A polarization three-dimensional imaging method capable of characterizing the absolute depth of a target

Abstract

The present invention discloses a polarization three-dimensional imaging method capable of characterizing the absolute depth of a target, comprising: obtaining polarization images of the target at different polarization angles; obtaining the degree of polarization at each pixel point on the surface of the target object in the polarization image; obtaining the initial values of the azimuth angle and the initial values of the incident angle at different pixel points on the target surface; correcting the initial values of the azimuth angle and the incident angle; constructing a normal vector for each pixel point on the target surface using the corrected azimuth angle and the incident angle; reconstructing the three-dimensional contour of the target using the normal vector integral to obtain the surface function of the target; and combining the camera calibration and the surface function to obtain the absolute depth information of the target. The present invention recharacterizes the target surface normal vector using the component of the target surface normal vector along the depth direction as a unit. The three-dimensional model reconstructed based on the surface normal vector has a definite changing relationship with the absolute depth information of the target, breaking the limitation of traditional polarization three-dimensional imaging that displays depth as relative depth.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical imaging, and in particular relates to a polarization three-dimensional imaging method capable of characterizing the absolute depth of a target. Background Art

[0002] Currently, 3D imaging technology is increasingly being used in industrial production, facial recognition, and security detection. As an important branch of 3D imaging, polarization 3D imaging offers several advantages, including simple equipment, high precision, and limited application limitations. Therefore, research on polarization 3D imaging technology is of great significance. Previous polarization 3D imaging techniques interpreted the polarization information of the target's reflected light to determine the target surface normal vector, normalized the normal vector, and finally integrated it to obtain a 3D model of the target. However, the depth of the 3D model obtained by the traditional polarization 3D imaging method using normalized normal vectors is relative, which does not change with changes in the target detection distance and cannot represent the absolute depth of the target. Clearly, this does not conform to the future development trend of imaging.

[0003] A binocular polarization 3D reconstruction method has been proposed. This method realizes 3D reconstruction of the target by acquiring the polarization information of the target and combining it with the depth information obtained by Kinect. Using the target surface normal obtained by obtaining the depth information of the target through Kinect as a priori condition can effectively solve the multi-value problem of the target surface normal in polarization 3D imaging. Specifically, first, a camera is used to shoot polarization images at different angles, and the polarization degree of the target surface is calculated according to the Stokes vector method. The incident angle and azimuth of the target incident light can be calculated from the polarization degree. However, there is uncertainty in the incident angle and azimuth obtained by the polarization method. Therefore, it is necessary to use Kinect to obtain the surface normal information obtained by the depth information of the target for correction. The correction method is to make the component n of the target surface normal obtained by the polarization information on the x-axis and y-axis x and n y The direction of is consistent with the direction obtained from the depth information. Finally, after normalizing the corrected normal, the components of the normal on the x-axis and y-axis are integrated to reconstruct the 3D contour of the target.

[0004] The target depth information obtained by the binocular 3D imaging technology based on polarization information through Kinect is only used as the normal correction for the polarization 3D imaging. Since this method integrates the normalized surface normal vector to achieve the 3D reconstruction effect, the depth information obtained by the reconstructed 3D model is relative and can only reflect the distance between the target and the camera, but cannot represent the absolute depth information of the target. Summary of the Invention

[0005] To address the aforementioned problems in the prior art, the present invention provides a polarization-based three-dimensional imaging method capable of characterizing the absolute depth of a target. The technical problem to be solved by the present invention is achieved through the following technical solutions:

[0006] One aspect of the present invention provides a polarization three-dimensional imaging method capable of characterizing the absolute depth of a target, comprising:

[0007] Use the polarization three-dimensional imaging system to obtain polarization images of the target at different polarization angles;

[0008] Obtain the polarization degree at each pixel point on the surface of the target object in the polarization image;

[0009] Acquiring initial values of azimuth angles and initial values of incident angles at different pixel points on a target surface according to the polarization image and the polarization degree;

[0010] Correcting the initial value of the azimuth angle and the initial value of the incident angle to obtain corrected azimuth angle and incident angle;

[0011] Construct a normal vector for each pixel on the target surface using the corrected azimuth and incident angle;

[0012] The normal vector integral is used to reconstruct the three-dimensional contour of the target and obtain the surface height function of the target;

[0013] The absolute depth information of the target is obtained by combining the calibration of the camera in the polarization three-dimensional imaging system and the surface function.

[0014] In one embodiment of the present invention, obtaining polarization images of a target at different polarization angles using a polarization three-dimensional imaging system includes:

[0015] The reflected light from the surface of the target object is collected using a polarization 3D imaging system to obtain polarization images I′0, I′ of the target object scene at four polarization angles of 0°, 45°, 90°, and 135°. 45 , I′ 90 , I′ 135 ;

[0016] The polarization images I′0, I′ are segmented by using a threshold segmentation algorithm. 45 , I′ 90 , I′ 135 The target object and background are segmented to obtain the polarization images I0 and I 45 , I 90 , I 135 .

[0017] In one embodiment of the present invention, obtaining the polarization degree at each pixel point on the surface of the target object in the polarization image includes:

[0018] Acquiring the light intensity at each pixel in the polarization image;

[0019] The light intensity at each pixel is used to obtain the maximum light intensity, the minimum light intensity and the polarization phase angle of the reflected light at each pixel in the polarization image;

[0020] The polarization degree value at each pixel point is obtained using the maximum light intensity value and the minimum light intensity value.

[0021] In one embodiment of the present invention, obtaining initial values of azimuth angles and initial values of incident angles at different pixel points on the target surface according to the polarization image and the polarization degree includes:

[0022] Obtaining an initial azimuth angle value at pixel point u, where the initial azimuth angle value is equal to the polarization phase angle of the reflected light at pixel point u;

[0023] According to the relationship between the polarization degree and the incident angle of the incident light on the target surface, the initial value of the incident angle of the incident light at each pixel point is obtained.

[0024] In one embodiment of the present invention, correcting the initial value of the azimuth angle and the initial value of the incident angle includes:

[0025] The multi-value problem of the initial value of the incident angle is corrected to obtain a corrected incident angle. The correction formula is expressed as:

[0026]

[0027] G N G represents the gradient field change reference information obtained by solving the target surface intensity information. polar represents the gradient field parameter obtained by using the polarization information of the target reflected light, Represents the two-norm operation, if Then the initial value of the azimuth angle is the corrected azimuth angle. Then the initial azimuth angle value is flipped 180 degrees to obtain a corrected azimuth angle;

[0028] The initial value of the incident angle within a predetermined range near 90° is optimized using an optimization algorithm to avoid a maximum value of tanθ, thereby obtaining an optimized incident angle θ.

[0029] In one embodiment of the present invention, the optimization algorithm is:

[0030]

[0031] in, f(x) is the initial value of the input incident angle, and takes three values ​​recorded as T1, T0, T * , where T0 approaches T* ;

[0032] The optimized incident angle is:

[0033] in,

[0034] In one embodiment of the present invention, constructing a normal vector for each pixel point on the target surface using the corrected azimuth angle and incident angle includes:

[0035] According to the corrected azimuth and incident angle, the depth direction component is normalized to construct the normal vector:

[0036]

[0037] Among them, θ and are the optimized incident angle and corrected azimuth angle of the incident light on the target surface, respectively.

[0038] In one embodiment of the present invention, combining the calibration of the camera in the polarization 3D imaging system and the surface function to obtain the absolute depth information of the target includes:

[0039] According to the camera calibration, the change matrix of the pixel coordinates of the target surface in the world coordinate system, camera coordinate system, image coordinate system and pixel coordinate system is obtained, and the coordinates of the target surface in the world coordinate system (X w Y w Z w );

[0040] Establish the coordinates of the target in the world coordinate system and the three-dimensional coordinates (X r Y r Z r ) And calculate the scale change factor τ;

[0041] The coordinates of the target in the camera coordinate system are obtained according to the coordinates of the target in the world coordinate system (X c Y c Z c ), and obtain the absolute depth d at each point of the target abs :

[0042]

[0043] Among them, d abs (u) represents the absolute depth of the target at pixel u.

[0044] Another aspect of the present invention provides a storage medium storing a computer program for executing the steps of the polarization three-dimensional imaging method capable of characterizing the absolute depth of a target as described in any one of the above embodiments.

[0045] Another aspect of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor calls the computer program in the memory, it implements the steps of the polarization three-dimensional imaging method capable of characterizing the absolute depth of the target as described in any of the above embodiments.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] 1. The polarization 3D imaging method of the present invention recharacterizes the target surface normal using its depth component. Based on the relationship between the 3D model reconstructed from this surface normal and the target's actual detection distance, or absolute depth information, the method uses camera calibration and other methods to obtain a function that directly represents the target's absolute depth information within the reconstructed polarization 3D imaging model. This method overcomes the limitation of traditional polarization 3D imaging, which displays depth as relative depth, and provides a path for the diversification of future polarization 3D camera capabilities.

[0048] 2. The polarization 3D imaging method of the embodiment of the present invention proposes an optimization algorithm for the incident angle of incident light, which solves the problem that some points of the normal vector represented by the incident angle are not desirable.

[0049] 3. The polarization 3D model constructed by the polarization 3D imaging method of the present invention reconstructs a 3D surface that conforms to actual observation changes as the target detection distance changes, effectively showing the high-frequency changes of the target.

[0050] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a flow chart of a polarization three-dimensional imaging method capable of characterizing the absolute depth of a target provided by an embodiment of the present invention;

[0052] Figure 2 The azimuth angle of incident light on the surface of an object provided by an embodiment of the present invention is and a schematic diagram of the incident angle θ;

[0053] Figure 3 This is a schematic diagram of a change matrix of a target in a world coordinate system, a camera coordinate system, an image coordinate system, and a pixel coordinate system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following is a detailed description of a polarization three-dimensional imaging method capable of characterizing the absolute depth of a target proposed in accordance with the present invention, in conjunction with the accompanying drawings and specific embodiments.

[0055] The aforementioned and other technical contents, features, and effects of the present invention are clearly presented in the following detailed description of the specific embodiments in conjunction with the accompanying drawings. Through the description of the specific embodiments, a deeper and more specific understanding of the technical means and effects adopted by the present invention to achieve the intended purpose can be obtained. However, the accompanying drawings are provided for reference and illustration purposes only and are not intended to limit the technical solutions of the present invention.

[0056] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations are intended to cover non-exclusive inclusion, such that an article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the article or device comprising the element.

[0057] See Figure 1 , Figure 1 This is a flow chart of a polarization 3D imaging method capable of characterizing the absolute depth of a target, provided by an embodiment of the present invention. The polarization imaging method includes:

[0058] S1: Use the polarization 3D imaging system to obtain polarization images of the target at different polarization angles;

[0059] Specifically, the S1 includes:

[0060] S11: Use the polarization 3D imaging system to collect the reflected light from the surface of the target object, and obtain polarization images I′0, I′ of the target object scene at four polarization angles of 0°, 45°, 90° and 135° respectively. 45 , I′ 90 , I′ 135 .

[0061] Specifically, in a natural light environment, a CMOS (Complementary Metal Oxide Semiconductor) camera in the polarization detector of the polarization 3D imaging system is used to collect the reflected light from the object surface, thereby obtaining polarization images I′0, I′1 of the object scene at four angles of 0°, 45°, 90° and 135°.45 , I′ 90 , I′ 135 .

[0062] S12: Using a threshold segmentation algorithm to segment the polarization images I'0, I' 45 , I′ 90 , I′ 135 The target object and background are segmented to obtain the polarization images I0 and I 45 , I 90 , I 135 .

[0063] Specifically, in order to reduce the amount of calculation, it is necessary to segment the acquired polarization image of the target object scene and the scene depth information respectively, separate the target object from the background, and remove the background part. Therefore, this embodiment uses a threshold segmentation algorithm to separate the polarization image object and the background of the object scene. The polarization images after threshold segmentation are I0, I 45 , I 90 , I 135 .

[0064] S2: Obtain the polarization degree at each pixel point on the surface of the target object in the polarization image.

[0065] First, the light intensity at each pixel in the polarization image is obtained, and then the light intensity at each pixel is used to obtain the maximum light intensity, the minimum light intensity and the polarization phase angle of the reflected light at each pixel in the polarization image.

[0066] Specifically, the light intensity and the rotation angle ν of the polarizer in the polarization 3D imaging system (ν is the angle between the polarizer's transmission axis and the starting position) have the following relationship:

[0067]

[0068] Where u represents the pixel number, I(u,v) represents the light intensity at pixel u when the polarizer is rotated at an angle ν in the polarization 3D imaging system; I max and I min They represent the maximum and minimum light intensities observed by the CMOS camera at pixel u when the polarizer is rotated continuously for one circle, and φ represents the polarization phase angle of the reflected light at pixel u, that is, the polarization angle corresponding to the maximum light intensity observed at each pixel. Polarization images under multiple polarization angles are collected and recorded as I i (u) represents the intensity value at each pixel u of the polarization image collected by rotating the polarizer for the i-th time, then the above formula (1) can be rewritten as:

[0069]

[0070] in, v i Indicates the angle of the polarizer when it is rotated for the i-th time.

[0071] When the polarizer is rotated at least three times, Polarizer rotation angle v in the matrix i It is known that the matrix I i The light intensity value of each pixel in the collected polarization image represented by (u) is also known, so the coefficient matrix x can be obtained and is denoted as x = [x1x2x3] T , therefore, the maximum light intensity, minimum light intensity and polarization phase angle at each pixel in the polarization image are:

[0072]

[0073] Then, the polarization degree is calculated by the formula The polarization degree value at each pixel can be obtained.

[0074] S3: Acquire initial values of azimuth angles and initial values of incident angles at different pixel points on the target surface according to the polarization image and the polarization degree.

[0075] Specifically, the S3 includes:

[0076] S31: Obtaining an initial azimuth angle value at pixel point u, where the initial azimuth angle value is equal to the polarization phase angle φ of the reflected light at pixel point u.

[0077] Specifically, the polarization images I0, I 45 , I 90 , and I 135 According to step S2, the initial value of the azimuth angle of pixel u can be obtained, which is equal to the polarization phase angle φ of the reflected light at pixel u. However, since the polarization image intensity obtained by two rotation angles 180° apart is the same during the rotation of the polarizer, there is an uncertainty of 180° between the incident azimuth angle of the incident light to be reconstructed and the actual incident light azimuth angle, that is: Here, Λ represents a binary operator with a value of {0, 1}. In this embodiment, the value of Λ needs to be determined to eliminate the multi-value problem, which will be described in detail below.

[0078] S32: The incident angle can be calculated based on the relationship between the polarization degree and the incident angle of the incident light on the target surface. The calculation formula is:

[0079]

[0080] Here, n represents the refractive index of the surface of the target object. In this embodiment, the refractive index of the surface of the target object is generally set to 1.5.

[0081] S4: Correcting the initial value of the azimuth angle and the initial value of the incident angle to obtain corrected azimuth angle and incident angle.

[0082] Specifically, step S4 of this embodiment includes:

[0083] S41: Correcting the multi-value problem of the initial value of the incident angle to obtain a corrected incident angle.

[0084] Specifically, the relationship between the intensity gradient field of the target reflected light and its contour is used to convert the azimuth angle information of the target object surface into gradient field information. The correction process of the azimuth angle multi-value problem in the polarization 3D imaging scene is expressed as follows:

[0085]

[0086] G N G represents the gradient field change reference information obtained by solving the target surface intensity information. polar represents the gradient field parameter obtained by using the polarization information of the target reflected light, Represents the two-norm operation, if Then the initial value of the azimuth angle is the corrected azimuth angle. The initial azimuth angle value is then flipped 180° to obtain a corrected azimuth angle.

[0087] Specifically, the azimuth information correction of the target surface microfacet can be expressed by the following formula:

[0088]

[0089] in, represents the azimuth angle of the corrected target microfacet (where one pixel on the target surface represents one microfacet), and φ represents the polarization phase angle of the reflected light at pixel u. Ultimately, the normal vector information of the target surface microfacet can be uniquely solved.

[0090] S42: Optimizing the initial value of the incident angle within a predetermined range near 90° using an optimization algorithm to avoid a maximum value of tanθ, thereby obtaining an optimized incident angle θ.

[0091] Specifically, in the following process of obtaining the normal vector, the parameter tanθ needs to be introduced into the expression of the normal vector, which causes tanθ to be infinite when the incident angle θ is around 90°, and the incident angle has some undesirable points. In view of this, this embodiment proposes an optimization algorithm, which is as follows:

[0092] Assume the optimization algorithm is:

[0093]

[0094] in, f(x) is the initial value of the input incident angle, and takes three values ​​recorded as T1, T0, T * , where T0 approaches T * .

[0095] The optimized incident angle is:

[0096] in,

[0097] In this embodiment, T1 is selected as 85.5°, T * If 85° is selected and T0 is selected as 84.5°, then after the above algorithm optimizes all initial values of the incident angle, the original initial values of the incident angle between 85° and 90° are reduced to 85°, and the original angles around 85° and below are maintained. Therefore, this algorithm effectively solves the following problem of some points on the normal vector being undesirable.

[0098] S5: Construct a normal vector for each pixel on the target surface using the corrected azimuth and incident angle.

[0099] Specifically, according to the corrected azimuth and incident angle, the depth direction, i.e., the z-axis direction component, is normalized to construct the normal vector:

[0100]

[0101] Among them, θ and are the optimized incident angle and corrected azimuth angle of the incident light on the target surface, respectively.

[0102] S6: Use the normal vector integral to reconstruct the three-dimensional contour of the target and obtain the surface height function of the target.

[0103] Specifically, after obtaining the corrected target surface normal vector, the target surface is reconstructed by surface integral, and the normal vector is represented by the gradient field. At this time, the normal vector is represented as:

[0104]

[0105] Where z(u) = f(x,y) represents the height of the target surface. For globally continuous normal gradient fields, global or local integration methods are usually used. Global integration is more robust to noise and the reconstructed surface is smoother. However, for general target surfaces, there are discrete non-integrable regions, and it is impossible to directly reconstruct the three-dimensional surface by integration. In this case, the non-integrable region in the target normal vector gradient field information can be projected onto the integrable surface slope subspace, and the distance function between the normal gradient field and the continuous integrable micro-surface element can be defined as:

[0106] D{(p,q),(z x ,z y )}=∫∫|z x -p| 2 +|z y -q| 2 dxdy (10)

[0107] Among them, p and q represent the gradient field of the target microfacet normal vector in the x and y directions respectively, and z x and z y represent the partial derivatives of the surface function z(u) in the x and y directions respectively.

[0108] When the solution of the above equation reaches the minimum, the normal vector gradient field and the integrable microsurface are orthogonal projections. The surface function z(u) can be expressed as a linear combination of a series of orthogonal basis functions ψ(x, y; w), w=(w x ,w y ) represents a two-dimensional index. Then the surface function can be expressed as:

[0109]

[0110] Where C(w) is the expansion coefficient of z(u). Then the surface function gradient field is expressed as follows:

[0111]

[0112]

[0113] Among them, p g (u) represents the gradient field of the surface function along the x direction, q g (u) represents the gradient field of the surface function along the y-axis direction, and Represent the optimal coefficient sets in the x and y directions respectively. Select the Fourier coefficient expansion with a complete orthogonal basis, then the basis function can be represented as:

[0114]

[0115] Where M and N represent the dimensions of the target two-dimensional image along the x-axis and along the y-axis respectively. The surface function gradient field is expressed as follows:

[0116]

[0117]

[0118] Where α and β represent the Fourier coefficients of the discrete differential operator in the x and y directions, respectively.

[0119] Substituting equations (12), (13), and (14) into (11), we can establish a connection between the discrete gradient field information in the spatial domain and the surface function information in the Fourier domain to reconstruct the surface function and construct a three-dimensional polarization model of the object surface. The calculation formula is as follows:

[0120]

[0121] Among them, F{} and F -1 {} denote discrete Fourier transform and inverse transform respectively.

[0122] S7: Combining the calibration of the camera in the polarization 3D imaging system and the surface function, obtaining absolute depth information of the target.

[0123] In order to characterize the absolute depth of the target, this embodiment combines the camera calibration method. The purpose of camera calibration is to determine the relationship between the three-dimensional geometric position of a point on the surface of the spatial target and its corresponding point in the image. That is, the change matrix of the target in the world coordinate system, camera coordinate system, image coordinate system and pixel coordinate system can be calculated, such as Figure 3 shown.

[0124] The specific camera calibration process is as follows:

[0125] (a) Prepare calibration images

[0126] Calibration images are taken using a calibration plate at different positions, angles, and postures. A minimum of three images are taken, and 10-20 images are ideal. The calibration plate uses a checkerboard pattern consisting of alternating black and white rectangles.

[0127] (b) Extract corner information for each calibration image

[0128] Corner points are the inner corner points on the calibration board. These corner points do not touch the edge of the calibration board. The corner points are extracted by using the findChessboardCorners function.

[0129] (c) For each calibration image, further extract sub-pixel information

[0130] By using the cornerSubPix function, sub-pixel information is further extracted based on the initially extracted corner information to reduce the camera calibration error.

[0131] (d) Draw the found inner corner points on the chessboard calibration map

[0132] Use the drawChessboardCorners function to draw the successfully calibrated corner points.

[0133] (e) Camera calibration

[0134] Use the calibrateCamera function to calibrate and calculate the camera's intrinsic and extrinsic parameters. Through the above steps, the camera's intrinsic and extrinsic parameters are obtained, where the intrinsic parameters are:

[0135]

[0136] External parameters are:

[0137]

[0138] Among them, d x and d y Represents the distance represented by one pixel on the horizontal and vertical coordinates respectively, in mm, f represents the focal length of the camera, R represents a third-order rotation matrix, and T represents a third-order translation matrix.

[0139] The transformation from the world coordinate system to the pixel coordinate system is expressed as follows:

[0140]

[0141] Among them, (X w Y w Z w ) represents the coordinates in the target world coordinate system obtained by camera calibration.

[0142] Furthermore, the surface function Z(u) reconstructed in step S6 in units of the depth direction component has a certain scale change relationship with the coordinates of the target in the world coordinate system, namely:

[0143]

[0144] Among them, (X r Y r Z r ) is the three-dimensional coordinate of the target reconstructed by the surface function Z(u). By substituting it into formula (16), the scale change factor τ can be obtained from the internal and external parameters and the corresponding coordinates in the pixel coordinate system, so that the three-dimensional world coordinates of the target can be represented. Then, the world coordinate system is transformed to the camera coordinate system through the external parameter matrix, and the corresponding point on the camera coordinate system is obtained. Its coordinates are (X c Y c Z c ), and finally the absolute depth d at each point of the target can be obtained abs :

[0145]

[0146] Among them, d abs(u) represents the absolute depth of the target at pixel u. It should be noted that when the camera moves away from Δd in the depth direction, the scale change factor τ does not change, but the translation matrix T in the extrinsic parameter matrix will change accordingly, and the changed value is: Therefore, the absolute depth information will also change accordingly. In other words, the polarization 3D model constructed by the polarization 3D imaging method of the embodiment of the present invention reconstructs a 3D surface that conforms to the actual observed changes as the target detection distance changes, effectively showing the high-frequency changes of the target.

[0147] In summary, the polarization 3D imaging method of the embodiments of the present invention recharacterizes the target surface normal vector using its depth component as the unit. A 3D model reconstructed based on this surface normal vector and the actual detection distance of the target, i.e., the absolute depth information, are determined by the relationship function. This relationship function is then derived through camera calibration and other means. This allows the absolute depth information of the target to be directly represented in the reconstructed polarization 3D imaging model. This method overcomes the limitation of traditional polarization 3D imaging, which displays depth as relative depth, and provides a direction for the diversification of polarization 3D camera functions in the future.

[0148] Another embodiment of the present invention provides a storage medium storing a computer program for executing the steps of the polarization three-dimensional imaging method capable of characterizing the absolute depth of a target as described in the above-mentioned embodiment. Another aspect of the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor calls the computer program in the memory, the steps of the polarization three-dimensional imaging method capable of characterizing the absolute depth of a target as described in the above-mentioned embodiment are executed. Specifically, the above-mentioned integrated module implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium and includes several instructions for causing an electronic device (which can be a personal computer, server, or network device, etc.) or a processor to execute some of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0149] The above is a further detailed description of the present invention in conjunction with specific preferred embodiments, and the specific implementation of the present invention should not be considered to be limited to these descriptions. For those skilled in the art of the present invention, without departing from the concept of the present invention, several simple deductions or substitutions can be made, which should be considered to fall within the scope of protection of the present invention.

Claims

1. A polarization three-dimensional imaging method capable of characterizing the absolute depth of a target, characterized in that: include: Use the polarization three-dimensional imaging system to obtain polarization images of the target at different polarization angles; Obtain the polarization degree at each pixel point on the surface of the target object in the polarization image; Acquiring initial values of azimuth angles and initial values of incident angles at different pixel points on a target surface according to the polarization image and the polarization degree; Correcting the initial value of the azimuth angle and the initial value of the incident angle to obtain corrected azimuth angle and incident angle; Construct a normal vector for each pixel on the target surface using the corrected azimuth and incident angle; The normal vector integral is used to reconstruct the three-dimensional contour of the target and obtain the surface function of the target; Combining the calibration of the camera in the polarization 3D imaging system and the surface function, the absolute depth information of the target is obtained. The step of correcting the initial value of the azimuth angle and the initial value of the incident angle comprises: The multi-value problem of the initial azimuth angle is corrected to obtain a corrected azimuth angle. The correction formula is expressed as: Among them, G N G represents the gradient field change reference information obtained by solving the target surface intensity information. polar represents the gradient field parameter obtained by using the polarization information of the target reflected light, Represents the two-norm operation, if Then the initial value of the azimuth angle is the corrected azimuth angle. Then the initial azimuth angle value is flipped 180° to obtain a corrected azimuth angle; The optimization algorithm is used to optimize the initial value of the incident angle within a predetermined range near 90° to avoid the maximum value of tanθ, and the optimized incident angle θ is obtained. The optimization algorithm is: in, f(x) is the initial value of the input incident angle, and takes three values ​​recorded as T1, T0, T * , T1 is selected as 85.5°, T * Select 85° and T0 select 84.5°, where T0 is close to T * ; The optimized incident angle is: in, The optimization algorithm is used to reduce the initial value of the incident angle input between 85° and 90° to 85°, and keep the value around 85° and less than 85° unchanged.

2. The polarization three-dimensional imaging method capable of characterizing the absolute depth of a target according to claim 1, characterized in that: The polarization 3D imaging system is used to obtain polarization images of the target at different polarization angles, including: The reflected light from the surface of the target object is collected using a polarization 3D imaging system to obtain polarization images I′0, I′ of the target object scene at four polarization angles of 0°, 45°, 90°, and 135°. 45 , I′ 90 , I′ 135 ; The polarization images I′0, I′ are segmented by using a threshold segmentation algorithm. 45 , I′ 90 , I′ 135 The target object and background are segmented to obtain the polarization images I0 and I 45 , I 90 , I 135 .

3. The polarization three-dimensional imaging method capable of characterizing the absolute depth of a target according to claim 1, characterized in that: Obtain the polarization degree at each pixel on the surface of the target object in the polarization image, including: Acquiring the light intensity at each pixel in the polarization image; The light intensity at each pixel is used to obtain the maximum light intensity, the minimum light intensity and the polarization phase angle of the reflected light at each pixel in the polarization image; The polarization degree value at each pixel point is obtained using the maximum light intensity value and the minimum light intensity value.

4. The polarization three-dimensional imaging method capable of characterizing the absolute depth of a target according to claim 3, characterized in that: Acquiring initial values of azimuth angles and initial values of incident angles at different pixel points on a target surface according to the polarization image and the polarization degree, including: Obtaining an initial azimuth angle value at pixel point u, where the initial azimuth angle value is equal to the polarization phase angle of the reflected light at pixel point u; According to the relationship between the polarization degree and the incident angle of the incident light on the target surface, the initial value of the incident angle of the incident light at each pixel point is obtained.

5. The polarization three-dimensional imaging method capable of characterizing the absolute depth of a target according to claim 1, characterized in that: The normal vector is constructed for each pixel point on the target surface using the corrected azimuth and incident angle, including: According to the corrected azimuth and incident angle, the depth direction component is normalized to construct the normal vector: Among them, θ and are the optimized incident angle and corrected azimuth angle of the incident light on the target surface, respectively.

6. The polarization three-dimensional imaging method capable of characterizing the absolute depth of a target according to claim 1, characterized in that: Combining the calibration of the camera in the polarization 3D imaging system and the surface function to obtain absolute depth information of the target includes: According to the camera calibration, the change matrix of the pixel coordinates of the target surface in the world coordinate system, camera coordinate system, image coordinate system and pixel coordinate system is obtained, and the coordinates of the target surface in the world coordinate system (X w Y w Z w ); Establish the coordinates of the target in the world coordinate system and the three-dimensional coordinates (X r Y r Z r ) And the scale change factor τ is calculated, where the expression of the surface function Z(u) is: Among them, F{} and F -1 {} denotes discrete Fourier transform and inverse transform, α, β Represent the Fourier coefficients of the discrete differential operator in the x and y directions, M and N represent the dimensions of the target two-dimensional image along the x axis and along the y axis, respectively. g represents the gradient field of the surface function along the x direction, q g Represents the gradient field of the surface function along the y-axis; The coordinates of the target in the camera coordinate system are obtained according to the coordinates of the target in the world coordinate system (X c Y c Z c ), and obtain the absolute depth d at each point of the target abs : Among them, d abs (u) represents the absolute depth of the target at pixel u.

7. A storage medium, characterized in that: The storage medium stores a computer program, which is used to execute the steps of the polarization three-dimensional imaging method capable of characterizing the absolute depth of a target as described in any one of claims 1 to 6.

8. An electronic device, characterized in that: The method comprises a memory and a processor, wherein a computer program is stored in the memory, and when the processor calls the computer program in the memory, the steps of the polarization three-dimensional imaging method capable of characterizing the absolute depth of a target as claimed in any one of claims 1 to 6 are implemented.

Images