A forward-looking three-dimensional imaging method and system for non-scanning single-channel terahertz radar

Abstract

The present application relates to a method and system for forward-looking three-dimensional imaging of a scanless single-channel terahertz radar. The method comprises: obtaining a one-dimensional range image through a single-transmitting and single-receiving terahertz radar, collecting a three-dimensional depth image through a time-of-flight camera, taking a large number of one-dimensional range images as input and the corresponding three-dimensional depth images as output, training a neural network model to obtain a trained neural network model; obtaining a one-dimensional range image of a scene to be detected, inputting it into the trained neural network model, and obtaining a three-dimensional image of the scene to be detected according to the output of the trained neural network model. The present invention only uses a single-channel terahertz radar with a single transmit and single receive function, avoids the use of an antenna array and wavefront modulation, does not rely on aperture accumulation or relative motion to achieve scanless single-channel forward-looking three-dimensional imaging, greatly simplifies the system and reduces costs; proposes a forward-looking three-dimensional imaging algorithm based on deep learning, improves the imaging efficiency of the system, and can meet the requirements of high-resolution and high-frame-rate imaging.

Description

Technical Field

[0001] The present application relates to the technical field of terahertz radar three-dimensional imaging, and in particular to a scanning-free single-channel terahertz radar forward-looking three-dimensional imaging method and system. Background Art

[0002] The term "terahertz" officially appeared in 1970, referring to electromagnetic waves with frequencies between 0.1 and 10 THz. Terahertz waves are in the transition frequency band from electronics to optics. Compared with microwave radar, terahertz radar has a short wavelength, large bandwidth, and extremely high spatial resolution, so the imaging resolution is extremely high and can accurately depict the target features.

[0003] Classic radar imaging technology is usually based on the principles of tomography and range Doppler, and is realized through a virtual synthetic aperture formed by the relative motion of the target and the radar; in the field of forward-looking and staring imaging, such as aperture coded imaging, electromagnetic vortex imaging, etc., it is mainly through the modulation of the directional pattern and wavefront space of the electromagnetic wave to obtain a rich equivalent illumination mode, so that the target information in the echo is richer, and super-resolution imaging is performed on this basis. However, the above methods either require relative motion between the radar and the target, or require modulation of the electromagnetic wave or wavefront space to make the imaging system complex, and each has its own limitations. In addition, the existing radar forward-looking imaging technology uses more array systems, scanning systems, and wavefront space modulation systems. The radar imaging system of the array system and the wavefront space modulation system has a complex structure and is therefore more expensive; while the radar imaging system of the scanning system requires a long data accumulation process, and it is difficult to achieve high frame rate imaging. Therefore, the existing technology has the problem of poor adaptability. Summary of the invention

[0004] Based on this, it is necessary to provide a scanless single-channel terahertz radar forward-looking three-dimensional imaging method, system, computer equipment and storage medium that can simplify the radar imaging system and meet the requirements of high-resolution and high-frame rate imaging in response to the above-mentioned technical problems.

[0005] A non-scanning single-channel terahertz radar forward-looking three-dimensional imaging method, the method comprising:

[0006] The single-transmitting and single-receiving terahertz radar is controlled to transmit and receive signals through two synchronous clock signals to obtain a one-dimensional range image of the detection scene, and the time-of-flight camera is controlled to collect signals to obtain a three-dimensional depth image corresponding to the detection scene, and a training set consisting of multiple groups of one-dimensional range image and three-dimensional depth image pairs is obtained by further changing the posture and position of the target in the detection scene;

[0007] The one-dimensional range image in the training set is converted into a column vector and then input into a pre-designed neural network model, and the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image is used as an output to train the neural network model to obtain a trained neural network model;

[0008] A one-dimensional range image of the scene to be detected is obtained, converted into a column vector and input into the trained neural network model, and a three-dimensional image of the scene to be detected is obtained according to the column vector output by the trained neural network model.

[0009] In one of the embodiments, the method further includes: controlling a single-transmitting and single-receiving terahertz radar to collect signals in a trigger mode by using a first clock signal, so that the single-transmitting and single-receiving terahertz radar transmits and receives signals and saves data each time a falling edge is detected, thereby obtaining a one-dimensional range image of the detection scene;

[0010] The time-of-flight camera is controlled to collect signals in a trigger mode by a second clock signal, so that the time-of-flight camera collects and saves data each time a falling edge is detected, thereby obtaining a three-dimensional depth image corresponding to the detection scene; the three-dimensional depth image includes the horizontal and vertical coordinates and depth information of each point in the detection scene;

[0011] The first clock signal is synchronized with the second clock signal.

[0012] In one of the embodiments, it also includes: converting the one-dimensional range image in the training set into a column vector and inputting it into a pre-designed neural network model, using the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image as the output, training the neural network model to obtain a trained neural network model; the neural network is a neural network of a multi-layer perceptron; the neural network of the multi-layer perceptron includes an input layer, three fully connected layers and an output layer; the fully connected layer is used to connect each point of the input data with each point of the output data; each of the fully connected layers is followed by an activation function layer.

[0013] In one of the embodiments, the method further includes: converting the one-dimensional distance image in the training set into a column vector and inputting the column vector into a pre-designed neural network model;

[0014] Flattening the three-dimensional depth image corresponding to the one-dimensional range image in M×N dimensions in horizontal and vertical directions into a column vector with a length of MN;

[0015] The column vector with a length of MN is used as the output of the neural network model, and the neural network model is trained to obtain a trained neural network model.

[0016] In one of the embodiments, the method further includes: restoring the column vector with a length of MN output by the trained neural network model to M×N dimensions to obtain a three-dimensional image of the scene to be detected.

[0017] A non-scanning single-channel terahertz radar forward-looking three-dimensional imaging system, the system comprising:

[0018] A data collection module is used to control the single-transmitting and single-receiving terahertz radar to transmit and receive signals through two synchronous clock signals to obtain a one-dimensional range image of the detection scene, and to control the flight time camera to collect signals to obtain a three-dimensional depth image corresponding to the detection scene, and further obtain a training set consisting of multiple groups of one-dimensional range image and three-dimensional depth image pairs by changing the posture and position of the target in the detection scene;

[0019] A model training module, used for converting a one-dimensional range image in a training set into a column vector and inputting the converted column vector into a pre-designed neural network model, and using the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image as an output to train the neural network model and obtain a trained neural network model;

[0020] The model application module is used to obtain a one-dimensional range image of the scene to be detected, convert it into a column vector and input it into the trained neural network model, and obtain a three-dimensional image of the scene to be detected according to the column vector output by the trained neural network model.

[0021] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0022] The single-transmitting and single-receiving terahertz radar is controlled to transmit and receive signals through two synchronous clock signals to obtain a one-dimensional range image of the detection scene, and the time-of-flight camera is controlled to collect signals to obtain a three-dimensional depth image corresponding to the detection scene, and a training set consisting of multiple groups of one-dimensional range image and three-dimensional depth image pairs is obtained by further changing the posture and position of the target in the detection scene;

[0023] The one-dimensional range image in the training set is converted into a column vector and then input into a pre-designed neural network model, and the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image is used as an output to train the neural network model to obtain a trained neural network model;

[0024] A one-dimensional range image of the scene to be detected is obtained, converted into a column vector and input into the trained neural network model, and a three-dimensional image of the scene to be detected is obtained according to the column vector output by the trained neural network model.

[0025] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0026] The single-transmitting and single-receiving terahertz radar is controlled to transmit and receive signals through two synchronous clock signals to obtain a one-dimensional range image of the detection scene, and the time-of-flight camera is controlled to collect signals to obtain a three-dimensional depth image corresponding to the detection scene, and a training set consisting of multiple groups of one-dimensional range image and three-dimensional depth image pairs is obtained by further changing the posture and position of the target in the detection scene;

[0027] The one-dimensional range image in the training set is converted into a column vector and then input into a pre-designed neural network model, and the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image is used as an output to train the neural network model to obtain a trained neural network model;

[0028] A one-dimensional range image of the scene to be detected is obtained, converted into a column vector and input into the trained neural network model, and a three-dimensional image of the scene to be detected is obtained according to the column vector output by the trained neural network model.

[0029] The above-mentioned scanless single-channel terahertz radar forward-looking three-dimensional imaging method and system obtains a one-dimensional range image of the detection scene by transmitting and receiving signals of a single-transmitting and single-receiving terahertz radar, collects a three-dimensional depth image by a time-of-flight camera, and obtains a training set consisting of a pair of one-dimensional range images and three-dimensional depth image data; a large number of one-dimensional range images are used as input, and the corresponding three-dimensional depth images are used as output to train the neural network model to obtain a trained neural network model; the one-dimensional range image of the scene to be detected is obtained, input into the trained neural network model, and a three-dimensional image of the scene to be detected is obtained according to the output of the trained neural network model. The present invention only uses a single-channel terahertz radar with a single transmission and a single reception, avoids the use of an antenna array and wavefront modulation, does not rely on aperture accumulation or relative motion to achieve scanless single-channel forward-looking three-dimensional imaging, greatly simplifies the system and reduces costs; a forward-looking three-dimensional imaging algorithm based on deep learning is proposed, which improves the imaging efficiency of the system and can meet the requirements of high-resolution and high-frame rate imaging. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a diagram showing an application scenario of a non-scanning single-channel terahertz radar forward-looking three-dimensional imaging method in an embodiment;

[0031] Figure 2 It is a schematic diagram of the process of a non-scanning single-channel terahertz radar forward-looking three-dimensional imaging method in one embodiment;

[0032] Figure 3 A multi-layer perceptron model framework used in one embodiment;

[0033] Figure 4 It is a structural diagram of a non-scanning single-channel terahertz radar forward-looking three-dimensional imaging system in one embodiment;

[0034] Figure 5 It is a structural block diagram of a non-scanning single-channel terahertz radar forward-looking three-dimensional imaging system in one embodiment;

[0035] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0037] The non-scanning single-channel terahertz radar forward-looking three-dimensional imaging method provided in this application can be applied to Figure 1 In the application environment shown. Among them, under the condition of forward imaging, the terahertz radar system emits a linear frequency modulation pulse waveform and obtains a one-dimensional range image of the scene, and the time-of-flight camera is used to simultaneously obtain the depth information of the entire scene, that is, the real three-dimensional image of the scene. Through the control and processing terminal, two synchronous clock signals are separated to control the terahertz radar system and the time-of-flight camera to work synchronously, and a one-dimensional range image-depth image pair is obtained, the former is used as the input for training the neural network model, and the latter is used as the output. Among them, the control and processing terminal can be but not limited to various personal computers, laptops, tablet computers and portable wearable devices.

[0038] In one embodiment, Figure 2 As shown, a non-scanning single-channel terahertz radar forward-looking three-dimensional imaging method is provided, and the method is applied to Figure 1 The control and processing terminal in the example is used to illustrate, including the following steps:

[0039] Step 202, using two synchronized clock signals to respectively control the single-transmitting and single-receiving terahertz radar to transmit and receive signals to obtain a one-dimensional range image of the detection scene, and to control the flight time camera to collect signals to obtain a three-dimensional depth image corresponding to the detection scene, and further by changing the posture and position of the target in the detection scene, a training set consisting of multiple groups of one-dimensional range image and three-dimensional depth image pairs is obtained.

[0040] The present invention uses only a single-transmitting and single-receiving single-channel terahertz radar, which can avoid the use of an antenna array and wavefront modulation, and achieves non-scanning single-channel forward-looking three-dimensional imaging without relying on aperture accumulation or relative motion, thereby greatly simplifying the system and reducing costs.

[0041] The terahertz radar transmits a linear frequency modulated pulse signal to the target scene. The general expression of the signal is:

[0042]

[0043] where w(t) is the real envelope, is the signal modulation phase. For the standard LFM signal, we have:

[0044]

[0045] Where T p represents the pulse width of the signal, and K represents the signal modulation frequency.

[0046] Assuming the received echo delay is t0, the expression of the echo is:

[0047] s r (t) = s (t - t0) = rect ((t - t0) / T p )exp(jπK(t-t0) 2 ) (3)

[0048] The spectrum is in the form of:

[0049]

[0050] According to the set linear frequency modulation characteristics, the time domain and frequency domain responses of the matched filter can be obtained as follows:

[0051]

[0052] After matched filtering, the frequency domain expression of the output signal is:

[0053]

[0054] The Fourier transform of the rectangular window function is the sinc function. According to the properties of the Fourier transform, the time domain expression of the output signal after matched filtering can be obtained as follows:

[0055] s out (t) = IFFT[S out (f)]=T p sinc[πKT p (t-t0)] (7)

[0056] Observing the above formula, according to the properties of the sinc function, we can know that the output signal has a peak at t=t0. Multiplying t0 by c / 2 can obtain the position information of the detected target, that is, the one-dimensional range image.

[0057] A time-of-flight camera can measure the distance between itself and objects in the detection scene, thereby obtaining 3D information of the scene.

[0058] Step 204, convert the one-dimensional range image in the training set into a column vector and input it into a pre-designed neural network model, use the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image as the output, train the neural network model, and obtain a trained neural network model.

[0059] The back propagation algorithm is usually used to train the neural network model. By adjusting the parameters, the loss function is reduced to obtain the optimal network model.

[0060] Step 206, obtain a one-dimensional range image of the scene to be detected, convert it into a column vector and input it into the trained neural network model, and obtain a three-dimensional image of the scene to be detected according to the column vector output by the trained neural network model.

[0061] Since the radar imaging system of the array system and the wavefront spatial modulation system has a complex structure and thus a high cost, in view of this shortcoming, the present invention uses a single-transmitting and single-receiving radar system without wavefront modulation, which can greatly simplify the system and reduce costs; while the radar imaging system of the scanning system requires a long data accumulation process and is difficult to achieve high frame rate imaging. In view of this shortcoming, the present invention proposes a forward-looking three-dimensional imaging algorithm based on deep learning. After the neural network model training phase is completed, the single-channel one-dimensional range image echo is directly input into the trained network model to invert the three-dimensional image to achieve real-time imaging. The present invention realizes a non-scanning single-channel terahertz radar forward-looking three-dimensional imaging system that is simple and small in system, low in cost, and can meet the needs of high-resolution and high-frame rate imaging.

[0062] In the above-mentioned scanless single-channel terahertz radar forward-looking three-dimensional imaging method, a one-dimensional range image of the detection scene is obtained by transmitting and receiving signals of a single-transmitting and single-receiving terahertz radar, and a three-dimensional depth image is collected by a time-of-flight camera to obtain a training set consisting of a pair of one-dimensional range images and three-dimensional depth image data; a large number of one-dimensional range images are used as input, and the corresponding three-dimensional depth images are used as output to train the neural network model to obtain a trained neural network model; a one-dimensional range image of the scene to be detected is obtained, input into the trained neural network model, and a three-dimensional image of the scene to be detected is obtained according to the output of the trained neural network model. The present invention only uses a single-channel terahertz radar with a single transmission and a single reception, avoids the use of an antenna array and wavefront modulation, does not rely on aperture accumulation or relative motion to achieve scanless single-channel forward-looking three-dimensional imaging, greatly simplifies the system and reduces costs; a forward-looking three-dimensional imaging algorithm based on deep learning is proposed, which improves the imaging efficiency of the system and can meet the requirements of high-resolution and high-frame rate imaging.

[0063] In one of the embodiments, it also includes: controlling a single-transmitting and single-receiving terahertz radar to collect signals in a trigger mode through a first clock signal, so that the single-transmitting and single-receiving terahertz radar receives and sends signals and saves data every time a falling edge is detected, thereby obtaining a one-dimensional distance image of the detection scene; controlling a time-of-flight camera to collect signals in a trigger mode through a second clock signal, so that the time-of-flight camera collects and saves data every time a falling edge is detected, thereby obtaining a three-dimensional depth image corresponding to the detection scene; the three-dimensional depth image includes the horizontal and vertical coordinates and depth information of each point in the detection scene; and the first clock signal is synchronized with the second clock signal.

[0064] In one of the embodiments, it also includes: converting the one-dimensional range image in the training set into a column vector and then inputting it into a pre-designed neural network model, using the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image as the output, training the neural network model to obtain a trained neural network model; the neural network is a neural network of a multi-layer perceptron; the neural network of the multi-layer perceptron includes an input layer, three fully connected layers and an output layer; the fully connected layer is used to connect each point of the input data with each point of the output data, and the number of nodes in the three fully connected layers is 1024, 512, and 256 respectively; each fully connected layer is followed by an activation function layer, and the activation function used is a tanh function.

[0065] Specifically, the multi-layer perceptron model framework used in the present invention is as follows: Figure 3 As shown. Multilayer Perceptron (MLP) is an artificial neural network (ANN) with a forward structure that maps a set of input vectors into a set of output vectors. It can be abstracted into a directed graph consisting of multiple hierarchical node layers, and each layer of nodes is fully connected to the adjacent nodes of the next layer. In MLP, except for the input layer, each node in the remaining layers has an activation function, which is usually trained using the back propagation algorithm. By analyzing its structure, it can be seen that it can connect each point of the input data with each point of the output data. Each distance unit of the one-dimensional range image obtained by the single-transmitter and single-receiver terahertz radar contains information from all imaging grids of the three-dimensional scene. It can be seen that the two have certain similarities in structure.

[0066] In one of the embodiments, it also includes: converting the one-dimensional range image in the training set into a column vector and inputting it into a pre-designed neural network model; flattening the three-dimensional depth image corresponding to the one-dimensional range image in the horizontal and vertical directions into a column vector with a length of MN; using the column vector with a length of MN as the output of the neural network model, training the neural network model to obtain a trained neural network model.

[0067] Specifically, the data collected by the single-transmitter single-receiver terahertz radar and the time-of-flight camera are preprocessed to obtain their main information. Then each one-dimensional range image is stored in the form of a column vector, and each depth image (dimension is M×N) is also flattened into a column vector (dimension is 1×MN), and the two are matched one by one, and each pair of data is a group. 90% of all data groups are randomly selected as training sets, and the remaining 10% are used as test sets. The one-dimensional range image in each group of data is used as the input for training the neural network, and the column vector expanded from the depth image is used as the output. By adjusting the parameters, the loss function is reduced to obtain the optimal network model.

[0068] In one of the embodiments, the method further includes: restoring the column vector with a length of MN output by the trained neural network model to M×N dimensions to obtain a three-dimensional image of the scene to be detected.

[0069] It should be understood that although Figure 2 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0070] In one embodiment, Figure 5 As shown, a non-scanning single-channel terahertz radar forward-looking three-dimensional imaging system is provided, including: a data collection module 502, a model training module 504 and a model application module 506, wherein:

[0071] The data collection module 502 is used to control the single-transmitting and single-receiving terahertz radar to transmit and receive signals through two synchronous clock signals to obtain a one-dimensional range image of the detection scene, and to control the flight time camera to collect signals to obtain a three-dimensional depth image corresponding to the detection scene, and further obtain a training set consisting of multiple groups of one-dimensional range image and three-dimensional depth image pairs by changing the posture and position of the target in the detection scene;

[0072] A model training module 504 is used to convert the one-dimensional range image in the training set into a column vector and input it into a pre-designed neural network model, and use the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image as the output to train the neural network model to obtain a trained neural network model;

[0073] The model application module 506 is used to obtain a one-dimensional range image of the scene to be detected, convert it into a column vector and input it into the trained neural network model, and obtain a three-dimensional image of the scene to be detected according to the column vector output by the trained neural network model.

[0074] The data collection module 502 is also used to control the single-transmit single-receive terahertz radar to collect signals in a trigger mode through a first clock signal, so that the single-transmit single-receive terahertz radar receives and sends signals and saves data every time a falling edge is detected, thereby obtaining a one-dimensional distance image of the detection scene; control the time-of-flight camera to collect signals in a trigger mode through a second clock signal, so that the time-of-flight camera collects and saves data every time a falling edge is detected, thereby obtaining a three-dimensional depth image corresponding to the detection scene; the three-dimensional depth image includes the horizontal and vertical coordinates and depth information of each point in the detection scene; the first clock signal is synchronized with the second clock signal.

[0075] The model training module 504 is also used to convert the one-dimensional range image in the training set into a column vector and input it into a pre-designed neural network model, and use the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image as the output to train the neural network model to obtain a trained neural network model; the neural network is a neural network of a multi-layer perceptron; the neural network of the multi-layer perceptron includes an input layer, three fully connected layers and an output layer; the fully connected layer is used to connect each point of the input data with each point of the output data; each fully connected layer is followed by an activation function layer.

[0076] The model training module 504 is also used to convert the one-dimensional range image in the training set into a column vector and input it into a pre-designed neural network model; flatten the three-dimensional depth image corresponding to the one-dimensional range image in the horizontal and vertical directions into a column vector of length MN; use the column vector of length MN as the output of the neural network model, train the neural network model, and obtain a trained neural network model.

[0077] The model training module 504 is also used to train the neural network model through a back propagation algorithm to obtain a trained neural network model.

[0078] The model application module 506 is also used to restore the column vector with a length of MN output by the trained neural network model to M×N dimensions to obtain a three-dimensional image of the scene to be detected.

[0079] For the specific definition of the scanless single-channel terahertz radar forward-looking three-dimensional imaging system, please refer to the definition of the scanless single-channel terahertz radar forward-looking three-dimensional imaging method above, which will not be repeated here. Each module in the above-mentioned scanless single-channel terahertz radar forward-looking three-dimensional imaging system can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0080] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a non-scanning single-channel terahertz radar forward-looking three-dimensional imaging method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0081] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0082] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the above method embodiment when executing the computer program.

[0083] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiment are implemented.

[0084] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0085] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0086] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A non-scanning single-channel terahertz radar forward-looking three-dimensional imaging method, characterized in that: The method comprises: The single-transmitting and single-receiving terahertz radar is controlled to transmit and receive signals through two synchronous clock signals to obtain a one-dimensional range image of the detection scene, and the time-of-flight camera is controlled to collect signals to obtain a three-dimensional depth image corresponding to the detection scene, and a training set consisting of multiple groups of one-dimensional range image and three-dimensional depth image pairs is obtained by further changing the posture and position of the target in the detection scene; The one-dimensional range image in the training set is converted into a column vector and then input into a pre-designed neural network model, and the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image is used as an output to train the neural network model to obtain a trained neural network model; Obtaining a one-dimensional range image of the scene to be detected, converting it into a column vector and inputting it into the trained neural network model, and obtaining a three-dimensional image of the scene to be detected according to the column vector output by the trained neural network model; The method of controlling the single-transmitting and single-receiving terahertz radar to transmit and receive signals to obtain a one-dimensional range image of the detection scene and controlling the time-of-flight camera to collect signals to obtain a three-dimensional depth image corresponding to the detection scene by respectively controlling the single-transmitting and single-receiving terahertz radar to transmit and receive signals by two synchronous clock signals, includes: controlling the single-transmitting and single-receiving terahertz radar to collect signals in a trigger mode by a first clock signal, so that the single-transmitting and single-receiving terahertz radar transmits and receives signals and saves data every time a falling edge is detected, so as to obtain a one-dimensional range image of the detection scene; controlling the time-of-flight camera to collect signals in a trigger mode by a second clock signal, so that the time-of-flight camera collects and saves data every time a falling edge is detected, so as to obtain a three-dimensional depth image corresponding to the detection scene; the three-dimensional depth image includes the horizontal and vertical coordinates and depth information of each point in the detection scene; the first clock signal is synchronized with the second clock signal; The one-dimensional range image in the training set is converted into a column vector and then input into a pre-designed neural network model, and the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image is used as an output, and the neural network model is trained to obtain a trained neural network model, including: converting the one-dimensional range image in the training set into a column vector and then input into a pre-designed neural network model, and the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image is used as an output, and the neural network model is trained to obtain a trained neural network model; the neural network is a neural network of a multi-layer perceptron; the neural network of the multi-layer perceptron includes an input layer, three fully connected layers and an output layer; the fully connected layer is used to connect each point of the input data with each point of the output data; each of the fully connected layers is followed by an activation function layer.

2. The method according to claim 1, characterized in that The one-dimensional range image in the training set is converted into a column vector and then input into a pre-designed neural network model. The column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image is used as an output, and the neural network model is trained to obtain a trained neural network model, including: The one-dimensional distance image in the training set is converted into a column vector and then input into the pre-designed neural network model; The three-dimensional depth image corresponding to the one-dimensional range image is The length of the flattened Column vector of ; The length is The column vector of is used as the output of the neural network model, and the neural network model is trained to obtain a trained neural network model.

3. The method according to claim 2, characterized in that Obtaining a three-dimensional image of the scene to be detected according to the column vector output by the trained neural network model, including: The length of the trained neural network model output is The column vector of dimensional, and obtain a three-dimensional image of the scene to be detected.

4. A non-scanning single-channel terahertz radar forward-looking three-dimensional imaging system, characterized in that: The system comprises: A data collection module is used to control the single-transmitting and single-receiving terahertz radar to transmit and receive signals through two synchronous clock signals to obtain a one-dimensional range image of the detection scene, and to control the flight time camera to collect signals to obtain a three-dimensional depth image corresponding to the detection scene, and further obtain a training set consisting of multiple groups of one-dimensional range image and three-dimensional depth image pairs by changing the posture and position of the target in the detection scene; A model training module, used for converting a one-dimensional range image in a training set into a column vector and inputting the converted column vector into a pre-designed neural network model, and using the column vector obtained by flattening the three-dimensional depth image corresponding to the one-dimensional range image as an output to train the neural network model and obtain a trained neural network model; A model application module, used to obtain a one-dimensional range image of the scene to be detected, convert it into a column vector and input it into the trained neural network model, and obtain a three-dimensional image of the scene to be detected according to the column vector output by the trained neural network model; The data collection module is also used to: control the single-transmit single-receive terahertz radar to collect signals in a trigger mode through a first clock signal, so that the single-transmit single-receive terahertz radar receives and sends signals and saves data every time a falling edge is detected, thereby obtaining a one-dimensional distance image of the detection scene; control the time-of-flight camera to collect signals in a trigger mode through a second clock signal, so that the time-of-flight camera collects and saves data every time a falling edge is detected, thereby obtaining a three-dimensional depth image corresponding to the detection scene; the three-dimensional depth image includes the horizontal and vertical coordinates and depth information of each point in the detection scene; the first clock signal is synchronized with the second clock signal; The model training module is also used to: convert the one-dimensional range image in the training set into a column vector and then input it into a pre-designed neural network model; convert the three-dimensional depth image corresponding to the one-dimensional range image in the horizontal and vertical directions The length of the flattened Column vector of The column vector of is used as the output of the neural network model, and the neural network model is trained to obtain a trained neural network model.

Images