Multi-band integrated terahertz array imaging system and method

Abstract

The invention discloses a multi-band integrated terahertz array imaging system and method, and belongs to the technical field of detection. The system comprises a multi-port vector network analyzer, a terahertz sensor array, a mechanical arm and a server. The terahertz sensor array adopts a multi-layer integrated chip structure integrating receiving and transmitting, a traditional discrete terahertz link is replaced, and the size is remarkably reduced. An array is driven through a multi-port vector network analyzer, an S parameter is collected, an equivalent dense array element is formed in combination with precise scanning of a mechanical arm, data are processed by a server, and three-dimensional tomography is achieved. According to the invention, through a method of combining'global rapid detection 'and'target area fine detection', the detection efficiency and precision are effectively balanced, and the problems of large array integration difficulty, inconsistent phase centers, low imaging resolution and incapability of flexibly adapting to different detection requirements in the prior art are solved; and high-resolution and high-efficiency identification and positioning of the internal defects of the non-metallic material are realized.

Description

Technical Field

[0001] This invention belongs to the field of detection technology and terahertz imaging, specifically relating to a multi-band integrated terahertz array imaging system and method. Background Technology

[0002] New materials, such as composite materials and thermal insulation materials, are increasingly widely used in high-end manufacturing. However, they are prone to defects such as pores, delamination, and inclusions during preparation and use, necessitating reliable non-destructive testing technologies. Terahertz waves have excellent penetrability to non-metallic materials, and their low photon energy does not cause ionizing damage to organisms or materials. Therefore, terahertz imaging technology is one of the preferred technologies for achieving large field-of-view, high-resolution, and rapid non-destructive testing of non-metallic materials.

[0003] Currently, most common terahertz array imaging methods are based on linear frequency modulated continuous wave (LFM) systems. This system has the following inherent drawbacks: First, the system's phase noise and frequency modulation nonlinearity severely impact testing accuracy, especially in the terahertz band, where nonlinear errors are difficult to calibrate, leading to poor pixel consistency and a noticeable "snowflake" noise effect in the final imaging results. Second, LFM test bandwidth is typically narrow, and the number of frequency points is relatively fixed, making it difficult to flexibly configure sweep parameters. Therefore, it is challenging to effectively balance range resolution, detection range, and detection speed, limiting its ability to detect multiple defects at different depths.

[0004] Furthermore, traditional terahertz signal generation and detection links typically consist of discrete components such as frequency dividers, frequency multipliers, amplifiers, and antennas, resulting in a large and complex system that is difficult to integrate at high density on a large scale. Even if integration is achieved, the phase center inconsistency between array elements will lead to a decrease in the spatial resolution of the imaging.

[0005] Therefore, there is an urgent need in this field for a terahertz array imaging solution that can maintain high precision and wide measurement range, achieve miniaturization and high integration, and possess flexible and efficient detection capabilities. Summary of the Invention

[0006] To address the aforementioned technical problems in existing technologies, this invention provides a multi-band integrated terahertz array imaging system and method to solve the technical problems in existing terahertz imaging technologies, such as large system size, difficulty in array integration, significant phase noise impact, single detection mode, and inability to balance efficiency and accuracy.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: A multi-band integrated terahertz array imaging system includes: a multi-port vector network analyzer, a terahertz sensor array, a robotic arm, a server, a power module, and a display; The power modules provide stable power to the multi-port vector network analyzer, terahertz sensor array, robotic arm, and server, respectively. The multiport vector network analyzer, connected via lines to the power module, terahertz sensor array, and server, operates in the radio frequency band and is configured to provide RF excitation and local oscillator signals to the terahertz sensor array and receive intermediate frequency signals from the terahertz sensor array to obtain its S-parameters. The terahertz sensor array, connected to the power module and robotic arm via wiring, includes multiple transceiver sensor units operating in the terahertz frequency band, covering a frequency range of 90GHz to 500GHz and divided into multiple sub-band arrays. Each sensor unit employs a multi-layer integrated structure, including a source-measurement layer, a directional coupling layer, and an antenna layer, to generate and radiate terahertz signals while receiving reflected signals from the sample under test. The robotic arm, connected to the power module and server via wiring, is configured to control the terahertz sensor array to perform scanning motion in a two-dimensional plane according to a preset trajectory and stepping accuracy. The server is connected to a multi-port vector network analyzer, a robotic arm, and a display via wires. It is configured to control the robotic arm to move along a predetermined route, acquire S-parameters obtained from the multi-port vector network analyzer in real time, and use the S-parameters for 3D image reconstruction and defect detection.

[0008] The monitor, connected to the power module and server via cables, is configured to display images generated by the server in real time.

[0009] Preferably, the terahertz sensor array comprises at least four sub-band arrays, covering the frequency bands of 90GHz-140GHz, 140GHz-220GHz, 220GHz-330GHz and 330GHz-500GHz respectively; each sub-band array is arranged in multiple rows, and the sensor units within the array are distributed at equal intervals.

[0010] Preferably, the source and measurement layers of a single terahertz sensor unit are fully integrated on a single chip, including a power amplifier chip, a frequency multiplier chip, and a frequency mixer chip.

[0011] Preferably, the step interval for the robotic arm to control the terahertz sensor array to perform scanning motion is ≤0.5mm.

[0012] Preferably, the server is configured to perform two detection modes: Full-area rapid detection mode: Uses wide-frequency stepping or point-frequency signals for rapid scanning to initially locate defect areas; Target area fine-grained detection mode: Uses full-band, narrow-sweep-step signals for fine scanning to obtain information on the depth, size, and type of defects.

[0013] Preferably, the terahertz sensor array adopts a sparse array configuration, forming an equivalent dense array element through the alternating transmission and reception of sensor units; when processing data, the server filters multiple S-parameters at the same equivalent array element location, selecting the S-parameter with the largest amplitude for imaging, thereby improving image quality.

[0014] Furthermore, this invention also mentions a multi-band integrated terahertz array imaging method, which employs the multi-band integrated terahertz array imaging system described above, characterized by comprising the following steps: Step 1: Test Preparation; specifically includes the following steps: Step 1.1: Power on all hardware devices in the system and preheat them for 30 minutes; Step 1.2: Use a multi-port vector network analyzer to check whether all terahertz sensor arrays are working properly. The standard for judging whether they are working properly is to place a smooth metal plate in the calibration position and obtain S11, S22, S33, ..., SNN parameter values ​​that are all between 0dB and -10dB. Step 2: Rapid full-area inspection to quickly locate potential defects in the test material; specifically including the following steps: Step 2.1: Set the first set of test parameters; the first set of test parameters has a wider test bandwidth and / or a larger sweep step; Step 2.2: Place the smooth metal plate in the calibration position, use a robotic arm to move the terahertz sensor array to the center of the metal plate, use a multi-port vector network analyzer to obtain the S-parameters in this fixed state, transmit them to the server for storage, and wait for subsequent processing. At this time, the array matrix is ​​obtained and denoted as A. Step 2.3: Place the test sample in the test area, control the robotic arm to drive the terahertz sensor array to the starting point of the test, perform a full-area scan of the test sample, and collect the first S-parameter dataset; Step 2.4: Perform calculations on the first S-parameter dataset and the reference matrix A to generate the first data matrix, and use the linear array imaging algorithm to obtain a two-dimensional or three-dimensional image of the test material. Determine whether there are defects based on the imaging results. If there are no defects, the test ends. If there are defects, determine the approximate area of ​​the defects and proceed with the test. Step 3: Detailed inspection of the target area; precise analysis of information within the material containing the defect, including depth, size, and type; specifically including the following steps: Step 3.1: If a defect is detected in step 2, set a second set of test parameters. The second set of test parameters has a wider test bandwidth and / or a smaller sweep step. Step 3.2: Place the smooth metal plate in the calibration position, use a robotic arm to move the terahertz sensor array to the center of the metal plate, use a multi-port vector network analyzer to obtain the S-parameters in this fixed state, transmit them to the server for storage, and wait for subsequent processing. The resulting array matrix is ​​denoted as A. Step 3.3: Place the test sample in the test area, control the robotic arm to perform a fine scan in the located defect area, and collect the second S-parameter dataset; Step 3.4: Process the second S-parameter dataset to generate the second data matrix, use the linear array imaging algorithm to obtain the three-dimensional tomographic image of the test material, and then determine the depth location, size and type of the defect; Step 4: Output the detection results and end the test.

[0015] Preferably, the process of processing the S-parameter dataset and generating the data matrix in steps 2 and 3 includes: The S-parameter matrix B acquired at each scanning position is compared with the reference matrix A to obtain matrix C; When there are multiple S-parameters at the same physical location, the S-parameter with the largest amplitude is selected as the final representative value for that location, forming the dimensionality-reduced data matrix D. The data matrices D from multiple scanning positions are merged according to preset rules to form the final data matrix E used for imaging.

[0016] Preferably, the first set of test parameters includes spot frequency, narrowband signal or broadband signal with a sweep step ≥1GHz; the second set of test parameters is a full-band scanning signal with a sweep step <1MHz.

[0017] Preferably, by integrating multi-band signals and performing multiple relative measurements, the nonlinear error of the system is suppressed, and the distance resolution of three-dimensional tomography is improved by widening the overall test bandwidth.

[0018] The beneficial technical effects of this invention are as follows: (1) Multi-band integrated terahertz array imaging uses a terahertz sensor array to replace the traditional terahertz signal generation and detection link, reducing the volume by more than 200 times. This solves the problems of poor imaging spatial resolution caused by the difficulty of array integration and the inconsistency of the phase center of the equivalent array elements. The terahertz sensor realizes the acquisition of reflected signals while generating signals. The equivalent phase centers of the arrays are independent of each other and do not interfere with each other, which significantly reduces the signal crosstalk problem between the array elements of the large-scale sparse array. (2) A multi-port vector network analyzer is used as the driver and receiver of the terahertz sensor array. By configuring parameters such as the number of sweep points and the sweep time, the rapid detection of a single target can be achieved. At the same time, parameters such as the number of sweep points and the sweep time can be increased to realize the reuse of information from different frequency points, which is convenient for multi-sub-band processing, increases the detection depth, and thus realizes the detection of multiple targets. Through multi-band integration, the test bandwidth is broadened and the resolution of three-dimensional tomographic imaging is improved. At the same time, through multiple relative quantity tests, nonlinear errors are suppressed, and high-resolution rapid imaging of the three-dimensional structure of the material and detection and identification of internal defects are realized. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a multi-band integrated terahertz array imaging system provided in an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of a multi-band terahertz sensor array.

[0021] Figure 3 This is a schematic diagram of the structure of a single terahertz sensor.

[0022] Figure 4 This is a schematic diagram of the internal structure of a terahertz sensor.

[0023] Figure 5 A schematic diagram of the equivalent array elements obtained by transmitting data from one sensor and receiving data from other array elements.

[0024] Figure 6 This is an equivalent array element diagram of a terahertz sensor array.

[0025] Figure 7 This is a schematic diagram of the robotic arm's motion trajectory.

[0026] Figure 8 This is a flowchart of the imaging method of the present invention. Detailed Implementation

[0027] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments: like Figure 1 As shown, the system in this embodiment mainly consists of a multi-port vector network analyzer, a multi-band terahertz sensor array, a multi-axis scanning robotic arm, a server, a display, and a power module.

[0028] The multiport vector network analyzer operates at frequencies ranging from 10MHz to 26.5GHz and its functions include: firstly, providing radio frequency and local oscillator signals for terahertz sensor arrays; and secondly, obtaining the S-parameters of terahertz sensor arrays. The multi-band terahertz sensor array operates in frequency bands covering 90GHz-500GHz, such as... Figure 2As shown, the terahertz sensor arrays are 90GHz-140GHz, 140GHz-220GHz, 220GHz-330GHz, and 330GHz-500GHz, respectively. The arrays are divided into four rows according to their operating frequency bands: the first row is for the 90GHz-140GHz band, the second row for the 140GHz-220GHz band, the third row for the 220GHz-330GHz band, and the fourth row for the 330GHz-500GHz band. The terahertz sensors in each band are arranged at equal intervals. Taking a 64-channel vector network analyzer as an example, it can control 32 terahertz sensor arrays. Ports 1 to 16 control 8 terahertz sensors in the 90GHz-140GHz band, ports 17 to 32 control 8 terahertz sensors in the 140GHz-220GHz band, ports 33 to 48 control 8 terahertz sensors in the 220GHz-330GHz band, and ports 49 to 64 control 8 terahertz sensors in the 330GHz-500GHz band.

[0029] The dimensions (length * width * height) of a single terahertz sensor are 20mm * 20mm * 15mm, such as... Figure 3 As shown. The internal structure of the sensor is as follows. Figure 4 As shown, the internal structure employs a three-layer vertical integration: the source and measurement layer integrates a power amplifier chip, a frequency multiplier chip, and a frequency mixer chip; the directional coupling layer enables signal feeding and reference signal coupling; and the antenna layer uses a stepped horn antenna for radiation and reception. This chip-integrated approach to the terahertz sensor reduces its length by more than 20 times and its width by more than 10 times, facilitating array deployment and achieving ultimate miniaturization.

[0030] Under the excitation of the radio frequency signal, the terahertz signal of the corresponding frequency band is obtained by frequency doubling, amplification and frequency doubling, and then fed to the free space of the stepped horn antenna radiation value through the MEMS stacked coupler. The MEMS stacked coupler feeds the signal to the antenna and simultaneously couples a signal as a reference signal, which is received by the mixer and down-converted to obtain the reference intermediate frequency signal, which is received by the multi-port vector network analyzer. At the same time, the terahertz signal reflected by the test sample is received by the stepped horn antenna, coupled through the coupler, and down-converted to obtain the test signal, which is received by the multi-port vector network analyzer to obtain the test intermediate frequency signal.

[0031] Terahertz sensors employ an integrated transceiver design, enabling the simultaneous radiation and reception of reflected terahertz signals. For example, consider eight sensors arranged in a linear array. Figure 5As shown, the sensors are denoted as X1, X2, X3, X4, X5, X6, X7, and X8; the distance between the antenna apertures of adjacent sensors is 20 mm; the equivalent array elements obtained by X1 transmitting a terahertz signal and X1 and other sensors receiving the signal are denoted as Y1, Y2, Y3, Y4, Y5, Y6, Y7, and Y8, with a distance of 10 mm between the equivalent array elements; the corresponding S-parameters are S11, S12, S13, S14, S15, S16, S17, and S18, respectively. When all terahertz sensors operate alternately, the obtained equivalent array elements are as follows... Figure 6 As shown, taking an 8-sensor linear array as an example, 15 equivalent array elements can be obtained. Multiple S-parameters correspond to some of these equivalent array elements. Taking equivalent array element Y8 as an example, the obtained S-parameters include S18, S27, S36, S45, S54, S65, S76, and S81. By processing these S-parameters and selecting appropriate data, the imaging quality can be significantly improved. For a terahertz sensor array with four frequency bands, a total of 8*8*4 scattering parameters can be obtained. For a larger number of sensors, it is only necessary to appropriately increase the number of ports on the multi-port vector network analyzer, or to use multiple multi-port vector network analyzers connected in series with a common time base.

[0032] The robotic arm's function is to control the terahertz sensor array to perform two-dimensional movement. As can be seen from the above description, the equivalent array element spacing is 10mm, which is insufficient to meet the requirement of being less than half the wavelength. Using a center frequency of 295GHz in the 90GHz-500GHz range, the equivalent array element spacing should be less than 0.77mm. Since the hardware cannot meet these requirements, the robotic arm is used to control the terahertz sensor array according to... Figure 7 The trajectory shown has a step interval of 0.5 mm. This yields more encrypted data for subsequent processing.

[0033] The server serves three purposes: first, to control the robotic arm to move along a predetermined route; second, to acquire S-parameters from a multi-port vector network analyzer in real time; and third, to use the S-parameters for 3D image reconstruction and defect detection.

[0034] A multi-band integrated terahertz array imaging method, the process of which is as follows: Figure 8 As shown, it includes the following steps: Step 1, Test Preparation: 1. Power on all hardware devices in the system and allow them to preheat for 30 minutes; 2. Use a multi-port vector network analyzer to check that all terahertz sensor arrays are working properly. The standard for judging that they are working properly is to place a smooth metal plate in the calibration position and obtain S11, S22, S33, ..., SNN parameter values ​​that are all between 0dB and -10dB. Step 2, Rapid Full-Area Testing: Quickly locate potential defects in the test material. 1. Set the system operating frequency, signal power, robotic arm movement speed and other test parameters, and select a point frequency, narrowband signal or wideband signal in the 90GHz-500GHz frequency band, but with a sweep step of GHz or even larger. 2. Place a smooth metal plate in the calibration position, use a robotic arm to move the terahertz sensor array to the center of the metal plate, use a multi-port vector network analyzer to obtain the S-parameters under this fixed state, transmit them to the server for storage, and wait for subsequent processing. The resulting array matrix is ​​denoted as A. 3. Place the test sample in the test area. Use a robotic arm to control the terahertz sensor array to reach the starting point of the test and obtain the S-parameters at this position, denoted as B1. Following the planned motion path along the X-axis, move to the next position with a set step value of 0.5mm, obtaining the S-parameters denoted as B2. The maximum number of movements along the X-axis is 40, and these are successively denoted as B1, B2, ..., B40. Compare the above array matrices with A to obtain C1, C2, C3, ..., C40. Merge these array matrices into a single array, following the following principles: First, for multiple S-parameters corresponding to the same position, the S-parameter corresponding to the largest amplitude is taken as the final S-parameter at that position, and other S-parameters are no longer retained. Based on this, C1, C2, C3, ..., C40 can all be reduced to M*N array matrices D1, D2, ... D1, D2, D3, ..., D40, where N represents the number of sensors and M represents the number of frequency points; secondly, D1, D2, D3, ..., D40 are merged into an M*(40*N) array E1, where the first column of E1 is the first column of D1, the second column of E1 is the first column of D2, the third column of E1 is the first column of D3, and so on up to the 40th column; the 41st column of E1 is the second column of D1, the 42nd column of E1 is the second column of D2, the 43rd column of E1 is the second column of D3, and so on..., finally resulting in an M*(40*N) array E1; according to the planned movement route, the robotic arm controls the terahertz sensor to move to the next position on the y-axis, resulting in a new array E2, and so on, resulting in q arrays, namely E1, E2, ..., Eq, where the value of q is the set number of steps on the y-axis.

[0035] 4. Using the array described above, obtain two-dimensional or three-dimensional images of the test material using a linear array imaging algorithm. Determine whether defects exist based on the imaging results. If no defects exist, the test ends. If defects exist, determine the approximate area of ​​the defects and proceed with further testing. Step 3, Target Area Detailed Testing: Precisely analyze the depth, size, and type of defects within the material. 1. Set test parameters such as system operating frequency, signal power, and robotic arm movement speed, select a full-band test bandwidth of 90GHz-500GHz, and a frequency sweep step of less than 1MHz or even smaller; 2. Place a smooth metal plate in the calibration position, use a robotic arm to move the terahertz sensor array to the center of the metal plate, use a multi-port vector network analyzer to obtain the S-parameters under this fixed state, transmit them to the server for storage, and wait for subsequent processing. The resulting array matrix is ​​denoted as A. 3. Place the test sample in the test area. Use a robotic arm to control the terahertz sensor array to reach the starting point of the test and obtain the S-parameters at this position, denoted as B1. Following the planned motion path along the X-axis, move to the next position with a set step value of 0.5mm, obtaining the S-parameters denoted as B2. The maximum number of movements along the X-axis is 40, and these are successively denoted as B1, B2, ..., B40. Compare the above array matrices with A to obtain C1, C2, C3, ..., C40. Merge these array matrices into a single array, following the following principles: First, for multiple S-parameters corresponding to the same position, the S-parameter corresponding to the largest amplitude is taken as the final S-parameter at that position, and other S-parameters are no longer retained. Based on this, C1, C2, C3, ..., C40 can all be reduced to M*N array matrices D1, D2, ... D1, D2, D3, ..., D40, where N represents the number of sensors and M represents the number of frequency points; secondly, D1, D2, D3, ..., D40 are merged into an M*(40*N) array E1, where the first column of E1 is the first column of D1, the second column of E1 is the first column of D2, the third column of E1 is the first column of D3, and so on up to the 40th column; the 41st column of E1 is the second column of D1, the 42nd column of E1 is the second column of D2, the 43rd column of E1 is the second column of D3, and so on..., finally resulting in an M*(40*N) array E1; according to the planned movement route, the robotic arm controls the terahertz sensor to move to the next position on the y-axis, resulting in a new array E2, and so on, resulting in q arrays, namely E1, E2, ..., Eq, where the value of q is the set number of steps on the y-axis.

[0036] 4. Using the array described above, a linear array imaging algorithm is used to obtain a three-dimensional tomographic image of the test material, thereby determining the depth, size, and type of the defect. Step four, test complete.

[0037] Of course, the above description is not intended to limit the present invention, and the present invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A multi-band integrated terahertz array imaging system, characterized in that, include: Multiport vector network analyzer, terahertz sensor array, robotic arm, server, power module, and display; The power modules provide stable power to the multi-port vector network analyzer, terahertz sensor array, robotic arm, and server, respectively. The multiport vector network analyzer, connected via lines to the power module, terahertz sensor array, and server, operates in the radio frequency band and is configured to provide RF excitation and local oscillator signals to the terahertz sensor array and receive intermediate frequency signals from the terahertz sensor array to obtain its S-parameters. The terahertz sensor array, connected to the power module and robotic arm via wiring, includes multiple transceiver sensor units operating in the terahertz frequency band, covering a frequency range of 90GHz to 500GHz and divided into multiple sub-band arrays. Each sensor unit employs a multi-layer integrated structure, including a source-measurement layer, a directional coupling layer, and an antenna layer, to generate and radiate terahertz signals while receiving reflected signals from the sample under test. The robotic arm, connected to the power module and server via wiring, is configured to control the terahertz sensor array to perform scanning motion in a two-dimensional plane according to a preset trajectory and stepping accuracy. The server is connected to the multi-port vector network analyzer, the robotic arm, and the display via wires. It is configured to control the robotic arm to move along a predetermined route, acquire the S-parameters obtained by the multi-port vector network analyzer in real time, and use the S-parameters to perform 3D image reconstruction and defect detection. The monitor, connected to the power module and server via cables, is configured to display images generated by the server in real time.

2. The multi-band integrated terahertz array imaging system according to claim 1, characterized in that, The terahertz sensor array comprises at least four sub-band arrays, covering the frequency bands of 90GHz-140GHz, 140GHz-220GHz, 220GHz-330GHz and 330GHz-500GHz respectively; each sub-band array is arranged in multiple rows, and the sensor units within the array are distributed at equal intervals.

3. The multi-band integrated terahertz array imaging system according to claim 1, characterized in that, The source and measurement layers of a single terahertz sensor unit are fully integrated on a single chip, including a power amplifier chip, a frequency multiplier chip, and a frequency mixer chip.

4. The multi-band integrated terahertz array imaging system according to claim 1, characterized in that, The step interval for the robotic arm to control the terahertz sensor array to perform scanning motion is ≤0.5mm.

5. The multi-band integrated terahertz array imaging system according to claim 1, characterized in that, The server is configured to perform two detection modes: Full-area rapid detection mode: Uses wide-frequency stepping or point-frequency signals for rapid scanning to initially locate defect areas; Target area fine-grained detection mode: Uses full-band, narrow-sweep-step signals for fine scanning to obtain information on the depth, size, and type of defects.

6. The multi-band integrated terahertz array imaging system according to claim 1, characterized in that, The terahertz sensor array adopts a sparse array method, forming an equivalent dense array element by alternating the operation of sensor units for transmitting and receiving. When processing data, the server filters multiple S-parameters at the same equivalent array element location and selects the S-parameter with the largest amplitude for imaging to improve image quality.

7. A multi-band integrated terahertz array imaging method, employing the multi-band integrated terahertz array imaging system as described in any one of claims 1 to 6, characterized in that, Includes the following steps: Step 1: Test Preparation; specifically includes the following steps: Step 1.1: Power on all hardware devices in the system and preheat them for 30 minutes; Step 1.2: Use a multi-port vector network analyzer to check whether all terahertz sensor arrays are working properly. The standard for judging whether they are working properly is to place a smooth metal plate in the calibration position and obtain S11, S22, S33, ..., SNN parameter values ​​that are all between 0dB and -10dB. Step 2: Rapid full-area inspection to quickly locate potential defects in the test material; specifically including the following steps: Step 2.1: Set the first set of test parameters; the first set of test parameters has a wider test bandwidth and / or a larger sweep step; Step 2.2: Place the smooth metal plate in the calibration position, use a robotic arm to move the terahertz sensor array to the center of the metal plate, use a multi-port vector network analyzer to obtain the S-parameters in this fixed state, transmit them to the server for storage, and wait for subsequent processing. At this time, the array matrix is ​​obtained and denoted as A. Step 2.3: Place the test sample in the test area, control the robotic arm to drive the terahertz sensor array to the starting point of the test, perform a full-area scan of the test sample, and collect the first S-parameter dataset; Step 2.4: Perform calculations on the first S-parameter dataset and the reference matrix A to generate the first data matrix, and use the linear array imaging algorithm to obtain a two-dimensional or three-dimensional image of the test material. Determine whether there are defects based on the imaging results. If there are no defects, the test ends. If there are defects, determine the approximate area of ​​the defects and proceed with the test. Step 3: Detailed inspection of the target area; precise analysis of information within the material containing the defect, including depth, size, and type; specifically including the following steps: Step 3.1: If a defect is detected in step 2, set a second set of test parameters. The second set of test parameters has a wider test bandwidth and / or a smaller sweep step. Step 3.2: Place the smooth metal plate in the calibration position, use a robotic arm to move the terahertz sensor array to the center of the metal plate, use a multi-port vector network analyzer to obtain the S-parameters in this fixed state, transmit them to the server for storage, and wait for subsequent processing. The resulting array matrix is ​​denoted as A. Step 3.3: Place the test sample in the test area, control the robotic arm to perform a fine scan in the located defect area, and collect the second S-parameter dataset; Step 3.4: Process the second S-parameter dataset to generate the second data matrix, use the linear array imaging algorithm to obtain the three-dimensional tomographic image of the test material, and then determine the depth location, size and type of the defect; Step 4: Output the detection results and end the test.

8. The multi-band integrated terahertz array imaging method according to claim 7, characterized in that, The process of processing the S-parameter dataset and generating the data matrix in steps 2 and 3 includes: The S-parameter matrix B acquired at each scanning position is compared with the reference matrix A to obtain matrix C; When there are multiple S-parameters at the same physical location, the S-parameter with the largest amplitude is selected as the final representative value for that location, forming the dimensionality-reduced data matrix D. The data matrices D from multiple scanning positions are merged according to preset rules to form the final data matrix E used for imaging.

9. The multi-band integrated terahertz array imaging method according to claim 7, characterized in that, The first set of test parameters includes spot frequency, narrowband signal, or broadband signal with a sweep step ≥ 1 GHz; the second set of test parameters is a full-band sweep signal with a sweep step < 1 MHz.

10. The multi-band integrated terahertz array imaging method according to claim 7, characterized in that, By integrating multi-band signals and performing multiple relative measurements, the nonlinear error of the system is suppressed, and the distance resolution of three-dimensional tomography is improved by widening the overall test bandwidth.

Images