Non-visual imaging system based on distributed multi-channel parallel detection

Abstract

This invention relates to the field of photoelectric detection technology and discloses a non-line-of-sight imaging system based on distributed multi-channel parallel detection. A pulsed laser emits a laser beam, which is then split to obtain a transmitted light portion. This transmitted light portion is irradiated onto the surface of a relay wall by a scanning galvanometer. After diffuse reflection from the relay wall surface, it irradiates the surface of a hidden object. The light is then reflected back to the relay wall surface and collected by a single-photon detector array via a scanning galvanometer and a distributed fiber optic detection array. A processing unit calculates the required imaging information based on the detection signals. The distributed fiber optic detection array includes a central detection channel and multiple edge detection channels distributed outside the sidewalls of the central detection channel. Each edge detection channel contains an optical fiber connected to the single-photon detector array. The distance between the edge detection point corresponding to the edge detection channel and the irradiation point on the relay wall surface is a predetermined value to improve the signal-to-noise ratio of the detection signal for the corresponding non-line-of-sight three-echo beam.

Description

Technical Field

[0001] This invention relates to the field of photoelectric detection technology, specifically to a non-line-of-sight imaging system based on distributed multi-channel parallel detection. Background Technology

[0002] In modern warfare, disaster relief, and other events, the obstruction of streets and buildings makes it difficult to effectively observe suspicious targets and activity areas using direct imaging methods. There is an urgent need for methods that can bypass obstructing objects to track and observe targets behind them. Non-line-of-sight imaging methods based on photon time-of-flight can reconstruct the surface morphology of hidden objects by utilizing the diffuse reflection of relay surfaces and the time of flight of photons between the hidden object and the relay surface. This has significant application prospects in many fields such as autonomous driving, disaster relief, and medical diagnosis. However, traditional confocal non-line-of-sight imaging systems are limited by the dead time and afterpulse effect of single-photon detection caused by direct reflection light, as well as the stacking effect of time-correlated single-photon counts. This results in limited signal-to-noise ratio and temporal accuracy of the detected signal, thus restricting the reconstruction accuracy of hidden objects in the non-line-of-sight range.

[0003] The afterpulse effect refers to the phenomenon where, when a single-photon avalanche diode detector experiences an avalanche, some charge carriers remain in the multiplication layer. When these retained charge carriers are subsequently released, they also trigger an avalanche, generating a non-photon detection pulse. This pulse is called the afterpulse. The afterpulse can cause false counting, and an increase in the count rate will lead to an increase in afterpulse noise.

[0004] The stacking effect, also known as the photon stacking effect, occurs when the incident flux of photons increases, the ability to detect newly arriving photons decreases, and the overall detection efficiency of the system decreases. The shorter the recovery time, the more photons can be detected. Increasing the count rate will make the stacking effect more pronounced. Summary of the Invention

[0005] The purpose of this invention is to overcome the problem of limited reconstruction accuracy in existing reconstruction technologies and to provide a non-line-of-sight imaging system based on distributed multi-channel parallel detection.

[0006] To achieve the above objectives, the present invention provides a non-line-of-sight imaging system based on distributed multi-channel parallel detection, comprising: a pulsed laser, a laser linear polarization beam splitter, a scanning galvanometer, a distributed fiber optic detection array, a single-photon detector array, a relay wall, and a processing unit.

[0007] The laser emitted by the pulsed laser is incident on a laser linear polarization beam splitter to obtain a transmitted light portion. The transmitted light portion is then illuminated by a scanning mirror on the surface of the relay wall to form a two-dimensional spatial orthogonal scanning illumination. The surface of the relay wall undergoes diffuse reflection based on the two-dimensional spatial orthogonal scanning illumination and is transmitted to the surface of a hidden object outside the field of view. After being reflected again from the surface of the hidden object back to the surface of the relay wall, the non-field-of-view triple echo beam, after being diffusely reflected again on the surface of the relay wall, is collected by the single-photon detector array through the scanning mirror and the distributed fiber optic detection array. The processing unit calculates the required imaging information based on the detection signals of the corresponding non-field-of-view triple echo beams collected by the distributed fiber optic detection array.

[0008] The distributed optical fiber detection array includes a central detection channel and multiple edge detection channels distributed outside the sidewall of the central detection channel. The multiple edge detection channels, except for the central detection channel, contain optical fibers and are connected to the single-photon detector array via optical fibers. The non-line-of-sight triple echo beam is collected by the single-photon detector array through the multiple edge detection channels. The position of the central detection point on the relay wall surface corresponding to the central detection channel corresponds to the position of the illumination point on the relay wall surface corresponding to the non-line-of-sight triple echo beam. The distance between the edge detection point on the relay wall surface corresponding to the edge detection channel and the illumination point on the relay wall surface corresponding to the non-line-of-sight triple echo beam is a predetermined value, thereby improving the signal-to-noise ratio of the detection signal of the corresponding non-line-of-sight triple echo beam.

[0009] As one possible implementation, it also includes a single-mode fiber, a laser collimator, and a first lens. The laser emitted by the pulsed laser is transmitted through the single-mode fiber, collimated by the laser collimator, and then incident on a laser linear polarization beam splitter for beam splitting. The non-field-of-view triple echo beam passes through the scanning galvanometer and is then focused by the first lens and input to the distributed fiber optic detection array.

[0010] As one possible implementation, a light stopper is also included. After the laser beam is split by the laser linear polarization beam splitter, a reflected light portion is obtained, which is collected by the light stopper.

[0011] As one possible implementation, the pulsed laser is a pulsed laser with picosecond resolution, and the single-photon detector array is a detector array with picosecond resolution.

[0012] As one possible implementation, the distance between the edge detection point on the relay wall surface corresponding to the edge detection channel and the illumination point on the relay wall surface corresponding to the non-viewpoint triple echo beam is adjusted to a predetermined value by making the ratio of the intensity value of the direct reflected light generated by the two-dimensional spatial orthogonal scanning illumination on the relay wall surface to the intensity value of the non-viewpoint triple echo beam 20:1 to 25:1.

[0013] As one possible implementation, the inscribed circle diameter of the intermediate detection channel and the edge detection channel is equal to the diameter of the optical fiber. By selecting optical fibers of different diameters, the distance between the central axis of the edge detection channel and the intermediate detection channel is adjusted, thereby adjusting the distance between the edge detection point of the edge detection channel on the relay wall surface and the illumination point of the non-field-of-view third echo beam on the relay wall surface to a predetermined value.

[0014] As one possible implementation, both the edge detection channel and the intermediate detection channel are regular hexagonal prism structures and have optical fibers inside. There are 6 edge detection channels, which are distributed on the six sides of the intermediate detection channel.

[0015] As one possible implementation, the processing unit includes a counting module, which analyzes and processes the detection signals of the corresponding non-line-of-sight three-echo beams collected by a single-photon detector array with picosecond resolution to obtain picosecond resolution photon count-time-of-flight data; wherein, the time of flight is the round-trip flight time of the photon between the relay wall and the surface of the hidden object.

[0016] As one possible implementation, the processing unit further includes a reconstruction module, which reconstructs and restores the reflectivity information of the surface of the hidden object based on the Wiener filter deconvolution non-view reconstruction algorithm of view cone transformation, the obtained photon number-time of flight data with picosecond resolution, and other relevant parameters, thereby realizing the non-view reconstruction of the hidden object.

[0017] As one possible implementation method, the process of reconstructing and restoring the reflectivity information of the surface of the hidden object using a Wiener filter deconvolution non-view reconstruction algorithm based on view cone transform, along with the obtained photon number-time-of-flight data and other relevant parameters, specifically includes:

[0018] Establish a standard forward model for the confocal case; the formula for the standard forward model is:

[0019]

[0020] In the formula, τ represents the photon count-time of flight data at each scanning point on the repeater surface, (x′,y′) is the scanning point position on the repeater wall surface, (x,y,z) represents the spatial position point on the surface of the hidden object, r is the distance between the spatial position point on the surface of the hidden object and the corresponding scanning point on the repeater wall surface, ρ represents the reflectivity of the hidden object surface, δ represents the Dirac function, Ω represents the three-dimensional space where the hidden object is located, t represents the time of flight, and c represents the speed of light;

[0021] The standard forward model is converted into a 3D convolutional form using a view cone transformation, resulting in a 3D convolutional forward model. The view cone transformation formula is as follows:

[0022] v = (tc / 2) 2 In the formula, Let z represent z in the new coordinate system, and from this, we can deduce that v represents (tc / 2) in the new coordinate system. 2 ;

[0023] The formula for obtaining the forward model in the form of 3D convolution is:

[0024] R t {τ}(x′,y′,v)=∫∫∫ Ω R z {ρ}(x,y,u)h(x′-x,y′-y,vu)dxdydu; where, R t {τ} represents the photon number-time-of-flight data in the form of a three-dimensional convolution, R z {ρ} represents the reflectivity of the hidden object's surface in the form of a 3D convolution;

[0025] get: A non-view reconstruction formula is established based on Wiener filtering and 3D deconvolution, and R... t {τ}、R z The reflectivity information of the hidden object's surface is reconstructed and restored by substituting {ρ} and other relevant parameters into the non-view reconstruction formula; wherein, the non-view reconstruction formula is:

[0026] In the formula, F represents the three-dimensional Fourier transform, and ρ represents the hidden surface reflectivity information of the object. This represents the focusing matrix containing the Fourier coefficients of the three-dimensional convolution kernel, and α represents the signal-to-noise ratio.

[0027] The beneficial effects of this invention are as follows: This invention discloses a non-line-of-sight imaging system based on distributed multi-channel parallel detection. After the pulsed laser emits laser light, it is split to obtain a transmitted light portion. The transmitted light portion is diffusely reflected by a scanning galvanometer onto the surface of a relay wall and then onto the surface of a hidden object. After being reflected back to the surface of the relay wall, it is collected by a single-photon detector array through a scanning galvanometer and a distributed fiber optic detection array. The processing unit calculates the required imaging information based on the detection signal. The distributed fiber optic detection array includes a central detection channel and multiple edge detection channels distributed outside the sidewall of the central detection channel. The multiple edge detection channels are equipped with optical fibers and are connected to the single-photon detector array through optical fibers. The distance between the edge detection point corresponding to the edge detection channel and the illumination point on the surface of the relay wall is a predetermined value, which improves the signal-to-noise ratio of the detection signal of the corresponding non-line-of-sight three-echo beam.

[0028] This invention replaces traditional single-channel detection with distributed multi-channel detection, thereby reducing the photon count rate on a single detection channel and suppressing the stacking effect of time-dependent single-photon counts, thus improving detection time accuracy while ensuring overall detection efficiency.

[0029] This invention improves detection performance by adjusting the distance between the edge detection point corresponding to the edge detection channel and the illumination point on the relay wall surface to a predetermined value, thereby increasing the signal-to-noise ratio of the detection signal of the corresponding non-field-of-view third echo beam and suppressing the single-photon detection dead time and after-pulse effect caused by direct reflected light. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of a non-line-of-sight imaging system with distributed multi-channel parallel detection according to an embodiment of the present invention;

[0031] Figure 2 This is a schematic diagram of a distributed fiber optic detection array in a non-line-of-sight imaging system with distributed multi-channel parallel detection according to an embodiment of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] See Figure 1 This embodiment provides a technical solution: a non-line-of-sight imaging system based on distributed multi-channel parallel detection, including: a pulsed laser, a laser linear polarization beam splitter, a scanning galvanometer, a distributed fiber optic detection array, a single-photon detector array, a relay wall, and a processing unit;

[0034] The laser emitted by the pulsed laser is incident on a laser linear polarization beam splitter to obtain a transmitted light portion. The transmitted light portion is then illuminated by a scanning mirror on the surface of the relay wall to form a two-dimensional spatial orthogonal scanning illumination. The surface of the relay wall undergoes diffuse reflection based on the two-dimensional spatial orthogonal scanning illumination and is transmitted to the surface of a hidden object outside the field of view. After being reflected again from the surface of the hidden object back to the surface of the relay wall, the non-field-of-view triple echo beam, after being diffusely reflected again on the surface of the relay wall, is collected by the single-photon detector array through the scanning mirror and the distributed fiber optic detection array. The processing unit calculates the required imaging information based on the detection signals of the corresponding non-field-of-view triple echo beams collected by the distributed fiber optic detection array.

[0035] The distributed optical fiber detection array includes a central detection channel and multiple edge detection channels distributed outside the sidewall of the central detection channel. The multiple edge detection channels, except for the central detection channel, contain optical fibers and are connected to the single-photon detector array via optical fibers. The non-line-of-sight triple echo beam is collected by the single-photon detector array through the multiple edge detection channels. The position of the central detection point on the relay wall surface corresponding to the central detection channel corresponds to the position of the illumination point on the relay wall surface corresponding to the non-line-of-sight triple echo beam. The distance between the edge detection point on the relay wall surface corresponding to the edge detection channel and the illumination point on the relay wall surface corresponding to the non-line-of-sight triple echo beam is a predetermined value, thereby improving the signal-to-noise ratio of the detection signal of the corresponding non-line-of-sight triple echo beam.

[0036] Furthermore, it also includes a single-mode fiber, a laser collimator, and a first lens. The laser emitted by the pulsed laser is transmitted through the single-mode fiber, collimated by the laser collimator, and then incident on the laser linear polarization beam splitter for beam splitting. The non-field-of-view triple echo beam passes through the scanning galvanometer and is then focused by the first lens and input to the distributed fiber optic detection array.

[0037] Furthermore, it also includes a light stopper. After the laser beam is split by the laser linear polarization beam splitter, a reflected light portion is obtained, which is collected by the light stopper.

[0038] The pulsed laser is a pulsed laser with high temporal resolution, and the single-photon detector array is a detector array with high temporal resolution, specifically, it can be at the picosecond level. That is, in this embodiment, the time-of-flight data of photons required for imaging is recorded by using a pulsed laser with high temporal resolution. The high temporal resolution of the single-photon detector array is obtained by synchronizing the pulsed laser with a high temporal resolution synchronization signal. The number of detectors in the single-photon detector array in this embodiment corresponds to the number of edge detection channels.

[0039] In this embodiment, the scanning galvanometer is controlled by a galvanometer control module. The galvanometer control module inputs different voltage signals to the scanning galvanometer to adjust the position of the transmitted light portion illuminating the surface of the relay wall. This allows the reflected light from the relay wall surface illumination to illuminate the surface of the hidden object, providing a data basis for the reconstruction of the hidden object.

[0040] This embodiment does not explicitly describe the predetermined value. However, as an example, the ratio of the intensity of the directly reflected light generated by the two-dimensional spatial orthogonal scanning illumination on the relay wall surface to the intensity of the non-viewpoint triple echo beam can be set to 20:1 to 25:1. This adjusts the distance between the edge detection point on the relay wall surface corresponding to the edge detection channel and the illumination point on the relay wall surface corresponding to the non-viewpoint triple echo beam to a predetermined value. The ratio of the intensity of the directly reflected light to the intensity of the non-viewpoint triple echo beam can be measured using existing software for measuring intensity ratios during the distance adjustment process, which will not be described in detail in this embodiment. This suppresses the single-photon detection dead time and after-pulse effect caused by the directly reflected light, thereby improving the signal-to-noise ratio of the detection signal corresponding to the non-viewpoint triple echo beam and thus improving detection performance. However, other suitable intensity ratios can also be obtained experimentally.

[0041] Finally, the multi-channel detection signals corresponding to multiple edge detection channels collected by the single-photon detector array are averaged to obtain the photon flight time data within the overall detection range. The three-dimensional reflectivity distribution information of the hidden object is then obtained by the non-line-of-sight reconstruction algorithm.

[0042] In this embodiment, the inscribed circle diameters of the intermediate detection channel and the edge detection channel are equal to the diameter of the optical fiber. Therefore, by selecting optical fibers of different diameters within the edge detection channel, the distance between the central axis of the edge detection channel and the intermediate detection channel can be adjusted, thereby adjusting the distance between the edge detection point on the relay wall surface corresponding to the edge detection channel and the illumination point on the relay wall surface corresponding to the non-field-of-view third echo beam to a predetermined value.

[0043] like Figure 1 As shown, this embodiment provides a system schematic diagram, including a pulsed laser 1, a single-mode fiber 2, a laser collimator 3, a laser linear polarization beam splitter 4, a light stopper 5, a scanning galvanometer 6, a first lens 7, a pinhole 8, a distributed fiber optic detection array 9, a single-photon detector array 10, a processing unit 11, a relay wall 12, and a hidden object 13; the scanning galvanometer 6 is connected to a linear power supply 14 and a galvanometer control module 15.

[0044] When the system is working, the laser emitted by the high time resolution pulsed laser 1 is collimated and emitted after passing through the single-mode fiber 2 and the laser collimator 3, and then incident on the laser linear polarization beam splitter 4 for beam splitting. The reflected light part of the laser after beam splitting is collected by the light stopper 5, and the transmitted light part is scanned and illuminated in two-dimensional space orthogonally on the surface of the repeater wall by the scanning galvanometer 6. The galvanometer control module 15 controls the spatial position of the illumination point on the surface of the repeater wall 12 by controlling the input voltage of the scanning galvanometer 6, and generates a clock signal of the dot-line frame to synchronize with the signal of the processing unit 11, so as to realize the correspondence between the scanning point position and the detection signal timing.

[0045] The scanning beam is diffusely reflected on the surface of the relay wall 12 and then transmitted to the surface of the hidden object 13 outside the field of view. After being reflected again on the surface of the hidden object 13, it returns to the surface of the relay wall 12. Subsequently, after being diffusely reflected again on the surface of the relay wall 12, some photons, namely the non-field-of-view triple echo beam, are reflected by the scanning galvanometer 6 to the laser linear polarization beam splitter 4. The beam is then focused onto the pinhole 8 by the lens 7 and received by the distributed fiber optic detection array 9 and the high temporal resolution single-photon detector array 10, thus obtaining high temporal resolution imaging information, namely the round-trip flight time of the photons between the relay wall 12 and the surface of the hidden object 13, as well as the number of photons. The non-field-of-view triple echo beam passing through the laser linear polarization beam splitter 4 is only one implementation method to make the system more compact. In other embodiments, the beam splitter may not be used. Then, combined with the non-field-of-view reconstruction algorithm, the surface reflectivity information of the hidden object 13 is reconstructed and restored.

[0046] As an example, the distributed optical fiber detection array 9 in this embodiment can be adopted... Figure 2 As shown in the distribution, both the edge detection channel 10 and the intermediate detection channel 20 are regular hexagonal prism structures with internal optical fibers 11. There are six edge detection channels, which are distributed on the six sides of the intermediate detection channel. Except for the intermediate detection channel, all edge detection channels are connected to the single-photon detector array 10 via optical fibers. The core diameter of the optical fiber is equal to the inscribed circle of the edge detection channel. Since the distance between the central axis of the edge detection channel and the intermediate detection channel is positively correlated with the distance between the edge detection point corresponding to the edge detection channel and the intermediate detection point corresponding to the intermediate detection channel, the distance between the central axis of the edge detection channel and the intermediate detection channel can be adjusted by selecting optical fibers with different core diameters. This, in turn, changes the intensity ratio of directly reflected light and non-line-of-sight third-echo beams in the detection signal.

[0047] This invention optimizes the traditional non-line-of-sight imaging detection method by replacing the traditional single-channel detection with distributed multi-channel detection, thereby reducing the photon count rate on a single detection channel and suppressing the time-dependent single-photon count stacking effect, thus improving the detection time accuracy while ensuring overall detection efficiency.

[0048] Specifically, the principle of multi-channel distributed control of the distributed fiber optic detection array is as follows: In a traditional confocal system, the detection point coincides with the laser illumination point, causing the detector to receive a large amount of directly reflected light signals from the repeater wall surface. This leads to dead time and after-pulse effects in the single-photon detector, suppressing the detection of subsequent non-line-of-sight third echo beams and significantly reducing the signal-to-noise ratio. Therefore, in this embodiment, the detection point on the repeater wall surface is slightly offset from the illumination point on the repeater wall surface to avoid the adverse effects of directly reflected light. However, the offset distance should not be too large, otherwise the imaging process will not satisfy the confocal model, resulting in a decrease in reconstruction accuracy. It should be noted that the problem of excessive distance causing the imaging process to fail to satisfy the confocal model and reduce reconstruction accuracy is caused by the hardware structure, and this embodiment will not elaborate further.

[0049] In summary, there is an optimal distance between the detection point and the illumination point, and the optimal distribution of the detection points is a circular trajectory centered on the illumination point. On this trajectory, the intensity ratio of the directly reflected light signal to the non-visual area third echo beam signal is a fixed value. Therefore, by adjusting the intensity ratio of the directly reflected light to the non-visual area third echo beam, the distribution of the detection fiber can be indirectly optimized, so that the detection points are all located on the optimal circular trajectory, thereby improving the signal-to-noise ratio while ensuring the overall detection efficiency.

[0050] It should be noted that the directly reflected light refers to the light emitted by the pulsed laser, which, after passing through the laser linear polarization beam splitter and scanning mirror, is reflected off the surface of the relay wall and then directly passes through the laser linear polarization beam splitter and distributed fiber optic detection array without being reflected by the hidden object, and is then collected by the single-photon detector array.

[0051] The non-line-of-sight triple echo beam refers to the laser emitted by the pulsed laser, which, after passing through the laser linear polarization beam splitter, is diffusely reflected on the surface of the relay wall and then transmitted to the surface of a hidden object in the non-line-of-sight range. After being reflected again on the surface of the hidden object, it returns to the surface of the relay wall, and after being diffusely reflected again on the surface of the relay wall, it passes through the laser linear polarization beam splitter and the distributed fiber optic detection array before being collected by the single-photon detector array.

[0052] Specifically, in photon number-time of flight detection data, since there is a certain time difference between the detection signal of the directly reflected light and the detection signal of the non-line-of-sight third echo beam, and after adjusting the intensity, it is possible to clearly distinguish between the directly reflected light and the non-line-of-sight third echo beam.

[0053] In this embodiment, the workflow of multi-channel distribution control of the probe fiber is as follows: A zero-voltage signal is input to the scanning galvanometer, so that the laser irradiation point on the repeater wall surface illuminated by the pulsed laser through the scanning galvanometer is located at the center of the scanning area; such as Figure 1 The distributed fiber optic detection array 9 is replaced with an indicator light source 16. First, the indicator light source 16 is connected to the detection fiber interface of the intermediate detection channel. The beam emitted by the indicator light source 16 passes through a lens, a laser linear polarization beam splitter, and a galvanometer to form an illumination point corresponding to the intermediate detection channel on the repeater wall surface, simulating the intermediate detection point of the distributed fiber optic detection array on the repeater wall surface. Then, the illumination point corresponding to the intermediate detection channel is adjusted to coincide with the laser illumination point. Next, the indicator light source is connected to the detection fiber interface of the edge detection channel. The beam emitted by the indicator light source 16 also passes through a lens, a laser linear polarization beam splitter, and a galvanometer to form an illumination point corresponding to the edge detection channel on the repeater wall surface, simulating the edge detection point of the distributed fiber optic detection array on the repeater wall surface. This makes the intermediate detection point of the intermediate detection channel and the edge detection point of the edge detection channel on the repeater wall surface visible on the repeater wall surface, facilitating the adjustment of the distance between the edge detection point and the illumination point.

[0054] Specifically, firstly, the illumination point of the corresponding intermediate detection channel formed by the indicator light source on the surface of the relay wall is adjusted to completely coincide with the laser illumination point of the pulsed laser. At this time, the non-view imaging satisfies the complete confocal model. The complete confocal model means that the laser illumination point on the surface of the relay wall and the detection focus of the detector in the non-view imaging system are completely concentric. The purpose of this step is to make the light intensity on the edge detection channel uniform.

[0055] Then, the edge detection channel of the distributed fiber optic detection array 9 is reconnected to the detection fiber interface. The time-of-flight signal of the photons in the edge detection channel is collected using the time-correlated photon counting module. In single-point mode, the intensity ratio of the direct reflected light signal and the non-line-of-sight third echo beam is recorded within a fixed time. The central axis of the edge detection channel and the middle detection channel are adjusted by using optical fibers of different diameters, and the above operation is repeated until the intensity ratio is optimal.

[0056] Furthermore, the principle of adjusting the intensity ratio of the direct reflected light and the non-line-of-sight third echo beam in the detection signal by changing the fiber core diameter in this embodiment is as follows: the intensity distribution of the illumination light on the repeater wall surface is approximately a Gaussian distribution with the illumination point as the Gaussian center. The central axis of the edge detection channel corresponds to the edge detection point on the repeater wall surface. Since the distance between the central axis of the edge detection point and the middle detection channel is approximately equal to half the fiber core diameter, changing the fiber core diameter can change the distance between the edge detection point and the Gaussian center, i.e., the illumination point, thereby changing the intensity of the direct reflected light in the detection signal. Since the non-line-of-sight third echo beam does not change significantly due to the change in the distance between the edge detection point and the illumination point, the intensity ratio of the direct reflected light and the non-line-of-sight third echo beam in the detection signal can be directly adjusted.

[0057] This invention optimizes the multi-channel distribution of the detection fiber to modulate the intensity of the directly reflected light signal and the non-line-of-sight third echo beam signal, thereby suppressing the single-photon detection dead time effect and after-pulse effect caused by the directly reflected light and improving the signal-to-noise ratio of the detection signal.

[0058] In this embodiment, the processing unit includes a counting module. The processing unit calculates the required imaging information based on the detection signals of the corresponding non-line-of-sight three-echo beams collected by the distributed fiber optic detection array. Specifically, the counting module analyzes and processes the detection signals of the corresponding non-line-of-sight three-echo beams collected by the detector array with picosecond resolution to obtain picosecond resolution photon number-time-of-flight data; wherein, the time of flight is the round-trip flight time of photons between the relay wall and the surface of the hidden object.

[0059] It should be noted that the technique used in this embodiment to obtain photon number-time of flight data based on the detection signal of the corresponding non-line-of-sight three echo beams is existing technology and will not be described in detail in this embodiment.

[0060] Furthermore, the processing unit also includes a reconstruction module, which reconstructs and restores the reflectivity information of the surface of the hidden object based on the Wiener filter deconvolution non-view reconstruction algorithm of view cone transform, the obtained photon number-time of flight data with picosecond resolution, and other relevant parameters, thereby realizing the non-view reconstruction of the hidden object.

[0061] In this embodiment, the reconstruction algorithm is based on an orthogonal grid. Therefore, this embodiment forms a two-dimensional spatial orthogonal scanning illumination with an orthogonal grid structure, which facilitates subsequent algorithm reconstruction.

[0062] The process of reconstructing and restoring the reflectivity information of the hidden object's surface using a Wiener filter deconvolution non-view reconstruction algorithm based on view cone transform, along with the obtained photon number-time-of-flight data and other relevant parameters, specifically includes:

[0063] Establish a standard forward model for the confocal case; the formula for the standard forward model is:

[0064]

[0065] In the formula, τ represents the photon count-time of flight data at each scanning point on the repeater surface, (x′,y′) is the scanning point position on the repeater wall surface, (x,y,z) represents the spatial position point on the surface of the hidden object, r is the distance between the spatial position point on the surface of the hidden object and the corresponding scanning point on the repeater wall surface, ρ represents the reflectivity of the hidden object surface, δ represents the Dirac function, Ω represents the three-dimensional space where the hidden object is located, t represents the time of flight, and c represents the speed of light. These data can be obtained directly through measurement or other methods.

[0066] The standard forward model is converted into a 3D convolutional form using a view cone transformation, resulting in a 3D convolutional forward model. The view cone transformation formula is as follows:

[0067] v = (tc / 2) 2 In the formula, Let z represent z in the new coordinate system, and from this, we can deduce that v represents (tc / 2) in the new coordinate system. 2 ;

[0068] The formula for obtaining the forward model in the form of 3D convolution is:

[0069] R t {τ}(x′,y′,v)=∫∫∫ Ω R z {ρ}(x,y,u)h(x′-x,y′-y,vu)dxdydu; where, R t {τ} represents the photon number-time-of-flight data in the form of a three-dimensional convolution, R z {ρ} represents the reflectivity of the hidden object's surface in the form of a 3D convolution;

[0070] The calculation yielded: h(x′-x,y′-y,vu)=δ((x′-x) 2 +(y′-y) 2 +uv);

[0071] A non-view reconstruction formula is established based on Wiener filtering and 3D deconvolution, and R... t {τ}、R z The reflectivity information of the hidden object's surface is reconstructed and restored by substituting {ρ} and other relevant parameters into the non-view reconstruction formula, thereby achieving non-view reconstruction of the hidden object; wherein, the non-view reconstruction formula is:

[0072] In the formula, F represents the three-dimensional Fourier transform, and ρ represents the hidden surface reflectivity information of the object. This represents the focusing matrix containing the Fourier coefficients of the three-dimensional convolution kernel, and α represents the signal-to-noise ratio. These data can be obtained directly.

[0073] This invention utilizes multi-channel photon time-of-flight data fusion and a Wiener filter deconvolution non-view reconstruction algorithm based on view cone transformation to achieve high signal-to-noise ratio and high temporal resolution non-view imaging.

[0074] In this embodiment of the invention, scanning and detection are achieved through the scanning module and the distributed multi-channel detection array. The multi-channel photon time-of-flight signals of different spatial illumination points are recorded. Combined with data fusion and non-view reconstruction algorithms, the three-dimensional reflectivity information of hidden objects in the non-view range is restored.

[0075] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications to the technical solutions of the present invention by utilizing the methods and techniques disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solutions of the present invention shall fall within the protection scope of the technical solutions of the present invention.

Claims

1. A non-line-of-sight imaging system based on distributed multi-channel parallel detection, characterized in that, include: Pulsed laser, laser linear polarization beam splitter, scanning galvanometer, distributed fiber optic detection array, single-photon detector array, repeater wall, processing unit; The laser emitted by the pulsed laser is incident on a laser linear polarization beam splitter to obtain a transmitted light portion. The transmitted light portion is irradiated on the surface of the relay wall by a scanning galvanometer to form a two-dimensional spatial orthogonal scanning illumination. The surface of the relay wall undergoes diffuse reflection based on the two-dimensional spatial orthogonal scanning illumination and is transmitted to the surface of the hidden object outside the field of view. After the beam of non-line-of-sight third echo beams is reflected back to the relay wall surface from the surface of the hidden object, the beams are diffusely reflected again from the relay wall surface and then collected by the single-photon detector array through the scanning galvanometer and the distributed fiber optic detection array. The processing unit calculates the required imaging information based on the detection signal of the corresponding non-line-of-sight third echo beams collected by the distributed fiber optic detection array. The distributed optical fiber detection array includes a central detection channel and multiple edge detection channels distributed outside the sidewall of the central detection channel. The multiple edge detection channels, except for the central detection channel, contain optical fibers and are connected to the single-photon detector array via optical fibers. The non-line-of-sight triple echo beam is collected by the single-photon detector array through the multiple edge detection channels. The position of the central detection point on the relay wall surface corresponding to the central detection channel corresponds to the position of the illumination point on the relay wall surface corresponding to the non-line-of-sight triple echo beam. The distance between the edge detection point on the relay wall surface corresponding to the edge detection channel and the illumination point on the relay wall surface corresponding to the non-line-of-sight triple echo beam is a predetermined value, thereby improving the signal-to-noise ratio of the detection signal of the corresponding non-line-of-sight triple echo beam.

2. The non-line-of-sight imaging system based on distributed multi-channel parallel detection according to claim 1, characterized in that, It also includes a single-mode fiber, a laser collimator, and a first lens. The laser emitted by the pulsed laser is transmitted through the single-mode fiber, collimated by the laser collimator, and then incident on the laser linear polarization beam splitter for beam splitting. The non-field-of-view triple echo beam passes through the scanning galvanometer and is then focused by the first lens and input to the distributed fiber optic detection array.

3. The non-line-of-sight imaging system based on distributed multi-channel parallel detection according to claim 1, characterized in that, It also includes a light stopper. After the laser beam is split by the laser linear polarization beam splitter, a reflected light portion is obtained, which is collected by the light stopper.

4. The non-line-of-sight imaging system based on distributed multi-channel parallel detection according to claim 1, characterized in that, The pulsed laser is a pulsed laser with picosecond resolution, and the single-photon detector array is a detector array with picosecond resolution.

5. The non-line-of-sight imaging system based on distributed multi-channel parallel detection according to claim 1, characterized in that, By setting the ratio of the intensity of the direct reflected light generated on the surface of the relay wall based on the two-dimensional spatial orthogonal scanning illumination to the intensity of the non-viewpoint triple echo beam to 20:1 to 25:1, the distance between the edge detection point on the surface of the relay wall corresponding to the edge detection channel and the illumination point on the surface of the relay wall corresponding to the non-viewpoint triple echo beam is adjusted to a predetermined value.

6. The non-line-of-sight imaging system based on distributed multi-channel parallel detection according to claim 1, characterized in that, The diameter of the inscribed circle of the intermediate detection channel and the edge detection channel is equal to the diameter of the optical fiber. By selecting optical fibers of different diameters, the distance between the central axis of the edge detection channel and the intermediate detection channel is adjusted, thereby adjusting the distance between the edge detection point of the edge detection channel on the relay wall surface and the illumination point of the non-field-of-view third echo beam on the relay wall surface to a predetermined value.

7. The non-line-of-sight imaging system based on distributed multi-channel parallel detection according to claim 1, characterized in that, Both the edge detection channel and the middle detection channel are regular hexagonal prism structures with internal optical fibers. There are 6 edge detection channels, which are distributed on the six sides of the middle detection channel.

8. The non-line-of-sight imaging system based on distributed multi-channel parallel detection according to claim 1, characterized in that, The processing unit includes a counting module, which analyzes and processes the detection signals of the corresponding non-line-of-sight three-echo beams collected by a single-photon detector array with picosecond resolution to obtain picosecond resolution photon number-time-of-flight data; wherein, the time of flight is the round-trip flight time of the photon between the relay wall and the surface of the hidden object.

9. The non-line-of-sight imaging system based on distributed multi-channel parallel detection according to claim 8, characterized in that, The processing unit further includes a reconstruction module, which reconstructs and restores the reflectivity information of the surface of the hidden object based on the Wiener filter deconvolution non-view reconstruction algorithm of view cone transformation and the obtained photon number-time of flight data with picosecond resolution, thereby realizing the non-view reconstruction of the hidden object.

10. The non-line-of-sight imaging system based on distributed multi-channel parallel detection according to claim 9, characterized in that, The process of reconstructing and restoring the reflectivity information of the hidden object's surface based on the Wiener filter deconvolution non-view reconstruction algorithm using the obtained photon number-time-of-flight data specifically includes: Establish a standard forward model for the confocal case; the formula for the standard forward model is: ； In the formula, This represents the photon count-time-of-flight data for each scanning point on the repeater surface. (x′,y′) represents the scanning point position on the repeater wall surface, (x,y,z) represents the spatial position point on the hidden object surface, and r is the distance between the spatial position point on the hidden object surface and the corresponding scanning point on the repeater wall surface. This represents the reflectivity of the hidden object's surface. Represents the Dirac function, The hidden object is located in three-dimensional space, t represents the time of flight, and c represents the speed of light. The standard forward model is converted into a 3D convolutional form using a view cone transformation, resulting in a 3D convolutional forward model. The view cone transformation formula is as follows: In the formula, Let z represent the new coordinate system, and then we can deduce that... Represents the new coordinate system ; The formula for obtaining the forward model in the form of 3D convolution is: In the formula, Represents photon number-time-of-flight data in the form of a three-dimensional convolution. Represents the reflectivity of the hidden object's surface in the form of a 3D convolution; get: , ; A non-view reconstruction formula is established based on Wiener filtering and 3D deconvolution. , The reflectivity information of the hidden object's surface is reconstructed and restored by substituting the non-view domain reconstruction formula; wherein, the non-view domain reconstruction formula is: In the formula, F represents the three-dimensional Fourier transform. This indicates that the surface reflectivity information of the hidden object is being concealed. This represents the focus matrix containing the Fourier coefficients of the three-dimensional convolution kernel. This indicates the signal-to-noise ratio.

Images