A non-confocal body scattering imaging and target three-dimensional reconstruction method

Abstract

The application discloses a kind of non-confocal body scattering imaging and target three-dimensional reconstruction method, comprising the following steps: S1: the establishment of the forward double optical path scattering transmission model of adaptive non-confocal imaging system, the transmission model input imaging three-dimensional target, output three-dimensional image;S2: based on target characteristics, scene hypothesis is proposed, and the forward double optical path scattering transmission model is simplified based on the assumption;S3: the boundary conversion model of non-confocal scattering field is used to inversely solve forward model, realize the two-dimensional topography reconstruction of scattering scene;S4: using depth focusing method realizes the three-dimensional reconstruction of scattering scene.The method provided by the application can realize data acquisition and calculation reconstruction under single exposure using non-confocal imaging system, greatly improves the imaging speed while ensuring the reconstruction quality in strong scattering environment, and simultaneously provides an effective three-dimensional reconstruction method, which is beneficial to the practical process of scattering imaging.

Description

Technical Field

[0001] This invention relates to the fields of computer vision and digital image processing, and in particular to a method for nonfocal scattering imaging and target 3D reconstruction. Background Technology

[0002] Scattering media are widely present in nature, such as clouds, smoke, and turbid water. Clear imaging through scattering media is beneficial for many practical applications, such as autonomous driving in dense fog, scientific research and exploration in murky deep water, and search and rescue operations at fire scenes with dense smoke. However, because strong scattering media cause nonlinear perturbations to light, imaging through strong scattering media faces significant challenges. Achieving both clear and rapid imaging is crucial to effectively improving the practicality of imaging technology.

[0003] To address this issue, current solutions include imaging methods based on spatial characteristics, imaging methods based on gating or temporal correlation, and methods that model the spatiotemporal propagation of scattered light. Spatial characteristic-based imaging methods utilize the spatial invariance of speckle patterns caused by the scattering medium or chromaticity space characteristics to achieve scattering imaging. These methods offer fast imaging speeds, but because they only consider the two-dimensional spatial characteristics of scattering, they cannot handle strong scattering scenarios, such as dense fog with extremely low visibility or underwater environments. The second type of imaging method, based on gating or temporal correlation, uses gating techniques to filter out photons that have undergone only a few scattering events or compares the correlation between images with and without scattering to achieve scattering imaging. While these methods can achieve rapid imaging with ultrafast detector arrays, they still cannot handle imaging problems in strong scattering scenarios. The third type of method, modeling the spatiotemporal propagation of scattered light, fully utilizes the spatial and temporal characteristics of scattered photons, accurately describing the spatiotemporal propagation of scattered light. Therefore, it can clearly reconstruct target objects in strong scattering scenarios. However, these methods require long exposure times to capture more photons or the use of scanning confocal systems to enhance the effective signal, hindering rapid imaging and impeding the practical application of this technology.

[0004] It should be noted that the information disclosed in the background section above is only for understanding the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to overcome the problem that the above-mentioned background technology cannot improve the imaging speed while ensuring the imaging quality of strong scattering imaging, and to provide a nonfocal scattering imaging and target three-dimensional reconstruction method.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A nonfocal scattering imaging and target 3D reconstruction method includes the following steps:

[0008] S1: Establish a forward dual-path scattering transmission model adapted to the non-confocal imaging system. This transmission model takes the imaging three-dimensional target as input and outputs a three-dimensional image.

[0009] S2: Based on the target characteristics, scenario assumptions are proposed, and the forward dual-path scattering transmission model is simplified based on the proposed assumptions;

[0010] S3: The forward double optical path scattering transmission model is solved in reverse by using the boundary transformation model of the non-confocal scattering field to realize the two-dimensional topography reconstruction of the scattering scene;

[0011] S4: Utilize depth focusing method to achieve 3D reconstruction of scattering scene.

[0012] Furthermore:

[0013] The input to this transmission model is the imaging three-dimensional target. The output is a 3D image acquired by an ultrafast detector. Where X, Y, and Z represent the spatial dimensions of the target voxel space in three mutually perpendicular directions in three-dimensional space, M×M represents the spatial pixel size of the ultrafast detector, and T represents the temporal size of the image acquired by the ultrafast detector; the mapping relationship between the three-dimensional target O and the three-dimensional image P in the model is represented by A(·); the noise interference W of the scene is also modeled in the model.

[0014] The mapping factor A(·) in the transmission model is established by a two-optical-path transmission process; the two-optical-path transmission process refers to light emitted from a pulsed laser being forward-scattered to the target object, then reflected by the target object and back-scattered to the detector; wherein, the forward-scattered field is determined by the function φ i The backscattered field is characterized by (x, y, z, t), where x, y, and z represent the three components of three-dimensional space, t represents the time dimension, and the backscattered field is represented by the function φ. e The surface of the target object is characterized by (x, y, z, t), and the effect of the target object's surface on the light is characterized by the function g(x, y, z). The function g(x, y, z) describes the influence of the target object's surface normal, albedo, bidirectional scattering distribution, and other characteristics on the propagation direction and intensity of the light incident on the target surface. The mapping factor A(·) is:

[0015] A(·)=[φ i (x, y, z, t) × g T (x, y, z)]*φ e (x, y, z, t)

[0016] Where * represents convolution operation, g T (x, y, z) is a function formed by repeating the three-dimensional function g(x, y, z) T times in the fourth dimension.

[0017] The forward scattering field φ i (x, y, z, t) and the backscattered field φ e The scattered field (x, y, z, t) satisfies the differential equation expressed by the time-domain diffusion equation, namely:

[0018]

[0019] Let represent the partial derivative with respect to the t component, D represent the diffusion coefficient of the scattering medium, and μa represent the absorption coefficient of the scattering medium.

[0020] The scenario assumption based on target characteristics assumes that the surface structure of the target does not introduce new optical path differences. Therefore, the time-domain starting point of the backscattered field is time-steady, and it is expressed as the time-domain integral of the product of the forward scattered field and the surface characteristic function of the target:

[0021] φ e (x, y, z, t = 0) = ∫φ i (x, y, z, t) × g T (x, y, z)dt

[0022] The target object is approximated as the time-domain starting point of the backscattered field, i.e., O = φ. e (x, y, z, t = 0);

[0023] The acquired three-dimensional transient image is represented as the spatial starting point of the backscattered field, i.e., P = φ e (x, y, z = 0, t).

[0024] In the simplified forward light transmission model, the forward scattering imaging problem is represented by the boundary values ​​of the scattering field as follows:

[0025] φ e (x, y, z = 0, t) = A(φ) e (x, y, z, t = 0)) + W.

[0026] The boundary transformation method for the non-confocal scattering field includes the following steps:

[0027] S31: The acquired 3D transient image φ e After performing a two-dimensional Fourier transform on (x, y) and a one-dimensional numerical transform on t, we obtain (x, y, z = 0, t). This indicates that the backscattered field is in (k x k y Fourier representation of the domain f);

[0028] S32: Based on the equation of light dispersion right Perform from f to k zVariable substitution and interpolation operations are used to obtain Φ′. e (k x k y k z ), indicating that the backscattered field is in (k x k y k z Fourier representation of the domain;

[0029] S33: For Φ′ e (k x k y k z φ is obtained by performing a three-dimensional inverse Fourier transform. e (x, y, z, t = 0), which is the target object.

[0030] The one-dimensional numerical transformation related to t converts the time-domain information to the frequency domain. The conversion method is as follows:

[0031]

[0032] The depth focusing method described above uses a discrete transformation matrix H, composed of a time-domain transformation expression, that exhibits depth focusing characteristics. Matrix H is:

[0033]

[0034] Where, [f1, ..., f T ] and [t1, ..., t T [ ] represent different frequencies and different times in discrete form. By using H matrices of different sizes, target information focused at different depths can be reconstructed to achieve three-dimensional reconstruction of the target object.

[0035] The three-dimensional reconstruction includes the following steps:

[0036] S41: Set the size of the H matrix to T×T, take the acquired transient image of size M×M×T as input, repeat steps S31-S33 to obtain the target image O1 focused to a certain depth;

[0037] S42: Set the size of the H matrix to (T+ΔT)×(T+ΔT), perform time-domain zero-padding on the transient image in S41, and fill the size of the transient image to M×M×(T+ΔT);

[0038] S43: Using the transient image filled in step S42 as input, repeat steps S31-S33 to obtain the target image O2 focused to another depth;

[0039] S44: Repeat steps S41-S43 to obtain a series of focused two-dimensional images O1, O2, ..., O nEach of the resulting images is normalized and then merged into the final 3D image O.

[0040] The present invention has the following beneficial effects:

[0041] The proposed dual-optical-path coupling optical transmission model for non-confocal imaging systems quantitatively describes the spatiotemporal transmission of scattered light in non-confocal systems, providing accurate model support for subsequent inverse processes. Secondly, the non-confocal boundary value transformation model provided by this invention effectively avoids the inverse process of ill-conditioned problems; the entire solution process is free of computational approximations, thus enabling accurate and clear imaging under strong scattering. Finally, the target 3D reconstruction method provided by this invention utilizes the depth-focusing characteristics of the transmission matrix to achieve target 3D reconstruction. Since the non-confocal imaging system used in this invention can complete data acquisition within millisecond-level exposure times, combined with the invented forward imaging model and inverse method, high-quality scattering imaging can be achieved even with extremely weak signal strength. This ensures high-quality strong scattering imaging while achieving rapid imaging, accelerating the practical application of scattering imaging technology.

[0042] To address the problem that existing scattering imaging techniques cannot simultaneously guarantee imaging quality and imaging speed in strong scattering environments, the method provided by this invention utilizes a non-confocal imaging system to achieve data acquisition and computational reconstruction under a single exposure, significantly improving imaging speed while ensuring reconstruction quality in strong scattering environments. Furthermore, it provides an effective three-dimensional reconstruction method, which is beneficial to the practical application of scattering imaging. Attached Figure Description

[0043] Figure 1 This is a flowchart of the nonfocal scattering imaging and target 3D reconstruction method in an embodiment of the present invention;

[0044] Figure 2a This is a reference image of the imaging target in a volume scattering scenario according to an embodiment of the present invention;

[0045] Figure 2b These are three-dimensional transient images acquired in a volume scattering scene in an embodiment of the present invention;

[0046] Figures 2c-2d This is the target reconstruction process in a volume scattering scenario according to an embodiment of the present invention;

[0047] Figure 2e This is a reconstruction result diagram of an embodiment of the present invention;

[0048] Figure 3a This is a schematic diagram of the imaging target according to an embodiment of the present invention;

[0049] Figures 3b-3c This is a schematic diagram illustrating how the reconstruction results are focused on U and T respectively in an embodiment of the present invention;

[0050] Figures 3d-3f These are the three-view diagrams of the final merged three-dimensional imaging result after the focused sub-images are normalized according to the embodiments of the present invention. Detailed Implementation

[0051] The embodiments of the present invention will be described in detail below. It should be emphasized that the following description is merely exemplary and not intended to limit the scope and application of the present invention.

[0052] It should be noted that when a component is referred to as "fixed to" or "set on" another component, it can be directly on or indirectly on that other component. When a component is referred to as "connected to" another component, it can be directly connected to or indirectly connected to that other component. Furthermore, a connection can be used for fixing, coupling, or communication.

[0053] It should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention.

[0054] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of embodiments of the present invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0055] See Figure 1 This invention discloses a nonfocal scattering imaging and target 3D reconstruction method, comprising the following steps:

[0056] S1: Establish a forward dual-path scattering transmission model adapted to the non-confocal imaging system. This transmission model takes the imaging three-dimensional target as input and outputs a three-dimensional image.

[0057] S2: Based on the target characteristics, scenario assumptions are proposed, and the forward dual-path scattering transmission model is simplified based on the proposed assumptions;

[0058] S3: The forward double optical path scattering transmission model is solved in reverse by using the boundary transformation model of the non-confocal scattering field to realize the two-dimensional topography reconstruction of the scattering scene;

[0059] S4: Utilize depth focusing method to achieve 3D reconstruction of scattering scene.

[0060] In some embodiments, the input to the dual optical path scattering and transmission model of the non-confocal imaging system is the imaging three-dimensional target. The output is a 3D image acquired by an ultrafast detector. Where X, Y, and Z represent the spatial dimensions of the target voxel space in three mutually perpendicular directions in three-dimensional space, M×M represents the spatial pixel size of the ultrafast detector, and T represents the temporal dimension of the image acquired by the ultrafast detector. The mapping relationship between the three-dimensional target O and the three-dimensional image P in the model is represented by A(·). The model also models the noise interference W of the scene. In the non-confocal imaging system, the ultrafast detector has an array of photosensitive elements, which can complete the acquisition of three-dimensional transient images in a single exposure. Figure 2a An example of an imaging target is shown. Figure 2b This paper presents a 3D transient image with dimensions of 32×32×128 acquired in this imaging scene.

[0061] In some embodiments, the mapping factor A(·) in the dual-optical-path scattering transmission model adapted to a non-confocal imaging system is established by the dual-optical-path transmission process. The dual-optical-path transmission process refers to light emitted from a pulsed laser being forward-scattered to the target object, then reflected by the target object and back-scattered back to the detector. The forward-scattering field can be represented by the function φ. i The backscattered field is characterized by (x, y, z, t), where x, y, and z represent the three components of three-dimensional space, and t represents the time dimension. e The surface of the target object is characterized by (x, y, z, t), and the effect of the surface on light is represented by the function g(x, y, z). The function g(x, y, z) describes the influence of the surface normal, albedo, and bidirectional scattering distribution on the propagation direction and intensity of light incident on the target surface. Therefore, the mapping factor A(·) can be expressed as:

[0062] A(·)=[φ i (x, y, z, t) × g T (x, y, z)]*φ e (x, y, z, t)

[0063] Where * represents convolution operation, g T (x, y, z) is a function formed by repeating the three-dimensional function g(x, y, z) T times in the fourth dimension.

[0064] In some embodiments, the forward scattered field φ i (x, y, z, t) and the backscattered field φ e The scattered field (x, y, z, t) satisfies the differential equation expressed by the time-domain diffusion equation, namely:

[0065]

[0066] This represents the partial derivative with respect to the t component, where D represents the diffusion coefficient of the scattering medium, and μ... a This represents the absorption coefficient of the scattering medium.

[0067] In some embodiments, based on the scenario assumptions of the target characteristics, assuming that the surface structure of the target does not introduce new optical path differences, the time-domain starting point of the backscattered field is time-steady and can be expressed as the time-domain integral of the product of the forward scattered field and the surface characteristic function of the target:

[0068] φ e (x, y, z, t = 0) = ∫φ i (x, y, z, t) × g T Since the starting point of the time domain describes the main characteristics of the target object, the target object can be approximated as the starting point of the time domain of the backscattered field, i.e., O = φ. e (x, y, z, t = 0). Since the spatial position of the detector is fixed, taking the detector's location as the spatial origin, the acquired three-dimensional transient image can be represented as the spatial origin of the backscattered field, i.e., P = φ. e (x, y, z = 0, t).

[0069] In some embodiments, the simplified forward light transmission model, the forward scattering imaging problem can be represented by the boundary values ​​of the scattering field as follows:

[0070] φ e (x, y, z = 0, t) = A(φ) e (x, y, z, t = 0)) + W.

[0071] In some embodiments, the boundary transformation method for a non-confocal scattering field includes the following steps:

[0072] S31: The acquired 3D transient image φ e After performing a two-dimensional Fourier transform on (x, y) and a one-dimensional numerical transform on t, we obtain (x, y, z = 0, t). This indicates that the backscattered field is in (k x k y Fourier representation of the domain f), such as Figure 2c As shown;

[0073] S32: Based on the equation of light dispersion right Perform the process from f to k z Variable substitution and interpolation operations are used to obtain Φ′. e (k x k y kz ), indicating that the backscattered field is in (k x k y k z Fourier representation of the domain, such as Figure 2d As shown;

[0074] S33: For Φ′ e (k x k y k z φ is obtained by performing a three-dimensional inverse Fourier transform. e (x, y, z, t = 0), i.e., the target object, such as Figure 2e As shown.

[0075] In some embodiments, the above-described one-dimensional numerical transformation of t can convert time-domain information to the frequency domain, and the transformation method can be expressed as follows:

[0076]

[0077] The aforementioned depth-focusing method utilizes a discrete transformation matrix H, constructed from the time-domain transformation expression, which exhibits depth-focusing characteristics. The H matrix can be expressed as:

[0078]

[0079] Where, [f1, ..., f T ] and [t1, ..., t T [ ] represent different frequencies and different times in discrete form. By using H matrices of different sizes, target information focused at different depths can be reconstructed, and finally, the three-dimensional reconstruction of the target object can be achieved.

[0080] In some embodiments, a dual-depth target object such as Figure 3a As shown in the example, the above-mentioned three-dimensional reconstruction method includes the following steps:

[0081] S41: Set the size of the H matrix to T×T. Using a transient image of size M×M×T as input, repeat steps S31-S33 to obtain the target image O1 focused at a certain depth, as shown below. Figure 3b As shown;

[0082] S42: Set the size of the H matrix to (T+ΔT)×(T+ΔT), perform time-domain zero-padding on the transient image in S41, and fill the size of the transient image to M×M×(T+ΔT);

[0083] S43: Using the transient image filled in S42 as input, repeat steps S31-S33 to obtain the target image O2 focused at another depth, as shown. Figure 3c As shown;

[0084] S44: Repeat steps S41-S43 to obtain a series of focused two-dimensional images O1, O2, ..., O n Each of the resulting images is normalized, and then they are merged into the final 3D image O, such as... Figure 3d -f is shown.

[0085] Based on the steps of the above-described nonfocal scattering imaging and target 3D reconstruction method, the transmission of light in the polyethylene foam scattering medium is modeled and the imaging target located within it is reconstructed; the imaging target is a bicycle, such as... Figure 2a As shown, it is located inside a volume scattering medium. Non-confocal three-dimensional transient images of this volume scattering scene were acquired using a non-confocal imaging system, as shown... Figure 2b As shown. Using the acquired image as input, the invented non-confocal boundary value transformation method is used for target reconstruction. The reconstruction process is as follows: Figure 2c As shown in -d, the reconstruction result is as follows Figure 2e As shown in the figure, the reconstruction result of the method proposed in this invention is very close to the real structure of the imaging target, proving the accuracy of the modeling method and the accuracy of the reconstruction method proposed in this invention. Figures 3a-3f The reconstruction process of the invented 3D target reconstruction method is demonstrated, in which... Figure 3a This is a schematic diagram of the imaging target, which consists of the letters U and T, located at different depths. Using the invented depth-focusing reconstruction method, the reconstruction results can be focused onto U and T respectively, as shown below. Figure 3b As shown in -c, the focused sub-images are then normalized and merged. The final merged 3D imaging result has three views as shown. Figure 3d -f is shown.

[0086] The proposed dual-optical-path coupling optical transmission model for non-confocal imaging systems quantitatively describes the spatiotemporal transmission of scattered light in non-confocal systems, providing accurate model support for subsequent inverse processes. Secondly, the non-confocal boundary value transformation model provided by this invention effectively avoids the inverse process of ill-conditioned problems; the entire solution process is free of computational approximations, thus enabling accurate and clear imaging under strong scattering. Finally, the target 3D reconstruction method provided by this invention utilizes the depth-focusing characteristics of the transmission matrix to achieve target 3D reconstruction. Since the non-confocal imaging system used in this invention can complete data acquisition within millisecond-level exposure times, combined with the invented forward imaging model and inverse method, high-quality scattering imaging can be achieved even with extremely weak signal strength. This ensures high-quality strong scattering imaging while achieving rapid imaging, accelerating the practical application of scattering imaging technology.

[0087] To address the problem that existing scattering imaging techniques cannot simultaneously guarantee imaging quality and imaging speed in strong scattering environments, the method provided by this invention utilizes a non-confocal imaging system to achieve data acquisition and computational reconstruction under a single exposure, significantly improving imaging speed while ensuring reconstruction quality in strong scattering environments. Furthermore, it provides an effective three-dimensional reconstruction method, which is beneficial to the practical application of scattering imaging.

[0088] Existing non-confocal scattering light modeling methods are inaccurate, and existing scattering imaging methods have poor reconstruction capabilities and slow imaging speeds in strong scattering environments. In contrast, our method can reconstruct complex targets in volume scattering environments, and its imaging speed is much higher than other algorithms.

[0089] The key advantages of this invention are:

[0090] Compared to existing methods that use confocal systems for scanning imaging, which have long data acquisition times, this invention uses a non-confocal system to accelerate data acquisition time.

[0091] Compared to existing non-confocal imaging methods, this invention has superior imaging capabilities in strong scattering environments.

[0092] This invention can ensure high-quality imaging in strong scattering environments while accelerating the imaging speed, which is beneficial to the practical application of scattering technology.

[0093] Specific application scenarios of this invention include:

[0094] It can enhance the imaging capabilities of autonomous vehicles in extreme weather conditions, such as fog and haze, thereby assisting them in obstacle recognition and path planning;

[0095] It can be used in medical observation to detect diseased biological tissues through biological tissues with strong scattering ability;

[0096] It can be used for scientific research and exploration in turbid deep seas, enhancing the visual capabilities of underwater robots, and enabling deep-sea exploration vessels to obtain clear images through turbid waters;

[0097] It can be used to search for trapped people in fire rescue scenes filled with thick smoke;

[0098] Underwater rescue can be carried out in turbid water.

[0099] This invention has significant market value, particularly for the autonomous driving industry. Currently, autonomous vehicles perform poorly in harsh weather conditions, and imaging problems in strong scattering environments hinder their practical application. This invention can improve their imaging capabilities in extreme environments, thus promoting their development. The method proposed in this invention can be applied to LiDAR systems for autonomous driving, improving LiDAR's imaging capabilities in foggy or hazy weather, assisting autonomous vehicles in safe driving under challenging weather conditions, and advancing the development of autonomous vehicles.

[0100] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0101] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0102] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0103] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0104] The background section of this invention may include background information about the problems or environment in which the invention is being developed, and is not necessarily a description of prior art. Therefore, the content included in the background section does not constitute an admission of prior art by the applicant.

[0105] The above description provides a further detailed explanation of the present invention in conjunction with specific / preferred embodiments, and it should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various substitutions or modifications can be made to these described embodiments without departing from the concept of the present invention, and all such substitutions or modifications should be considered within the scope of protection of the present invention. In the description of this specification, the reference to terms such as "an embodiment," "some embodiments," "preferred embodiment," "example," "specific example," or "some examples," etc., indicates that the specific features, structures, materials, or characteristics described in connection with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in one or more embodiments or examples. Without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification and the features of different embodiments or examples. Although the embodiments of the present invention and their advantages have been described in detail, it should be understood that various changes, substitutions, and modifications can be made herein without departing from the scope of protection of the patent application.

Claims

1. A method for nonfocal scattering imaging and target 3D reconstruction, characterized in that, Includes the following steps: S1: Establish a forward dual-path scattering transmission model adapted to the non-confocal imaging system. This transmission model takes the imaging three-dimensional target as input and outputs a three-dimensional image. S2: Based on the target characteristics, scenario assumptions are proposed, and the forward dual-path scattering transmission model is simplified based on the proposed assumptions; S3: The forward double optical path scattering transmission model is solved in reverse using the boundary transformation model of the non-confocal scattering field to realize the two-dimensional topography reconstruction of the scattering scene; S4: Utilize depth focusing method to achieve 3D reconstruction of scattering scene; The input to this transmission model is the imaging three-dimensional target. The output is a three-dimensional image acquired by an ultrafast detector. , in These represent the spatial dimensions of the target voxel space in three mutually perpendicular directions in three-dimensional space. The spatial pixel size represents the ultrafast detector. Represents the temporal dimension of the image captured by the ultrafast detector; the three-dimensional target in the model. and 3D images The mapping relationship between them is The model also models the noise interference in the scene. ; Mapping factor in the transmission model Established by a two-path optical path transmission process; the two-path optical path transmission process refers to light emitted from a pulsed laser being forward-scattered to the target object, then reflected by the target object and back-scattered to the detector; wherein, the forward-scattered field is determined by a function Characterization, in which, These represent the three components of three-dimensional space. Representing the time dimension, the backscattered field is represented by the function Characterization, the effect of the target object's surface on light is determined by a function Representation; Function Describe the influence of the target surface normal, albedo, and bidirectional scattering distribution characteristics on the propagation direction and intensity of light incident on the target surface; mapping factor. for: ， in, This represents the convolution operation. To make three-dimensional functions Repeat in the fourth dimension The function formed in this way.

2. The nonfocal scattering imaging and target 3D reconstruction method as described in claim 1, characterized in that, The forward scattering field and backscattered field The scattered field satisfies the differential equation expressed by the time-domain diffusion equation, namely: Indicates to Find the partial derivatives of the components. This represents the diffusion coefficient of the scattering medium. This represents the absorption coefficient of the scattering medium.

3. The nonfocal scattering imaging and target 3D reconstruction method as described in claim 1, characterized in that, The scenario assumption based on target characteristics assumes that the surface structure of the target does not introduce new optical path differences. Therefore, the time-domain starting point of the backscattered field is time-steady, and it is expressed as the time-domain integral of the product of the forward scattered field and the surface characteristic function of the target: The target object is approximately the starting point of the backscattered field in the time domain, i.e. ; The acquired three-dimensional transient image is represented as the spatial starting point of the backscattered field, i.e. .

4. The nonfocal scattering imaging and target 3D reconstruction method as described in claim 1, characterized in that, In the simplified forward light transmission model, the forward scattering imaging problem is represented by the boundary values ​​of the scattering field as follows: 。 5. The nonfocal scattering imaging and target 3D reconstruction method as described in claim 1, characterized in that, The boundary transformation method for the non-confocal scattering field includes the following steps: S31: Acquired 3D transient images Conduct related Two-dimensional Fourier transform and related After a one-dimensional numerical transformation, we obtain This indicates that the backscattered field is in Fourier representation of the domain; S32: Based on the equation of light dispersion right Perform from arrive Variable substitution and interpolation operations yield the following results: This indicates that the backscattered field is in Fourier representation of the domain; S33: Yes Performing a three-dimensional inverse Fourier transform yields , that is, the target object.

6. The nonfocal scattering imaging and target 3D reconstruction method as described in claim 5, characterized in that, The relevant A one-dimensional numerical transformation is used to convert time-domain information to the frequency domain. The transformation method is as follows: 。 7. The nonfocal scattering imaging and target 3D reconstruction method as described in claim 1, characterized in that, The depth focusing method comprises a discrete transformation matrix consisting of a time-domain transformation expression. It has the characteristic of deep focusing, matrix for: in, and Representing different frequencies and different times in discrete form, using different sizes The matrix reconstruction extracts target information focused at different depths, enabling the three-dimensional reconstruction of the target object.

8. The nonfocal scattering imaging and target 3D reconstruction method as described in claim 7, characterized in that, The three-dimensional reconstruction includes the following steps: S41: Settings The size of the matrix is Based on the size of the collected data Using the transient image as input, repeat steps S31-S33 to obtain a target image focused at a certain depth. ; S42: Settings The size of the matrix is The transient image in S41 is zero-padding in the temporal domain, and the size of the transient image is padded to 1. ; S43: Using the transient image filled in step S42 as input, repeat steps S31-S33 to obtain the target image focused to another depth. ; S44: Repeat steps S41-S43 to obtain a series of focused two-dimensional images. Each of the resulting images is normalized, and then they are merged into the final 3D image. .

Images