WO2020143352A1 - 安检设备及其图像检测方法 - Google Patents

安检设备及其图像检测方法 Download PDF

Info

Publication number
WO2020143352A1
WO2020143352A1 PCT/CN2019/121756 CN2019121756W WO2020143352A1 WO 2020143352 A1 WO2020143352 A1 WO 2020143352A1 CN 2019121756 W CN2019121756 W CN 2019121756W WO 2020143352 A1 WO2020143352 A1 WO 2020143352A1
Authority
WO
WIPO (PCT)
Prior art keywords
dimensional
resolution
image
security inspection
inspection device
Prior art date
Application number
PCT/CN2019/121756
Other languages
English (en)
French (fr)
Inventor
徐利民
祁春超
谭信辉
Original Assignee
深圳市华讯方舟太赫兹科技有限公司
华讯方舟科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳市华讯方舟太赫兹科技有限公司, 华讯方舟科技有限公司 filed Critical 深圳市华讯方舟太赫兹科技有限公司
Publication of WO2020143352A1 publication Critical patent/WO2020143352A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V8/00—Prospecting or detecting by optical means
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods

Definitions

  • the present application relates to the field of image processing, in particular to a security inspection device and an image detection method thereof.
  • Millimeter-wave security detectors use radar near-field imaging, and the system uses the performance of human bodies or objects in the millimeter wave band to reflect, reflect, and scatter on the human body under test. Or the object forms a three-dimensional image, and the hidden objects have different emission, reflection and scattering properties from the human body, and will appear in a different form from the human body on the image.
  • the imaging resolution of the millimeter wave security inspection instrument is generally not high, and the signal-to-noise ratio is relatively low.
  • the security inspection process it is easy to ignore the necessary information, so it is very important to improve the imaging resolution of the security inspection millimeter wave security inspection instrument.
  • the cost is too high.
  • the present application provides a security inspection device and an image detection method thereof to solve the problems of low resolution and low signal-to-noise ratio of the millimeter wave security inspection instrument in the prior art.
  • the present application proposes an image detection method applied to security inspection equipment.
  • the detection method includes: the security inspection equipment acquires at least one three-dimensional scanned image; the security inspection equipment projects at least one three-dimensional scanned image to obtain two-dimensional Low-resolution images; the security inspection equipment performs super-resolution reconstruction on the two-dimensional low-resolution images to obtain a two-dimensional high-resolution image, wherein the super-resolution reconstruction includes reverse iterative projection; the security equipment displays the two-dimensional high-resolution images.
  • this application proposes a security inspection device, the security inspection device at least includes: a scanning arm, a memory, a processor and a display; wherein the scanning arm scans the detection area and obtains the same A three-dimensional scanned image of the scene is stored in a memory; the processor extracts the three-dimensional scanned image from the memory and performs multi-angle projection processing to obtain multiple low-resolution two-dimensional images, the processor pairs The low-resolution two-dimensional image is subjected to super-resolution reconstruction processing to obtain a high-resolution two-dimensional image, and the high-resolution two-dimensional image is sent to the display; the display displays the high-resolution two-dimensional image on a display Interface.
  • the application of the present application to the image detection method of security inspection equipment includes: the security inspection equipment acquires at least one three-dimensional scanned image; the security inspection equipment projects at least one three-dimensional scanned image to obtain two-dimensional Low-resolution images; security inspection equipment performs reverse iterative projection reconstruction on two-dimensional low-resolution images to obtain two-dimensional high-resolution images; security inspection equipment displays two-dimensional high-resolution images.
  • the application improves the resolution of the image of the security inspection equipment, improves the detection and recognition rate of foreign objects by the security inspection equipment, reduces the false alarm rate, and improves the security inspection efficiency.
  • FIG. 1 is a schematic flowchart of an image detection method in an embodiment of this application.
  • FIG. 2 is a schematic structural view of the security inspection device in the embodiment of FIG. 1;
  • FIG. 3 is a schematic diagram of a flow of acquiring a three-dimensional scanned image by the security inspection device in the embodiment of FIG. 1;
  • FIG. 4 is a schematic flow chart of the projection of a three-dimensional scanned image by a security inspection device in the embodiment of FIG. 1;
  • FIG. 5 is a schematic flowchart of sub-pixel displacement between low-resolution images in the embodiment of FIG. 4;
  • FIG. 6 is another schematic flow chart of the projection of the three-dimensional scan image by the security inspection device in the embodiment of FIG. 4;
  • FIG. 7 is a schematic flow chart of image registration of a two-dimensional image by a security inspection device in the embodiment of FIG. 6;
  • FIG. 8 is a schematic flowchart of iteratively performing back projection iteration on a two-dimensional image by a security inspection device in the embodiment of FIG. 7;
  • FIG. 9 is an effect diagram of the back projection algorithm in the embodiment of FIG. 8;
  • FIG. 10 is a schematic flow chart of super-resolution reconstruction of a two-dimensional image by a security inspection device in the embodiment of FIG. 1;
  • FIG. 11 is a schematic diagram of the structure of the case equipment in the embodiment of the present application.
  • the millimeter wave human body security instrument is a kind of commonly used equipment in the field of human body security inspection.
  • the millimeter wave human body security instrument uses radar near-field imaging, and the system uses the performance of human body or objects in the millimeter wave band to emit, reflect and scatter Or the object forms a three-dimensional image, and the hidden objects have different emission, reflection and scattering properties from the human body, and will appear in a different form from the human body on the image.
  • the imaging resolution of the millimeter-wave security detector is generally not high, and the signal-to-noise ratio is relatively low. The cost of improving the imaging resolution by improving the security detector hardware is too high.
  • this application provides an image detection method applied to a millimeter wave security inspection instrument, specifically applying a super-resolution image processing technology to the millimeter wave security inspection instrument to improve image resolution and signal noise Than to improve the recognition rate of foreign object detection.
  • the image detection method adopted in this application is also suitable for nondestructive detection imaging based on terahertz radar array.
  • the security inspection equipment obtains multiple low-resolution radar images through multiple scans of the same scene, and then reconstructs multiple low-resolution images to obtain high-resolution images.
  • FIG. 1 is a schematic flowchart of an image detection method in an embodiment of the present application.
  • This embodiment provides an image detection method applied to a security inspection device.
  • the detection method includes:
  • the security inspection equipment acquires at least one three-dimensional scanned image
  • the security inspection equipment uses a radar near-field imaging method to scan the detection area to obtain at least one three-dimensional scan image; wherein, the security inspection equipment rotates back and forth multiple times around the detection area to scan the same scene in the detection area To obtain at least one three-dimensional image of the same scene.
  • the security inspection device projects at least one three-dimensional scanned image to obtain a two-dimensional low-resolution image
  • the security inspection device stores the scanned at least one three-dimensional scanned image to project the three-dimensional scanned image information onto the two-dimensional plane.
  • the security equipment can scan the same scene only once to obtain a three-dimensional image, and the three-dimensional image is projected onto a two-dimensional plane to obtain a low-resolution two-dimensional image; the security equipment carries out the low-resolution two-dimensional image Rotation displacement, through image registration, obtains multiple two-dimensional low-resolution images with sub-pixel displacement.
  • the security inspection equipment can perform multiple scans on the same scene to obtain multiple three-dimensional images.
  • the millimeter wave security inspection equipment has undergone multiple reciprocal scans due to errors in mechanical positioning accuracy and slight movement of the human body to be inspected.
  • Each three-dimensional radar imaging is different, so after projecting the multiple three-dimensional images onto the two-dimensional plane, multiple two-dimensional low-resolution images with sub-pixel displacement are obtained.
  • the security inspection device performs super-resolution reconstruction on the two-dimensional low-resolution image to obtain a two-dimensional high-resolution image, where the super-resolution reconstruction includes reverse iterative projection;
  • the security inspection equipment performs super-resolution reconstruction processing on the stored two-dimensional low-resolution images with sub-pixel displacement, which can pass through three other processes: image registration, interpolation and image restoration. Among them, in the image restoration process, the reverse Projection iteration algorithm.
  • the security inspection device displays a two-dimensional high-resolution image.
  • the security inspection equipment displays the two-dimensional high-resolution images calculated by the iterative algorithm of reverse projection, so that the security personnel can make timely manual judgments on the security inspection results; at the same time, the obtained two-dimensional high-resolution images are used for the subsequent depth-based The improvement of the accuracy of the learned image recognition algorithm is also helpful.
  • the application of this embodiment to the image detection method of security inspection equipment includes: the security inspection equipment acquires at least one three-dimensional scanned image; the security inspection equipment projects at least one three-dimensional scanned image to obtain a two-dimensional low-resolution image; the security inspection equipment Super-resolution reconstruction of the two-dimensional low-resolution image to obtain a two-dimensional high-resolution image, including reverse iterative projection; security inspection equipment displays a two-dimensional high-resolution image.
  • the application improves the resolution of the image of the security inspection equipment, improves the detection and recognition rate of foreign objects by the security inspection equipment, reduces the false alarm rate, and improves the security inspection efficiency.
  • FIG. 2 is a schematic structural diagram of the security inspection device in the embodiment of FIG. 1
  • FIG. 3 is a schematic flowchart of acquiring a three-dimensional scanned image by the security inspection device in the embodiment of FIG.
  • the steps of the security inspection device 200 acquiring at least one three-dimensional scanned image include:
  • S301 Two oppositely disposed scanning arms 210 of the security inspection device 200 rotate around the detection area to scan the same scene at least once.
  • the security inspection device 200 may be a cylindrical security scanner.
  • the security inspection device 200 includes two oppositely disposed scanning arms 210 and a detection area 211.
  • the scanning arm 210 may rotate around the detection area 211 to target persons to be tested in the detection area 211 To scan.
  • the scanning arm 210 can reciprocate around the detection area 211 and perform multiple scans to obtain multiple three-dimensional scanned images of the same scene.
  • the scanning arm 210 rotates and scans around the detection area 211, wherein the rotation angle of the single arm is not less than 120°, and the scanning coverage angle is not less than 120°, so as to establish at least one three-dimensional scan image of the same scene.
  • the security inspection device 210 can record the rotation angle at the same time, wherein the single arm in the scanning arm 210 needs to rotate at least 120° around the detection, so that the scanning angle covered by the scanning arm 210 is not less than 120 ° to create an effective and complete 3D scanned image.
  • two oppositely arranged scan arms are provided to scan the same scene at least once, and at the same time, the coverage angle of the single-arm scan is not less than 120°, so that the obtained three-dimensional scan image has a comparison
  • Complete information is conducive to subsequent image registration and super-resolution reconstruction of 3D scanned images.
  • FIG. 4 is a schematic flowchart of the projection of the three-dimensional scanned image by the security inspection device in the embodiment of FIG. 1.
  • Related reference numerals in other drawings are cited in FIG. 4, and the flow charts indicated by the reference signs are the same as those indicated by the same reference signs in other drawings.
  • the steps of the security inspection device projecting at least one three-dimensional scanned image to obtain multiple low-resolution two-dimensional images include:
  • the security inspection device can project multiple three-dimensional scanned images obtained by performing multiple scans on the same scene into multiple low-resolution two-dimensional images.
  • the security inspection device scans the detection area through the scanning arm to obtain at least one three-dimensional scanned image, and projects the at least one three-dimensional image. Among them, in this embodiment, the security inspection device performs projection processing on a plurality of scanned images.
  • each radar 3D imaging is different, so after projecting multiple 3D images onto a 2D plane, the Multiple 2D low-resolution images with sub-pixel displacement can be obtained.
  • the security inspection device may also project a three-dimensional scanned image obtained by scanning the same scene once into a low-resolution two-dimensional image, and the low-resolution two-dimensional image is rotated and slightly displaced to obtain multiple low-resolution two-dimensional images.
  • the security inspection device may also process the three-dimensional image obtained after one scan.
  • the low-resolution two-dimensional image obtained after projection of the three-dimensional image obtained after one scan must have accurate registration after the rotation displacement. In order to obtain multiple two-dimensional low-resolution images with sub-pixel displacement.
  • the security inspection equipment performs reverse projection reconstruction on the two-dimensional low-resolution image with sub-pixel displacement after precision registration to obtain a high-resolution two-dimensional image, and displays it, so that the security personnel can observe the high-resolution image ,Improve work efficiency.
  • FIG. 5 is a schematic flowchart of sub-pixel displacement between low-resolution images in the embodiment of FIG. 4;
  • FIG. 6 is another schematic flowchart of projection of a three-dimensional scanned image by a security inspection device in the embodiment of FIG. 4.
  • Related reference numerals in other drawings are cited in FIG. 6, and the flow charts indicated by the reference signs are the same as the processes indicated by the same reference signs in other drawings.
  • the step of the security inspection device projecting the three-dimensional scanned image to obtain multiple low-resolution two-dimensional images further includes:
  • the security inspection device sets a first rotation angle and a first projection angle
  • the scanning arm rotates around the detection area, and each time a fixed angle is rotated, the security inspection device stores the three-dimensional scanned image at the corresponding angle and projects it into a two-dimensional low-resolution image, where the fixed angle is the first rotation angle.
  • the scanning arm rotates around the detection area, and the rotation angle of the single arm is not less than 120°.
  • the security inspection equipment is provided with a first projection angle.
  • the rotation angle of the scanning arm is within the first projection angle, the scanning arm continues to rotate to scan the detection area
  • the security inspection equipment projects the scanned three-dimensional image information onto a two-dimensional plane to obtain a two-dimensional low-resolution image;
  • the security inspection equipment controls the scanning arm to stop scanning, or, the security inspection equipment controls The scanning arm performs reciprocating scanning.
  • S602 The scanning arm of the security inspection device rotates to be within the first projection angle, and the security inspection device sets the forward projection of the corresponding three-dimensional scanned image onto the two-dimensional plane to obtain multiple two-dimensional low-resolution images.
  • the first rotation angle can be set to 10°.
  • the security inspection device stores the three-dimensional image scanned by the scanning arm every 10°, and the three-dimensional image information is corrected. To the two-dimensional plane.
  • the security inspection device performs image registration on multiple low-resolution two-dimensional images to satisfy that multiple low-resolution two-dimensional images have sub-pixel displacement for the same scene.
  • the security inspection equipment projects three-dimensional scanned images into multiple low-resolution two-dimensional images according to the accuracy of the image registration algorithm.
  • FIG. 7 is a schematic flowchart of image registration performed by a security inspection device on a two-dimensional image in the embodiment of FIG. 6. Relevant reference numerals in other drawings are referenced in FIG. 7, and the flow charts indicated by the reference signs are consistent with the processes indicated by the same reference signs in other drawings.
  • Image registration includes:
  • the security inspection device is provided with a first threshold; wherein, the first threshold is the sum of squared maximum average errors of displacement distances of single pixels that satisfy the sub-pixel displacement condition.
  • the security inspection device uses an image registration algorithm based on gray scale and template to perform image registration processing on the three-dimensional scanned image to obtain multiple two-dimensional low-resolution images with sub-pixel displacement, which is specifically expressed as, The average error square sum algorithm is used.
  • the security inspection device controls the average square error of the displacement between corresponding pixels of at least two two-dimensional low-resolution images to be less than the first threshold.
  • the average error square sum algorithm is specifically implemented as follows: when projecting, the maximum displacement between corresponding pixels of two two-dimensional low-resolution images is set, where the correspondence between any two low-resolution two-dimensional images is set The sum of the squared average errors of the pixel displacements is less than the first threshold, thus ensuring that the registration accuracy of the two low-resolution images is high, that is, the sub-pixel displacement conditions for super-resolution reconstruction are met.
  • FIG. 8 is a schematic flowchart of the back projection iteration performed by the security inspection device on the two-dimensional image in the embodiment of FIG.
  • the security inspection equipment performs back projection algorithm processing on multiple two-dimensional low-resolution images to obtain two-dimensional high-resolution images.
  • the core algorithm is the reverse iteration algorithm.
  • the pixel size of the target high-resolution image is first determined, and according to the pixel size requirements of the target high-resolution image, a low-resolution image obtained by blurring, displacement, and downsampling of the high-resolution image is calculated.
  • the process of blurring, displacement and downsampling can be represented by matrix W.
  • the low-resolution image obtained in the embodiment of the present application is set to x
  • the calculated low-resolution image is set to y
  • the noise is n.
  • the relationship between each calculated low-resolution image and the low-resolution image obtained in this embodiment is as shown in equation (1).
  • W can be set to a two-dimensional Gaussian function or mean distribution.
  • the error between the low-resolution image obtained by the simulation calculation and the low-resolution image obtained in this embodiment is continuously projected to the HR image to achieve the correction effect.
  • the iteration stop condition ends with the difference of y-Wx reaching an allowable range.
  • hBP is the back projection kernel function.
  • the processing steps of the back projection algorithm include: the security inspection equipment sets an appropriate back projection kernel function h to perform back projection algorithm processing on multiple two-dimensional low-resolution images; wherein, the security inspection equipment sets the value of the back projection kernel function It is a matrix with all elements 1.
  • the back projection kernel function can be obtained arbitrarily, and the choice of the kernel function affects the iterative convergence speed.
  • a suitable function distribution that is, a matrix with all elements being 1, is selected to ensure the response speed requirements of the algorithm.
  • its kernel function may be set to Gaussian distribution.
  • FIG. 9 is an effect diagram of the back projection algorithm in the embodiment of FIG. 8; wherein, FIG. a is the original image, and FIG. b is the effect image of the four low-resolution images after the first back projection iteration, and FIG. c It is an effect picture after super-resolution reconstruction. It can be seen from the figure that the resolution and signal-to-noise ratio of the high-resolution image converged by the back projection algorithm have been significantly improved.
  • the high-resolution image obtained by super-resolution reconstruction The number of pixels is 4 times that of the original low-resolution image pixels.
  • FIG. 10 is a schematic diagram of a process of super-resolution reconstruction of a two-dimensional image by a security inspection device in the embodiment of FIG. 1; in an embodiment of the present application, the security inspection device performs super-resolution reconstruction processing on a two-dimensional low-resolution image with sub-pixel displacement, which can be Through the other three processes: image registration, interpolation and image restoration, in the image restoration process, iterative back projection algorithm can be used.
  • the security inspection device scans and projects the resulting three-dimensional image into a two-dimensional low-resolution image with sub-pixel displacement, and then iteratively projects the two-dimensional low-resolution image to obtain a two-dimensional high-resolution image, where,
  • the imaging effect of the two-dimensional high-resolution image is related to the number of low-resolution images obtained by scanning, the reconstruction algorithm, and the pixel size of the target high-resolution image.
  • FIG. 11 is a schematic structural diagram of a security inspection device in an embodiment of the present application.
  • An embodiment of the present application provides a security inspection device 200.
  • the security inspection device at least includes: a scanning arm 210, a memory 220, a processor 230, and a display 240; wherein, the scanning arm 210 scans the detection area and acquires a three-dimensional scan image of the same scene, Stored in the memory 220; the processor 230 extracts the three-dimensional scanned image from the memory 220 and performs multi-angle projection processing to obtain multiple low-resolution two-dimensional images, and the processor 230 performs super-resolution reconstruction processing on the low-resolution two-dimensional image to The high-resolution two-dimensional image is obtained, and the high-resolution two-dimensional image is sent to the display 240; the display 240 displays the high-resolution two-dimensional image on the display interface.
  • the security inspection device 200 in this embodiment cooperates with the scanning arm 210, the memory 220, the processor 230, and the display 240.
  • the three-dimensional scanned image obtained by the scanning arm is projected to have sub-pixels Displaced two-dimensional low-resolution image, which is calculated by reverse iterative projection calculation and becomes a two-dimensional high-resolution image; in this way, the security inspection equipment does not need to improve the resolution and signal-to-noise ratio by improving the security instrument hardware , Reducing costs, and at the same time, the application of super-resolution image processing technology to millimeter-wave security detectors, image resolution and signal-to-noise ratio, thereby improving the recognition rate of foreign object detection, reducing the false alarm rate, greatly improving the efficiency of security personnel .

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Geophysics (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)
  • Image Processing (AREA)
  • Emergency Alarm Devices (AREA)

Abstract

本申请公开了一种安检设备及其图像检测方法，该检测方法包括：安检设备获取至少一幅三维扫描图像；安检设备对至少一幅三维扫描图像进行投影，以得到二维低分辨图像；安检设备对二维低分辨图像进行超分辨重构，以获得二维高分辨图像，其中超分辨重构包括反向迭代投影；安检设备显示二维高分辨图像。通过这种方式，本申请在不改变安检设备硬件情况下，提高了安检设备图像的分辨率，提升了安检设备对异物的检测识别率，同时降低了虚警率，提高了安检效率。

Description

安检设备及其图像检测方法 技术领域
本申请涉及图像处理领域，特别是涉及一种安检设备及其图像检测方法。
背景技术
目前，在人体安检领域，毫米波人体安检仪的应用越来越广泛了，毫米波安检仪通过雷达近场成像，系统通过人体或物品在毫米波段的发射、反射和散射等性能对被测人体或物体形成三维图像，隐匿的物品由于具有与人体不同的发射、反射和散射性质，在图像上会以不同于人体的形态展现出来。
但是，现有技术中，毫米波安检仪成像一般分辨率不高，且信噪比比较低，在安检过程中，容易忽略必要信息，所以提高安检毫米波安检仪成像分辨率非常重要。然而通过改善安检仪硬件来提高成像分辨率，成本又太高。
技术解决方案
本申请提供一种安检设备及其图像检测方法，以解决现有技术中毫米波安检仪分辨率不高、信噪比比较低的问题。
为解决上述技术问题，本申请提出一种应用于安检设备图像检测方法，该检测方法包括：安检设备获取至少一幅三维扫描图像；安检设备对至少一幅三维扫描图像进行投影，以得到二维低分辨图像；安检设备对二维低分辨图像进行超分辨重构，以获得二维高分辨图像，其中所述超分辨重构包括反向迭代投影；安检设备显示二维高分辨图像。
为解决上述技术问题，本申请提出一种安检设备，所述安检设备至少包括：扫描臂，存储器，处理器和显示器；其中，所述扫描臂对所述检测区进行扫描，并获取所述同一场景的三维扫描图像，存储至存储器中；所述处理器从所述存储器中提取所述三维扫描图像并进行多角度的投影处理以获得多幅所述低分辨二维图像，所述处理器对所述低分辨二维图像进行超分辨重构处理以获得高分辨二维图像，并将所述高分辨二维图像发送至所述显示器；所述显示器将所述高分辨二维图像显示在显示界面上。
本申请的有益效果是：区别于现有技术，本申请应用于安检设备图像检测方法包括：安检设备获取至少一幅三维扫描图像；安检设备对至少一幅三维扫 描图像进行投影，以得到二维低分辨图像；安检设备对二维低分辨图像进行反向迭代投影重构，以获得二维高分辨图像；安检设备显示二维高分辨图像。通过这种方式，本申请在不改变安检设备硬件情况下，提高了安检设备图像的分辨率，提升了安检设备对异物的检测识别率，同时降低了虚警率，提高了安检效率。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本申请实施例中图像检测方法的流程示意图；
图2是图1实施例中安检设备的结构示意图；
图3是图1实施例中安检设备获取三维扫描图像的流程示意图；
图4是图1实施例中安检设备对三维扫描图像进行投影的流程示意图；
图5是图4实施例中低分辨图像之间的亚像素位移的流程示意图；
图6是图4实施例中安检设备对三维扫描图像进行投影的另一流程示意图；
图7是图6实施例中安检设备对二维图像进行图像配准的流程示意图；
图8是图7实施例中安检设备对二维图像进行反向投影迭代的流程示意图；
图9是图8实施例中反向投影算法的效果图；
图10是图1实施例中安检设备对二维图像进行超分辨重构的流程示意图；
图11是本申请实施例中案件设备的结构示意图。
本发明的实施方式
毫米波人体安检仪是人体安检领域的一种常用设备，在实际应用中，毫米波安检仪通过雷达近场成像，系统通过人体或物品在毫米波段的发射、反射和散射等性能对被测人体或物体形成三维图像，隐匿的物品由于具有与人体不同的发射、反射和散射性质，在图像上会以不同于人体的形态展现出来。而在现有技术中，毫米波安检仪成像一般分辨率不高，且信噪比比较低，通过改善安检仪硬件来提高成像分辨率的成本太高。
为解决现有技术中存在的问题，本申请提供一种应用于毫米波安检仪的图 像检测方法，具体将一种超分辨图像处理技术应用于毫米波安检仪，以提高图像分辨率和信噪比，从而提高异物检测识别率。同时，本申请所采用的图像检测方法同样适用于基于太赫兹雷达阵列的无损检测成像。本申请中，安检设备通过对同一场景的多次扫描，获得多幅低分辨的雷达成像，再对多幅低分辨图像进行重构，得到高分辨图像。
为使本领域的技术人员更好地理解本申请的技术方案，下面结合附图和具体实施方式对发明所提供的一种应用于安检设备的图像检测方法做进一步详细描述。
参阅图1，图1是本申请实施例中图像检测方法的流程示意图。本实施例提供一种应用于安检设备的图像检测方法，该检测方法包括：
S101：安检设备获取至少一幅三维扫描图像；
在本实施例中，安检设备采用雷达近场成像方法，对检测区进行扫描，以得到至少一幅三维扫描图像；其中，安检设备绕检测区往复旋转多次以对检测区内同一场景进行扫描，获得所述同一场景的至少一幅所述三维图像。
S102：安检设备对至少一幅三维扫描图像进行投影，以得到二维低分辨图像；
在本实施例中，安检设备将扫描得到的至少一幅三维扫描图像存储，以将该三维扫描图像信息投影至二维平面上。
可选的是，安检设备可对同一场景只进行一次扫描，得到一幅三维图像，该三维图像投影至二维平面，得到一幅低分辨二维图像；安检设备对该低分辨二维图像进行旋转位移，通过图像配准，得到具备亚像素位移的多幅二维低分辨图像。
可选的是，安检设备可对同一场景进行多次扫描，得到多幅三维图像，其中，毫米波安检设备在多次往复扫描中，由于机械定位精度的误差和待检测人员人体的轻微移动，每一次雷达三维成像都不一样，因此该多幅三维图像投影至二维平面后，得到具备亚像素位移的多幅二维低分辨图像。
S103：安检设备对二维低分辨图像进行超分辨重构，以获得二维高分辨图像，其中所述超分辨重构包括反向迭代投影；
安检设备对存储的具有亚像素位移的二维低分辨图像进行超分辨重构处理，可通过以外三个流程：图像配准、插值和图像复原，其中，在图像复原过程中，可采用反向投影迭代算法。
S104：安检设备显示二维高分辨图像。
安检设备将经过反向投影迭代算法计算后得到的二维高分辨图像显示出来，以使安检人员可以对安检结果做出及时的人工判断；同时，得到的二维高分辨图像对于后续的基于深度学习的图像识别算法的准确率的提升也有帮助。
区别于现有技术，本实施例应用于安检设备图像检测方法包括：安检设备获取至少一幅三维扫描图像；安检设备对至少一幅三维扫描图像进行投影，以得到二维低分辨图像；安检设备对二维低分辨图像进行超分辨重构，以获得二维高分辨图像，其中包括反向迭代投影；安检设备显示二维高分辨图像。通过这种方式，本申请在不改变安检设备硬件情况下，提高了安检设备图像的分辨率，提升了安检设备对异物的检测识别率，同时降低了虚警率，提高了安检效率。
参阅图2和图3，图2是图1实施例中安检设备的结构示意图，图3是图1实施例中安检设备获取三维扫描图像的流程示意图。安检设备200获取至少一幅三维扫描图像的步骤包括：
S301：安检设备200的两个相对设置的扫描臂210环绕检测区旋转，以对同一场景进行至少一次扫描。
安检设备200可为圆柱形安检扫描仪，其中，安检设备200包括两个相对设置的扫描臂210以及检测区211，扫描臂210可绕检测区211旋转，对位于检测区211内的待测人员进行扫描。扫描臂210可绕检测区211往复旋转，进行多次扫描，以得到同一场景的多个三维扫描图像。
S302：扫描臂210绕检测区211旋转扫描，其中，单臂旋转的角度不小于120°，扫描覆盖角度不小于120°，以建立同一场景的至少一幅三维扫描图像。
扫描臂210绕检测区211旋转扫描时，安检设备210可同时记录旋转角度，其中，扫描臂210中的单臂绕检测需旋转不小于120°，以使扫描臂210覆盖的扫描角度不小于120°，以建立有效完整的三维扫描图像。
区别于现有技术，本实施例通过设置有两个相对设置的扫描臂，对同一场景进行至少一次扫描，同时，单臂扫描覆盖角度不小于120°，以使得所获得的三维扫描图像具有比较完整的信息，有利于后续对三维扫描图像进行图像配准和超分辨重构。
参阅图4，图4是图1实施例中安检设备对三维扫描图像进行投影的流程示意图。图4中引用了其他附图中的相关标号，该标号指代的流程图与其他附图 中相同标号指代的流程一致。安检设备对至少一幅三维扫描图像进行投影得到多幅低分辨二维图像步骤包括：
S401：安检设备可将对同一场景进行多次扫描得到的多幅三维扫描图像投影成多幅低分辨二维图像。
安检设备通过扫描臂对检测区进行扫描，获得至少一幅三维扫描图像，并对至少一幅三维图像进行投影。其中，在本实施例中，安检设备对多幅扫描图像进行投影处理。
毫米波安检设备在多次往复扫描中，由于机械定位精度的误差和待检测人员人体的轻微移动，每一次雷达三维成像都不一样，因此获得的多幅三维图像投影至二维平面后，就能得到具备亚像素位移的多幅二维低分辨图像。
S402：安检设备还可将对同一场景的进行一次扫描后得的三维扫描图像投影成低分辨二维图像，低分辨二维图像进行旋转、微小位移，以得到多幅低分辨二维图像。
在本实施例中，安检设备还可以对一次扫描后得到的三维图像进行处理，其中，一次扫描后得到的三维图像投影后得到的低分辨二维图像，其旋转位移必须具有精度配准后，才能得到具备亚像素位移的多幅二维低分辨图像。
安检设备对进行精度配准后的具备亚像素位移的二维低分辨图像进行反向投影重构，得到高分辨的二维图像，并将其显示出来，以使安检人员能通过观察高分辨图像，提高工作效率。
参阅图5和图6，图5是图4实施例中低分辨图像之间的亚像素位移的流程示意图；图6是图4实施例中安检设备对三维扫描图像进行投影的另一流程示意图。图6中引用了其他附图中的相关标号，该标号指代的流程图与其他附图中相同标号指代的流程一致。安检设备对所述三维扫描图像进行投影得到多幅所述低分辨二维图像的步骤进一步包括：
S601：所述安检设备设置第一旋转角度和第一投影角度；
扫描臂绕检测区旋转，每旋转一个固定角度，安检设备将对应角度的三维扫描图像存储，并投影成二维低分辨图像，其中，该固定角度为第一旋转角度。
扫描臂绕检测区旋转，单臂旋转角度不小于120°，其中，安检设备设置有第一投影角度，当扫描臂转动角度在第一投影角度内时，扫描臂继续转动，对检测区进行扫描，同时，安检设备将扫描所得的三维图像信息投射至二维平面上，以获得二维低分辨图像；当扫描臂转动至第一投影角度，安检设备控制扫 描臂停止扫描，或者，安检设备控制扫描臂进行往复扫描。
S602：安检设备的扫描臂旋转至在第一投影角度内，安检设备设置将对应的三维扫描图像正向投影至二维平面，以获得多幅二维低分辨图像。
可选的是，第一旋转角度可设置为10°，当扫描臂开始绕检测区旋转时，扫描臂每旋转10°，安检设备将扫描臂扫描所得的三维图像存储，并将三维图像信息正向透射至二维平面。
S603：安检设备对多幅低分辨二维图像进行图像配准，以满足多幅低分辨二维图像对同一场景具备亚像素位移。
安检设备根据图像配准算法精度，将三维扫描图像投影成多个低分辨二维图像。
参阅图7，图7是图6实施例中安检设备对二维图像进行图像配准的流程示意图。图7中引用了其他附图中的相关标号，该标号指代的流程图与其他附图中相同标号指代的流程一致。图像配准包括：
S701：安检设备设置有第一阈值；其中，第一阈值为满足亚像素位移条件的单像素的位移距离的最大平均误差平方和。
在本申请实施例中，安检设备采用基于灰度和模板的图像配准算法，对三维扫描图像进行图像配准处理，以获得具备亚像素位移的多幅二维低分辨图像，具体表现为，采用平均误差平方和算法。
S702：安检设备在对三维扫描图像进行投影时，控制至少两幅二维低分辨图像的对应像素点之间的位移的平均误差平方和小于第一阈值。
平均误差平方和算法具体实施为：在进行投影时，对两幅二维低分辨图像的对应像素点之间的最大位移进行设置，其中，设定任意两幅低分辨二维图像之间的对应像素点的位移的平均误差平方和小于第一阈值，从而保证了两幅低分辨图像的配准精度较高，即满足超分辨重构的亚像素位移条件。
参阅图8，图8是图7实施例中安检设备对二维图像进行反向投影迭代的流程示意图。安检设备对多幅二维低分辨图像进行反向投影算法处理，以得到二维高分辨图像。
在本实施例中，采用核心算法为反向迭代算法。
在本实施例中，先确定目标高分辨图像的像素大小，根据目标高分辨图像的像素大小要求，计算求得高分辨率图像经过模糊、位移、降采样得到的低分辨率图像。其中，模糊、位移和降采样的过程可以用矩阵W表示。
设定本申请实施例中获得的低分辨图像设定为x，计算得到的低分辨率图像设为y，噪声为n。每一幅计算得到的低分辨率的图像与本实施例中获得的低分辨图像之间的关系如式(1)。
y k＝Wx+n k,1≤k≤p   (1)
在毫米波成像系统中，W可以设定为一个二维分布的高斯函数或均值分布。
再者，将模拟计算得到的低分辨率图像与本实施例中获得的低分辨率图像之间误差不断投射到HR图像以达到修正效果。其中，迭代停止条件以y-Wx的差值达到一个可允许的范围内为止。用反向迭代算法对高分辨图像的估计表达式如下：
Figure PCTCN2019121756-appb-000001
其中，hBP是反向投影核函数。
反向投影算法处理步骤包括：安检设备设定合适的反向投影核函数h，以对多幅二维低分辨图像进行反向投影算法处理；其中，安检设备设定反向投影核函数取值为全部元素为1的矩阵。
在理论上，反向投影核函数可任意获取，其核函数的选取影响迭代收敛速度，在本实施例中，选择合适的函数分布，即全部元素为1的矩阵，确保算法的响应速度要求。在其他实施例中，其核函数可设置为高斯分布。
参阅图9，图9是图8实施例中反向投影算法的效果图；其中，图a是原图，图b是四幅低分辨图像经过第一次反向投影迭代后的效果图，图c是经过超分辨重构后的效果图，从图中可见，经过反向投影算法所收敛的高分辨图像的分辨率和信噪比有了明显的改善，经过超分辨重构获得的高分辨图像的像素点数是原有低分辨图像像素的4倍。
图10是图1实施例中安检设备对二维图像进行超分辨重构的流程示意图；本申请实施例中，安检设备对具有亚像素位移的二维低分辨图像进行超分辨重构处理，可通过以外三个流程：图像配准、插值和图像复原，其中，在图像复原过程中，可采用反向投影迭代算法。
在本实施例中，安检设备扫描将所得的三维图像投影成具备亚像素位移的二维低分辨图像，并将该二维低分辨图像经过反向迭代投影，得到二维高分辨图像，其中，该二维高分辨图像的成像效果与扫描获得的低分辨图像的数量、重构算法以及目标高分辨图像的像素大小有关。
参阅图11，图11是本申请实施例中安检设备的结构示意图。本申请实施例提供一种安检设备200，安检设备至少包括：扫描臂210，存储器220，处理器230和显示器240；其中，扫描臂210对检测区进行扫描，并获取同一场景的三维扫描图像，存储至存储器220中；处理器230从存储器220中提取三维扫描图像并进行多角度的投影处理以获得多幅低分辨二维图像，处理器230对低分辨二维图像进行超分辨重构处理以获得高分辨二维图像，并将高分辨二维图像发送至显示器240；显示器240将高分辨二维图像显示在显示界面上。
区别于现有技术，本实施例中的安检设备200通过扫描臂210，存储器220，处理器230和显示器240的互相配合，经过图像配准，将通过扫描臂获得三维扫描图像投影成具备亚像素位移的二维低分辨图像，该二维低分辨图像经过反向迭代投影计算，变成二维高分辨图像；通过这种方式，安检设备无需通过改善安检仪硬件来提高分辨率和信噪比，降低了成本，同时，将超分辨图像处理技术应用于毫米波安检仪，图像分辨率和信噪比提高，从而提高异物检测识别率，降低了虚警率，大大提高了安检人员的工作效率。
以上所述仅为本申请的实施方式，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (10)

  1. 一种应用于安检设备的图像检测方法，其特征在于，所述检测方法包括：
    所述安检设备获取至少一幅三维扫描图像；
    所述安检设备对所述至少一幅三维扫描图像进行投影，以得到二维低分辨图像；
    所述安检设备对所述二维低分辨图像进行超分辨重构，以获得二维高分辨图像，其中超分辨重构包括反向迭代投影；
    所述安检设备显示所述二维高分辨图像。
  2. 根据权利要求1所述的图像检测方法，其特征在于，所述安检设备获取至少一幅三维扫描图像的步骤包括：
    所述安检设备的两个相对设置的扫描臂环绕检测区往复旋转，以对同一场景进行至少一次扫描。
  3. 根据权利要求2所述的图像检测方法，其特征在于，所述扫描臂绕所述检测区旋转时，单臂旋转的角度小于120°，扫描覆盖角度不小于120°，以建立所述同一场景的所述至少一幅三维扫描图像。
  4. 根据权利要求2所述的图像检测方法，其特征在于，所述安检设备对所述至少一幅三维扫描图像进行投影得到所述多幅低分辨二维图像步骤包括：
    所述安检设备将对所述同一场景进行多次扫描得到的所述多幅三维扫描图像投影成多幅所述低分辨二维图像；或者，
    所述安检设备将对所述同一场景的进行一次扫描后得到的三维扫描图像投影成所述低分辨二维图像，所述低分辨二维图像进行旋转、微小位移，以得到多幅所述低分辨二维图像。
  5. 根据权利要求4所述的图像检测方法，其特征在于，所述安检设备对所述三维扫描图像进行投影得到多幅所述低分辨二维图像的步骤进一步包括：
    所述安检设备设置第一旋转角度和第一投影角度；
    所述安检设备的所述扫描臂旋转至在第一投影阈限内，所述安检设备设置将所述三维扫描图像正向投影至二维平面，以获得多幅所述二维低分辨图像；
    所述安检设备对多幅所述低分辨二维图像进行图像配准，以满足多幅所述低分辨二维图像对所述同一场景具备亚像素位移。
  6. 根据权利要求5所述的图像检测方法，其特征在于，所述图像配准包括：
    所述安检设备设置有第一阈值；
    所述安检设备在对所述三维扫描图像进行投影时，控制至少两幅所述二维低分辨图像的对应像素点之间的位移的平均误差平方和小于所述第一阈值；
    其中，所述第一阈值为满足亚像素位移条件的单像素的位移距离的最大平均误差平方和。
  7. 权利要求6所述的图像检测方法，其特征在于，所述安检设备对多幅所述二维低分辨图像进行反向投影算法处理，以得到所述二维高分辨图像。
  8. 权利要求7述的图像检测方法，其特征在于，所述反向投影算法处理步骤包括：
    所述安检设备设定合适的反向投影核函数h，以对多幅所述二维低分辨图像进行反向投影算法处理。
  9. 权利要求8的图像检测方法，其特征在于，所述反向投影核函数取值为全部元素为1的矩阵。
  10. 一种安检设备，其特征在于，所述安检设备至少包括：
    扫描臂，存储器，处理器和显示器；
    其中，所述扫描臂对检测区进行扫描，并获取同一场景的三维扫描图像，存储至所述存储器中；所述处理器从所述存储器中提取所述三维扫描图像并进行多角度的投影处理以获得多幅低分辨二维图像，所述处理器对所述低分辨二维图像进行超分辨重构处理以获得高分辨二维图像，并将所述高分辨二维图像发送至所述显示器；所述显示器将所述高分辨二维图像显示在显示界面上。
PCT/CN2019/121756 2019-01-10 2019-11-28 安检设备及其图像检测方法 WO2020143352A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910024567.5A CN109886872B (zh) 2019-01-10 2019-01-10 安检设备及其图像检测方法
CN201910024567.5 2019-01-10

Publications (1)

Publication Number Publication Date
WO2020143352A1 true WO2020143352A1 (zh) 2020-07-16

Family

ID=66925858

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/121756 WO2020143352A1 (zh) 2019-01-10 2019-11-28 安检设备及其图像检测方法

Country Status (2)

Country Link
CN (1) CN109886872B (zh)
WO (1) WO2020143352A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109886872B (zh) * 2019-01-10 2023-05-16 深圳市重投华讯太赫兹科技有限公司 安检设备及其图像检测方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003011764A2 (en) * 2001-08-03 2003-02-13 Volker Westphal Real-time imaging system and method
CN102359971A (zh) * 2011-09-14 2012-02-22 上海英迈吉东影图像设备有限公司 一种实现单源多视角安全检查的方法及系统
CN102460703A (zh) * 2009-05-26 2012-05-16 拉皮斯坎系统股份有限公司 用于高数据率的x 射线层析扫描仪的成像、数据捕获、数据传输和数据分发方法以及系统
CN107228867A (zh) * 2017-06-21 2017-10-03 同方威视技术股份有限公司 安检图像显示方法、设备和安检系统
CN109886872A (zh) * 2019-01-10 2019-06-14 深圳市华讯方舟太赫兹科技有限公司 安检设备及其图像检测方法

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102681023B (zh) * 2012-04-19 2014-05-07 首都师范大学 一种太赫兹波成像系统
CN103870836A (zh) * 2014-01-10 2014-06-18 浙江工业大学 基于凸集投影超分辨率重构的合成孔径雷达图像目标识别方法
US9639973B2 (en) * 2014-04-04 2017-05-02 Decision Sciences International Corporation Muon tomography imaging improvement using optimized limited angle data
CN104008539B (zh) * 2014-05-29 2017-02-15 西安理工大学 基于多尺度几何分析的图像超分辨率重建方法
CN105069748B (zh) * 2015-07-16 2017-11-10 哈尔滨工业大学 一种基于微小卫星物方扫描技术获取高分辨率图像的方法
CN105974486A (zh) * 2016-04-27 2016-09-28 华讯方舟科技有限公司 鞋内物品检测设备
CN107155096B (zh) * 2017-04-19 2019-07-12 清华大学 一种基于半误差反向投影的超分辨率重建方法及装置
CN108549112A (zh) * 2018-04-24 2018-09-18 貊大卫 一种高速实时x射线计算机断层扫描安检系统
CN108845367A (zh) * 2018-05-30 2018-11-20 深圳市华讯方舟太赫兹科技有限公司 一种安检系统以及安检方法
CN109116433B (zh) * 2018-09-03 2020-02-04 深圳市华讯方舟太赫兹科技有限公司 一种安检设备及其成像方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003011764A2 (en) * 2001-08-03 2003-02-13 Volker Westphal Real-time imaging system and method
CN102460703A (zh) * 2009-05-26 2012-05-16 拉皮斯坎系统股份有限公司 用于高数据率的x 射线层析扫描仪的成像、数据捕获、数据传输和数据分发方法以及系统
CN102359971A (zh) * 2011-09-14 2012-02-22 上海英迈吉东影图像设备有限公司 一种实现单源多视角安全检查的方法及系统
CN107228867A (zh) * 2017-06-21 2017-10-03 同方威视技术股份有限公司 安检图像显示方法、设备和安检系统
CN109886872A (zh) * 2019-01-10 2019-06-14 深圳市华讯方舟太赫兹科技有限公司 安检设备及其图像检测方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
WU, XUANPEI ET AL: "Super-resolution of Color Images Based on Iterative Back Projection ", COMPUTER & DIGITAL ENGINEERING, vol. 43, no. 6, 30 June 2015 (2015-06-30), pages 1113 - 1117, XP009522049, ISSN: 1672-9722 *

Also Published As

Publication number Publication date
CN109886872B (zh) 2023-05-16
CN109886872A (zh) 2019-06-14

Similar Documents

Publication Publication Date Title
US7950849B2 (en) Method and device for geometry analysis and calibration of volumetric imaging systems
US8938111B2 (en) Computed tomography imaging process and system
CN109580630A (zh) 一种机械零部件缺陷的视觉检测方法
CN100581479C (zh) 三维超声影像的重建方法
WO2018126335A1 (zh) 基于小球模体的锥束ct系统几何参数评价及校正方法
US20220182593A1 (en) Method for high-precision true color three-dimensional reconstruction of a mechanical component
US20190197764A1 (en) Three-dimensional imaging method and system
CN114372992A (zh) 一种基于动平台的边缘角点检测四目视觉算法
CN114998448B (zh) 一种多约束双目鱼眼相机标定与空间点定位的方法
WO2020143352A1 (zh) 安检设备及其图像检测方法
CN105319225B (zh) 一种实现板状样品高分辨率大视野cl成像的扫描方法
CN117115272A (zh) 用于降水粒子多角度成像的远心相机标定和三维重构方法
CN116310131A (zh) 一种考虑多视图融合策略的三维重建方法
CN102062740B (zh) 锥束ct扫描成像方法及系统
CN112233156B (zh) 微纳米ct投影数据的中心切片对齐方法
CN111028337B (zh) 一种改善有限视角问题的三维光声成像方法
US10096148B1 (en) Portable x-ray computed tomography
Huang et al. A 3D reconstruction method based on one-dimensional galvanometer laser scanning system
US11175242B2 (en) Geometric alignment, sample motion correction, and intensity normalization of computed tomography projections using pi-line optimization
CN109959933A (zh) 一种基于压缩感知的多基线圆迹合成孔径雷达成像方法
CN112233157B (zh) 微纳米ct系统转台误差的频域校正方法
CN114820927A (zh) 一种基于数据重排和共轭射线的工业ct三维图像重建方法
CN113052929A (zh) 一种基于投影视角加权的直线扫描cl重建方法
CN117859151A (zh) 用于电子冷冻断层扫描重建的系统和方法
Luo et al. Geometric calibration based on a simple phantom for multi-lens microscopic CT

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19909604

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19909604

Country of ref document: EP

Kind code of ref document: A1