A method for eliminating interference from strong scattering bright targets in millimeter wave human body imaging

Abstract

The present invention provides a method for eliminating interference from strong scattered bright targets in millimeter wave human body imaging, including three-dimensional image reconstruction processing, three-dimensional image filtering, human body contour extraction and generation of two-dimensional visual images, utilizing the characteristics that the imaging results of dihedral angles or trihedral angles formed by natural structures such as human bones and skin are not on the surface of the human body contour, and eliminating strong bright spots and strong bright lines by extracting human body surface contour data. The present invention can greatly reduce the strong bright spots and strong bright lines in the two-dimensional millimeter wave visual image of the human body, make the image texture uniform, more realistically reflect the uniform characteristics of the human skin surface, avoid the influence of strong bright spots and strong bright lines on the judgment of suspicious items, and make the staff more accurate in identifying suspicious items on the human skin surface through human eyes or artificial intelligence, reduce the time waste caused by the staff's uncertainty or misjudgment of suspicious items, and improve the traffic efficiency of security inspection equipment.

Description

Technical Field

[0001] The invention relates to the field of measurement and testing technology, and in particular to a method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging. Background Art

[0002] In recent years, the types of items that cause harm have increased, and traditional human body security inspection methods can no longer meet current security needs. Traditional handheld metal detectors and metal detection gates can only detect metal contraband carried by the human body, and are powerless against non-metallic contraband such as plastic bombs and ceramic knives; although X-ray security inspection equipment can penetrate clothing to image all prohibited items, it poses a certain threat to human health and is not the best human body security inspection method. At present, millimeter wave three-dimensional imaging technology can achieve human body imaging through clothing, which is an effective method to replace traditional human body security inspection methods.

[0003] Millimeter wave three-dimensional imaging technology transmits broadband millimeter wave signals at multiple locations in space to illuminate human targets, receives and collects echo signals reflected by human targets, and obtains a three-dimensional millimeter wave human image of the human body through three-dimensional image reconstruction. In order to adapt to human visual habits or for automatic detection and processing of suspicious objects based on artificial intelligence, the three-dimensional image is usually projected into a two-dimensional space by taking the maximum value in the distance or summing the distance to obtain a two-dimensional millimeter wave visual image of the human body. However, natural structures such as human bones and skin often form dihedral angles or trihedral angles, which can easily cause strong bright spots or strong bright lines in the two-dimensional millimeter wave visual image of the human body, seriously interfering with the screening effect of dangerous goods carried by the human body. Summary of the invention

[0004] The present invention aims to solve the interference of strong bright spots or strong bright lines in millimeter wave three-dimensional imaging technology, and provides a method for eliminating interference of strong scattered bright targets in millimeter wave human body imaging. The method utilizes the characteristics that the imaging results of dihedral angles or trihedral angles formed by natural structures such as human bones and skin are not on the surface of the human body contour, and the strong bright spots and strong bright lines are eliminated by extracting the human body surface contour data.

[0005] The present invention provides a method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging, comprising the following steps:

[0006] S1. 3D image reconstruction processing: Use millimeter wave 3D imaging equipment to obtain human target echo data, and use 3D holographic imaging algorithm to process the human target echo data to obtain 3D imaging result S ijk , where 1≤i≤N y , 1≤j≤N x , 1≤k≤N z , N y N is the number of imaging points in the y direction of the three-dimensional imaging result. xN is the number of imaging points in the x direction of the three-dimensional imaging result. z is the number of imaging points in the z direction from the three-dimensional imaging result, the y direction is the vertical direction, and the x direction is the horizontal direction;

[0007] S2, 3D image filtering: 3D imaging result S ijk Perform filtering and normalization to obtain the normalized human body three-dimensional imaging result S′ ijk ;

[0008] S3. Human body contour extraction: Establishing the human body contour two-dimensional interpolation model matrix R mn , for the normalized human body 3D imaging result S′ ijk The distance to the maximum value ρ′ ij Perform two-dimensional interpolation to obtain the estimated human body contour position The three-dimensional imaging result S ijk The middle position is And optimize to get the human body contour data

[0009] S4. Generate two-dimensional visual image: extract three-dimensional imaging result S ijk The middle distance is Data from the image to obtain a two-dimensional visual image The interference of strong scattering bright targets is eliminated.

[0010] The method for eliminating interference of strongly scattered bright targets in millimeter wave human body imaging of the present invention is, as a preferred embodiment, in step S1, the human body target echo data is the human body target front echo data and / or the human body target back echo data;

[0011] The three-dimensional imaging result is a front three-dimensional imaging result and / or a back three-dimensional imaging result.

[0012] The method for eliminating interference of strong scattered bright targets in millimeter wave human body imaging of the present invention is, as a preferred embodiment, in step S2, the filtering method includes: median filtering and smoothing filtering, and the median filtering is used to filter out the interference of the three-dimensional imaging result S ijk The salt and pepper noise in the image is removed by smoothing filtering. ijk Random noise and coherent speckle noise in .

[0013] The method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging described in the present invention, as a preferred embodiment, step S3 includes the following steps:

[0014] S31. Establishing the human body contour two-dimensional interpolation model matrix R mn , where 1≤m≤M=α y N y , 1≤n≤N=α x Nx , M is the number of points of the two-dimensional interpolation model in the y direction, α y is the interpolation coefficient in the vertical direction, α y <1, N is the number of points in the x direction of the two-dimensional interpolation model, α x is the interpolation coefficient in the horizontal direction, α x <1;

[0015] S32, normalizing the human body three-dimensional imaging result S′ ijk The distance direction takes the maximum value ρ′ ij , record the maximum value ρ′ ij The position of the maximum value ρ′ ij Perform two-dimensional interpolation to obtain the initial value Where t is the number of iterations, 0≤t≤T;

[0016] S33, initial value Perform two-dimensional interpolation to obtain the estimated human body contour position

[0017] S34, based on the three-dimensional imaging results of the human body ijk Middle position The sum of the amplitude data is used as the objective function to calculate the objective function

[0018] S35. Judgment Is it true? If so, keep the matrix If not, the matrix is ​​not retained. The current value of the element in; changes the matrix The elements in return to step S32 until the matrix All elements in the calculation are completed;

[0019] S36, t=t+1, return to step S32, until the number of iterations t is T times, the optimization is completed, and the human body contour data is obtained

[0020] The method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging of the present invention is, as a preferred embodiment, in step S31, the human body contour two-dimensional interpolation model matrix R mn The value range is: 1≤R mn ≤N z .

[0021] The method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging described in the present invention is, as a preferred embodiment, in step S34, the objective function for:

[0022]

[0023] The method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging described in the present invention is, as a preferred embodiment, in step S4, the two-dimensional visual image for:

[0024]

[0025] The present invention has the following advantages:

[0026] The present invention can greatly reduce strong bright spots and strong bright lines in the two-dimensional millimeter wave visual image of the human body, make the image texture uniform, more truly reflect the uniform characteristics of the human skin surface, avoid the influence of strong bright spots and strong bright lines on the judgment of suspicious objects, and enable staff to judge suspicious objects on the surface of human skin with higher accuracy through human eyes or artificial intelligence, reduce time wasted due to staff uncertainty or misjudgment of suspicious objects, and improve the passage efficiency of security inspection equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A flow chart of a method for eliminating interference from strong scattering bright targets in millimeter wave human body imaging;

[0028] Figure 2 A method for eliminating interference from strong scattering bright targets in millimeter-wave human body imaging. A two-dimensional millimeter-wave visual image of the back of the human body;

[0029] Figure 3 A method for eliminating interference from strong scattering bright targets in millimeter wave human body imaging is a horizontal x-distance z section diagram of the bright spot in the human body three-dimensional imaging result;

[0030] Figure 4 A schematic diagram of human body contour data for a method of eliminating interference from strong scattering bright targets in millimeter wave human body imaging;

[0031] Figure 5 A two-dimensional visual diagram of a method for eliminating interference from strong scattering bright targets in millimeter-wave human imaging. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0033] Example 1

[0034] like Figure 1 As shown, a method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging includes the following steps:

[0035] S1. 3D image reconstruction processing: Use millimeter wave 3D imaging equipment to obtain human target echo data, and use 3D holographic imaging algorithm to process the human target echo data to obtain 3D imaging result S ijk , where 1≤i≤N y , 1≤j≤N x , 1≤k≤N z , N y N is the number of imaging points in the y direction of the three-dimensional imaging result. x N is the number of imaging points in the x direction of the three-dimensional imaging result. z is the number of imaging points in the z direction from the three-dimensional imaging result, the y direction is the vertical direction, and the x direction is the horizontal direction;

[0036] The human target echo data is the human target front echo data and / or the human target back echo data, and the three-dimensional imaging result is the front three-dimensional imaging result and / or the back three-dimensional imaging result.

[0037] S2, 3D image filtering: 3D imaging result S ijk Perform filtering and normalization to obtain the normalized human body three-dimensional imaging result S′ ijk ;

[0038] The filtering methods include: median filtering and smoothing filtering. The median filtering is used to filter out the three-dimensional imaging result S ijk The salt and pepper noise in the image is removed by smoothing filtering. ijk Random noise and coherent speckle noise in;

[0039] S3. Human body contour extraction: Establishing the human body contour two-dimensional interpolation model matrix R mn , for the normalized human body 3D imaging result S′ ijk The distance to the maximum value ρ′ ij Perform two-dimensional interpolation to obtain the estimated human body contour position The three-dimensional imaging result S ijk The middle position is And optimize to get the human body contour data

[0040] S31. Establishing the human body contour two-dimensional interpolation model matrix R mn , where 1≤m≤M=α y N y , 1≤n≤N=α x N x , M is the number of points of the two-dimensional interpolation model in the y direction, α y is the interpolation coefficient in the vertical direction, α y <1, N is the number of points in the x direction of the two-dimensional interpolation model, α x is the interpolation coefficient in the horizontal direction, αx <1;

[0041] Human body contour two-dimensional interpolation model matrix R mn The value range is: 1≤R mn ≤N z ;

[0042] S32, normalizing the human body three-dimensional imaging result S′ ijk The distance direction takes the maximum value ρ′ ij , record the maximum value ρ′ ij The position of the maximum value ρ′ ij Perform two-dimensional interpolation to obtain the initial value Where t is the number of iterations, 0≤t≤T;

[0043] S33, initial value Perform two-dimensional interpolation to obtain the estimated human body contour position

[0044] S34, based on the three-dimensional imaging results of the human body ijk Middle position The sum of the amplitude data is used as the objective function to calculate the objective function

[0045] Objective Function for:

[0046]

[0047] S35. Judgment Is it true? If so, keep the matrix If not, the matrix is ​​not retained. The current value of the element in; changes the matrix The elements in return to step S32 until the matrix All elements in the calculation are completed;

[0048] S36, t=t+1, return to step S32, until the number of iterations t is T times, the optimization is completed, and the human body contour data is obtained

[0049] S4. Generate two-dimensional visual image: extract three-dimensional imaging result S ijk The middle distance is Data from the image to obtain a two-dimensional visual image 2D visual image for:

[0050] The interference of strong scattering bright targets is eliminated.

[0051] Example 2

[0052] like Figure 1 As shown, a method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging includes the following steps:

[0053] 1. 3D image reconstruction processing

[0054] Use millimeter wave three-dimensional imaging equipment to obtain the echo data of the front (back) of the human target, use the three-dimensional holographic imaging algorithm to process the original echo data, take the amplitude of the processing result, and obtain the three-dimensional imaging result S of the front (back) of the human target ijk , where 1≤i≤N y , 1≤j≤N x , 1≤k≤N z , N y N is the number of imaging points in the vertical y direction of the three-dimensional imaging result. x N is the number of imaging points in the horizontal x direction of the three-dimensional imaging result. z is the number of imaging points in the distance direction z of the three-dimensional imaging result.

[0055] 2. 3D Image Filtering

[0056] (1) Using the median filtering method to filter out the salt and pepper noise in the 3D imaging results;

[0057] (2) Using smoothing filtering method to filter out random noise and coherent speckle noise in the three-dimensional imaging results;

[0058] (3) The filtered three-dimensional imaging result is normalized to obtain the normalized three-dimensional human body imaging result S′ ijk .

[0059] 3. Human body contour extraction

[0060] (1) Establish the human body contour two-dimensional interpolation model matrix R mn , where 1≤m≤M=α y N y , 1≤n≤N=α x N x , M is the number of points of the two-dimensional interpolation model in the vertical y direction, α y <1 is the interpolation coefficient in the vertical direction, N is the number of points in the horizontal x direction of the two-dimensional interpolation model, α x <1 is the interpolation coefficient in the horizontal direction. The matrix elements of the two-dimensional interpolation model usually have a value range of 1≤R mn ≤N z ;

[0061] (2) The filtered human body three-dimensional imaging result S′ ijk Take the maximum value in the distance direction and record the position of the maximum value ρ′ij . For ρ′ ij Perform two-dimensional interpolation to obtain the initial value

[0062] (3) Yes Perform two-dimensional interpolation to obtain the estimated human body contour position

[0063] (3) Based on the three-dimensional imaging results of the human body S ijk Middle position The sum of the amplitude data is used as the objective function to calculate the objective function

[0064]

[0065] (4) Change the matrix For each element in , execute steps (2) and (3). If The element's current value is retained.

[0066] (5) Repeat steps (2)(3)(4) T times, where T is the number of searches. The optimization is completed and the human body contour data is obtained.

[0067] 4. Generate 2D visual images

[0068] Extract the human body three-dimensional imaging results, the distance is The data is then used to obtain a two-dimensional image of the human target.

[0069]

[0070] like Figure 2 Shown is a two-dimensional millimeter-wave visual image of the back of a human body. The two bright spots and bright lines on the left and right within the white dotted frame in the image are the results of trihedral angle (bright spots) and dihedral angle (bright lines) imaging formed by natural structures such as human bones and skin.

[0071] like Figure 3 The figure shows the horizontal x-distance z section where the bright spot is located in the human body 3D imaging result. The imaging result of the bright spot is not at the same distance from the human body contour surface (within the human body contour surface). The white dotted line in the figure is the maximum position of the human body 3D imaging result at different horizontal x positions and distance z.

[0072] like Figure 4 The white dotted line is the human body contour data obtained by the method proposed by the present invention. Figure 5 The figure shows a two-dimensional visual image generated by the method proposed by the present invention, in which the two bright spots and bright lines on the left and right sides within the white dotted frame in the image are removed.

[0073] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for eliminating interference from strong scattering bright targets in millimeter wave human body imaging, characterized by: The following steps are involved: S1. 3D image reconstruction processing: using millimeter wave 3D imaging equipment to obtain human target echo data, and using 3D holographic imaging algorithm to process the human target echo data to obtain 3D imaging result S ijk , where 1≤i≤N y , 1≤j≤N x , 1≤k≤N z , N y N is the number of imaging points in the y direction of the three-dimensional imaging result. x N is the number of imaging points in the x direction of the three-dimensional imaging result. z is the number of imaging points in the z direction from the three-dimensional imaging result, the y direction is the vertical direction, and the x direction is the horizontal direction; S2, three-dimensional image filtering: filtering the three-dimensional imaging result S ijk Perform filtering and normalization to obtain the normalized human body three-dimensional imaging result S′ ijk ; S3. Human body contour extraction: Establishing the human body contour two-dimensional interpolation model matrix R mn , for the normalized human body three-dimensional imaging result S′ ijk The distance to the maximum value ρ′ ij Perform two-dimensional interpolation to obtain the estimated human body contour position The three-dimensional imaging result S ijk The middle position is And optimize to get the human body contour data S31, establish the human body contour two-dimensional interpolation model matrix R mn , where 1≤m≤M=α y N y , 1≤n≤N=α x N x , M is the number of points of the two-dimensional interpolation model in the y direction, α y is the interpolation coefficient in the vertical direction, α y <1, N is the number of points in the x direction of the two-dimensional interpolation model, α x is the interpolation coefficient in the horizontal direction, α x <1; S32, normalizing the human body three-dimensional imaging result S′ ijk The distance direction takes the maximum value ρ′ ij , record the maximum value ρ′ ij The position of the maximum value ρ′ ij Perform two-dimensional interpolation to obtain the initial value Where t is the number of iterations, 0≤t≤T; S33, initial value Perform two-dimensional interpolation to obtain the estimated human body contour position S34, based on the three-dimensional imaging results of the human body ijk Middle position The sum of the amplitude data is used as the objective function to calculate the objective function S35. Judgment Is it true? If so, keep the matrix If not, the matrix is ​​not retained. The current value of the element in; changes the matrix The elements in return to step S32 until the matrix All elements in the calculation are completed; S36, t=t+1, return to step S32, until the number of iterations t is T times, the optimization is completed, and the human body contour data is obtained S4, generating a two-dimensional visual image: extracting the three-dimensional imaging result S ijk The middle distance is Data from the image to obtain a two-dimensional visual image The interference of strong scattering bright targets is eliminated.

2. The method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging according to claim 1, characterized in that: In step S1, the human body target echo data is human body target front echo data and / or human body target back echo data; The three-dimensional imaging result is a front three-dimensional imaging result and / or a back three-dimensional imaging result.

3. The method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging according to claim 1, characterized in that: In step S2, the filtering method includes: median filtering and smoothing filtering, wherein the median filtering is used to filter out the three-dimensional imaging result S ijk The salt and pepper noise in the smoothing filter is used to filter out the three-dimensional imaging result S ijk Random noise and coherent speckle noise in .

4. The method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging according to claim 1, characterized in that: In step S31, the human body contour two-dimensional interpolation model matrix R mn The value range is: 1≤R mn ≤N z .

5. The method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging according to claim 1, characterized in that: In step S34, the objective function for:

6. The method for eliminating interference of strong scattering bright targets in millimeter wave human body imaging according to claim 1, characterized in that: In step S4, the two-dimensional visual image for:

Images