[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

CN111553310A - Security inspection image acquisition method and system based on millimeter wave radar and security inspection equipment - Google Patents

Security inspection image acquisition method and system based on millimeter wave radar and security inspection equipment Download PDF

Info

Publication number
CN111553310A
CN111553310A CN202010394789.9A CN202010394789A CN111553310A CN 111553310 A CN111553310 A CN 111553310A CN 202010394789 A CN202010394789 A CN 202010394789A CN 111553310 A CN111553310 A CN 111553310A
Authority
CN
China
Prior art keywords
image
gray
security inspection
millimeter wave
wave radar
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
CN202010394789.9A
Other languages
Chinese (zh)
Other versions
CN111553310B (en
Inventor
夏勇
陈帅
田西兰
蔡红军
王曙光
夏鹏
张江辉
刘丽莎
王斌
吴颖
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
CETC 38 Research Institute
Original Assignee
CETC 38 Research Institute
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 CETC 38 Research Institute filed Critical CETC 38 Research Institute
Priority to CN202010394789.9A priority Critical patent/CN111553310B/en
Publication of CN111553310A publication Critical patent/CN111553310A/en
Application granted granted Critical
Publication of CN111553310B publication Critical patent/CN111553310B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/107—Static hand or arm
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • G01S13/887—Radar or analogous systems specially adapted for specific applications for detection of concealed objects, e.g. contraband or weapons
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V3/00—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation
    • G01V3/12—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation operating with electromagnetic waves
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G06V10/457—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by analysing connectivity, e.g. edge linking, connected component analysis or slices
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Remote Sensing (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Electromagnetism (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Human Computer Interaction (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Environmental & Geological Engineering (AREA)
  • Geology (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Geophysics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

The invention provides a security check image acquisition method based on a millimeter wave radar, which comprises the steps of carrying out human identification on an original image, then carrying out denoising processing on the human image, obtaining a gray threshold value based on gray histogram statistics, carrying out normalization processing on the image based on the gray threshold value, and obtaining the definition of the image through a model trained in the early stage so as to obtain the imaging quality; the invention also provides a security inspection image acquisition system and security inspection equipment based on the millimeter wave radar; the millimeter wave radar-based security inspection image acquisition method, system and security inspection equipment provided by the invention have the advantages that: whether people exist is judged based on the gray value sum, the operation rule is simple, and the accuracy is high; selecting high N-bit data to perform denoising, and effectively removing noise under the condition of keeping data integrity; and the gray level threshold is determined based on the gray level histogram statistics, so that the contrast of the image is improved, and the subsequent image processing and identification are facilitated.

Description

Security inspection image acquisition method and system based on millimeter wave radar and security inspection equipment
Technical Field
The invention relates to the technical field of image processing, in particular to a security inspection image acquisition method and system based on a millimeter wave radar and security inspection equipment.
Background
In recent years, millimeter wave radar imaging technology is increasingly applied to security inspection equipment in public places, with requirements on security inspection passing efficiency and accuracy are higher and higher, and the quality of millimeter wave imaging directly affects the security inspection efficiency, at present, an automatic detection technology is mainly used for automatically detecting whether images meeting quality requirements include forbidden articles, so that the images transmitted to a system are required to meet the requirements, and quality evaluation of millimeter wave imaging becomes a problem to be solved urgently.
Due to the fact that millimeter wave echo data are extremely poor and wide in distribution, when the millimeter wave echo data are converted into images in the prior art, the image contrast is very poor. The image definition is an important evaluation factor of data quality of millimeter wave equipment, and the existing method mainly comprises the steps of manually designing descriptors such as SIFT (scale invariant feature transform) and the like as definition evaluation features, and performing definition judgment by using an SVM (support vector machine), but the descriptors are easily influenced by noise and are difficult to rapidly and effectively remove and alarm fuzzy images.
The attitude of the target also determines the data quality, and the current attitude assessment methods mainly include two types: the first method is to extract HOG features based on a traditional algorithm and then train an SVM for evaluation, but the method needs a large amount of data and is low in accuracy; the other method is based on deep learning, such as an Open-Pose network, but the network needs to calibrate a large number of characteristic points of human bodies, so that the workload is huge, the network is huge, and the operation efficiency is influenced.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a method for rapidly processing millimeter wave images and judging whether the image quality meets the requirements, so as to obtain a security inspection image meeting the requirements.
The invention solves the technical problems through the following technical scheme: the method for acquiring security inspection images based on the millimeter wave radar comprises the following steps
Step A: acquiring original data of a target image obtained by scanning of a millimeter wave radar on line;
and B: judging whether the target image is occupied or not based on the gray value sum, if not, sending out unmanned prompt information, and if so, skipping to the step C;
and C: preprocessing the image of the person to form a gray-scale image;
the pretreatment comprises the following steps:
step i: extracting high N-bit original data aiming at a manned image to manufacture a human body mask, smoothing the human body mask, and mapping the human body mask to the original data for denoising;
step ii: determining a gray threshold value of the denoised data based on gray histogram statistics, assigning a gray value larger than the gray threshold value as the gray threshold value, and carrying out normalization processing on the assigned data to obtain the gray image;
step D: inputting the gray-scale image into a trained two-classification convolutional neural network classifier, outputting definition representing image quality, and giving a prompt for re-acquiring the image if the definition is smaller than a preset threshold; and if the definition is greater than a preset threshold value, obtaining a security inspection image meeting the requirement.
Preferably, the method for determining whether the target image is a person in the step B includes: calling gray values of all pixel points in a certain number of manned images and unmanned images in an off-line mode, summing the gray values, and determining a threshold value T of the gray value sum of the unmanned images based on the statistical valuespCalculating to obtain the gray value sum T of the current target imageallIf T isall≤TpIf no person exists in the current scanning area, the system sends out no-person prompt information, if T is up toall＞TpAnd the current scanning area has people.
Preferably, the method for smoothing the human body mask in step i is to process the human body mask based on a corrosion expansion algorithm, map the expanded human body mask to the original data, retain the original data in the expanded human body mask region, and remove other noises.
Preferably, in step ii, gray histogram statistics is performed on the denoised raw data, and the gray value with the percentage exceeding P% is set as the gray threshold ThUpdating the gray value T of the pixel pointdataThe method comprises
Figure BDA0002482497550000021
I.e. the grey value is greater than the grey threshold ThThe gray value of the pixel point is set as Th；
Then, the data is normalized by the method
Figure BDA0002482497550000022
Wherein, TmaxAnd TminRespectively representing the maximum value and the minimum value of the gray value of the pixel point.
Preferably, the training method of the two-class convolutional neural network classifier used in step D is as follows: and D, collecting the clear images and the fuzzy images processed in the steps A-C to construct a training set, inputting the training set into a convolutional neural network, performing five-fold cross validation and adjusting model parameters to obtain the two-classification convolutional neural network classifier, wherein the number of the clear images in the training set is larger than that of the fuzzy images.
Preferably, the method further comprises the step of determining whether the posture of the detected person in the clear image is correct or not based on the relative positions of the head and the hand, and the step of adding mosaic processing to the head of the detected person in the image meeting the requirements.
Preferably, the method for determining whether the posture of the detected person is correct includes the following steps:
step I: collecting a certain number of clear images, and marking the positions of the head and the hands in the images to obtain a training set;
step II: inputting the training set into a detection network for training to obtain a head and hand position detection model;
step III: d, inputting the clear image obtained in the step D into a head and hand position detection model to obtain a head and hand position and a central point coordinate in the image; if the recognition result is one head and two hands and the coordinate of the central point meets the yhand1-yhead|＞X，|yhand2-yhead|＞X，|yhand1-yhand2|＜Y，S＜|xhand1-xhand2If the | is less than T, the posture meets the requirement, otherwise, the posture does not meet the requirement, and the system sends out a prompt of posture error;
wherein, the coordinates of the central points of the two hands and the head are respectively (x)hand1,yhand1)，(xhand2,yhand2)，(xhead,yhead) (ii) a X, Y, S, T are distance thresholds.
The invention also provides a security check image acquisition system based on the millimeter wave radar, which comprises
An image acquisition module: acquiring original data of a target image obtained by scanning of a millimeter wave radar on line;
a human body recognition module: judging whether the target image is occupied or not based on the gray value sum, if not, sending out unmanned prompt information, and if so, transmitting the unmanned prompt information to the next module;
a denoising processing module: extracting high N-bit original data aiming at a manned image to manufacture a human body mask, smoothing the human body mask, and mapping the human body mask to the original data for denoising;
a normalization processing module: determining a gray threshold value of the denoised data based on gray histogram statistics, assigning a gray value larger than the gray threshold value as the gray threshold value, and carrying out normalization processing on the assigned data to obtain a gray image;
a definition recognition module: inputting the gray-scale image into a trained two-classification convolutional neural network classifier, outputting definition representing image quality, and giving a prompt for re-acquiring the image if the definition is smaller than a preset threshold; and if the definition is greater than a preset threshold value, obtaining a security inspection image meeting the requirement.
Preferably, the system further comprises a posture detection module for judging whether the posture of the detected person in the clear image is correct or not and a mosaic processing privacy protection module for adding the head of the detected person in the image meeting the requirements.
The invention also provides security inspection equipment using the security inspection image acquisition method based on the millimeter wave radar.
The millimeter wave radar-based security inspection image acquisition method, system and security inspection equipment provided by the invention have the advantages that: based on the gray value sum of the original image, whether a person exists is judged, the operation rule is simple, the algorithm response speed is high, and the accuracy is high; selecting high N-bit data to denoise original data, and effectively removing noise under the condition of keeping data integrity; determining a gray threshold value based on gray histogram statistics, improving the contrast of an image, and facilitating subsequent image processing and identification; compared with the prior art using an SVM classifier, the method has the advantages that the algorithm complexity is obviously reduced, the result is more accurate and efficient, and the difficulty of manually extracting the features is reduced by training the convolutional neural network to recognize the definition of the image; whether the human body posture meets the requirements or not is monitored and recognized based on the head and hand positions, the recognition rule is simple, the algorithm operation speed is high, the result is accurate, the correction direction can be directly determined, and the use is convenient; and the final imaging result is subjected to mosaic processing, so that the privacy of the user is emphasized.
Drawings
Fig. 1 is a flowchart of a security inspection image acquisition method based on a millimeter wave radar according to an embodiment of the present invention;
FIG. 2 provides image comparison of a manned situation and an unmanned situation according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating a state change of an image in a denoising process according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a sharp image and a blurred image provided by an embodiment of the present invention;
FIG. 5 is a schematic diagram of a head and hand detection provided by an embodiment of the present invention;
fig. 6 is a clear image result after final mosaic processing provided by an embodiment of the present invention.
Detailed Description
In order that the objects, technical solutions and advantages of the present invention will become more apparent, the present invention will be further described in detail with reference to the accompanying drawings in conjunction with the following specific embodiments.
As shown in fig. 1, the embodiment provides a security inspection image acquisition method based on millimeter wave radar, which includes
Step A: acquiring original data of a target image obtained by scanning of a millimeter wave radar on line, wherein the extracted original data has M bits, and the millimeter wave radar imaging data used in the embodiment is 16 bits;
and B: judging whether the target image is occupied or not based on the gray value sum, if not, sending out unmanned prompt information, confirming whether the detection area is occupied or not through a worker, and if not, prompting the next detector to enter the detection area; if yes, jumping to the step C;
referring to fig. 2, gray values of all pixel points in a certain number of statistical images of the human images and the unmanned images are called off-line and summed, and a threshold T of the gray value sum of the unmanned images is determined based on the statistical valuespAnd obtaining the gray value and T of the current target image through online calculation in the security inspection processallIf T isall≤TpIf no person exists in the current scanning area, sending out no-person prompt information, if T is the numberall＞TpIf yes, the current scanning area is occupied, and the subsequent steps are continuously executed.
And C: preprocessing the image of the person to obtain a gray-scale image; the method specifically comprises the following steps;
step i: referring to fig. 3, a human body mask is manufactured by extracting high N-bit raw data for a human body image, and the raw data is denoised after the human body mask is smoothed;
the smoothing method comprises the steps of processing a human body mask based on a corrosion expansion algorithm in the prior art, mapping the expanded human body mask to original data, reserving the original data in an expanded human body mask area, and removing other noise data; where N should not be greater than M, in this embodiment, N is made 8, that is, the first 8 bits of data are taken for processing;
step ii: carrying out gray histogram statistics on the denoised original data, and setting the gray value with the proportion exceeding P% as a gray threshold value ThWherein P is an empirical value, and those skilled in the art can determine the influence of P on the image quality based on experience or experiment and determine a specific numerical value; updating gray value T of pixel pointdataThe method comprises
Figure BDA0002482497550000051
I.e. the grey value is greater than the grey threshold ThThe gray value of the pixel point is set as Th；
Then, the data is normalized by the method
Figure BDA0002482497550000052
Wherein, TmaxAnd TminRespectively representing the maximum value and the minimum value of the gray value of the pixel point.
Step D: inputting the gray-scale image into a trained two-classification convolutional neural network classifier, outputting definition representing image quality, and giving a prompt for re-acquiring the image if the definition is smaller than a preset threshold; if the definition is greater than a preset threshold value, obtaining a security inspection image meeting the requirement; the preset threshold of the definition is an empirical value and is determined empirically or experimentally.
The method for training the two-classification convolutional neural network classifier comprises the following steps: collecting the clear images and the blurred images processed in the steps A-C to construct a training set, wherein the number of the clear images in the training set is larger than that of the blurred images, the ratio of the clear images to the blurred images is 2:1, referring to FIG. 4, model training is mainly carried out by using the clear images, and the blurred images are recognition processing of a training model on typical blurred situations; and verifying the training result and adjusting algorithm parameters by a five-fold intersection method in the training process, and training to obtain the two-classification convolutional neural network classifier for identifying the image definition.
After the clear image is obtained, the target image can be further processed to detect whether the illegal object is carried, but the good imaging quality does not mean that the method is applicable to identification of the illegal object, and generally, the detected person is required to keep a certain posture, so that the embodiment further comprises the step of judging whether the posture of the detected person in the clear image is correct based on the relative positions of the head and the hand, and the method specifically comprises the following steps:
step I: collecting a certain number of clear images, and marking the positions of the head and the hands in the images to obtain a training set;
step II: inputting the training set into a detection network for training to obtain a head and hand position detection model;
step III: inputting the clear image obtained in the step D into a head and hand position detection model, and referring to FIG. 5, obtaining a head and hand position and a central point coordinate in the image; if the recognition result is one head and two hands and the coordinate of the central point meets the yhand1-yhead|＞X，|yhand2-yhead|＞X，|yhand1-yhand2|＜Y，S＜|xhand1-xhand2If < T, the posture meets the requirement and can be used for subsequent security check, otherwise, the posture does not meet the requirement and the system sends out a wrong postureCarrying out error prompt, and re-acquiring the image after the posture of the detected person is adjusted and returning to the step A for processing;
wherein, the coordinates of the central points of the two hands and the head are respectively (x)hand1,yhand1)，(xhand2,yhand2)，(xhead,yhead) (ii) a X, Y, S and T are distance thresholds which are preset rules of workers in the previous period.
After the head and hand positions of the detected person are effectively identified, mosaic processing can be added to the head of the detected person for privacy protection, and finally an imaging result as shown in fig. 6 is obtained.
The embodiment also provides a security check image acquisition system based on the millimeter wave radar, which comprises
An image acquisition module: acquiring original data of a target image obtained by scanning of a millimeter wave radar on line;
a human body recognition module: judging whether the target image is occupied or not based on the gray value sum, if not, sending out unmanned prompt information, and if so, transmitting the unmanned prompt information to the next module;
a denoising processing module: extracting high N-bit original data aiming at a manned image to manufacture a human body mask, smoothing the human body mask, and mapping the human body mask to the original data for denoising;
a normalization processing module: determining a gray threshold value of the denoised data based on gray histogram statistics, assigning a gray value larger than the gray threshold value as the gray threshold value, and carrying out normalization processing on the assigned data to obtain a gray image;
a definition recognition module: inputting the gray-scale image into a trained two-classification convolutional neural network classifier, outputting definition representing image quality, and giving a prompt for re-acquiring the image if the definition is smaller than a preset threshold; and if the definition is greater than a preset threshold value, obtaining a security inspection image meeting the requirement.
The security inspection image acquisition system based on the millimeter wave radar further comprises a posture detection module for judging whether the posture of the detected person in the clear image is correct or not and a mosaic processing privacy protection module for adding a mosaic to the head of the detected person in the image meeting the requirements.
The embodiment further provides a security inspection device, which comprises a millimeter wave radar, and the millimeter wave radar can be directly added on the basis of the common security inspection device; acquiring a millimeter wave image meeting the requirement based on the security inspection image acquisition method based on the millimeter wave radar provided by the embodiment; referring to fig. 1, a display control subsystem is configured for the equipment to display imaging conditions, prompt information sent by the display system is displayed, suggestions are given according to user gestures, and the gestures of people are corrected in an auxiliary manner to obtain millimeter wave images meeting requirements.

Claims (10)

1. The security inspection image acquisition method based on the millimeter wave radar is characterized by comprising the following steps: comprises that
Step A: acquiring original data of a target image obtained by scanning of a millimeter wave radar on line;
and B: judging whether the target image is occupied or not based on the gray value sum, if not, sending out unmanned prompt information, and if so, skipping to the step C;
and C: preprocessing the image of the person to form a gray-scale image;
the pretreatment comprises the following steps:
step i: extracting high N-bit original data aiming at a manned image to manufacture a human body mask, smoothing the human body mask, and mapping the human body mask to the original data for denoising;
step ii: determining a gray threshold value of the denoised data based on gray histogram statistics, assigning a gray value larger than the gray threshold value as the gray threshold value, and carrying out normalization processing on the assigned data to obtain the gray image;
step D: inputting the gray-scale image into a trained two-classification convolutional neural network classifier, outputting definition representing image quality, and giving a prompt for re-acquiring the image if the definition is smaller than a preset threshold; and if the definition is greater than a preset threshold value, obtaining a security inspection image meeting the requirement.
2. The millimeter wave radar-based security inspection image acquisition method according to claim 1, whereinIn the following steps: the method for judging whether the target image contains a person in the step B comprises the following steps: calling gray values of all pixel points in a certain number of manned images and unmanned images in an off-line mode, summing the gray values, and determining a threshold value T of the gray value sum of the unmanned images based on the statistical valuespCalculating to obtain the gray value sum T of the current target imageallIf T isall≤TpIf no person exists in the current scanning area, the system sends out no-person prompt information, if T is up toall＞TpAnd the current scanning area has people.
3. The millimeter wave radar-based security inspection image acquisition method according to claim 2, characterized in that: and i, processing the human body mask based on a corrosion expansion algorithm, mapping the expanded human body mask to original data, reserving the original data in the expanded human body mask region, and removing other noises.
4. The millimeter wave radar-based security inspection image acquisition method according to claim 3, wherein: in step ii, gray histogram statistics is carried out on the de-noised original data, and the gray value with the proportion exceeding P% is set as a gray threshold value ThUpdating the gray value T of the pixel pointdataThe method comprises
Figure FDA0002482497540000011
I.e. the grey value is greater than the grey threshold ThThe gray value of the pixel point is set as Th；
Then, the data is normalized by the method
Figure FDA0002482497540000021
Wherein, TmaxAnd TminRespectively representing the maximum value and the minimum value of the gray value of the pixel point.
5. The millimeter wave radar-based security inspection image acquisition method according to claim 4, wherein: the training method of the two-classification convolutional neural network classifier used in the step D comprises the following steps: and D, collecting the clear images and the fuzzy images processed in the steps A-C to construct a training set, inputting the training set into a convolutional neural network, performing five-fold cross validation and adjusting model parameters to obtain the two-classification convolutional neural network classifier, wherein the number of the clear images in the training set is larger than that of the fuzzy images.
6. The millimeter wave radar-based security inspection image acquisition method according to claim 1, wherein: the method also comprises a step of judging whether the posture of the detected person in the clear image is correct or not based on the relative positions of the head and the hands, and a step of adding mosaic processing to the head of the detected person in the image meeting the requirements.
7. The millimeter wave radar-based security inspection image acquisition method according to claim 1, wherein: the method for judging whether the posture of the detected person is correct comprises the following steps:
step I: collecting a certain number of clear images, and marking the positions of the head and the hands in the images to obtain a training set;
step II: inputting the training set into a detection network for training to obtain a head and hand position detection model;
step III: d, inputting the clear image obtained in the step D into a head and hand position detection model to obtain a head and hand position and a central point coordinate in the image; if the recognition result is one head and two hands and the coordinate of the central point meets the yhand1-yhead|＞X，|yhand2-yhead|＞X，|yhand1-yhand2|＜Y，S＜|xhand1-xhand2If the | is less than T, the posture meets the requirement, otherwise, the posture does not meet the requirement, and the system sends out a prompt of posture error;
wherein, the coordinates of the central points of the two hands and the head are respectively (x)hand1,yhand1)，(xhand2,yhand2)，(xhead,yhead) (ii) a X, Y, S, T are distance thresholds.
8. Security installations image acquisition system based on millimeter wave radar, its characterized in that: comprises that
An image acquisition module: acquiring original data of a target image obtained by scanning of a millimeter wave radar on line;
a human body recognition module: judging whether the target image is occupied or not based on the gray value sum, if not, sending out unmanned prompt information, and if so, transmitting the unmanned prompt information to the next module;
a denoising processing module: extracting high N-bit original data aiming at a manned image to manufacture a human body mask, smoothing the human body mask, and mapping the human body mask to the original data for denoising;
a normalization processing module: determining a gray threshold value of the denoised data based on gray histogram statistics, assigning a gray value larger than the gray threshold value as the gray threshold value, and carrying out normalization processing on the assigned data to obtain a gray image;
a definition recognition module: inputting the gray-scale image into a trained two-classification convolutional neural network classifier, outputting definition representing image quality, and giving a prompt for re-acquiring the image if the definition is smaller than a preset threshold; and if the definition is greater than a preset threshold value, obtaining a security inspection image meeting the requirement.
9. The millimeter wave radar-based security inspection image acquisition system according to claim 8, wherein: the system also comprises a gesture detection module for judging whether the gesture of the detected person in the clear image is correct or not and a mosaic processing privacy protection module for adding the head of the detected person in the image meeting the requirements.
10. A security inspection apparatus, characterized in that: the millimeter wave radar-based security inspection image acquisition method according to any one of claims 1 to 7 is used.
CN202010394789.9A 2020-05-08 2020-05-08 Security inspection image acquisition method and system based on millimeter wave radar and security inspection equipment Active CN111553310B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010394789.9A CN111553310B (en) 2020-05-08 2020-05-08 Security inspection image acquisition method and system based on millimeter wave radar and security inspection equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010394789.9A CN111553310B (en) 2020-05-08 2020-05-08 Security inspection image acquisition method and system based on millimeter wave radar and security inspection equipment

Publications (2)

Publication Number Publication Date
CN111553310A true CN111553310A (en) 2020-08-18
CN111553310B CN111553310B (en) 2023-04-07

Family

ID=72008046

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010394789.9A Active CN111553310B (en) 2020-05-08 2020-05-08 Security inspection image acquisition method and system based on millimeter wave radar and security inspection equipment

Country Status (1)

Country Link
CN (1) CN111553310B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112099001A (en) * 2020-09-18 2020-12-18 欧必翼太赫兹科技(北京)有限公司 Control method of three-dimensional special-shaped planar aperture holographic imaging security inspection radar
CN112733722A (en) * 2021-01-11 2021-04-30 深圳力维智联技术有限公司 Gesture recognition method, device and system and computer readable storage medium
CN112861608A (en) * 2020-12-30 2021-05-28 浙江万里学院 Detection method and system for distracted driving behaviors
WO2022121695A1 (en) * 2020-12-09 2022-06-16 同方威视技术股份有限公司 Three-dimensional imaging method and apparatus, and three-dimensional imaging device

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105513035A (en) * 2014-09-22 2016-04-20 北京计算机技术及应用研究所 Method and system for detecting human body hidden item in passive millimeter wave image
US20180329053A1 (en) * 2016-08-25 2018-11-15 Shenzhen Cct Thz Technology Co., Ltd. Human-body foreign-matter detection method and system based on millimetre-wave image
CN109543582A (en) * 2018-11-15 2019-03-29 杭州芯影科技有限公司 Human body foreign body detection method based on millimeter-wave image
CN109711331A (en) * 2018-12-25 2019-05-03 山东雷诚电子科技有限公司 A kind of millimetre-wave radar safety check instrument foreign matter detecting method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105513035A (en) * 2014-09-22 2016-04-20 北京计算机技术及应用研究所 Method and system for detecting human body hidden item in passive millimeter wave image
US20180329053A1 (en) * 2016-08-25 2018-11-15 Shenzhen Cct Thz Technology Co., Ltd. Human-body foreign-matter detection method and system based on millimetre-wave image
CN109543582A (en) * 2018-11-15 2019-03-29 杭州芯影科技有限公司 Human body foreign body detection method based on millimeter-wave image
CN109711331A (en) * 2018-12-25 2019-05-03 山东雷诚电子科技有限公司 A kind of millimetre-wave radar safety check instrument foreign matter detecting method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
穆森;单海婧;周锦源;邱桂苹;: "一种被动毫米波图像中人体隐匿物品的检测方法" *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112099001A (en) * 2020-09-18 2020-12-18 欧必翼太赫兹科技(北京)有限公司 Control method of three-dimensional special-shaped planar aperture holographic imaging security inspection radar
CN112099001B (en) * 2020-09-18 2021-09-03 欧必翼太赫兹科技(北京)有限公司 Control method of three-dimensional special-shaped planar aperture holographic imaging security inspection radar
WO2022121695A1 (en) * 2020-12-09 2022-06-16 同方威视技术股份有限公司 Three-dimensional imaging method and apparatus, and three-dimensional imaging device
GB2616181A (en) * 2020-12-09 2023-08-30 Nuctech Co Ltd Three-dimensional imaging method and apparatus, and three-dimensional imaging device
CN112861608A (en) * 2020-12-30 2021-05-28 浙江万里学院 Detection method and system for distracted driving behaviors
CN112733722A (en) * 2021-01-11 2021-04-30 深圳力维智联技术有限公司 Gesture recognition method, device and system and computer readable storage medium

Also Published As

Publication number Publication date
CN111553310B (en) 2023-04-07

Similar Documents

Publication Publication Date Title
CN111553310B (en) Security inspection image acquisition method and system based on millimeter wave radar and security inspection equipment
JP7113657B2 (en) Information processing device, information processing method, and program
CN106778586B (en) Off-line handwritten signature identification method and system
CN112819772B (en) High-precision rapid pattern detection and recognition method
WO2018216629A1 (en) Information processing device, information processing method, and program
CN106934795B (en) A kind of automatic testing method and prediction technique of glue into concrete beam cracks
CN111784747B (en) Multi-target vehicle tracking system and method based on key point detection and correction
CN112446896B (en) Conveying material falling monitoring method, device and system based on image recognition
CN111564015B (en) Method and device for monitoring perimeter intrusion of rail transit
US8090151B2 (en) Face feature point detection apparatus and method of the same
CN112419298B (en) Bolt node plate rust detection method, device, equipment and storage medium
CN109389105B (en) Multitask-based iris detection and visual angle classification method
JP4858612B2 (en) Object recognition system, object recognition method, and object recognition program
CN111340041B (en) License plate recognition method and device based on deep learning
CN111611907A (en) Image-enhanced infrared target detection method
CN109255336A (en) Arrester recognition methods based on crusing robot
CN116311079A (en) Civil security engineering monitoring method based on computer vision
KR20080079798A (en) Method of face detection and recognition
CN114972871A (en) Image registration-based few-sample image anomaly detection method and system
CN117115415B (en) Image marking processing method and system based on big data analysis
JP5080416B2 (en) Image processing apparatus for detecting an image of a detection object from an input image
CN111553437A (en) Neural network based image classification method
CN117623031A (en) Elevator non-inductive control system and method
CN116912945A (en) Gesture recognition method, device, equipment and computer program product
CN112329796A (en) Infrared imaging cirrus cloud detection method and device based on visual saliency

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant