Method, system, device, and medium for occlusion removal when generating a multi-depth diffraction field

Abstract

The application discloses a method, system, device and medium for eliminating occlusion when generating a multi-depth diffraction field, and relates to the technical field of holograms. The method comprises the following steps: calculating the Nth diffraction field complex amplitude according to a preset target plane depth value, a maximum value in a preset depth interval, a corresponding part of the Nth depth plane in an intensity image of a target three-dimensional scene, a diffraction field complex amplitude of the N-1th depth plane, and a depth value of the Nth depth plane; wherein, when 1 < n < N, the diffraction field complex amplitude of the nth depth plane is calculated according to a preset depth interval, a corresponding part of the nth depth plane in the intensity image of the target three-dimensional scene, the diffraction field complex amplitude of the n-1th depth plane, and the depth value of the nth depth plane; and when n = 1, the diffraction field complex amplitude of the first depth plane is obtained according to the preset depth interval and the intensity image of the target three-dimensional scene. The application can solve the problem of calculating multi-depth diffraction field occlusion.

Description

Technical Field

[0001] This invention relates to the field of holographic technology, and in particular to a method, system, device and medium for eliminating occlusion when generating multi-depth diffraction fields. Background Technology

[0002] When generating multi-depth diffraction fields from 3D datasets, handling the occlusion relationships between different layers is crucial. Sampling methods are divided into two categories: complete sampling and incomplete sampling. Complete sampling includes obtaining all occluded objects, while incomplete sampling does not.

[0003] The existing solution is to calculate the diffraction field of multi-depth scenes using the standard method, the specific process of which is as follows: Figure 6 As shown, where, Figure 1 The image is a 3D scene featuring "playing cards". It consists of four layers to illustrate the occlusion process. The first layer is the tablecloth, the second layer is the 10 of clubs, the third layer is the 9 of diamonds, and the fourth layer is the 8 of spades. Figure 2 for Figure 1 Intensity image, Figure 3 for Figure 1 Depth image, Figure 4 Part (a) is Figure 3 Intensity image of the first depth plane in the middle, Figure 4 Part (b) is Figure 3 Intensity image of the second depth plane in the middle, Figure 4 Part (c) is Figure 3 Intensity image of the third depth plane in the middle. Figure 4 The middle (d) part is Figure 3 Intensity image of the fourth depth plane in the middle. Figure 5 Part (a) is Figure 3 The mask image of the first depth plane in the middle. Figure 5 Part (b) is Figure 3 The mask image of the first depth plane in the middle. Figure 5 Part (c) is Figure 3 The mask image of the first depth plane in the middle. Figure 5 The middle (d) part is Figure 3 The mask image of the first depth plane in the middle. Figure 6 To calculate using the standard method Figure 1 Flowchart of multi-depth diffraction field. According to... Figure 6 It can be seen that the first layer propagates forward, generating the diffraction field of the first depth plane (layer). Subsequently, the second mask blocks the first diffraction field, allowing it to propagate forward together with the superposition of the second layer, generating the second diffraction field. Similarly, each superposition of the diffraction field and each layer of the slice is replaced by the next layer at the corresponding position after diffraction. According to angular spectrum theory, the complex amplitude of the diffraction field calculated by the layer-by-layer replacement method can be expressed as:

[0004]

[0005] In the calculation process of the diffraction field, the lack of information of the occluded object seriously affects the reconstruction quality of the hologram, especially severe dark lines will appear at the boundaries of different layers. Therefore, solving the occlusion problem in the calculation of multi-depth diffraction fields is very important for the reconstruction quality of the hologram. Contents of the Invention

[0006] The objective of the present invention is to provide an occlusion elimination method, system, device and medium for generating a multi-depth diffraction field, which can solve the occlusion problem in the calculation of multi-depth diffraction fields.

[0007] To achieve the above objective, the present invention provides the following schemes:

[0008] An occlusion elimination method for generating a multi-depth diffraction field, comprising:

[0009] acquiring an intensity image and a depth image of a target three-dimensional scene;

[0010] discretizing the depth image of the target three-dimensional scene according to a preset depth interval and a preset total number of depth planes of the discrete depth image to obtain the discrete depth image; the discrete depth image includes the preset total number of depth planes, and depth values of the same depth plane are identical;

[0011] based on the angular spectrum theory, calculating the diffraction field complex amplitude of the Nth depth plane according to a preset target plane depth value, the maximum value in the preset depth interval, the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the diffraction field complex amplitude of the (N-1)th depth plane and the depth value of the Nth depth plane, wherein N is the preset total number of depth planes of the discrete depth image; wherein when 1<n<N, based on the angular spectrum theory, calculating the diffraction field complex amplitude of the nth depth plane according to a depth interval, the corresponding part of the nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the diffraction field complex amplitude of the (n-1)th depth plane and the depth value of the nth depth plane; when n=1, based on the angular spectrum theory, obtaining the diffraction field complex amplitude of the first depth plane according to the depth interval and the intensity image of the target three-dimensional scene; the depth interval is calculated according to the preset depth interval and the preset total number of depth planes of the discrete depth image.

[0012] Optionally, discretizing the depth image of the target three-dimensional scene according to a preset depth interval and a preset total number of depth planes of the discrete depth image to obtain the discrete depth image specifically comprises:

[0013] setting the preset depth interval as [D1, D2], and setting the preset total number of depth planes of the discrete depth image as N;

[0014] According to the formula Calculate the depth value d of the nth depth plane in a discrete depth image. n ;

[0015] The grayscale value of the depth image of the target 3D scene is d n-1 to d n The values ​​are all set to d n A discrete depth image is obtained.

[0016] Optionally, the depth interval is based on the formula Calculate, where Δz is the depth interval and [D1, D2] is the preset depth range.

[0017] Optionally, based on angular spectrum theory, the complex amplitude of the diffraction field in the first depth plane is obtained according to the depth interval and the intensity image of the target 3D scene, specifically:

[0018] The intensity distribution of the target 3D scene is obtained based on the intensity image of the target 3D scene;

[0019] Based on angular spectrum theory, the complex amplitude of the diffraction field of the first depth plane is calculated according to the depth interval and the intensity distribution of the target three-dimensional scene.

[0020] Optionally, based on angular spectrum theory, the complex amplitude of the diffraction field of the nth depth plane is calculated according to the depth interval, the corresponding part of the nth depth plane in the intensity image of the target 3D scene, the complex amplitude of the diffraction field of the (n-1)th depth plane, and the depth value of the nth depth plane. Specifically, this includes:

[0021] The intensity distribution of the target 3D scene at the nth depth plane is calculated based on the corresponding part of the nth depth plane in the intensity image of the target 3D scene.

[0022] Based on angular spectrum theory, the complex amplitude of the diffraction field in the nth depth plane is calculated according to the intensity distribution of the target 3D scene in the nth depth plane, the depth value of the nth depth plane, the depth interval, and the complex amplitude of the diffraction field in the (n-1)th depth plane.

[0023] Optionally, based on angular spectrum theory, the complex amplitude of the diffraction field of the Nth depth plane is calculated according to the preset target plane depth value, the maximum value in the preset depth interval, the corresponding part of the Nth depth plane in the intensity image of the target 3D scene, the complex amplitude of the diffraction field of the (N-1)th depth plane, and the depth value of the Nth depth plane. Specifically, this includes:

[0024] Calculate the depth difference based on the preset target plane depth value and the maximum value in the preset depth range;

[0025] calculating the intensity distribution of the target three-dimensional scene at the Nth depth plane according to the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene;

[0026] based on the angular spectrum theory, calculating the complex amplitude of the diffraction field at the Nth depth plane according to the intensity distribution of the target three-dimensional scene at the Nth depth plane, the depth value of the Nth depth plane, the depth difference, and the complex amplitude of the diffraction field at the (N-1)th depth plane.

[0027] An occlusion elimination system for generating a multi-depth diffraction field, comprising:

[0028] an acquisition module, configured to acquire an intensity image and a depth image of a target three-dimensional scene;

[0029] a discretization module, configured to discretize the depth image of the target three-dimensional scene according to a preset depth interval and a preset total number of depth planes of the discrete depth image to obtain the discrete depth image; the discrete depth image includes the preset total number of depth planes, and depth values of the same depth plane are the same;

[0030] a diffraction field complex amplitude calculation module, configured to calculate the complex amplitude of the diffraction field at the Nth depth plane based on the angular spectrum theory according to a preset target plane depth value, the maximum value in the preset depth interval, the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the complex amplitude of the diffraction field at the (N-1)th depth plane, and the depth value of the Nth depth plane, where N is the preset total number of depth planes of the discrete depth image; wherein when 1<n<N, based on the angular spectrum theory, the complex amplitude of the diffraction field at the nth depth plane is calculated according to the depth interval, the corresponding part of the nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the complex amplitude of the diffraction field at the (n-1)th depth plane, and the depth value of the nth depth plane; when n=1, based on the angular spectrum theory, the complex amplitude of the diffraction field at the 1st depth plane is obtained according to the depth interval and the intensity image of the target three-dimensional scene; the depth interval is calculated according to the preset depth interval and the preset total number of depth planes of the discrete depth image.

[0031] Optionally, the discretization module specifically comprises:

[0032] a preset unit, configured to set the preset depth interval as [D1, D2], and set the preset total number of depth planes of the discrete depth image as N;

[0033] a depth value calculation unit, configured to according to the formula calculate the depth value d of the nth depth plane of the discrete depth image n ;

[0034] a discretization unit, configured to assign the gray value d in the depth image of the target three-dimensional scenen-1 to d n are all set to d n to obtain a discrete depth image.

[0035] An electronic device, comprising:

[0036] a memory and a processor, wherein the memory is configured to store a computer program, and the processor runs the computer program to cause the electronic device to execute the above-mentioned occlusion elimination method when generating a multi-depth diffraction field.

[0037] A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the above-mentioned occlusion elimination method when generating a multi-depth diffraction field.

[0038] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0039] Based on the angular spectrum theory, the complex amplitude of the diffraction field of the Nth depth plane is calculated according to the preset depth value of the target plane, the maximum value in the preset depth interval, the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the complex amplitude of the diffraction field of the (N-1)th depth plane, and the depth value of the Nth depth plane, where N is the total number of preset depth planes of the discrete depth image; wherein, when 1<n<N, based on the angular spectrum theory, the complex amplitude of the diffraction field of the nth depth plane is calculated according to the depth interval, the corresponding part of the nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the complex amplitude of the diffraction field of the (n-1)th depth plane, and the depth value of the nth depth plane; when n=1, based on the angular spectrum theory, the complex amplitude of the diffraction field of the first depth plane is obtained according to the depth interval and the intensity image of the target three-dimensional scene, which can solve the problem of calculating occlusion in the multi-depth diffraction field and improve the reconstruction quality of the hologram. Brief Description of the Drawings

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can also be obtained based on these drawings without any creative effort.

[0041] Figure 1 is a 3D scene diagram featuring "playing cards";

[0042] Figure 2 is Figure 1 the intensity image of;

[0043] Figure 3 is Figure 1Depth image;

[0044] Figure 4 for Figure 3 Intensity images of planes at various depths;

[0045] Figure 5 for Figure 3 Mask images of each depth plane in the middle;

[0046] Figure 6 To calculate using the standard method Figure 1 Flowchart of multi-depth diffraction field;

[0047] Figure 7 This is a comparative diagram showing the complex amplitude results of the diffraction field obtained by using the standard method and the occlusion elimination method provided in the embodiments of the present invention when generating multi-depth diffraction fields;

[0048] Figure 8 This is a flowchart of the diffraction field of the 2nd to Nth depth planes (layers) calculated using the occlusion elimination method for generating multi-depth diffraction fields provided in the embodiments of the present invention.

[0049] Figure 9 This is a flowchart of calculating the diffraction field of the first depth plane (layer) using the occlusion elimination method for generating multi-depth diffraction fields provided in the embodiments of the present invention;

[0050] Figure 10 A flowchart illustrating the occlusion elimination method for generating multi-depth diffraction fields provided in this embodiment of the invention. Detailed Implementation

[0051] 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.

[0052] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0053] When performing incomplete sampling on 3D scenes, noticeable dark boundaries appear in hologram reconstruction, disrupting the visual continuity of the 3D scene. To address this issue, this invention proposes a layer-by-layer replacement method (L2RM), an occlusion removal method for generating multi-depth diffraction fields, to smooth the boundaries between different layers. In practice, incomplete sampling is more common and convenient than complete sampling, making L2RM suitable for handling occlusion relationships in larger scenes and also suitable for scenarios such as data preprocessing for multi-depth hologram neural network models. Figure 7 Part (a) in the figure is a schematic diagram of the complex amplitude of the diffraction field obtained by the standard method, Figure 7 Part (b) in the figure is a schematic diagram of the complex amplitude of the diffraction field obtained by the L2RM method provided in the embodiment of the present invention, as Figure 7 Shown in part (b) of the figure, the first step of the L2RM method involves replacing the occluded pixels in the first layer with the known pixels at the corresponding positions in the sample layer. Next, the diffraction field of each layer is calculated, and the pixels in the diffraction field are replaced again with the pixels at the corresponding positions of this layer. Due to the filling of missing pixels, the boundary of the calculated diffraction field becomes smoother, which can be regarded as a continuation of the sample slice.

[0054] As Figure 10 Shown in the figure, an embodiment of the present invention provides an occlusion elimination method for generating a multi-depth diffraction field, comprising:

[0055] Step 101: Acquire an intensity image and a depth image of a target three-dimensional scene.

[0056] Step 102: Discretize the depth image of the target three-dimensional scene according to a preset depth interval and a preset total number of depth planes of the discrete depth image to obtain a discrete depth image; the discrete depth image includes the preset total number of depth planes, and depth values of the same depth plane are the same.

[0057] Step 103: Calculate the complex amplitude of the diffraction field of the Nth depth plane based on the angular spectrum theory according to the preset target plane depth value, the maximum value in the preset depth interval, the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the complex amplitude of the diffraction field of the (N-1)th depth plane and the depth value of the Nth depth plane, where N is the preset total number of depth planes of the discrete depth image; wherein, when 1<n<N, calculate the complex amplitude of the diffraction field of the nth depth plane based on the angular spectrum theory according to the depth interval, the corresponding part of the nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the complex amplitude of the diffraction field of the (n-1)th depth plane and the depth value of the nth depth plane; when n=1, obtain the complex amplitude of the diffraction field of the first depth plane based on the angular spectrum theory according to the depth interval and the intensity image of the target three-dimensional scene; the depth interval is calculated according to the preset depth interval and the preset total number of depth planes of the discrete depth image.

[0058] In practical applications, discretizing the depth image of the target three-dimensional scene according to the preset depth interval and the preset total number of depth planes of the discrete depth image to obtain the discrete depth image specifically comprises:

[0059] Generally, the depth range of a depth image is 0-255. Based on actual needs, a preset depth range is set to [D1, D2], where D1 represents the minimum depth value and D2 represents the maximum depth value. The total number of preset depth planes for the discrete depth image is set to N.

[0060] According to the formula Calculate the depth value d of the nth depth plane in a discrete depth image. n .

[0061] The grayscale value of the depth image of the target 3D scene is d n-1 to d n The values ​​are all set to d n Obtain a discrete depth image; specifically, merge the depth values ​​of each depth plane, that is, set the gray values ​​of the 3D scene's depth image belonging to 0-d1 as d1, and subsequently set the gray values ​​from d1 to d2 as d2, and the gray values ​​of d1 to d2 as d2. n-1 to d n The values ​​are all set to d n Finally, N grayscale values ​​are obtained.

[0062] In practical applications, the depth interval is based on the formula Calculate, where Δz is the depth interval and [D1, D2] is the preset depth range.

[0063] In practical applications, based on angular spectrum theory, the complex amplitude of the diffraction field in the first depth plane is obtained according to the depth interval and the intensity image of the target 3D scene, specifically:

[0064] The intensity distribution of the target 3D scene is obtained based on the intensity image of the target 3D scene.

[0065] Based on angular spectrum theory, the complex amplitude of the diffraction field in the first depth plane is calculated according to the depth interval and the intensity distribution of the target 3D scene; specifically, it is calculated according to the formula... Calculate the complex amplitude of the diffraction field in the first depth plane, where F{·} is the two-dimensional Fourier transform, F -1 {·} represents the two-dimensional inverse Fourier transform, I(x, y) is the intensity distribution of the original image (the target 3D scene), and H... Δz (f x f y Let be the optical transfer function (OTF) during the propagation of the diffraction field, expressed as: f x and f y These are the spatial frequencies of the diffraction field along the x-axis and y-axis, respectively, according to |f x |,|f y |<1 / 2Δp sets the sampling frequency, where Δp is the pixel interval of the SLM, and j is the imaginary unit.2 =-1, λ is the wavelength of the laser used, k = 2π / λ. For example... Figure 9 As shown, the intensity image (instead of the first layer image) propagates forward, resulting in the generation of a diffraction field in the first depth plane (layer).

[0066] In practical applications, based on angular spectrum theory, the complex amplitude of the diffraction field of the nth depth plane is calculated according to the depth interval, the corresponding part of the nth depth plane in the intensity image of the target 3D scene, the complex amplitude of the diffraction field of the (n-1)th depth plane, and the depth value of the nth depth plane. Specifically, this includes:

[0067] The intensity distribution of the target 3D scene at the nth depth plane is calculated based on the corresponding portion of the nth depth plane in the intensity image of the target 3D scene.

[0068] Based on angular spectrum theory, the complex amplitude of the diffraction field in the nth depth plane is calculated according to the intensity distribution of the target 3D scene in the nth depth plane, the depth value of the nth depth plane, the depth interval, and the complex amplitude of the diffraction field in the (n-1)th depth plane.

[0069] In practical applications, based on angular spectrum theory, the complex amplitude of the diffraction field of the Nth depth plane is calculated using the preset target plane depth value, the maximum value in the preset depth interval, the corresponding part of the Nth depth plane in the intensity image of the target 3D scene, the complex amplitude of the diffraction field of the (N-1)th depth plane, and the depth value of the Nth depth plane. Specifically, this includes:

[0070] The depth difference is calculated based on the preset target plane depth value and the maximum value in the preset depth range.

[0071] The intensity distribution of the target 3D scene at the Nth depth plane is calculated based on the corresponding portion of the Nth depth plane in the intensity image of the target 3D scene.

[0072] Based on angular spectrum theory, the complex amplitude of the diffraction field in the Nth depth plane is calculated according to the intensity distribution of the target 3D scene in the Nth depth plane, the depth value of the Nth depth plane, the depth difference, and the complex amplitude of the diffraction field in the (N-1)th depth plane.

[0073] In practical applications, the calculation of the complex amplitude of the diffraction field in the 2nd to Nth depth planes is performed using a superimposed masking method, such as... Figure 8 As shown, the second mask blocks the diffraction field of the first depth plane, and the two fields are superimposed and propagate forward to produce the diffraction field of the second depth plane. Similarly, the nth mask blocks the diffraction field of the (n-1)th depth plane, and the two fields are superimposed and propagate forward to produce the diffraction field of the nth depth plane.

[0074] Based on angular spectrum theory, calculate the complex amplitude U of the diffraction field in the 2nd to Nth depth planes of a multi-depth scene. n (x, y) can be represented as:

[0075]

[0076] U n-1 (x, y) is the complex amplitude of the diffraction field at the (n-1)th depth plane, I n The intensity distribution of the (x, y) scene at the nth depth plane.

[0077] Operating method Represented as

[0078] Where M n (x, y) represents the mask corresponding to the nth layer. represents the depth distribution, or depth value, of the scene in the nth depth plane, expressed as:

[0079] In the mask, "0" represents the diffraction field blocked by each layer, and "1" represents the diffraction field not blocked by each layer. Calculate the complex amplitude U of the diffraction field of the multi-depth scene on the target plane. N When (x, y), the depth difference z target Represented as: z target =d target -D2,d target This represents the preset target plane depth value.

[0080] Based on the above method, the complex amplitude U of the diffraction field of the target plane at multiple depths can be obtained. N (x, y).

[0081] In view of the above method, this embodiment of the invention provides an occlusion cancellation system for generating multi-depth diffraction fields, comprising:

[0082] The acquisition module is used to acquire intensity and depth images of the target 3D scene.

[0083] The discrete module is used to discretize the depth image of the target 3D scene according to the preset depth range and the preset total number of depth planes of the discrete depth image to obtain a discrete depth image; the discrete depth image includes the preset total number of depth planes, and the depth value of the same depth plane is the same.

[0084] Diffraction field complex amplitude calculation module, configured to calculate the diffraction field complex amplitude of the Nth depth plane based on the angular spectrum theory according to a preset target plane depth value, the maximum value in a preset depth interval, the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the diffraction field complex amplitude of the (N-1)th depth plane, and the depth value of the Nth depth plane, where N is the total number of preset depth planes of the discrete depth image; wherein when 1<n<N, the diffraction field complex amplitude of the nth depth plane is calculated based on the angular spectrum theory according to the depth interval, the corresponding part of the nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the diffraction field complex amplitude of the (n-1)th depth plane, and the depth value of the nth depth plane; when n=1, the diffraction field complex amplitude of the first depth plane is obtained based on the angular spectrum theory according to the depth interval and the intensity image of the target three-dimensional scene; the depth interval is calculated according to the preset depth interval and the total number of preset depth planes of the discrete depth image.

[0085] In practical applications, the discretization module specifically includes:

[0086] A preset unit, configured to set the preset depth interval as [D1, D2], and set the total number of preset depth planes of the discrete depth image as N.

[0087] A depth value calculation unit, configured to according to the formula calculate the depth value d of the nth depth plane of the discrete depth image n .

[0088] A discretization unit, configured to set all values with gray levels from d in the depth image of the target three-dimensional scene n-1 to d n to d, so as to obtain the discrete depth image n .

[0089] An embodiment of the present invention further provides an electronic device, including:

[0090] a memory and a processor, wherein the memory is configured to store a computer program, and the processor runs the computer program to cause the electronic device to execute the occlusion elimination method when generating a multi-depth diffraction field according to the above embodiments.

[0091] An embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the occlusion elimination method when generating a multi-depth diffraction field according to the above embodiments is implemented.

[0092] This invention proposes an occlusion removal method for calculating the diffraction field of a multi-depth scene. First, the 3D scene is transformed into a multi-depth scene; then, the layer-by-layer replacement method (L2RM) is introduced to smoothly reconstruct the occlusion boundaries between different layers of the scene, and the forward-backward bandwidth-limited angular spectrum method is used to calculate the multi-depth diffraction field.

[0093] The present invention has the following advantages:

[0094] (1) Occlusion Removal: When performing incomplete sampling of a 3D scene, due to the lack of information about occluded objects, obvious dark boundaries appear in hologram reconstruction, especially severe dark lines at the boundaries of different layers, which disrupts the visual continuity of the 3D scene. When using the L2RM method to achieve occlusion removal in diffraction calculation to generate holograms, the method provided in this invention effectively alleviates the dark boundary problem existing in the standard method and improves the reproduction quality of multi-depth holograms.

[0095] (2) Suitable for handling large-scale complex scenes: In the sampling process of 3D holography, it is very difficult, even unrealistic, to achieve complete sampling of large-scale and complex scenes. The L2RM method can effectively alleviate this contradiction.

[0096] (3) Suitable for data preprocessing of multi-depth holographic neural network models. In the training, testing and use stages of multi-depth holographic neural network models, the L2RM method can effectively improve the generation quality of holograms. Experiments show that the average intensity PSNR of the scene reconstructed using L2RM is 31.8dB and SSIM is 0.86, which are 4.0dB and 0.11 higher than the standard method, respectively.

[0097] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0098] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for eliminating occlusion when generating multi-depth diffraction fields, characterized in that, comprising: acquiring an intensity image and a depth image of a target three-dimensional scene; discretizing the depth image of the target three-dimensional scene according to a preset depth interval and a preset total number of depth planes of a discrete depth image to obtain the discrete depth image; the discrete depth image comprises the preset total number of depth planes, and depth values on the same depth plane are identical; based on angular spectrum theory, calculating the diffraction field complex amplitude of the Nth depth plane according to a preset target plane depth value, the maximum value in the preset depth interval, the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the diffraction field complex amplitude of the (N-1)th depth plane and the depth value of the Nth depth plane, wherein N is the preset total number of depth planes of the discrete depth image; wherein, when 1<n<N, based on the angular spectrum theory, calculating the diffraction field complex amplitude of the nth depth plane according to the depth interval, the corresponding part of the nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the diffraction field complex amplitude of the (n-1)th depth plane and the depth value of the nth depth plane; when n=1, based on the angular spectrum theory, obtaining the diffraction field complex amplitude of the first depth plane according to the depth interval and the intensity image of the target three-dimensional scene; the depth interval is calculated according to the preset depth interval and the preset total number of depth planes of the discrete depth image.

2. The occlusion elimination method for generating multi-depth diffraction fields according to claim 1, characterized in that, discretizing the depth image of the target three-dimensional scene according to the preset depth interval and the preset total number of depth planes of the discrete depth image to obtain the discrete depth image, specifically comprising: setting the preset depth interval as [D1, D2], and setting the preset total number of depth planes of the discrete depth image as N; According to the formula Calculate the depth value d of the nth depth plane in a discrete depth image. n ; The grayscale value of the depth image of the target 3D scene is d n-1 to d n The values ​​are all set to d n A discrete depth image is obtained.

3. The occlusion elimination method for generating multi-depth diffraction fields according to claim 1, characterized in that, The depth interval is based on the formula Calculate, where Δz is the depth interval and [D1, D2] is the preset depth range.

4. The occlusion elimination method for generating multi-depth diffraction fields according to claim 1, characterized in that, based on the angular spectrum theory, obtaining the diffraction field complex amplitude of the first depth plane according to the depth interval and the intensity image of the target three-dimensional scene, specifically as: obtaining the intensity distribution of the target three-dimensional scene according to the intensity image of the target three-dimensional scene; based on the angular spectrum theory, calculating the diffraction field complex amplitude of the first depth plane according to the depth interval and the intensity distribution of the target three-dimensional scene.

5. The occlusion elimination method for generating multi-depth diffraction fields according to claim 1, characterized in that, based on the angular spectrum theory, calculating the diffraction field complex amplitude of the nth depth plane according to the depth interval, the corresponding part of the nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the diffraction field complex amplitude of the (n-1)th depth plane and the depth value of the nth depth plane, specifically comprising: calculating the intensity distribution of the target three-dimensional scene on the nth depth plane according to the corresponding part of the nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene; based on the angular spectrum theory, calculating the diffraction field complex amplitude of the nth depth plane according to the intensity distribution of the target three-dimensional scene on the nth depth plane, the depth value of the nth depth plane, the depth interval and the diffraction field complex amplitude of the (n-1)th depth plane.

6. The occlusion elimination method for generating multi-depth diffraction fields according to claim 1, characterized in that, Based on the angular spectrum theory, the complex amplitude of the diffraction field of the Nth depth plane is calculated according to the preset depth value of the target plane, the maximum value in the preset depth interval, the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the complex amplitude of the diffraction field of the (N-1)th depth plane, and the depth value of the Nth depth plane, which specifically includes: Calculating the depth difference according to the preset depth value of the target plane and the maximum value in the preset depth interval; Calculating the intensity distribution of the target three-dimensional scene at the Nth depth plane according to the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene; Based on the angular spectrum theory, the complex amplitude of the diffraction field of the Nth depth plane is calculated according to the intensity distribution of the target three-dimensional scene at the Nth depth plane, the depth value of the Nth depth plane, the depth difference, and the complex amplitude of the diffraction field of the (N-1)th depth plane.

7. A system for eliminating occlusion when generating multi-depth diffraction fields, characterized in that, Comprising: An acquisition module, configured to acquire an intensity image and a depth image of a target three-dimensional scene; A discretization module, configured to discretize the depth image of the target three-dimensional scene according to a preset depth interval and a preset total number of depth planes of the discrete depth image to obtain a discrete depth image; the discrete depth image includes the preset total number of depth planes, and depth values of the same depth plane are the same; A diffraction field complex amplitude calculation module, configured to, based on the angular spectrum theory, calculate the complex amplitude of the diffraction field of the Nth depth plane according to the preset depth value of the target plane, the maximum value in the preset depth interval, the corresponding part of the Nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the complex amplitude of the diffraction field of the (N-1)th depth plane, and the depth value of the Nth depth plane, where N is the preset total number of depth planes of the discrete depth image; wherein when 1<n<N, based on the angular spectrum theory, the complex amplitude of the diffraction field of the nth depth plane is calculated according to the depth interval, the corresponding part of the nth depth plane of the discrete depth image in the intensity image of the target three-dimensional scene, the complex amplitude of the diffraction field of the (n-1)th depth plane, and the depth value of the nth depth plane; when n=1, based on the angular spectrum theory, the complex amplitude of the diffraction field of the first depth plane is obtained according to the depth interval and the intensity image of the target three-dimensional scene; the depth interval is calculated according to the preset depth interval and the preset total number of depth planes of the discrete depth image.

8. The occlusion elimination system for generating multi-depth diffraction fields according to claim 7, characterized in that, The discretization module specifically comprises: A preset unit, configured to set the preset depth interval as [D1, D2], and set the preset total number of depth planes of the discrete depth image as N; Depth value calculation unit, used to calculate depth value according to formula Calculate the depth value d of the nth depth plane in a discrete depth image. n ; Discrete unit, used to convert grayscale values ​​in the depth image of the target 3D scene into d n-1 to d n The values ​​are all set to d n A discrete depth image is obtained.

9. An electronic device, characterized in that, Comprising: A memory and a processor, wherein the memory is configured to store a computer program, and the processor runs the computer program to cause the electronic device to execute the occlusion elimination method for generating a multi-depth diffraction field according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, Which stores a computer program that, when executed by a processor, implements the occlusion elimination method for generating a multi-depth diffraction field according to any one of claims 1 to 6.

Images