represents a continuous movement of the sample along the x-direction. The x-coordinates of the sampling points contained in each segment constitute a group, and n is related to the movement of the sample in the y-direction. The sample moves n steps Ystep in the y-direction.

[0024] Specifically: In step (4), the number of sampling points in each group after regularization is set to n width , if the actual number of sampling points in the jth group is n j ≠n widt , using the nearest neighbor interpolation function f(n j *t)→f(n width *t), unify the length of the two-dimensional arrays in each group to obtain n two-dimensional matrices of the same size n width *t, then we combine these n two-dimensional matrices of the same size into a three-dimensional array, the size of the three-dimensional array is n widt *n*t, terahertz imaging data includes regularized two-dimensional arrays and three-dimensional arrays.

[0025] Specifically, step (5) inputs the three-dimensional array and the two-dimensional array into the mayavi module of python to perform three-dimensional tomography and cross-sectional image display respectively.

[0026] Specifically, the terahertz three-dimensional tomography method is written in Python. The host computer stores the coordinates of the sampling points and the terahertz time-domain spectrum signals corresponding to the coordinates in the memory. Python uses the contextlib module to obtain the three-dimensional array from the memory. After processing, the mayavi module is used to perform three-dimensional tomography and cross-sectional image display.

[0027] Compared with the prior art, the present invention has the following beneficial effects: a simple and efficient data processing method is proposed, and after clever processing, the contextlib module of Python is used for inter-memory communication, mayavi is used for imaging display, and the pyqt5 module is used for interface encapsulation and integrated into the host computer. Description of the drawings:

[0028] Figure 1It is a schematic diagram of the structure of the terahertz three-dimensional tomography system involved in Example 1.

[0029] Figure 2 This is a sample (standard card) diagram used in the terahertz three-dimensional tomography involved in Example 1. The intervals on the standard card are in mm.

[0030] Figure 3 This is a serpentine movement diagram of the sampling process involved in Example 1.

[0031] Figure 4 This is the change of the original x-coordinate of the sampling point involved in Example 1. There are obvious transverse multiple sampling points at the serpentine bend.

[0032] Figure 5 yes Figure 4 A partial enlarged view of .

[0033] Figure 6 It is the change of the x coordinate of the sampling point after the processing involved in Example 1.

[0034] Figure 7 yes Figure 6 A partial enlarged view of .

[0035] Figure 8 It is a three-dimensional tomographic stereogram involved in the embodiment.

[0036] Fig. 9 It is a three-dimensional cross-sectional view involved in the embodiment. Specific implementation method:

[0037] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0038] Example 1

[0039] like Figure 1 As shown, this embodiment uses a terahertz three-dimensional tomography system for data acquisition, and the terahertz three-dimensional tomography system includes a femtosecond laser 1, a beam splitter 2, an optical oscillation delay line 3, a bias source 4, a terahertz transmitting antenna 5, a terahertz detection antenna 10, a reflection optical path, a signal generation and acquisition processing unit 11, a host computer 12 and a two-dimensional scanning platform 13. The first collimating lens 6, the first focusing lens 7, the second focusing lens 8 and the second collimating lens 9 constitute the reflection optical path. The maximum acquisition area of ​​the two-dimensional scanning platform 13 is 100mm*100mm, and the moving step size can generally be set arbitrarily from 0.1mm to 2mm.

[0040] This embodiment uses a standard card as a sample for data collection. The standard card example is shown in FIG. Figure 2 The terahertz 3D tomography system uses 30 Hz, the acquisition time length is 90 ps, ​​and the 2D scanning platform movement step is set to 0.2 mm for data acquisition.

[0041] The specific sampling process is:

[0042] (1) Figure 3 As shown, the standard card is placed on the two-dimensional scanning platform 13, and the two-dimensional scanning platform 13 controls the sample to move in a serpentine manner along the x direction and the y direction, wherein the x direction is left-right movement (along the plane where the reflected light path is located), and the y direction is up-and-down movement (vertical to the plane where the reflected light path is located). Each time it moves to a sampling point, the terahertz time-domain spectrum signal of the sampling point is collected according to step (2), and at the same time, the two-dimensional scanning platform 13 also sends the coordinates of the sampling point to the host computer 12;

[0043] (2) The femtosecond pulse laser generated by the femtosecond laser 1 is split by the beam splitter 2, one beam is the pump light and the other is the detection light. Under the action of the bias source 4, the pump light is directly focused on the terahertz transmitting antenna 5 to generate a terahertz pulse. The terahertz pulse passes through the first collimating lens 6 and the first focusing lens 7 in sequence and then irradiates the standard card. The reflected terahertz pulse carrying the standard card information passes through the second focusing lens 8 and the second collimating lens 9 in sequence and is focused on the terahertz detection antenna 10, where it meets the detection light delayed by the optical oscillation delay line 3 and outputs a photocurrent signal. The photocurrent signal is processed by the signal generation and acquisition processing unit 11 and then sent to the host computer 12 to obtain the coordinates of the sampling point and the terahertz time domain spectrum signal corresponding to the coordinates. In the process, the optical oscillation delay line 3 and the bias source 4 are all connected to the signal generation and acquisition processing unit 11 and controlled by it.

[0044] (3) Three-dimensional tomography is performed based on the terahertz imaging data composed of the coordinates of the sampling points and the terahertz time-domain spectral signals corresponding to the coordinates.

[0045] The present embodiment relates to a terahertz three-dimensional tomography method, which specifically comprises the following steps:

[0046] (1) Assume that there are N collection points and the total coordinate set P of the collection points is {(x1, y1), (x2, y2), ..., (x N ,y N )}, and plot the x-coordinate variation of the sampling points according to the sampling order.

[0047] like Figure 4 is the coordinate change diagram of the sampling point in the x direction, where the horizontal axis represents the sampling point in the snake-like movement process, and the vertical axis represents the coordinate of the sampling point in the x direction. Figure 5It can be found that when the two-dimensional scanning platform moves laterally, the need to accelerate and decelerate at the serpentine turn will cause the platform to stay for too long, resulting in an abnormally large amount of data at the serpentine turn. The original terahertz imaging data is asymmetric and irregular, and cannot be directly processed for imaging.

[0048] (2) The first-order difference method is used to process the serpentine inflection point and remove the redundant sampling points at the inflection point in the sampling point x-coordinate change diagram. This effectively eliminates the excessively collected coordinate points and terahertz time domain signals during the startup and acceleration of the two-dimensional scanning platform.

[0049] Specifically:

[0050] (201) The x coordinates in the total set P are extracted in the order of collection to form a set X: (x1, x2, ..., x N ), subtract the x-coordinate value of the previous sampling point (the i-th sampling point) from the x-coordinate value of the next sampling point (the i+1-th sampling point) according to the following formula, and perform the first-order difference method. If Δx < x value , then delete the x coordinate of the corresponding sampling point (i+1th or ith sampling point) in the set X to obtain the optimized set X′;

[0051] Δx=x i+1 -x i , i∈(1,2,...,N-1)

[0052] Among them, x i+1 is the x coordinate value of the i+1th sampling point, x i is the x coordinate value of the i-th sampling point, x value is the x-direction threshold. Here, the threshold is set to 0.5 times the x-direction step. value =0.5*Xstep, Xstep is the step in the x direction;

[0053] Figure 6 and Figure 7 This is a graph showing the change in the x coordinate of the sampling point after processing in step (201). It can be seen that the two-dimensional scanning platform has achieved perfect integration at the serpentine turning point without any redundant stop points.

[0054] (3) Draw a new sampling point x coordinate change graph according to the order of collection for all sampling point x coordinates in set X′ ( Figure 6 ), using the cutting algorithm to cut at the inflection point of the snake to group the data;

[0055] Specifically, the peak-finding algorithm is used to find all the peaks and troughs in the x-coordinate change graph of the new sampling point, and the x-coordinate change graph is divided into n segments by cutting at the peaks and troughs. Each segment represents a continuous movement of the sample along the x-direction. The x-coordinates of the sampling points contained in each segment constitute a group, and n is related to the movement of the sample in the y-direction. The sample moves n steps Ystep in the y-direction.

[0056] (4) After grouping, the x-coordinates of each group of sampling points and their corresponding terahertz time domain signals form n groups of two-dimensional arrays, and the number of sampling points contained in the jth group is n. j , the length of the j-th two-dimensional array is n j *9000, j∈(1, 2, ..., n), after grouping, the number of sampling points (or the length of the two-dimensional array) in each group will be different, and the nearest neighbor interpolation method is further used to unify the number of sampling points (or the length of the two-dimensional array) in each group to obtain regularized terahertz imaging data;

[0057] Specifically: Set the number of sampling points in each group after regularization to n width , if the actual number of sampling points in the jth group is n j ≠n widt , using the nearest neighbor interpolation function f(n j *9000)→f(n width *9000), unify the length of the two-dimensional arrays in each group to obtain n two-dimensional matrices of the same size n width *9000, then we combine these n two-dimensional matrices of the same size into a three-dimensional array, the size of the three-dimensional array is n widt *n*9000, terahertz imaging data includes regularized two-dimensional arrays and three-dimensional arrays;

[0058] The width It can be (x max -x min ) / Xstep, where x max is the maximum value of the x coordinate in each group, x min is the minimum value of the x coordinate, which can also be n1, n2..., n n The majority of .

[0059] (5) The regularized terahertz imaging data is input into an imaging module to generate a three-dimensional image or an image of a certain cut surface.

[0060] Specifically, the three-dimensional array and the two-dimensional array are input into the mayavi module of python to perform three-dimensional tomography and cross-sectional display, respectively.

[0061] Specifically, the terahertz three-dimensional tomography method described in this embodiment is written in python, while the host computer software is written in C#. The host computer stores the coordinates of the sampling points and the terahertz time-domain spectrum signals corresponding to the coordinates in the memory. Python uses the contextlib module to obtain the three-dimensional array from the memory, and after processing, the mayavi module is used to perform three-dimensional tomography and cross-sectional display. Mayavi is encapsulated in the pyqt5 interface and then integrated with the host computer. Other relatively easy-to-use vtk and paraview visualization tools cannot be integrated with the host computer. The open source mayavi is more convenient to integrate. The contextlib module is used for communication transmission in the communication protocol, so that data is transferred between C# and python in the memory. This memory-based transmission method greatly speeds up the data transmission processing speed. The three-dimensional display image is as follows: Figure 7 and Figure 8 ,in Figure 7 It is a three-dimensional display. Figure 8 It is a three-dimensional cross-section diagram.

[0062] Example 2

[0063] Except for step (2), all other steps of this embodiment are the same as those of embodiment 1.

[0064] (2) The first-order difference method is used to process the serpentine inflection point and remove the redundant sampling points at the inflection point in the sampling point x-coordinate change diagram. This effectively eliminates the excessively collected coordinate points and terahertz time domain signals during the startup and acceleration of the two-dimensional scanning platform.

[0065] Specifically:

[0066] (201) The x coordinates in the total set P are extracted in the order of collection to form a set X: (x1, x2, ..., x N ), subtract the x-coordinate value of the previous sampling point (the i-th sampling point) from the x-coordinate value of the next sampling point (the i+1-th sampling point) according to the following formula, and perform the first-order difference method. If Δx < x value , then delete the x coordinate of the corresponding sampling point (i+1th or ith sampling point) in the set X to obtain the optimized set X1;

[0067] Δx=x i+1 -x i , i∈(1,2,...,N-1)

[0068] Among them, x i+1 is the x coordinate value of the i+1th sampling point, x i is the x coordinate value of the i-th sampling point, x value is the x-direction threshold. Here, the threshold is set to 0.5 times the x-direction step.value =0.5*Xstep, Xstep is the step in the x direction;

[0069] (202) In order to further remove the data at the inflection point, all the y coordinates in the total set P are extracted in the order of collection to form a set Y: (y1, y2, ..., y N ), subtract the y coordinate value of the previous sampling point (the i-th sampling point) from the y coordinate value of the next sampling point (the i+1-th sampling point) according to the following formula, and perform the first-order difference method. If Δy < y value , then the x coordinate of the corresponding sampling point (the i+1th or i-th sampling point) is deleted from the set X1 to obtain the further optimized set X′.

[0070] Δy=y i+1 -y i , i∈(1,2,...,N-1)

[0071] Among them, y i+1 is the y coordinate value of the i+1th sampling point, y i is the x coordinate value of the i-th sampling point, y value is the set y-direction threshold. Here, the threshold is set to twice the y-direction step, that is, y value =2*Ystep, Ystep is the step in the y direction.

[0072] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A terahertz three-dimensional tomography method, characterized in that: The specific steps include: (1) Obtain the coordinates of N sampling points and the terahertz time-domain spectrum signals corresponding to the coordinates from the host computer. The total coordinate set P of the sampling points is {(x1, y1), (x2, y2), …, (x N ,y N )}, the set X:(x1,x2,…,x N ) draws a graph of the x-coordinate changes of all sampling points in the sampling order; (2) Use the first-order difference to remove the redundant sampling points at the serpentine inflection points in the sampling point x coordinate change graph to obtain the optimized set X ′ ; (3) Set X ′ All the x-coordinates of the sampling points in the new sampling point x-coordinate change graph are drawn according to the collection order, and the cutting algorithm is used to cut at the serpentine inflection point to divide the data into n groups; (4) After grouping, the x-coordinates of each group of sampling points and their corresponding terahertz time-domain signals form a two-dimensional array. The number of sampling points in the jth group is n. j , the length of the j-th two-dimensional array is n j *t, j∈(1,2,…,n), t is the sampling time. After grouping, the length of the two-dimensional array in each group will be different. The nearest neighbor interpolation method is further used to unify the length of the two-dimensional array in each group to obtain regularized terahertz imaging data; (5) Input the regularized terahertz imaging data into the imaging module to generate a three-dimensional image or an image of a certain cut surface; Set X ′ The optimization is obtained by following the steps below: (201) Extract the x coordinates in the total set P in the order of collection to form a set X: (x1, x2, …, x N ), subtract the x-coordinate value of the i+1th sampling point from the x-coordinate value of the i-th sampling point according to the following formula, and perform the first-order difference method. If Δx <x value , then delete the x coordinate of the i+1 / ith sampling point in the set X to obtain the optimized set X ′ ; Δx=x i+1 -x i ,i∈(1,2,…,N-1) Among them, x i+1 is the x coordinate value of the i+1th sampling point, x i is the x coordinate value of the i-th sampling point, x value is the set x-direction threshold; or set X ′ The optimization is obtained by following the steps below: (201) Extract the x coordinates in the total set P in the order of collection to form a set X: (x1, x2, …, x N ), subtract the x-coordinate value of the i+1th sampling point from the x-coordinate value of the i-th sampling point according to the following formula, and perform the first-order difference method. If Δx <x value , then delete the x coordinate of the i+1 / ith sampling point in the set X to obtain the optimized set X1; Δx=x i+1 -x i ,i∈(1,2,…,N-1) Among them, x i+1 is the x coordinate value of the i+1th sampling point, x i is the x coordinate value of the i-th sampling point, x value is the set x-direction threshold; (202) In order to further remove the data at the inflection point, all the y coordinates in the total set P are extracted in the order of collection to form a set Y: (y1, y2, …, y N ), subtract the y coordinate value of the i+1th sampling point from the y coordinate value of the i-th sampling point according to the following formula, and perform the first-order difference method. If Δy <y value , then the x coordinate of the i+1 / ith sampling point is deleted from the set X1 to obtain the further optimized set X ′ ; Δy=y i+1 -y i ,i∈(1,2,…,N-1) Among them, y i+1 is the y coordinate value of the i+1th sampling point, y i is the x coordinate value of the i-th sampling point, y value The y-direction threshold is set.

2. The terahertz three-dimensional tomography method according to claim 1, characterized in that: In step (3), a peak-finding algorithm is used to find all the peaks and troughs in the x-coordinate change graph of the new sampling point, and the x-coordinate change graph is divided into n segments by cutting at the peaks and troughs. Each segment represents a continuous movement of the sample along the x-direction. The x-coordinates of the sampling points contained in each segment constitute a group, and n is related to the movement of the sample in the y-direction. The sample moves n steps Ystep in the y-direction.

3. The terahertz three-dimensional tomography method according to claim 1, characterized in that: In step (4), the number of sampling points in each group after regularization is set to n width , if the actual number of sampling points in the jth group is n j ≠n widt , using the nearest neighbor interpolation function f(n j *t)→f(n width *t), unify the length of the two-dimensional arrays in each group to obtain n two-dimensional matrices of the same size n width *t, then we combine these n two-dimensional matrices of the same size into a three-dimensional array, the size of the three-dimensional array is n widt *n*t, terahertz imaging data includes regularized two-dimensional arrays and three-dimensional arrays.

4. The terahertz three-dimensional tomography method according to claim 1, characterized in that: Step (5) inputs the three-dimensional array and the two-dimensional array into the mayavi module of python to perform three-dimensional tomography and cross-sectional image display respectively.

5. The terahertz three-dimensional tomography method according to claim 1, characterized in that: The terahertz three-dimensional tomography method is written in Python. The host computer stores the coordinates of the sampling points and the terahertz time-domain spectrum signals corresponding to the coordinates in the memory. Python uses the contextlib module to obtain the three-dimensional array from the memory. After processing through steps (1)-(4), the mayavi module is used to perform three-dimensional tomography and cross-sectional image display.

Images