CN102629315B - A kind of detection automatically hiding article and identification device - Google Patents
A kind of detection automatically hiding article and identification device Download PDFInfo
- Publication number
- CN102629315B CN102629315B CN201210050329.XA CN201210050329A CN102629315B CN 102629315 B CN102629315 B CN 102629315B CN 201210050329 A CN201210050329 A CN 201210050329A CN 102629315 B CN102629315 B CN 102629315B
- Authority
- CN
- China
- Prior art keywords
- human body
- human
- module
- image
- original image
- 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.)
- Expired - Fee Related
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 75
- 230000011218 segmentation Effects 0.000 claims abstract description 35
- 238000003384 imaging method Methods 0.000 claims description 41
- 230000000877 morphologic effect Effects 0.000 claims description 11
- 210000000707 wrist Anatomy 0.000 claims description 6
- 210000003484 anatomy Anatomy 0.000 claims description 5
- 210000003127 knee Anatomy 0.000 claims description 5
- 230000004807 localization Effects 0.000 claims description 5
- 238000003708 edge detection Methods 0.000 claims description 4
- 210000004247 hand Anatomy 0.000 claims description 4
- 230000036541 health Effects 0.000 claims description 4
- 210000003141 lower extremity Anatomy 0.000 claims description 4
- 230000004927 fusion Effects 0.000 claims description 3
- 210000000746 body region Anatomy 0.000 claims description 2
- 244000309466 calf Species 0.000 claims description 2
- 238000002224 dissection Methods 0.000 claims description 2
- 210000000689 upper leg Anatomy 0.000 claims description 2
- 238000000034 method Methods 0.000 abstract description 63
- 238000012545 processing Methods 0.000 description 12
- 238000005516 engineering process Methods 0.000 description 9
- 230000008569 process Effects 0.000 description 9
- 238000012958 reprocessing Methods 0.000 description 9
- 238000013461 design Methods 0.000 description 8
- 230000010339 dilation Effects 0.000 description 8
- 230000003628 erosive effect Effects 0.000 description 8
- 230000000694 effects Effects 0.000 description 6
- 238000001914 filtration Methods 0.000 description 6
- 239000002184 metal Substances 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 4
- 238000009499 grossing Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 230000002708 enhancing effect Effects 0.000 description 3
- 239000002360 explosive Substances 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000005855 radiation Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000002594 fluoroscopy Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 241000282326 Felis catus Species 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013016 damping Methods 0.000 description 1
- 230000005674 electromagnetic induction Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000000945 filler Substances 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 239000000843 powder Substances 0.000 description 1
- 201000007094 prostatitis Diseases 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a kind of detection automatically hiding article and identify device, including: scan module, adjusting module, segmentation locating module, bar combination model generation module, non-human target preliminary detection module, non-human target distribution module, classification recognition module.By the detection automatically of the concealment article of the present invention and knowledge method for distinguishing, it is achieved that by hiding the detection of article and identifying from manually becoming automatically, reduce the instructions for use of personnel, reduce personal error, shorten the detection interpretation time.
Description
Technical field
The present invention relates to the detection technique of field of safety check, more particularly it relates to hide automatically detecting and identifying device of article based on a kind of of mm-wave imaging.
Background technology
In field of safety check, for the detection of human body and concealment article thereof, there is following various ways: metal detector, x-ray fluoroscopy, infrared detection and millimeter wave detection etc..Metal detector is to be realized by electromagnetic induction, can only judge the presence or absence of metal object, it is impossible to imaging or determine object space.X-ray fluoroscopy equipment has very strong penetrance, is generally used for the detection for luggage and articles, as directly human detection to people's danger relatively greatly, being therefore generally rarely used in human detection in safety check.Infrared detection be the thermal radiation property utilizing object to carry out imaging, safety check can be used for the detection to human body.The brightness of the object in infrared image depends primarily on the surface emissivity characteristic of the temperature of object and the heat of radiation and object, it is characterized in there is no significant corner angle, marginal information, its edge lines are round and smooth, and grey scale change is slow, and shape details and small attitudes vibration to object are insensitive.These features make the human body in infrared image is detected have certain difficulty.
Millimeter wave (3GHz-300GHz) is a kind of electromagnetic wave between light wave and radio wave.Millimeter wave can penetrate all medicated clothing clothes, and the millimeter wave energy relatively metal of human body radiation, pottery, plastic explosive, powder type explosive and medicated clothing, insulant etc. are eager to excel in whatever one does, utilize active/passive millimeter-wave technology can detect prohibited items such as being hidden in the various cutters of human body surface, gun, explosive.Due to more higher than metal detection technical capability, more safer than ray technology, over nearly 10 years, human body millimeter wave safety check technology is rapidly developed.The focal plane arrays (FPA) scanning technique of passive-type, multi-beam frequency-scan technique and active 3D hologram millimeter-wave technology are in succession tested and are applied.Utilize active MMW rays safety detection apparatus to after human body imaging, the various article that display characteristics of human body that can be more visible in image and human body carry.
First, in millimeter wave safety check, the analysis of human body image is important composition link.After human body mm-wave imaging, how human body image is carried out detection and analyzes, be the safe examination system basis that realizes target detection automatization, be to concealment article position sign on human body and the basis to human body image secret protection in subsequent treatment.
Secondly, how concealment article are detected after mm-wave imaging and sign on human body, it prior art is the method by manual analysis, wherein image enhancement technique and multiframe comparison techniques are applied in manual analysis, but require over the interpretive analysis of professional person, the identification to concealment article and location can be realized.Although the image Segmentation Technology based on methods such as gray scale multi thresholds, Boundary Extraction, rim detection, region segmentation, wavelet transformation, morphology, fuzzy mathematics, genetic algorithm, neutral net, comentropies is attempted and is applied in the detection automatically of concealment article, but owing to disconnecting mutually with human vision mechanism, merely with the gray scale in image and spatial information, image is split, still can produce the situation inconsistent with human vision.And based on the method for positioning analyzing of human body prior model, the motion tracking of human body is applied, reduce the complexity of tracking, wherein mainly include the stick model etc. of bar band model as shown in figure 32, Figure 33, but only comprise human body contour outline information due to stick model, such as structure, shape, attitude etc., and stick model only comprises each articulare of human body, all can only be limited to the detection of human body, still can not directly solve concealment article detection and identification problem automatically at human body.
3rd, pass through millimeter wave scanning imaging, the concealment Item Information on human body can be detected, but exposure and the display of human body privacy (such as face and privacy places) can be caused simultaneously, how being analyzed processing to image after mm-wave imaging, the privacy information shielding human body before display concealment article is also the technical problem needing in safe examination system to solve.
Summary of the invention
Present invention aim at providing a kind of human body detecting method based on mm-wave imaging and device, it is achieved to the identification of partes corporis humani position and location in millimeter wave scans.
The method comprises the following steps: tested personnel carry out millimeter wave scanning and obtains original image；It is adjusted described original image obtaining target image；Segmentation and the location of human body is carried out according to described target image；Generate anthropometric dummy.
Further, described according to described target image carry out human body segmentation and location also include following sub-step: determine the vertical centrage of human body；Determine the coordinate of the described each key point of target image human body and obtain the horizontal division line between partes corporis humani position；Determine width and the slope of partes corporis humani position.
Further, described generation anthropometric dummy includes: according to the coordinate of described each key point, the width of described partes corporis humani position and slope, it is thus achieved that with the anthropometric dummy that rectangle and/or parallelogram form.
Further, described acquisition target image that described original image is adjusted also includes following sub-step: described original image carries out pretreatment and obtains preliminary denoising image；Described preliminary denoising image is carried out binaryzation and obtains preliminary bianry image；Described preliminary bianry image is carried out reprocessing and obtains described target image.
Further, the described pretreatment preliminary denoising image of acquisition that described original image is carried out farther includes following sub-step: described original image and background image gray value carry out difference operation；Picture smooth treatment；Linear gradation converts.
Further, the described binaryzation preliminary bianry image of acquisition that described preliminary denoising image is carried out is to utilize Pulse Coupled Neural Network algorithm to be criterion to the maximum with entropy to choose the threshold value of binaryzation.
Further, the described reprocessing described target image of acquisition that described preliminary bianry image is carried out is to pass through morphologic filtering.
Further, described included by morphologic filtering: the square structure element using the length of side to be 5 carries out erosion operation and eliminates the bright noise spot outside human body；The square structure element using the length of side to be 4 carries out eliminating while opening operation keeps image size isolated area and the burr at human body edge；The square structure element using the length of side to be 4 carries out filling the minuscule hole in human body and the edge of smooth human body while closed operation keeps image size；The square structure element using the length of side to be 5 carries out dilation operation makes image return to life size.
Further, the described pretreatment preliminary denoising image of acquisition that described original image is carried out also includes following sub-step: described original image is carried out image enhaucament.
Correspondingly, the human body detection device based on mm-wave imaging of the present invention, including: scan module, obtain original image for tested personnel being carried out millimeter wave scanning；Adjusting module, for being adjusted described original image obtaining target image；Segmentation locating module, for carrying out segmentation and the location of human body according to described target image；Anthropometric dummy generation module, is used for generating anthropometric dummy.
Further, described segmentation locating module also includes following submodule: vertically centrage module, for determining the vertical centrage of human body；Coordinate horizontal line module, for determining the coordinate of the described each key point of target image human body and obtaining the horizontal division line between partes corporis humani position；Width slope module, for determining width and the slope of partes corporis humani position.
Further, described anthropometric dummy generation module is further used for the coordinate according to described each key point, the width of described partes corporis humani position and slope, it is thus achieved that with the anthropometric dummy that rectangle and/or parallelogram form.
Further, described adjusting module also includes following submodule: pretreatment module, obtains preliminary denoising image for described original image carries out pretreatment；Binarization block, obtains preliminary bianry image for described preliminary denoising image carries out binaryzation；Reprocessing module, obtains described target image for described preliminary bianry image carries out reprocessing.
Further, described pretreatment module farther includes with lower unit:
Difference operation unit, for carrying out difference operation by described original image and background image gray value；Smoothing processing unit, is used for carrying out picture smooth treatment；Linear change unit, is used for carrying out linear gradation conversion.
Further, described binarization block is criterion further with Pulse Coupled Neural Network algorithm to the maximum with entropy and chooses the threshold value of binaryzation.
Further, described reprocessing module is reprocessed by morphologic filtering further.
Further, described included by morphologic filtering: the square structure element using the length of side to be 5 carries out erosion operation and eliminates the bright noise spot outside human body；The square structure element using the length of side to be 4 carries out eliminating while opening operation keeps image size isolated area and the burr at human body edge；The square structure element using the length of side to be 4 carries out filling the minuscule hole in human body and the edge of smooth human body while closed operation keeps image size；The square structure element using the length of side to be 5 carries out dilation operation makes image return to life size.
Further, described pretreatment module also includes: image enhancing unit, for described original image is carried out image enhaucament.
By the human body detecting method based on mm-wave imaging of the present invention and device, it is achieved that in millimeter-wave image to the identification of human body parts and process, hide article and secret protection for subsequent survey and provide the foundation.
The object of the invention also resides in the location providing a kind of detection automatically hiding article and recognition methods and device to realize in millimeter wave scans, concealment article being distributed on human body and identifies from manually becoming automatically, reduces the instructions for use of personnel.
Automatically detection and the recognition methods of described concealment article, comprises the following steps: tested personnel carry out millimeter wave scanning and obtains original image；It is adjusted described original image obtaining target image；Segmentation and the location of human body is carried out according to described target image；Generate bar combination model；According to described original image, non-human target is detected, it is thus achieved that non-human target distribution original image；Described bar combination model is utilized to obtain the described non-human target distribution original image location distribution information relative to human body；Described non-human target is carried out classification identification and shows the concealment article location distribution information relative to human body.
Further, described generation bar combination model includes following sub-step: generate the stick model providing each key point of human body；Generate the bar band model that human body contour outline information is provided；In conjunction with described stick model and described bar band model, generate bar combination model.
Further, described according to described original image, non-human target is detected, it is thus achieved that non-human target distribution original image includes following sub-step: described original image is carried out rim detection, preliminary identify non-human target；Non-human target distribution region is highlighted by mathematical morphological operation；Border according to described non-human target distribution region is chosen minimum circumscribed rectangle and is obtained non-human goal rule regional distribution chart；Merge described non-human goal rule regional distribution chart and described original image, it is thus achieved that described non-human target original image.
Further, described utilize described bar combination model obtain described non-human target distribution original image be by described non-human target distribution original image is input on described bar combination model relative to the location distribution information of human body.
Further, described described non-human target is carried out classification identification and show concealment article include following sub-step relative to the location distribution information of human body: the exposed position of human body is positioned；Will be distributed over the non-human target on the exposed position of human body and be defined as non-concealment article, will be distributed over the non-human target outside the exposed position of human body and be defined as concealment article；Reject the original image of described non-concealment article and show concealment article original image distributed intelligence on described bar combination model.
Correspondingly, automatically detecting and identification device of the concealment article of the present invention, including: scan module, obtain original image for tested personnel being carried out millimeter wave scanning；Adjusting module, for being adjusted described original image obtaining target image；Segmentation locating module, for carrying out segmentation and the location of human body according to described target image；Bar combination model generation module, is used for generating bar combination model；Non-human target preliminary detection module, for detecting non-human target according to described original image, it is thus achieved that non-human target distribution original image；Non-human target distribution module, for utilizing described bar combination model to obtain the described non-human target distribution original image location distribution information relative to human body；Classification recognition module, for carrying out classification identification and showing the concealment article location distribution information relative to human body to described non-human target.
Further, described bar combination model generation module includes following submodule: stick model generation module, for generating the stick model providing each key point of human body；Band model generation module, for generating the bar band model providing human body contour outline information；Binding modules, is used in conjunction with described stick model and described bar band model.
Further, described non-human target preliminary detection module includes following submodule: edge detection module, for described original image is carried out rim detection, and the non-human target of preliminary identification；Highlight module, for highlighting non-human target distribution region by mathematical morphological operation；Regularization module, obtains non-human goal rule regional distribution chart for choosing minimum circumscribed rectangle according to the border in described non-human target distribution region；Fusion Module, is used for merging described non-human goal rule regional distribution chart and described original image, it is thus achieved that described non-human target original image.
Further, described non-human target distribution module is by being input on described bar combination model by described non-human target distribution original image.
Further, described classification recognition module includes following submodule: exposed spots localization module, for the exposed position of human body is positioned；Sort module, is defined as non-concealment article for will be distributed over the non-human target on the exposed position of human body, will be distributed over the non-human target outside the exposed position of human body and is defined as concealment article；Display module, for rejecting the original image of described non-concealment article and showing concealment article original image distributed intelligence on described bar combination model.
Method and apparatus by detection and the identification automatically of the concealment article of the present invention, it is achieved that by hiding the detection of article and identifying from manually becoming automatically, reduce the instructions for use of personnel, reduce personal error, shorten the detection interpretation time.
The object of the invention also resides in offer a kind of method for secret protection based on mm-wave imaging and device, it is achieved that secret protection to tested personnel in millimeter wave scanning.
The method for secret protection based on mm-wave imaging of the present invention, comprises the following steps: tested personnel carry out millimeter wave scanning and obtains original image；Human detection and concealment Articles detecting is carried out according to described original image；Determine the privacy places of human body；The privacy places of human body is shielded and indicates the concealment Item Information on human body.
Further, the described privacy places determining human body includes: judge the sex of tested personnel, when tested personnel are male, the region of head zone and the downward torso width 1/2 of human body waist is defined as privacy places, when tested personnel are women, human body head region, the region of the downward torso width 1/2 of human body waist and trunk are privacy places from trunk upper end down to the region of trunk height 1/2.
Further, the concealment Item Information that the described privacy places to human body shields and indicates on human body includes: privacy places carries out Fuzzy processing on described original image and forms part obfuscation original image；Described concealment article sign frame is marked by described obscure portions original image.
Further, the concealment Item Information that the described privacy places to human body shields and indicates on human body includes: select the purpose image in described human detection；Judge whether concealment article are in the privacy places of human body, if it is, use the color blocks different from human body color represent concealment article and indicate on described purpose image；If it is not, then the original image by concealment article is shown on described purpose image.
Further, the concealment Item Information that the described privacy places to human body shields and indicates on human body includes: human body carries out whole Fuzzy processing on described original image and forms whole obfuscation original image；Judge whether concealment article are in the privacy places of human body, if it is, use the color blocks different from human body color represent concealment article and indicate on described whole obfuscation original images；If it is not, then the original image by concealment article is shown on described whole obfuscation original image.
Further, the concealment Item Information that the described privacy places to human body shields and indicates on human body includes: select the anthropometric dummy in described human detection；The color blocks different from anthropometric dummy color is used to represent concealment article and indicate in described anthropometric dummy.
Correspondingly, the privacy protection device based on mm-wave imaging of the present invention includes: scan module, obtains original image for tested personnel carry out millimeter wave scanning；Detecting device, for carrying out human detection and concealment Articles detecting according to described original image；Privacy places determines module, for determining the privacy places of human body；Privacy mask module, for shielding to the privacy places of human body and indicate the concealment Item Information on human body.
Further, described privacy places determines that module is further used for: judge the sex of tested personnel, when tested personnel are male, the region of head zone and the downward torso width 1/2 of human body waist is defined as privacy places, when tested personnel are women, human body head region, the region of the downward torso width 1/2 of human body waist and trunk are privacy places from trunk upper end down to the region of trunk height 1/2.
Further, described privacy mask module farther includes following submodule: obscure portions module, forms part obfuscation original image for privacy places carries out Fuzzy processing on described original image；First indicates module, for being marked by described concealment article sign frame on described obscure portions original image.
Further, described privacy mask module farther includes: select purpose image module, for selecting the purpose image in described human detection；Second indicates module, for judging whether concealment article are in the privacy places of human body, if it is, use the color blocks different from human body color represent concealment article and indicate on described purpose image；If it is not, then the original image by concealment article is shown on described purpose image.
Further, described privacy mask module farther includes: all obfuscation modules, forms whole obfuscation original image for human body carries out whole Fuzzy processing on described original image；3rd indicates module, for judging whether concealment article are in the privacy places of human body, if it is, use the color blocks different from human body color represent concealment article and indicate on described whole obfuscation original images；If it is not, then the original image by concealment article is shown on described whole obfuscation original image.
Further, described privacy mask module farther includes: preference pattern module, for selecting the anthropometric dummy in described human detection；4th indicates module, for using the color blocks different from anthropometric dummy color represent concealment article and indicate in described anthropometric dummy.
By the method for secret protection based on mm-wave imaging of the present invention and device, it is to avoid exposure to human body privacy during detection concealment article, it is achieved that the effective protection to privacy places of human body.
Accompanying drawing explanation
It is specifically described below with reference to accompanying drawings and in conjunction with the embodiments to the present invention.
Fig. 1 is the human body detecting method basic flow sheet based on mm-wave imaging；
Fig. 2 is the human body detection device basic block diagram based on mm-wave imaging；
Fig. 3 is original image；
Fig. 4 is preliminary denoising image；
Fig. 5 is preliminary bianry image；
Fig. 6 is based on the flow chart of step S2 in the human body detecting method of mm-wave imaging；
Fig. 7 is based on the structural representation of adjusting module in the human body detection device of mm-wave imaging；
Image for the purpose of Fig. 8；
Fig. 9 is human body backbone figure；
Figure 10 is human body segmentation effect figure；
Figure 11 is based on the flow chart of step S3 in the human body detecting method of mm-wave imaging；
Figure 12 is based on the structural representation splitting locating module in the human body detection device of mm-wave imaging；
Figure 13 is the anthropometric dummy figure obtained in the human body detecting method based on mm-wave imaging and device；
Figure 14 is the corresponding design sketch of anthropometric dummy figure and original image；
Figure 15 is the basic flow sheet of detection and the recognition methods automatically of concealment article；
Figure 16 is detection automatically and the flow chart of step S5 in recognition methods of concealment article；
Figure 17 is bar combination model figure；
Figure 18 is detection automatically and the flow chart of step S6 in recognition methods of concealment article；
Figure 19 is the non-human target image of preliminary identification；
Figure 20 is non-human target distribution administrative division map；
Figure 21 is non-human goal rule regional distribution chart；
Figure 22 is non-human target original image；
Figure 23 is non-human target original image scattergram on bar combination model；
Figure 24 is concealment article original image scattergram on bar combination model；
Figure 25 is detection automatically and the flow chart of step S8 in recognition methods of concealment article；
Figure 26 is the basic flow sheet of the method for secret protection based on mm-wave imaging；
Figure 27 is the structural representation of the privacy protection device based on mm-wave imaging；
Figure 28 is the design sketch of the first embodiment of privacy mask；
Figure 29 is the design sketch of the second embodiment of privacy mask；
Figure 30 is the design sketch of the third embodiment of privacy mask；
Figure 31 is the design sketch of the 4th kind of embodiment of privacy mask；
Figure 32 is the band model schematic in background technology；
Figure 33 is the stick model schematic diagram in background technology.
Detailed description of the invention
With reference to the accompanying drawings and by embodiments of the invention, technical scheme is described in detail.
A kind of human body detecting method based on mm-wave imaging in the present invention, comprises the following steps: S1, tested personnel are carried out millimeter wave scanning obtain original image；S2, described original image is adjusted obtain target image；S3, carry out segmentation and the location of human body according to described target image；S4, generation anthropometric dummy.As shown in Figure 1.
Correspondingly, as in figure 2 it is shown, the present invention also provides for a kind of human body detection device based on mm-wave imaging, including:
Scan module 1, is used for performing step S1, tested personnel carries out millimeter wave scanning and obtains original image；
Adjusting module 2, is used for performing step S2, is adjusted described original image obtaining target image；
Segmentation locating module 3, is used for performing step S3, carries out segmentation and the location of human body according to described target image；
Anthropometric dummy generation module 4, is used for performing step S4, generates anthropometric dummy.
In step sl, it is desirable to tested personnel enter millimeter wave scan test section, original image is obtained as shown in Figure 3 after being scanned detection by the mode of scan module 1 millimeter wave active/passive.Original image after scanning generally has a characteristic that image entirety is clear not, comprises much noise.
It is thus desirable to adjusting module 2 carries out step S2, original image is adjusted thus obtaining the target image being adapted for image operation and segmentation, as Fig. 6, step S2 include following sub-step: S21, original image is carried out pretreatment obtain preliminary denoising image；S22, preliminary denoising image is carried out binaryzation obtain preliminary bianry image；S23, preliminary bianry image is carried out reprocessing obtain described target image.
Correspondingly, such as Fig. 7, adjusting module 2 also includes following submodule:
Pretreatment module 21, is used for performing step S21, described original image carries out pretreatment and obtains preliminary denoising image；
Binarization block 22, is used for performing step S22, described preliminary denoising image carries out binaryzation and obtains preliminary bianry image；
Reprocessing module 23, is used for performing step S23, described preliminary bianry image carries out reprocessing and obtains described target image.
Further, pretreatment module 21 also includes image enhancing unit, difference operation unit, smoothing processing unit, linear change unit.Pretreatment module 21 performs step S21 to be needed to carry out following sub-step:
Image enhancing unit makes the contrast of human region and background area in original image increase for original image carries out image enhaucament, improves image visual effect.
Difference operation unit subtracts each other for the gray value carrying out difference operation original image in other words and empty background image by original image and the gray value of empty background image, thus eliminating system noise.Empty background image is exactly the image not having in millimeter wave scan test section to be scanned being formed during tested personnel.
Smoothing processing unit is used for the smoothing processing the carrying out image random noise to remove in image, by using in this unit Image is carried out low-pass filtering and realizes smooth operation by operator.
Linear change unit for carrying out gray scale stretching or claiming subregion linear transformation to image, the tonal range of background area uninterested in image is compressed, human region tonal range is extended, thus prominent human body parts, make human body parts overall clear, it is finally obtained preliminary denoising image, as shown in Figure 4.
Further, step S22 performed by binarization block 22, preliminary denoising image carries out binaryzation, and to obtain preliminary bianry image be utilize Pulse Coupled Neural Network (PCNN) algorithm to be criterion to the maximum with entropy to choose the threshold value of binaryzation, utilize this threshold value that the gray-scale map of preliminary denoising image is converted to the image of binaryzation, thus realizing the segmentation of human region and background area in image.
How more completely carrying out splitting by human body and background by selected threshold accurately is the key of problem, used here as comparatively ripe Pulse Coupled Neural Network (PCNN) technology, PCNN be the nineties Eckhorn etc. based on the mammiferous visual characteristics such as cat research and propose based on Pulse-coupled Neural Network Model, this model is used for the iterative process that the threshold value of image is chosen, and iterative formula is as follows:
F is exactly i-th, j neuronic n feed back input FI, j[n], II, j[n] is coefficient of connection, L for input stimulus signal (in the matrix constituted for image pixel here i-th, the gray value of j pixel), βI, j[n] connects item, TI, j[n] is dynamic threshold, i.e. the threshold value of required solution, Y in the present inventionI, j[n] is PCNN pulse output valve, UI, j[n] is internal activity item.The wherein internal general W=M of connection matrix M, W() MI, j, k, l、WI, j, k, lRespectively FI, j[n]、LI, jY in [n]I, jThe weight coefficient of [n], αF、αL、αTRespectively FI, j[n]、LI, j[n]、TI, jThe damping time constant of [n], VF、VL、VTRespectively FI, j[n]、LI, j[n]、TI, jIntrinsic electromotive force in [n].
Entropy is a kind of form of expression of image statistics, reflects the size that image comprises quantity of information.For image, after general segmentation, Image entropy is more big, obtains quantity of information from artwork more big after segmentation is described, segmentation image detail is more abundant, thus overall segmentation effect also should be more good.This patent uses entropy to be the criterion that criterion terminates as PCNN iteration to the maximum.The computing formula of entropy is:
H1(P)=-P1×log2P1-P0×log2P0。
Wherein, P1、P0Represent that pulse output valve Y [n] is 1, is the probability of 0 respectively.The present invention is by setting a very big iterations n, and such as n=100, use PCNN algorithm is iterated computing, and the entropy H of correspondence is obtained in each computing after terminating1(P), then compare the entropy that n computing obtains, obtain the value H that wherein entropy is maximum1max(P) iterations N timemax.Iterations is NmaxTime obtain threshold value T [Nmax], the Y [N of now PCNN outputmax] constitute under other parameters one stable condition of PCNN, the bianry image of overall segmentation best results.Wherein Y [Nmax] be 1 part be background parts, Y [Nmax] be 0 part be human body parts.
Being adapted in the PCNN formula of above-mentioned calculating process the span of each parameter is:
| αF | αL | αT | VF | VL | VT | β |
| 0.1～0.6 | 1～10 | 0.1～0.6 | 0.1～0.5 | 0.1～0.5 | 2～10 | 0.1～0.6 |
Two operators of W, M can use 1/r or 1/r2Element form constitute matrix, r represents the matrix length of side of operator.
Preferably, the present invention can take following parameter value: αF=0.2, αL=2, αT=0.1, VF=0.1, VL=0.5, VT=20, β=0.5, Be calculated, it is thus achieved that optimum efficiency preliminary bianry image as shown in Figure 5.
Further, reprocess the step S23 performed by module 23, described preliminary bianry image is carried out the reprocessing described target image of acquisition, be pass through morphologic filtering.Owing to Threshold segmentation can cause picture noise, this noise is mainly the isolated bright noise spot outside human body parts or the isolated dark noise point in human body.In order to remove these noises, the method that the present invention uses is that preliminary bianry image applied mathematics morphology operations method is filtered and is converted, obtain the image of a binaryzation with clear smooth profile and described target image, as shown in Figure 8, thus being beneficial to follow-up process.
Mathematical morphological operation method mainly includes erosion operation, dilation operation, opening operation and closed operation.
Erosion operation can weaken and even eliminates less than structural element bright areas, such that it is able to be used for effectively removing rough projection on isolated noise point border.
Dilation operation is the process closing in object by all background dots contacted with target object, can filling cavity and formed together with rough female on territory and flat image boundary.
Opening operation is image first to carry out erosion operation carry out dilation operation again, the isolated area in image and burr can be got rid of, utilization can eliminate the shape noise spot less than structural element, feature according to target noise, select suitable structural element, just can reject target noise, and background is remained.
Closed operation is image first to carry out dilation operation carry out erosion operation again, it is possible to the minuscule hole in filler body, connects contiguous thing and smooth object boundary.
Wherein, structural element is the basic operator of mathematical morphological operation, selects the structural element used to essentially consist in the shape and size of structural element.
Preferably, the morphology operations that this method uses can be following processing procedure: (1) uses the length of side to be the bright noise spot that image is carried out that erosion operation eliminates in image outside human body by the square structure element of 5；(2) using the length of side is that image is carried out opening operation by the square structure element of 4, keeps image size to eliminate isolated area and the burr at human body edge simultaneously；(3) using the length of side is that image is carried out closed operation by the square structure element of 4, fills the minuscule hole in human body and the border of smooth human body while keeping image size；(4) image is carried out expansive working and makes image return to life size by the square structure element using the length of side to be 5.By this process can remove long and wide be respectively less than 5 noise, fill up on human body long and wide be respectively less than 5 cavity, the target image formed after having processed comprises one and is similar to complete human body parts, make characteristics of human body become apparent from.
If it addition, image also has the White lnterfere region that area is not eliminated relatively greatly, it is possible to by calculating the area of each connected region in image, remove the region that area is less.
Further, as Figure 11, step S3, the segmentation carrying out human body according to described target image and location also include following sub-step: S31, determine the vertical centrage of human body；S32, determine the coordinate of the described each key point of target image human body and obtain the horizontal division line between partes corporis humani position；S33, the width determining partes corporis humani position and slope.
Correspondingly, as shown in figure 12, segmentation locating module 3 also includes following submodule:
Vertical centrage module 31, for performing step 31, determining the vertical centrage of human body；
Coordinate horizontal line module 32, for performing step 32, determining the coordinate of the described each key point of target image human body and obtain the horizontal division line between partes corporis humani position；
Width slope module 33, for performing step 33, determining width and the slope of partes corporis humani position.
When vertical centrage module 31 performs the position determining the vertical centrage of people in step S31, owing to the human region in target image has bilateral symmetry, therefore calculate the total pixel of the image of this human region and, as with S0Represent, then from human region left side edge start by the row of image calculate from left to right human body parts image pixel and, as with S1Represent, work as S1For S01/2 time, when prostatitis is the vertical centrage of human body.
Perform in step S32 in coordinate horizontal line module 32, the coordinate of each key point of human body is exactly the position coordinates of partes corporis humani position, such as the coordinate etc. of edge point coordinate, central point, human body includes: the crown, sole, neck, trunk upper end, trunk lower end (waist), crotch, knee, finger tip and ancon.Horizontal division line between coordinate and each position of each key point of human body is the process mutually calculated, specific as follows:
Downward from image top along the vertical centrage of human body, the first man body region boundary point found, if judging to continue to have downwards continuously from this point and length being not less than 1/10 same gray value of picture altitude, then determining the crown central point that this point is human body, this horizontal line residing for point is the crown horizontal line H2 of human body.
And the position of human foot is fixing position in the picture, thus may determine that sole coordinate and residing horizontal line are H9.Because the position that during imaging, people stands is fixing, deducts the vertical coordinate of sole with the vertical coordinate on the crown and can obtain the height H of people.
Judging according to human anatomy, head part accounts for the 15% of height, it may be determined that the coordinate of neck upper end and residing horizontal line H4.The height of neck accounts for the 45% of height of head, therefore may determine that the coordinate of the edge point of trunk upper end and residing horizontal line H5.
The represented target image of Fig. 8 is carried out refinement and can also obtain human body backbone figure, as shown in Figure 9.
In Fig. 9 human body backbone figure, the cross point of trunk and two lower limbs is as the position of trunk lower end, and its place horizontal line, as stringcourse H6, is the long H of lower limb along this line to the horizontal distance definition of footleg。
Judging according to human anatomy, the ratio of human calf and thigh is about 1:1.2, therefore may determine that the horizontal level of knee is for from foot upwards HlegThe position of × 5/11, therefore obtains knee level line H8.
On the target image shown in Fig. 8, from image base along vertical centrage upwards, with the first of human body image position of intersecting point as crotch, thus obtaining human body crotch horizontal line H7.
For cut-off rule, the image of human region being divided into left and right two halves, the peak of left one side of something and the position of left hand finger tip with vertical centrage, on image right, the peak of human body is exactly the position of right hand finger tip, thus obtaining finger tip horizontal line H0.Difference in height between two finger tips is ignored by the present invention.
Due to before tested personnel are scanned, requiring that its two-arm is flared out, therefore on target image, width between two elbows is the widest position of human body, and therefore finding the position of the human region leftmost side is exactly the left elbow of health, the rightmost side is exactly the right elbow of health, thus obtaining elbow lever line H3.Difference in height between two elbows is ignored by the present invention.
Length ratio according to human dissection theory, general human body hands and upper arm is 7:9, therefore may determine that the position of wrist according to the sharp position with elbow of hands, thus also obtain wrist horizontal line H1.
Human body segmentation's design sketch is as shown in Figure 10.
In step S33 performed by width slope module 33, intersecting, according to partes corporis humani position in the target image of Fig. 8, the width determining partes corporis humani position with each horizontal division line, the key point of recycling partes corporis humani position (two central points at two ends up and down such as a part) coordinate calculates the slope obtaining this position.
Further, step S4 performed by anthropometric dummy generation module 4 generates anthropometric dummy and includes respectively according to the coordinate of each position key point, width and slope, each position is represented with rectangle or parallelogram, all of position is linked together, namely an anthropometric dummy formed with rectangle and/or parallelogram is obtained, as shown in figure 13.This human body can carry out proportional corresponding with the human region in original image, and effect is as shown in figure 14.
It should be noted that the anthropometric dummy generated in the description of this method is the explanation carried out for bar band model.According to target image carry out human body segmentation and location after can also generate stick model.Further, bar band model and stick model all can be applied in following a kind of detection and recognition methods automatically hiding article.
On the other hand, the present invention also provides for a kind of detection automatically hiding article and recognition methods and a kind of detection automatically hiding article and identifies device, and one is characteristics of human body is extracted and positions, and two is that non-human target is identified.The process wherein characteristics of human body extracted and position mainly based on the aforesaid human body detecting method in mm-wave imaging and device but be not limited to this human body detecting method and device.
Therefore, such as Figure 15, a kind of detection and recognition methods automatically hiding article comprises the following steps: S1, tested personnel carry out millimeter wave scanning obtain original image；S2, described original image is adjusted obtain target image；S3, carry out segmentation and the location of human body according to described target image；S5, generation bar combination model；S6, according to described original image, non-human target is detected, it is thus achieved that non-human target distribution original image；S7, described bar combination model is utilized to obtain described non-human target distribution original image relative to the location distribution information of human body；S8, described non-human target is carried out classification identification and shows that concealment article are relative to the location distribution information of human body.
Correspondingly, a kind of automatically detection hiding article and identify device, including:
Scan module 1, is used for performing step S1, tested personnel carries out millimeter wave scanning and obtains original image；
Adjusting module 2, is used for performing step S2, is adjusted described original image obtaining target image；
Segmentation locating module 3, is used for performing step S3, carries out segmentation and the location of human body according to described target image；
Bar combination model generation module, is used for performing step S5, generates bar combination model；
Non-human target preliminary detection module, is used for performing step S6, according to described original image, non-human target is detected, it is thus achieved that non-human target distribution original image；
Non-human target distribution module, is used for performing step S7, utilizes described bar combination model to obtain the described non-human target distribution original image location distribution information relative to human body；
Classification recognition module, is used for performing step S8, described non-human target carries out classification identification and shows the concealment article location distribution information relative to human body.
Wherein step S1, S2 and S3 a kind of based in the human body detecting method of mm-wave imaging, scanning means 1, adjusting module 2, segmentation locating module 3 a kind of based on the human body detection device of mm-wave imaging in have be carried out illustrating, repeat no more herein.
Farther include following sub-step according to Figure 16, step S5: S51, generate provide each key point of human body stick model；S52, generate provide human body contour outline information bar band model；S53, in conjunction with described stick model and described bar band model, thus generating bar combination model.
Correspondingly, the bar combination model generation module performing step S5 includes following submodule:
Stick model generation module, is used for performing step S51, generates the stick model providing each key point of human body；
Band model generation module, is used for performing step S52, generates the bar band model providing human body contour outline information；
Binding modules, is used for performing step S53, in conjunction with described stick model and described bar band model.
Wherein, step S51 performed by stick model generation module is the step S32 by a kind of human body detecting method based on mm-wave imaging, determine the coordinate of the described each key point of target image human body and obtain the step of horizontal division line between partes corporis humani position, utilize each key point of human body therein to obtain each articulare building stick model, more just can generate stick model by each articulare of straight line connection.
In step S52 performed by band model generation module generate bar band model process with a kind of based on the human body detecting method of mm-wave imaging in S4 generation anthropometric dummy process identical.
Step S53 performed by binding modules is in conjunction with described stick model and described bar band model, generate bar combination model as shown in figure 17, thus completing extraction and the location of characteristics of human body in this method, circular node in Figure 17 represents articulare when characteristics of human body extracts, it is also possible to add the digitized representation sequence of extraction of each articulare in these articulares.
Further, the non-human target preliminary detection module of the step S6 of execution also includes: edge detection module, highlight module, regularization module and Fusion Module.Correspondingly, as shown in figure 18, step S6, according to described original image, non-human target is detected, it is thus achieved that non-human target distribution original image also includes following sub-step:
Described original image is carried out rim detection by S61, edge detection module, for instance make use of Sobel (Sobel) operator to carry out rim detection, tentatively identifies non-human target, as shown in figure 19；
S62, highlight module and highlight non-human target distribution region by mathematical morphological operation, such as, first Figure 19 is carried out erosion operation, then carries out dilation operation.The square structure element that wherein burn into dilation operation uses the length of side to be 2 and 4 respectively, thus the non-human target distribution administrative division map highlighted, such as Figure 20；
S63, regularization module choose minimum circumscribed rectangle according to the border in described non-human target distribution region, and making the irregular regioinvertions in Figure 20 is the non-human goal rule regional distribution chart in Figure 21；
S64, module merge Figure 21 and merge non-human goal rule regional distribution chart and described original image Fig. 3, it is thus achieved that the non-human target original image of Figure 22, this image also show the human body parts of non-human target position certainly.
nullHuman body contour outline information is only comprised due to stick model，There is no concrete articulare information，Stick model only comprises articulare information，Mask non-human target position on human body contour outline，So further，Step S7 performed by non-human target distribution module is by being input on described bar combination model by described non-human target distribution original image，The non-human target of human body contour outline information acquisition that in bar combination model, bar band model provides so both can have been utilized in the intramarginal distribution of human body contour outline，The human synovial dot position information that in bar combination model, stick model provides further is utilized to obtain the relative position relation between non-human target and articulare，Hence with bar combination model，Make non-human target be provided with on this object of reference of human body to position more accurately，The target original image as non-human in Figure 23 scattergram on bar combination model.
Further, classification recognition module includes following submodule: exposed spots localization module, sort module and display module.It is respectively used to perform step S8 described non-human target carries out class of risk identification and shows that concealment article are relative to each sub-steps in the location distribution information of human body, such as Figure 25:
The exposed position of human body is positioned by S81, exposed spots localization module, such as head, wrist, palm etc., it is possible to carry out so more targeted spots localization in the segmentation and location of aforesaid human body；
Non-human target is classified by S82, sort module, will be distributed over the non-human target on the exposed position of human body and is defined as non-concealment article, such as glasses, button, wrist-watch, ring etc.；Hiding with medicated clothing, the non-human target outside the exposed position of human body cannot be directly viewable by security staff, will be distributed over the non-human target outside the exposed position of human body and is defined as concealment article, it is therefore desirable to pays close attention to；
S83, display module are rejected the original image of described non-concealment article and show concealment article original image distributed intelligence on described bar combination model, as Figure 24 eliminates the design sketch shown again after the original image of the wrist-watch of wrist in Figure 23.
Detection and recognition methods and device automatically by the concealment article of the present invention, it is possible to reduce the instructions for use of personnel, reduce personal error, shortens the interpretation time that human body concealment dangerous materials check.
On the other hand, after human body mm-wave imaging, owing to original image ratio is more visible, detection and recognition methods automatically by the aforesaid human body detecting method based on mm-wave imaging and concealment article, may identify which and show the concealment article on human body and/or human body, but the exposure of privacy places of human body can be caused simultaneously.
In order to protect privacy, the present invention also provides for a kind of method for secret protection based on mm-wave imaging, such as Figure 26, including: S1, tested personnel carry out millimeter wave scanning obtains original image；A, carry out according to described original image human detection and concealment Articles detecting；B, determine the privacy places of human body；C, privacy places to human body shield and indicate the concealment Item Information on human body.
Correspondingly, as shown in figure 27, the present invention also provides for a kind of privacy protection device based on mm-wave imaging, including:
Scan module 1, obtains original image for tested personnel carry out millimeter wave scanning；
Detecting device, is used for performing step A, carries out human detection and concealment Articles detecting according to described original image；
Privacy places determines module, is used for performing step B, it is determined that the privacy places of human body；
Privacy mask module, is used for performing step C, the privacy places of human body is shielded and indicated the concealment Item Information on human body.
Wherein step A can be undertaken by detection and the recognition methods automatically of the aforesaid human body detecting method based on mm-wave imaging and concealment article, correspondingly, detecting device can include the detection automatically of the human body detection device based on mm-wave imaging and concealment article and identify device.
Perform the privacy places of step B and determine that module determines that privacy places is determined also according to the segmentation of aforementioned human body and location and human anatomy and positions by the privacy places of human body, including the sex judging tested personnel, when tested personnel are male, the region of downward to human body head region and human body waist center torso width 1/2 is defined as privacy places, when tested personnel are women, it is privacy places by the region of downward to human body head region and human body waist center torso width 1/2 and trunk from trunk upper end down to the region of trunk height 1/2.Mm-wave imaging for the tested personnel of male is illustrated by the present invention.
Perform step C, privacy places to human body shields and indicates the privacy mask module of the concealment Item Information on human body and following several detailed description of the invention can be adopted to carry out the protection of privacy:
(1) privacy mask module includes obscure portions module and the first sign module, privacy places is carried out Fuzzy processing on described original image and forms part obfuscation original image by obscure portions module, obfuscation can use morphology operations, use the mosaic area that the length of side is certain, or can directly use the rectangular block of solid color that privacy places is covered；First indicates module marks described concealment article sign frame on described obscure portions original image, the frame of such as highlight color, as shown in figure 28.This mode is adapted to concealment article and automatically detects situation about combining with manual detection.Or,
(2) privacy mask module includes selecting purpose image module and second to indicate module, because the image of the binaryzation of complete display can mask the Pixel Information of the privacy places of tested personnel, purpose image module is selected to select purpose graphical representation human body parts so present embodiment utilizes, this purpose image is obtained by the described human body detecting method based on mm-wave imaging, such as Fig. 8；Before this purpose image shows concealment Item Information, second indicates module first judges whether concealment article are in the privacy places of human body, if, the color blocks different from human body color is then used to represent concealment article and indicate on described purpose image, as used certain gray value, the grey rectangle block such as 128 indicates；If it is not, then directly the original image of concealment article is shown on described purpose image.Shielding the complete information of human body in original image so completely, reach again the purpose of secret protection, effect is as shown in figure 29.This method is only applicable under the concealment automatic detection mode of article, it is not necessary to the situation of display complete human body figure.Or,
(3) privacy mask module includes whole obfuscation module and the 3rd sign module.Human body is carried out whole Fuzzy processing on described original image and forms whole obfuscation original image by whole obfuscation modules, and the method for obfuscation can be identical with detailed description of the invention (1), and in present embodiment, obfuscation employs operator Before these whole obfuscation original images show concealment Item Information, 3rd indicates module judges whether concealment article are in the privacy places of human body, if, the color blocks different from human body color is then used to represent concealment article and indicate on described whole obfuscation original images, as used certain gray value, the grey rectangle block such as 128 indicates；If it is not, then the original image by concealment article is shown on described whole obfuscation original image, design sketch is Figure 30 such as.This method is only applicable under the concealment automatic detection mode of article, it is not necessary to the situation of display complete human body figure.Or,
(4) privacy mask module includes preference pattern module and the 4th sign module.Preference pattern module selects the anthropometric dummy in described human detection, and anthropometric dummy here can be bar band model, it is also possible to be bar combination model；4th indicates module uses the color blocks different from anthropometric dummy color, and such as redness, expression concealment article also indicate in described anthropometric dummy.So can also shield the privacy information of human body completely.On bar band model, the effect of display is as shown in figure 31.This method is only applicable to the complete information the not showing concealment article situation also without display complete human body figure.
Should be appreciated that above is illustrative and not restrictive by preferred embodiment to the detailed description that technical scheme carries out.Technical scheme described in each embodiment can be modified by those of ordinary skill in the art on the basis of reading description of the present invention, or wherein portion of techniques feature carries out equivalent replacement；And these amendments or replacement, do not make the essence of appropriate technical solution depart from the spirit and scope of various embodiments of the present invention technical scheme.Protection scope of the present invention is only limited by following claims.
Claims (2)
1. hide the detection automatically of article and identify device for one kind, it is characterised in that including:
Scanning means, obtains original image for tested personnel carry out millimeter wave scanning；
Adjusting module, for being adjusted described original image obtaining target image；
Segmentation locating module, for carrying out segmentation and the location of human body according to described target image, the step of described human body locating segmentation includes:
Downward from image top along the vertical centrage of human body, the first man body region boundary point found, if judging to continue to have downwards continuously from this point and length being not less than 1/10 same gray value of picture altitude, then determining the crown central point that this point is human body, this horizontal line residing for point is the crown horizontal line H2 of human body；
The position of human foot is fixing position in the picture, thus may determine that sole coordinate and residing horizontal line are H9, the position that during due to imaging, people stands is fixing, deducts the vertical coordinate of sole with the vertical coordinate on the crown and can obtain the height H of people；
Judge according to human anatomy, head part accounts for the 15% of height, may determine that the coordinate of neck upper end and residing horizontal line H4, the height of neck accounts for the 45% of height of head, therefore may determine that the coordinate of the edge point of trunk upper end and residing horizontal line H5；
Described target image carrying out refinement and obtains human body backbone figure, in human body backbone figure, the cross point of trunk and two lower limbs is as the position of trunk lower end, and its place horizontal line, as stringcourse H6, is the long Hleg of lower limb along this line to the horizontal distance definition of foot；
Judging according to human anatomy, the ratio of human calf and thigh is about 1:1.2, it is determined that the horizontal level of knee is the position from foot upwards Hleg × 5/11, therefore obtains knee level line H8；
On described target image, from image base along vertical centrage upwards, with the first of human body image position of intersecting point as crotch, thus obtaining human body crotch horizontal line H7；
For cut-off rule, the image of human region being divided into left and right two halves, the peak of left one side of something and the position of left hand finger tip with vertical centrage, on image right, the peak of human body is exactly the position of right hand finger tip, thus obtaining finger tip horizontal line H0；
Due to before tested personnel are scanned, requiring that its two-arm is flared out, therefore on target image, width between two elbows is the widest position of human body, and therefore finding the position of the human region leftmost side is exactly the left elbow of health, the rightmost side is exactly the right elbow of health, thus obtaining elbow lever line H3；
Difference in height between two finger tips is ignored, the difference in height between two elbows is ignored；
Length ratio according to human dissection theory, general human body hands and upper arm is 7:9, therefore may determine that the position of wrist according to the sharp position with elbow of hands, thus also obtain wrist horizontal line H1；
Bar combination model generation module, is used for generating bar combination model；Described bar combination model generation module includes following submodule:
Stick model generation module, for generating the stick model providing each key point of human body；
Band model generation module, for generating the bar band model providing human body contour outline information；
Binding modules, is used in conjunction with described stick model and described bar band model；
Non-human target preliminary detection module, for detecting non-human target according to described original image, it is thus achieved that non-human target distribution original image；This non-human target preliminary detection module includes
Edge detection module, for described original image is carried out rim detection, the non-human target of preliminary identification；
Highlight module, for highlighting non-human target distribution region by mathematical morphological operation；
Regularization module, obtains non-human goal rule regional distribution chart for choosing minimum circumscribed rectangle according to the border in described non-human target distribution region；
Fusion Module, is used for merging described non-human goal rule regional distribution chart and described original image, it is thus achieved that described non-human target original image；
Non-human target distribution module, for utilizing described bar combination model to obtain the described non-human target distribution original image location distribution information relative to human body；Described non-human target distribution module is by being input on described bar combination model by described non-human target distribution original image；
Classification recognition module, for carrying out classification identification and showing the concealment article location distribution information relative to human body to described non-human target.
2. the automatic of concealment article according to claim 1 detects and identifies device, it is characterised in that described classification recognition module includes following submodule:
Exposed spots localization module, for positioning the exposed position of human body；
Sort module, is defined as non-concealment article for will be distributed over the non-human target on the exposed position of human body, will be distributed over the non-human target outside the exposed position of human body and is defined as concealment article；
Display module, for rejecting the original image of described non-concealment article and showing concealment article original image distributed intelligence on described bar combination model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210050329.XA CN102629315B (en) | 2012-02-29 | 2012-02-29 | A kind of detection automatically hiding article and identification device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210050329.XA CN102629315B (en) | 2012-02-29 | 2012-02-29 | A kind of detection automatically hiding article and identification device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102629315A CN102629315A (en) | 2012-08-08 |
| CN102629315B true CN102629315B (en) | 2016-07-06 |
Family
ID=46587573
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210050329.XA Expired - Fee Related CN102629315B (en) | 2012-02-29 | 2012-02-29 | A kind of detection automatically hiding article and identification device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102629315B (en) |
Families Citing this family (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103942556B (en) * | 2014-04-22 | 2017-11-21 | 天津重方科技有限公司 | The identification of human body privacy and processing method based on X ray backscatter images |
| CN113759634B (en) | 2016-01-07 | 2024-07-12 | Lg伊诺特有限公司 | Lens driving device, camera module and optical equipment |
| CN106372583B (en) * | 2016-08-25 | 2018-09-14 | 华讯方舟科技有限公司 | A kind of human body foreign body detection method and system based on millimeter-wave image |
| EP3506148A4 (en) * | 2016-08-25 | 2019-07-24 | China Communication Technology Co., Ltd. | Millimetre wave image-based human body foreign matter detection method and system |
| CN106446819B (en) * | 2016-09-20 | 2019-04-16 | 上海理工大学 | The recognition methods of human body contour outline edge object in human body safety check imaging |
| CN106371148B (en) * | 2016-09-27 | 2019-05-03 | 华讯方舟科技有限公司 | A kind of human body foreign body detection method and system based on millimeter-wave image |
| CN106485077B (en) * | 2016-10-11 | 2018-12-14 | 徐州玖胜医疗器械有限公司 | A kind of supervision place medical monitoring systems |
| CN106454238A (en) * | 2016-10-11 | 2017-02-22 | 深圳万发创新进出口贸易有限公司 | Community anti-theft system based on the Internet of things |
| CN106373242A (en) * | 2016-10-11 | 2017-02-01 | 深圳万发创新进出口贸易有限公司 | Substation security entrance guard system |
| CN106448007A (en) * | 2016-10-11 | 2017-02-22 | 深圳万发创新进出口贸易有限公司 | Supermarket detection all-in-one machine |
| CN109117694A (en) * | 2017-06-26 | 2019-01-01 | 刘风华 | Contactless identification people, animal and plant |
| CN107589725B (en) * | 2017-08-29 | 2020-03-24 | 深圳市盛路物联通讯技术有限公司 | Smart home equipment control method and control equipment based on built-in antenna |
| CN108937869A (en) * | 2018-07-12 | 2018-12-07 | 芜湖博高光电科技股份有限公司 | A kind of the millimeter wave survival kit and operating method of hand-held perspective bone |
| CN108986023A (en) | 2018-08-03 | 2018-12-11 | 北京字节跳动网络技术有限公司 | Method and apparatus for handling image |
| CN109146883A (en) * | 2018-10-31 | 2019-01-04 | 中控智慧科技股份有限公司 | A kind of image processing method and screening machine |
| 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 |
| CN109597069A (en) * | 2018-12-25 | 2019-04-09 | 山东雷诚电子科技有限公司 | A kind of active MMW imaging method for secret protection |
| CN109799544B (en) * | 2018-12-28 | 2021-03-19 | 深圳市重投华讯太赫兹科技有限公司 | Intelligent detection method and device applied to millimeter wave security check instrument and storage device |
| CN111209793A (en) * | 2019-12-05 | 2020-05-29 | 重庆特斯联智慧科技股份有限公司 | Region shielding human body security check method and system based on artificial intelligence |
| CN111623878A (en) * | 2020-04-27 | 2020-09-04 | 深圳市国脉科技有限公司 | Body temperature measuring method and device, computer readable medium and electronic equipment |
| US11231498B1 (en) | 2020-07-21 | 2022-01-25 | International Business Machines Corporation | Concealed object detection |
| CN112597989B (en) * | 2020-12-18 | 2021-09-14 | 中国科学院上海微系统与信息技术研究所 | Millimeter wave three-dimensional holographic image concealed article detection method and system |
| CN113657497B (en) * | 2021-08-17 | 2024-03-08 | 上海亨临光电科技有限公司 | Method for realizing low signal-to-noise ratio image enhancement processing and human body image segmentation fusion |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101614722A (en) * | 2008-06-27 | 2009-12-30 | 同方威视技术股份有限公司 | A kind of personal security check system |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7123185B2 (en) * | 2004-04-14 | 2006-10-17 | Safeview, Inc. | Enhanced surveilled subject imaging |
| CN102028482B (en) * | 2009-09-30 | 2012-11-14 | 同方威视技术股份有限公司 | Human body detection image processing method and human body detection apparatus |
| CN101996317B (en) * | 2010-11-01 | 2012-11-21 | 中国科学院深圳先进技术研究院 | Method and device for identifying markers in human body |
-
2012
- 2012-02-29 CN CN201210050329.XA patent/CN102629315B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101614722A (en) * | 2008-06-27 | 2009-12-30 | 同方威视技术股份有限公司 | A kind of personal security check system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102629315A (en) | 2012-08-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102629315B (en) | A kind of detection automatically hiding article and identification device | |
| CN102697503B (en) | Human detection method based on millimeter wave imaging | |
| CN102708372B (en) | A kind of automatic detection and Identification method of hiding article | |
| CN102708560B (en) | A kind of method for secret protection based on mm-wave imaging | |
| CN102707322B (en) | Human body detection device based on millimeter wave imaging | |
| CN102708561B (en) | A kind of secret protection device based on mm-wave imaging | |
| CN103810478B (en) | Sitting posture detection method and device | |
| CN109447071A (en) | A kind of mm-wave imaging dangerous goods detection method based on FPGA and deep learning | |
| CN107169475B (en) | A kind of face three-dimensional point cloud optimized treatment method based on kinect camera | |
| CN110133739A (en) | A kind of X-ray rays safety detection apparatus and its drawing method is known automatically | |
| CN110186940B (en) | Security check identification method and device, computer equipment and storage medium | |
| CN108665487A (en) | Substation's manipulating object and object localization method based on the fusion of infrared and visible light | |
| CN109544563A (en) | A kind of passive millimeter wave image human body target dividing method towards violated object safety check | |
| US10884116B2 (en) | Human-body foreign-matter detection method and system based on millimetre-wave image | |
| CN106447634B (en) | A kind of privacy places positioning and guard method based on active mm-wave imaging | |
| CN106327459A (en) | Visible light and infrared image fusion algorithm based on UDCT (Uniform Discrete Curvelet Transform) and PCNN (Pulse Coupled Neural Network) | |
| CN102567733B (en) | Detection and shielding method for private parts of human bodies under non-cooperation condition in microwave images | |
| EP3506148A1 (en) | Millimetre wave image-based human body foreign matter detection method and system | |
| CN110245675A (en) | A kind of dangerous objects detection method based on millimeter-wave image human body contextual information | |
| CN108108740B (en) | Active millimeter wave human body image gender identification method | |
| CN109102003A (en) | A kind of small target detecting method and system based on Infrared Physics Fusion Features | |
| CN106326834A (en) | Human body gender automatic identification method and apparatus | |
| CN104361571B (en) | Infrared and low-light image fusion method based on marginal information and support degree transformation | |
| CN105118035A (en) | Self-adaptive optical spot signal extraction method based on sparse representation | |
| Yang et al. | A novel deformable body partition model for MMW suspicious object detection and dynamic tracking |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160706 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |