CN111985331A - Detection method and device for preventing secret of business from being stolen - Google Patents
Detection method and device for preventing secret of business from being stolen Download PDFInfo
- Publication number
- CN111985331A CN111985331A CN202010695613.7A CN202010695613A CN111985331A CN 111985331 A CN111985331 A CN 111985331A CN 202010695613 A CN202010695613 A CN 202010695613A CN 111985331 A CN111985331 A CN 111985331A
- Authority
- CN
- China
- Prior art keywords
- human body
- picture
- key points
- processed
- photographing device
- 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
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 110
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 46
- 238000012544 monitoring process Methods 0.000 claims abstract description 17
- 230000006399 behavior Effects 0.000 claims description 18
- 238000000034 method Methods 0.000 claims description 17
- 238000004364 calculation method Methods 0.000 claims description 12
- 210000000707 wrist Anatomy 0.000 claims description 4
- 230000011664 signaling Effects 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 2
- 238000003672 processing method Methods 0.000 claims description 2
- 238000012549 training Methods 0.000 description 11
- 238000013527 convolutional neural network Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 210000003423 ankle Anatomy 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 210000001508 eye Anatomy 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 230000036544 posture Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 210000000697 sensory organ Anatomy 0.000 description 1
- 238000011895 specific detection Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Signal Processing (AREA)
- Burglar Alarm Systems (AREA)
Abstract
The utility model provides a detection method and a device for preventing the secret of business from being stolen, which relates to the technical field of electronic information, and the specific technical proposal is as follows: the image acquisition equipment acquires a monitoring picture in a preset area in real time and sends the acquired monitoring picture to the anti-theft photo detection equipment; detecting the picture to be processed by adopting a target detection algorithm to obtain coordinate information and corresponding confidence degrees of all photographing devices in the picture to be processed; detecting the picture to be processed by using a target detection algorithm and a human body key point detection algorithm in combination with the photographing device and the human body key points, detecting a corresponding target area, and obtaining coordinate information of all human body key points in the picture to be processed; after coordinate information of all human body key points in the picture to be processed is obtained, according to the difference of human body parts where the human body key points are located, a weight value corresponding to each human body key point and a preset threshold value of confidence coefficient of the weight value are set, the existence of a steal photo behavior is determined, and an alarm signal is generated or a display screen is controlled to display an alarm picture.
Description
Technical Field
The invention relates to a detection method and a detection device for preventing a trade secret from being stolen in the technical field of electronic information. And more particularly to a detection method for preventing the theft of commercial secrets applied to a theft detection device.
Background
With the continuous development of high technology, the enhancement of computer functions and the popularization and application of computer applications, many matters now need to be processed by a computer, data displayed on a computer display screen can be taken into corresponding photos even without entering a confidential place, and if the data are not protected, the data are probably stolen in a mode of taking pictures and the like to cause data leakage. Computers are an indispensable tool for modern offices, some works cannot be carried out without computers, and many enterprises can use computers to store important information and even secrets of many companies. Whether the computer is a desktop computer or a laptop computer, many employees including managers often put the computer in a company after work or when going out, so that if an employee or other people with improper thinking possibly steal the computer to open the computer and check and copy the trade secrets in the computer. Therefore, a computer storing a trade secret without an access code would allow a thief to enter an unmanned environment. When the computer is in work, some people can leave the computer of the computer temporarily, some people forget to turn off the computer for a while, and even if the opening password is set, other people can still enter the computer in the time period. Because the special stealing-photograph equipment has the concealment or disguise property, the special stealing-photograph equipment has certain difficulty in finding or identifying the special stealing-photograph equipment, and cannot be directly detected by adopting a detection method of the set of special stealing-photograph equipment, so that the special stealing-photograph equipment becomes the difficulty in detection work of the special stealing-photograph equipment. Although the security management and control of computer equipment and networks are basically mature at present, the technical means for preventing screens from being stolen is still weak, and the events that the display information of the computer screens is shot by people and divulged are also frequent, thereby causing great loss to national security, enterprises and public institutions. In the prior art, the task of object detection is to find all objects of interest (objects) in an image, determine their positions and sizes, and is one of the core problems in the field of machine vision. Because various objects have different appearances, shapes and postures, and interference of factors such as illumination, shielding and the like during imaging is added, target detection is always the most challenging problem in the field of machine vision. Object detection is the basis of many computer vision tasks, which provide reliable information whether we need to implement image-to-text interaction or identify fine categories. In order to prevent the hand-held stealing shooting device from determining whether a stealing behavior occurs or not by identifying whether a shooting device exists in a monitoring picture or not, a picture or a video is given to judge what kind of target is contained in the picture or the video, and the position of the target and which target object or scene each pixel belongs to are located. The special equipment for preventing the stealing of photos and secret disclosure in the man-machine information visual interaction link can automatically hide display contents immediately when people take photos illegally with equipment such as mobile phones, cameras and video cameras, effectively prevent screen information from being shot and stolen, and can identify the identity of a user through face recognition, but the detection effect of the special equipment in a single detector is usually poor, and the position of a target is not accurate. If the photographing device is held in the hand of the user and does not take a photograph, the recognition method has the problem of easy misjudgment, and the reliability of user experience is reduced.
The traditional target detection method generally comprises three stages: a) selecting candidate regions on a given image; b) extracting features from the regions; c) the classification is performed using a trained classifier. In the stage b), it is often necessary to manually acquire the expression information related to the target in the original input, and further perform classifier learning on the extracted feature information related to the target, however, the manual feature extraction method has many limitations. On one hand, the manual method depends on a specific detection task to a great extent, for different targets or different forms of the same target, a designer needs to carefully think how to extract the characteristics of the targets, and the final recognition effect of the model is also limited by the experience of the designer; on the other hand, conventional detection models separate feature extraction and classification training. If the manually extracted features in the feature description are not sufficient to describe an object, some missing useful information can no longer be recovered from the classification training. These shortcomings prevent conventional detection models from obtaining signatures that are more consistent with target characteristics. Before the advent of the Deep Convolutional Neural Network (DCNN), the DPM algorithm was the best-shown algorithm in the field of target detection, and its basic idea was to extract DPM artificial features (as shown in the following figure) and then classify them by latentSVM. Firstly, the DPM feature calculation is complex and the calculation speed is slow; secondly, the artificial features have poor detection effects on objects that are rotated, stretched, and have varying viewing angles. These drawbacks greatly limit the application scenarios of the algorithms. Target tracking and target detection are both classical problems in the field of computer vision, but the problems solved by both are not the same. The input of target detection is generally a picture, and the output is the position of a frame containing an object in the picture. The inputs for target tracking are the video and the position of the object to be tracked in the first frame. The target detection is mostly applied to pictures and also to videos, and frame-by-frame detection is performed, but the problem is that the target detection algorithm cannot judge which object in the previous frame is the same object as which object in the previous frame. This is the core problem solved by the target tracking algorithm. Object detection algorithms are currently generally only able to identify objects of a particular class, i.e. the class present in the training set. Target tracking generally has no requirement on object categories, that is, an excellent target tracking algorithm can track objects of categories that have not been found in the training set. In terms of accuracy, the detection algorithm can find the targets frame by frame, but cannot solve the correlation between the targets, that is, the improved YOLOv33 algorithm can find one type of targets and detect the targets in real time basically, but does not know whether the type is the same object as the detected type of the frame, and the current mainstream tracking algorithm is specially used for the detection.
Disclosure of Invention
The invention aims to solve the problem of easy misjudgment in the prior art, provides a detection method for preventing the stealing of the trade secret, which can improve the detection accuracy and avoid the misjudgment problem caused by the independent detection of a photographing device in the prior art, and further aims to provide a detection device for preventing the stealing of the trade secret so as to solve the misjudgment problem caused by the detection of only the photographing device in the prior art.
The above object of the present invention can be achieved by the following technical solutions: a method of preventing hacking of a trade secret, comprising the steps of:
acquiring a picture to be processed: the image acquisition equipment monitors the preset area in real time, acquires a monitoring picture in the preset area, acquires the monitoring picture in the sensitive area acquired by the image acquisition equipment, and sends the acquired monitoring picture to the anti-theft photo detection equipment;
detecting whether there is a steal behavior: detecting the picture to be processed by adopting a target detection algorithm to obtain coordinate information and corresponding confidence degrees of all photographing devices in the picture to be processed;
detecting key points of a human body on the picture to be processed: detecting the picture to be processed by using a target detection algorithm and a human body key point detection algorithm in combination with the photographing device and the human body key points, detecting a corresponding target area, and obtaining coordinate information of all human body key points in the picture to be processed;
calculating the weight statistic value of the human body key points: after coordinate information of all human body key points in the picture to be processed is obtained, a weight value corresponding to each human body key point and a preset threshold value of the confidence coefficient of the weight value are set according to different human body parts where the human body key points are located. After the length of the preset area is determined, acquiring weight statistics of all human body key points in the preset area, calculating to obtain the weight statistics of all human body key points in the preset area according to the weight value of each human body key point in the preset area, judging whether the confidence coefficient of each photographing device is greater than a first preset threshold, acquiring all human body key points in the preset area by taking the center point coordinate of a boundary frame of the photographing device as an original point when the confidence coefficient of the photographing device is greater than the first preset threshold, and calculating to obtain the weight statistics of all human body key points in the preset area according to the weight value corresponding to each human body key point; and determining that the steal behavior exists according to the fact that the weight statistic value in the preset area is larger than or equal to the size of a second preset threshold value, and generating an alarm signal or controlling a display screen to display an alarm picture.
Compared with the prior art, the invention has the following beneficial effects:
the method comprises the steps of detecting a picture to be processed through a target detection algorithm and a human key point detection algorithm to obtain position information of a photographing device and coordinate information of human key points, acquiring a weight statistic value of the human key points in a preset area with the photographing device as the center when the confidence coefficient of the photographing device is larger than a first preset threshold value, and determining that a stealing behavior exists when the weight statistic value is larger than or equal to a second preset threshold value. Through combining together the device of shooing and human key point and detecting, can improve the degree of accuracy that detects, avoid detecting the erroneous judgement problem that the device of shooing brought alone among the prior art, improve user experience.
The detection device for preventing the commercial secrets from being stolen provided by the embodiment of the invention detects the picture to be processed through the target detection algorithm and the human key point detection algorithm to obtain the position information of the photographing device and the coordinate information of the human key points, obtains the weight statistic value of the human key points in the preset area taking the photographing device as the center when the confidence coefficient of the photographing device is greater than the first preset threshold value, and determines that the stealing behavior exists when the weight statistic value is greater than or equal to the second preset threshold value. Through combining together the device of shooing and human key point and detecting, can improve the degree of accuracy that detects, avoid detecting the erroneous judgement problem that the device of shooing brought alone among the prior art, improve user experience.
According to the detection method for preventing the commercial secret from being stolen, the detection is completed by a single-stage object detection classical algorithm SSD through a plurality of characteristic graphs. Detecting the picture to be processed through a target detection algorithm and a human key point detection algorithm to obtain position information of a photographing device and coordinate information of human key points, acquiring a weight statistic of human key points in a preset area taking the photographing device as a center when the confidence of the photographing device is greater than a first preset threshold, and determining that a steal action exists when the weight statistic is greater than or equal to a second preset threshold. Through combining together the device of shooing and human key point and detecting, can improve the degree of accuracy that detects, avoid detecting the erroneous judgement problem that the device of shooing brought alone among the prior art, improve user experience.
The invention is used for anti-theft illumination detection. The problem of misjudgment caused by the existing photographing device can be solved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.
FIG. 1 is a schematic flow diagram of a detection method of the present invention for preventing hacking of commercial secrets;
FIG. 2 is a schematic diagram of the structure of a detection device for preventing the theft of commercial secrets;
fig. 3 is a schematic structural view of the detection device of fig. 2.
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.
Detailed Description
See fig. 1. According to the invention, the following steps are adopted:
acquiring a picture to be processed: the image acquisition equipment monitors the preset area in real time, acquires a monitoring picture in the preset area, acquires the monitoring picture in the sensitive area acquired by the image acquisition equipment, and sends the acquired monitoring picture to the anti-theft photo detection equipment;
detecting whether there is a steal behavior: detecting the picture to be processed by adopting a target detection algorithm to obtain coordinate information and corresponding confidence degrees of all photographing devices in the picture to be processed;
detecting key points of a human body on the picture to be processed: detecting the picture to be processed by using a target detection algorithm and a human body key point detection algorithm in combination with the photographing device and the human body key points, detecting a corresponding target area, and obtaining coordinate information of all human body key points in the picture to be processed;
calculating the weight statistic value of the human body key points: after coordinate information of all human body key points in the picture to be processed is obtained, a weight value corresponding to each human body key point and a preset threshold value of the confidence coefficient of the weight value are set according to different human body parts where the human body key points are located. After the length of the preset area is determined, acquiring weight statistics of all human body key points in the preset area, calculating to obtain the weight statistics of all human body key points in the preset area according to the weight value of each human body key point in the preset area, judging whether the confidence coefficient of each photographing device is greater than a first preset threshold, acquiring all human body key points in the preset area by taking the center point coordinate of a boundary frame of the photographing device as an original point when the confidence coefficient of the photographing device is greater than the first preset threshold, and calculating to obtain the weight statistics of all human body key points in the preset area according to the weight value corresponding to each human body key point; and determining that the steal behavior exists according to the fact that the weight statistic value in the preset area is larger than or equal to the size of a second preset threshold value, and generating an alarm signal or controlling a display screen to display an alarm picture.
The embodiment of the disclosure provides a detection method for preventing a commercial secret from being stolen, which is applied to anti-theft detection equipment, and the detection method comprises the following steps:
101. and acquiring a picture to be processed.
In the embodiment of the present disclosure, acquiring a to-be-processed picture includes: and acquiring a monitoring picture in the sensitive area acquired by the image acquisition equipment. The sensitive area can be an area which is monitored by the image acquisition equipment and needs to be kept secret, and the sensitive area can be a place which needs to be kept secret, such as a display screen, a file, a commodity and the like. Taking the display screen as an example, the sensitive area is an area opposite to the display screen displaying the data to be protected, i.e., the sensitive area is an area where the display screen can be photographed.
In the embodiment of the present disclosure, the image capturing apparatus may be an apparatus having a shooting function, such as a camera or a camera. The image acquisition equipment can be a part of the anti-theft detection equipment, and also can be external image acquisition equipment which is connected and communicated with the anti-theft detection equipment through an external interface; the image acquisition device may also be part of the display screen or an external image acquisition device that communicates with the display screen through an external interface. The anti-theft detection device may be part of the display screen or an external device connected to the display screen.
In the embodiment of the disclosure, the image acquisition device monitors the preset area in real time, acquires the monitoring picture in the preset area, and sends the acquired monitoring picture to the anti-theft photo detection device, wherein the monitoring picture is a W × H color image.
102. And detecting the picture to be processed by adopting a target detection algorithm to obtain the position information and the corresponding confidence of all photographing devices in the picture to be processed.
In one embodiment, the coordinate information of the photographing apparatus includes coordinates of a center point of the bounding box of the photographing apparatus; the photographing device can be a mobile phone, a tablet, a camera and the like. In the embodiment of the present disclosure, the target detection algorithm may be a single-stage target detection classical algorithm SSD (the single shottector) algorithm, a yolo (younlookone) target detection algorithm, and the like, and a framework of the single-stage target detection classical algorithm SSD is on one basic CNN network, but may also be replaced by other networks, and some additional structures are added, so that the network has the following characteristics: detecting by using a multi-scale feature map; the real-time target detection algorithm YOLO algorithm adopts a single convolutional neural network to predict a plurality of bounding boxes and class probabilities, target region prediction and target class prediction are combined into a whole, and a target detection task is regarded as regression of the target region prediction and the class prediction. The real-time target detection algorithm YOLO first divides the image into an S × S grid. If the center of an object falls into a grid, the grid is responsible for detecting the object. B bounding boxes and confidence values (confidenccecore) are predicted in each grid. These confidence scores reflect the model's confidence in whether the box contains an object, and how accurately it predicts the box. Detecting different target types at different observation distances by a single detector, and detecting the target by sliding windows on the characteristic diagram by using window types with different sizes and aspect ratios; in Selective Search (SS), each pixel is taken as a group, then, texture of each group is calculated, and two closest groups are combined until all regions are combined together, a feature extractor (CNN) extracts features of the whole image first, and coordinate information and corresponding confidence of all photographing devices in the picture to be processed are obtained by performing target detection on the picture to be processed, wherein the coordinate information is represented by (x, y, w, h), (x, y) represents a center point coordinate of a boundary frame of the photographing device, and (w, h) represents a width and a height of the boundary frame of the photographing device.
The picture to be processed comprises at least one photographing device, and coordinate information and corresponding confidence of all photographing devices in the picture to be processed are obtained by performing target detection on the picture to be processed.
103. And detecting the key points of the human body scanning of the picture to be processed to obtain the coordinate information of all the key points of the human body in the picture to be processed.
The human body key points can be joints or positions of five sense organs, and exemplary key points which can be detected on the human body include: left and right eyes, nose, left and right ears, left and right shoulders, left and right elbows, left and right wrists, left and right hips, left and right knees, left and right ankles, etc. In the embodiment of the present disclosure, in order to detect whether there is a steal behavior, the key points of the human body are mainly the left and right shoulders, the left and right elbows, the left and right wrists, and the neck.
After coordinate information of all human body key points in the picture to be processed is obtained, a weight value corresponding to each human body key point is set according to different human body parts where the human body key points are located. Illustratively, the weight value for the wrist keypoint is 25, the weight value for the neck keypoint is 20, the weight value for the shoulder keypoint is 15, and the weight value for the elbow keypoint is 10. Of course, a default weight value may also be used for the weight value corresponding to the human body key point. The detector will make repeated detections for the same target. Non-maxima suppression is used to remove duplicate detections with low confidence, ranked from high to low confidence. If any default weight value is the same as the class predicted by the current default weight value and the default weight value is greater than 0.5, it is removed from the sequence.
104. When the confidence of the photographing device is larger than a first preset threshold, acquiring the weight statistic of all human key points in a preset area taking the photographing device as the center according to the coordinate information of the photographing device and the coordinate information of the human key points.
In the disclosed embodiment, the preset area may be a circle, a square, or the like. The length of the preset area (for a circle, the length is the radius of the circle; for a square, the length is the side length of the square) can be set according to historical experience, for example, the preset area is a circular area with the radius R, and the value range of the radius R can be 60-100 pixels; the probability of each category can also be calculated according to a preset human body key point training set.
The method for calculating according to the preset human body key point training set comprises the following steps: the preset human body key point training set is a training set for detecting human body key points. Specifically, calculating Euclidean distances between all elbow key points and corresponding shoulder key points in a preset human body key point training set; calculating the mean value and the standard deviation of all Euclidean distances according to the Euclidean distances between all elbow key points and corresponding shoulder key points; the sum of the mean and the standard deviation is twice determined as the length of the preset area.
After the length of the preset area is determined, acquiring the weight statistics of all human body key points in the preset area comprises the following steps:
when the confidence of the photographing device is larger than a first preset threshold, acquiring all human body key points in a preset area with the central point coordinate of the boundary frame of the photographing device as an original point;
and calculating to obtain a weight statistic value of all human body key points in the preset area according to the weight value of each human body key point in the preset area.
Specifically, whether the confidence of each photographing device is larger than a first preset threshold value or not is judged, when the confidence of each photographing device is larger than the first preset threshold value, all human key points in a preset area are obtained by taking the coordinates of the central point of the boundary frame of the photographing device as the original point, and then the weight statistics value of all human key points in the preset area is obtained through calculation according to the weight value corresponding to each human scanning key point.
105. And when the weight statistic value is greater than or equal to a second preset threshold value, determining that the steal behavior exists.
In the embodiment of the present disclosure, the higher the weight statistics value is, the higher the possibility that the photographing apparatus is used for stealing, so that when the weight statistics value in the preset area is greater than or equal to the second preset threshold value, it is determined that the photographing apparatus is used for stealing, that is, there is a stealing behavior. When it is determined that there is a steal behavior, an alarm signal may be generated or a display screen may be controlled to display an alarm screen.
The predetermined area is a circle with a radius R as an example. In one embodiment, the first preset threshold is 0.95, the second preset threshold is 40, the preset area is a circular area with a radius R, and the radius R ranges from 60 pixels to 100 pixels. When the confidence of the photographing device is larger than 0.95, the central point coordinate of the boundary box of the photographing device is used as the original point, the weight statistic value of all human body key points in the circular area with the radius R is calculated, and when the weight statistic value in the preset area is larger than or equal to 40, the photographing device is considered to be used for stealing the light, and the display screen is controlled to display a warning picture.
Based on the detection method for preventing the secret of the business from being stolen as described in the embodiment corresponding to fig. 1, the following is an embodiment of the disclosed device, which can be used to execute the embodiment of the disclosed method.
When the confidence of the photographing device is greater than a first preset threshold, the SSD completes detection through a plurality of characteristic graphs, and according to the coordinate information of the photographing device and the coordinate information of the human key points, the SSD calculates the weight statistics of all the human key points in a preset area taking the photographing device as the center, wherein the weight statistics comprises: when the confidence of the photographing device is larger than a first preset threshold, all human key points in a preset area with the central point coordinate of the boundary frame of the photographing device as the original point and a weight value corresponding to each human key point in the preset area are obtained, and a weight statistic value of all human key points in the preset area is obtained through calculation.
In one embodiment, before obtaining the picture to be processed, the method further includes: calculating Euclidean distances between all elbow key points and corresponding shoulder key points in a preset human body key point training set, calculating the mean value and standard deviation of the Euclidean distances, and determining the length of a preset area by twice of the sum of the mean value and standard deviation of the Euclidean distances.
See fig. 2. The disclosed embodiment also provides a detection device for preventing the trade secret from being stolen, which comprises: the image processing device comprises a receiver and a transmitter which are connected with an antenna, a processor which is connected with the receiver and the transmitter at the same end, a memory which stores at least one computer instruction, and a processor which at least comprises a detection module, wherein the processor is connected with the memory, loads and executes the at least one computer instruction, and detects a picture to be processed and human key points by adopting a target detection algorithm to obtain coordinate information and corresponding confidence degrees of all photographing devices in the picture to be processed and coordinate information of all human key points; according to the coordinate information of the photographing device and the coordinate information of the human key points, setting a first preset threshold value larger than the confidence coefficient and a second preset threshold value with the weight value corresponding to each human key point larger than or equal to the weight statistic value by utilizing the difference of the human body parts where the human key points are located, and calculating the weight statistic value of all the human key points in a preset area with the photographing device as the center; and generating an alarm signal for determining the existence of the steal behavior or controlling a display screen to display an alarm picture according to the fact that the weight statistic value in the preset area is larger than or equal to a second preset threshold value.
The embodiment of the present disclosure based on fig. 2 further provides a computer-readable storage medium, for example, the non-transitory computer-readable storage medium may be a Read Only Memory (ROM), a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.
An embodiment of the present disclosure provides a processor as shown in fig. 3, including: the image processing method comprises a first detection module 202 connected with an acquisition module 201, a calculation module 204 connected with a second detection module 203 and a setting module 206, and a determination module 205 connected with the calculation module 204, wherein the acquisition module 201 respectively transmits acquired images to be processed to the first detection module 202 and the second detection module 203 through behavior chains, and the first detection module 202 detects the images to be processed by adopting a target detection algorithm to obtain coordinate information and corresponding confidence degrees of all photographing devices in the images to be processed; the second detection module 203 detects the key points of the human body of the picture to be processed to obtain the coordinate information of all the key points of the human body in the picture to be processed; the first detection module 202 and the second detection module 203 respectively send the obtained confidence level and the coordinate information of the human key points to the calculation module 204, and when the confidence level of the photographing device is greater than a first preset threshold value, the calculation module 204 calculates a weight statistic value of all human key points in a preset area with the photographing device as the center according to the coordinate information of the photographing device and the coordinate information of the human key points; the setting module 206 sets a weight value corresponding to each human body key point according to the difference of the human body parts where the human body key points input by the second detecting module 203 are located, and the determining module 205 determines that there is a steal action by using the weight statistic value provided by the calculating module 204 to be greater than or equal to a second preset threshold value.
When the confidence of the photographing device is greater than a first preset threshold, the first detection module 202 acquires all human body key points in a preset area with the center point coordinate of the boundary frame of the photographing device as an origin; the calculating module 204 calculates a weight statistic of all human body key points in the preset region according to the weight value corresponding to each human body key point in the preset region.
In one embodiment, the calculating module 204 calculates euclidean distances between all elbow key points and corresponding shoulder key points in the preset human body key point training set, calculates a mean value and a standard deviation of the euclidean distances according to the euclidean distances, and the determining module 205 determines twice of the sum of the mean value and the standard deviation of the euclidean distances as the length of the preset region.
In one embodiment, the coordinate information of the photographing apparatus includes coordinates of a center point of the bounding box of the photographing apparatus;
it is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
Claims (10)
1. A method of preventing hacking of a trade secret, comprising the steps of:
acquiring a picture to be processed: the image acquisition equipment monitors the preset area in real time, acquires a monitoring picture in the preset area, acquires the monitoring picture in the sensitive area acquired by the image acquisition equipment, and sends the acquired monitoring picture to the anti-theft photo detection equipment;
detecting whether there is a steal behavior: detecting the picture to be processed by adopting a target detection algorithm to obtain coordinate information and corresponding confidence degrees of all photographing devices in the picture to be processed;
detecting key points of a human body on the picture to be processed: detecting the picture to be processed by using a target detection algorithm and a human body key point detection algorithm in combination with the photographing device and the human body key points, detecting a corresponding target area, and obtaining coordinate information of all human body key points in the picture to be processed;
calculating the weight statistic value of the human body key points: after coordinate information of all human key points in a picture to be processed is obtained, a weighted value corresponding to each human key point and a preset threshold of a confidence coefficient of each human key point are set according to different human body parts where the human key points are located, when the confidence coefficient of a photographing device is larger than a first preset threshold, a weight statistic value of the human key points in a preset area with the photographing device as the center is obtained, according to the fact that the weight statistic value in the preset area is larger than or equal to a second preset threshold, it is determined that a stealing behavior exists, and an alarm signal is generated or a display screen is controlled to display an alarm picture.
2. After the length of the preset area is determined, acquiring weight statistics of all human body key points in the preset area, calculating to obtain the weight statistics of all human body key points in the preset area according to the weight value of each human body key point in the preset area, judging whether the confidence coefficient of each photographing device is greater than a first preset threshold, acquiring all human body key points in the preset area by taking the center point coordinate of a boundary frame of the photographing device as an original point when the confidence coefficient of the photographing device is greater than the first preset threshold, and calculating to obtain the weight statistics of all human body key points in the preset area according to the weight value corresponding to each human body key point; and determining that the steal behavior exists according to the fact that the weight statistic value in the preset area is larger than or equal to the size of a second preset threshold value, and generating an alarm signal or controlling a display screen to display an alarm picture.
3. The method of preventing hacking of trade secrets of claim 1 wherein: the method for acquiring the picture to be processed comprises the following steps: the method comprises the steps of obtaining a monitoring picture in a sensitive area acquired by image acquisition equipment, wherein the sensitive area is an area which needs to be kept secret and is monitored by the image acquisition equipment, and the sensitive area is an area of a display screen, a file and a commodity place which are opposite to the display screen for displaying data to be protected and needs to be kept secret.
4. The method of preventing hacking of trade secrets of claim 1 wherein: the target detection algorithm comprises a single detector and an algorithm thereof including a real-time target detection algorithm YOLO, a single-stage target detection algorithm SSD and RetinaNet, wherein the single detector detects different target types at different observation distances, and the target is detected by sliding windows on a feature map by using window types with different sizes and aspect ratios; in Selective Search (SS), each pixel is taken as a group, then, texture of each group is calculated, and two closest groups are combined until all regions are combined together, a feature extractor (CNN) extracts features of the whole image first, and coordinate information and corresponding confidence of all photographing devices in the picture to be processed are obtained by performing target detection on the picture to be processed, wherein the coordinate information is represented by (x, y, w, h), (x, y) represents a center point coordinate of a boundary frame of the photographing device, and (w, h) represents a width and a height of the boundary frame of the photographing device.
5. The method of preventing hacking of trade secrets of claim 1 wherein: the detected key points include: the weight value of the key point of the wrist is 25, the weight value of the key point of the neck is 20, the weight value of the key point of the shoulder is 15, and the weight value of the key point of the elbow is 10.
6. The method of preventing hacking of trade secrets of claim 1 wherein: the first preset threshold value is 0.95, the second preset threshold value is 40, the preset area is a circular area with the radius R, and the value range of the radius R is 60-100 pixels.
7. The method of preventing hacking of trade secrets of claim 1 wherein: when the confidence of the photographing device is larger than 0.95, the central point coordinate of the boundary box of the photographing device is used as the original point, the weight statistic value of all human body key points in the circular area with the radius R is calculated, and when the weight statistic value in the preset area is larger than or equal to 40, the photographing device is considered to be used for stealing the light, and the display screen is controlled to display a warning picture.
8. A device for preventing the hacking of commercial secrets using the method of claim 1 comprising: the system comprises a receiver and a transmitter which are connected with an antenna, a processor which is connected with the receiver and the transmitter at the same end, and a memory which stores at least one computer instruction, wherein the processor loads and executes the at least one computer instruction, and detects the picture to be processed and the key points of the human body by adopting a target detection algorithm to obtain the coordinate information and the corresponding confidence of all photographing devices in the picture to be processed and the coordinate information of all key points of the human body; according to the coordinate information of the photographing device and the coordinate information of the human key points, setting a first preset threshold value larger than the confidence coefficient and a second preset threshold value with the weight value corresponding to each human key point larger than or equal to the weight statistic value by utilizing the difference of the human body parts where the human key points are located, and calculating the weight statistic value of all the human key points in a preset area with the photographing device as the center; and generating an alarm signal for determining the existence of the steal behavior or controlling a display screen to display an alarm picture according to the fact that the weight statistic value in the preset area is larger than or equal to a second preset threshold value.
9. The apparatus for detecting theft of a trade secret as claimed in claim 8, wherein: the processor includes: the image processing method comprises a first detection module 202 connected with an acquisition module 201, a calculation module 204 connected with a second detection module 203 and a setting module 206, and a determination module 205 connected with the calculation module 204, wherein the acquisition module 201 sends an acquired to-be-processed image into the first detection module 202 and the second detection module 203 through a behavior chain, and the first detection module 202 detects the to-be-processed image by adopting a target detection algorithm to obtain coordinate information and corresponding confidence degrees of all photographing devices in the to-be-processed image; the second detection module 203 detects the key points of the human body of the picture to be processed to obtain the coordinate information of all the key points of the human body in the picture to be processed; the first detection module 202 and the second detection module 203 respectively send the obtained confidence level and the coordinate information of the human key points to the calculation module 204, and when the confidence level of the photographing device is greater than a first preset threshold value, the calculation module 204 calculates a weight statistic value of all human key points in a preset area with the photographing device as the center according to the coordinate information of the photographing device and the coordinate information of the human key points; the setting module 206 sets a weight value corresponding to each human body key point according to the difference of the human body parts where the human body key points input by the second detecting module 203 are located, and the determining module 205 determines that there is a steal action by using the weight statistic value provided by the calculating module 204 to be greater than or equal to a second preset threshold value.
10. The apparatus for detecting theft of a trade secret as claimed in claim 1, wherein: when the confidence of the photographing device is greater than a first preset threshold, the first detection module 202 acquires all human body key points in a preset area with the center point coordinate of the boundary frame of the photographing device as an origin; the calculating module 204 calculates a weight statistic of all human body key points in the preset region according to the weight value corresponding to each human body key point in the preset region.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010695613.7A CN111985331B (en) | 2020-07-20 | 2020-07-20 | Detection method and device for preventing trade secret from being stolen |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010695613.7A CN111985331B (en) | 2020-07-20 | 2020-07-20 | Detection method and device for preventing trade secret from being stolen |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111985331A true CN111985331A (en) | 2020-11-24 |
| CN111985331B CN111985331B (en) | 2024-05-10 |
Family
ID=73438669
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010695613.7A Active CN111985331B (en) | 2020-07-20 | 2020-07-20 | Detection method and device for preventing trade secret from being stolen |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111985331B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114067441A (en) * | 2022-01-14 | 2022-02-18 | 合肥高维数据技术有限公司 | Shooting and recording behavior detection method and system |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002369125A (en) * | 2001-06-06 | 2002-12-20 | Canon Inc | Photographing apparatus and picture erasing method thereof |
| CN107066983A (en) * | 2017-04-20 | 2017-08-18 | 腾讯科技（上海）有限公司 | A kind of auth method and device |
| CN107294622A (en) * | 2016-04-01 | 2017-10-24 | 北京中密安信息安全技术有限公司 | Wireless detection method of stealing secret information |
| CN108140123A (en) * | 2017-12-29 | 2018-06-08 | 深圳前海达闼云端智能科技有限公司 | Face living body detection method, electronic device and computer program product |
| CN108182396A (en) * | 2017-12-25 | 2018-06-19 | 中国电子科技集团公司电子科学研究院 | A kind of automatic identification is taken pictures the method and device of behavior |
| US20180186452A1 (en) * | 2017-01-04 | 2018-07-05 | Beijing Deephi Technology Co., Ltd. | Unmanned Aerial Vehicle Interactive Apparatus and Method Based on Deep Learning Posture Estimation |
| US20190138791A1 (en) * | 2016-08-10 | 2019-05-09 | Tencent Technology (Shenzhen) Company Limited | Key point positioning method, terminal, and computer storage medium |
| CN110060441A (en) * | 2019-06-14 | 2019-07-26 | 三星电子（中国）研发中心 | Method and apparatus for terminal anti-theft |
| CN110070015A (en) * | 2019-04-12 | 2019-07-30 | 重庆科芮智能科技有限公司 | Image-recognizing method and device, system |
| CN110096972A (en) * | 2019-04-12 | 2019-08-06 | 重庆科芮智能科技有限公司 | Data guard method, apparatus and system |
| CN110134565A (en) * | 2019-04-12 | 2019-08-16 | 重庆科芮智能科技有限公司 | Environment light detection method and device |
| CN110163100A (en) * | 2019-04-17 | 2019-08-23 | 中国电子科技网络信息安全有限公司 | The anti-display of taking pictures of one kind |
| CN111062239A (en) * | 2019-10-15 | 2020-04-24 | 平安科技（深圳）有限公司 | Human body target detection method and device, computer equipment and storage medium |
| CN111200781A (en) * | 2018-11-19 | 2020-05-26 | 林桦 | Anti-photographing method and system based on computer vision and radio direction finding positioning |
| CN111259461A (en) * | 2020-01-20 | 2020-06-09 | 北京科芮智盈技术有限公司 | Anti-stealing-lighting method, terminal equipment and system |
-
2020
- 2020-07-20 CN CN202010695613.7A patent/CN111985331B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002369125A (en) * | 2001-06-06 | 2002-12-20 | Canon Inc | Photographing apparatus and picture erasing method thereof |
| CN107294622A (en) * | 2016-04-01 | 2017-10-24 | 北京中密安信息安全技术有限公司 | Wireless detection method of stealing secret information |
| US20190138791A1 (en) * | 2016-08-10 | 2019-05-09 | Tencent Technology (Shenzhen) Company Limited | Key point positioning method, terminal, and computer storage medium |
| US20180186452A1 (en) * | 2017-01-04 | 2018-07-05 | Beijing Deephi Technology Co., Ltd. | Unmanned Aerial Vehicle Interactive Apparatus and Method Based on Deep Learning Posture Estimation |
| CN107066983A (en) * | 2017-04-20 | 2017-08-18 | 腾讯科技（上海）有限公司 | A kind of auth method and device |
| CN108182396A (en) * | 2017-12-25 | 2018-06-19 | 中国电子科技集团公司电子科学研究院 | A kind of automatic identification is taken pictures the method and device of behavior |
| CN108140123A (en) * | 2017-12-29 | 2018-06-08 | 深圳前海达闼云端智能科技有限公司 | Face living body detection method, electronic device and computer program product |
| CN111200781A (en) * | 2018-11-19 | 2020-05-26 | 林桦 | Anti-photographing method and system based on computer vision and radio direction finding positioning |
| CN110070015A (en) * | 2019-04-12 | 2019-07-30 | 重庆科芮智能科技有限公司 | Image-recognizing method and device, system |
| CN110096972A (en) * | 2019-04-12 | 2019-08-06 | 重庆科芮智能科技有限公司 | Data guard method, apparatus and system |
| CN110134565A (en) * | 2019-04-12 | 2019-08-16 | 重庆科芮智能科技有限公司 | Environment light detection method and device |
| CN110163100A (en) * | 2019-04-17 | 2019-08-23 | 中国电子科技网络信息安全有限公司 | The anti-display of taking pictures of one kind |
| CN110060441A (en) * | 2019-06-14 | 2019-07-26 | 三星电子（中国）研发中心 | Method and apparatus for terminal anti-theft |
| CN111062239A (en) * | 2019-10-15 | 2020-04-24 | 平安科技（深圳）有限公司 | Human body target detection method and device, computer equipment and storage medium |
| CN111259461A (en) * | 2020-01-20 | 2020-06-09 | 北京科芮智盈技术有限公司 | Anti-stealing-lighting method, terminal equipment and system |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114067441A (en) * | 2022-01-14 | 2022-02-18 | 合肥高维数据技术有限公司 | Shooting and recording behavior detection method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111985331B (en) | 2024-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110188724B (en) | Method and system for helmet positioning and color recognition based on deep learning | |
| TWI686774B (en) | Human face live detection method and device | |
| WO2020215552A1 (en) | Multi-target tracking method, apparatus, computer device, and storage medium | |
| Tian et al. | Robust detection of abandoned and removed objects in complex surveillance videos | |
| Çetin et al. | Video fire detection–review | |
| Cucchiara et al. | Probabilistic posture classification for human-behavior analysis | |
| CN103824070B (en) | A kind of rapid pedestrian detection method based on computer vision | |
| CN108229335A (en) | It is associated with face identification method and device, electronic equipment, storage medium, program | |
| Lin et al. | Smoke detection in video sequences based on dynamic texture using volume local binary patterns | |
| US20130170711A1 (en) | Edge detection image capture and recognition system | |
| CN105868689A (en) | Cascaded convolutional neural network based human face occlusion detection method | |
| WO2020001216A1 (en) | Abnormal event detection | |
| KR101781358B1 (en) | Personal Identification System And Method By Face Recognition In Digital Image | |
| CN112287868B (en) | Human body action recognition method and device | |
| WO2019114145A1 (en) | Head count detection method and device in surveillance video | |
| Chang et al. | Localized detection of abandoned luggage | |
| CN112183472A (en) | Method for detecting whether test field personnel wear work clothes or not based on improved RetinaNet | |
| Zaidi et al. | Video anomaly detection and classification for human activity recognition | |
| CN112287823A (en) | Facial mask identification method based on video monitoring | |
| CN114140745A (en) | Method, system, device and medium for detecting personnel attributes of construction site | |
| CN111985331B (en) | Detection method and device for preventing trade secret from being stolen | |
| KR20200059643A (en) | ATM security system based on image analyses and the method thereof | |
| Mukto et al. | Design of a real-time crime monitoring system using deep learning techniques | |
| Yoon et al. | Tracking System for mobile user Based on CCTV | |
| Zhang et al. | A novel efficient method for abnormal face detection in ATM |
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 |