WO2020019673A1 - 基于图像分析的工地监控方法、装置及可读存储介质 - Google Patents

基于图像分析的工地监控方法、装置及可读存储介质 Download PDF

Info

Publication number
WO2020019673A1
WO2020019673A1 PCT/CN2018/124839 CN2018124839W WO2020019673A1 WO 2020019673 A1 WO2020019673 A1 WO 2020019673A1 CN 2018124839 W CN2018124839 W CN 2018124839W WO 2020019673 A1 WO2020019673 A1 WO 2020019673A1
Authority
WO
WIPO (PCT)
Prior art keywords
collected object
facial feature
monitoring area
collected
information
Prior art date
Application number
PCT/CN2018/124839
Other languages
English (en)
French (fr)
Inventor
温水
陈尚杰
孙道姣
万历
刘鑫
张猛
Original Assignee
深圳云天励飞技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳云天励飞技术有限公司 filed Critical 深圳云天励飞技术有限公司
Publication of WO2020019673A1 publication Critical patent/WO2020019673A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/172—Classification, e.g. identification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/26—Government or public services
    • G06Q50/265—Personal security, identity or safety
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
    • G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/168—Feature extraction; Face representation

Definitions

  • the present invention relates to the field of construction site monitoring and management, and in particular, to a construction site monitoring method, device, and readable storage medium based on image analysis.
  • biometric technology has been widely used in monitoring systems for various construction sites such as mining, subway construction, and building construction.
  • face recognition technology or fingerprint recognition technology is used for site safety management and control.
  • image analysis technology used by hats or work clothes for identification to escort construction site safety, it is undeniable that these have some effects.
  • it is far from enough to meet the requirements of engineering companies for the safety management of one person and one hat.
  • the existing technology still has some problems in the technology of combining biometrics and image analysis to achieve safety management and control.
  • the main purpose of the present invention is to provide a construction site monitoring method, device, and readable storage medium based on image analysis, so as to solve the problems that the construction site cannot implement one person, one cap, and multiple accidents because it cannot achieve one person, one cap, and control.
  • a first aspect of the present invention provides a method for monitoring a construction site based on image analysis, including:
  • a second aspect of the present invention also provides a site monitoring device based on image analysis, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor.
  • the processor implements the computer program when the computer program is executed. Steps of a site monitoring method based on image analysis.
  • the third aspect of the present invention also provides a computer-readable storage medium.
  • the computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the above-mentioned image analysis-based site monitoring method.
  • the facial feature information of the collected object By collecting the image information of the captured object when the captured object enters within the preset image acquisition range of the monitoring area, and acquiring the facial feature information of the captured object according to the image information, and then detecting the captured image based on the collected image information Whether the subject is wearing a safety helmet, and when it is detected that the collected object is wearing a safety helmet, obtain a code of the safety helmet, and according to the code of the safety helmet, obtain an authorized user in the monitoring area that has a one-to-one correspondence with the code. Finally, the facial feature information of the collected object is matched with the facial feature model of the authorized user in the monitoring area to obtain a matching result.
  • FIG. 1 is a schematic diagram of a network system architecture according to an embodiment of the present invention
  • FIG. 2 is a schematic flowchart of a method for monitoring a construction site based on image analysis according to an embodiment of the present invention
  • FIG. 3 is a schematic flowchart of interaction between a remote control center, an infrared sensor, and an image acquisition terminal according to an embodiment of the present invention
  • FIG. 4 is a schematic flowchart of another method for monitoring a construction site based on image analysis according to an embodiment of the present invention
  • FIG. 5 is a schematic flowchart of an application site monitoring method based on image analysis in an application scenario according to an embodiment of the present invention
  • FIG. 6 is a schematic structural diagram of a construction site monitoring device based on image analysis according to an embodiment of the present invention.
  • the network system architecture shown in FIG. 1 relates to an image acquisition terminal, a remote control center, and a display terminal.
  • the remote control center is the brain of the entire system, which can be understood as the server that implements the method of this solution, and controls the operation of other components of the system;
  • the image acquisition terminal is used to collect image information in the acquisition range, which can be a camera or a snapper , Camera, etc .;
  • the display terminal is used to display the image information collected by the image acquisition terminal.
  • the various parts communicate through the transmission component or the network to enable the entire image analysis-based monitoring method. Implementation can be supported by a system architecture.
  • FIG. 2 is a schematic flowchart of a site monitoring method based on image analysis provided by the present invention.
  • the method shown in FIG. 2 may be specifically implemented based on the network system architecture shown in the example in FIG. 1.
  • an aspect of the present invention provides a method for monitoring a construction site based on image analysis.
  • the method includes:
  • the monitoring area may be the entire construction site
  • the preset image acquisition range may be a range that can be covered by the image acquisition terminal, such as within a few meters, dozens of meters, dozens of meters centered on the entrance to and exit from the construction site. Within meters, it can even be a wider range.
  • the preset image acquisition range specifically, it can be detected by the way of infrared sensor induction
  • Distributed image acquisition terminal at the camera channel, collects image information including the face or face of the object being collected, including images, videos, etc., and automatically performs face recognition on the collected images or videos Tracking to extract the geometric composition relationship between the facial feature information points of the collected object, which includes the feature information points such as eyes, nose, mouth, forehead. According to the composition relationship between the feature information points, the facial feature information of the collected object that has entered the preset image collection range is determined. It should be noted that after obtaining the facial feature information of the collected object, it is not anxious to compare it with the The stored facial feature models of authorized users who can enter the monitoring area are matched.
  • the facial feature information of the collected object is obtained by using the face recognition technology in the biometric recognition technology, firstly, it is determined whether the collected object is wearing a safety helmet.
  • An image matching algorithm is used to detect the collected image information.
  • gray-level matching algorithms mainly include gray-level template matching algorithms.
  • Blocking matching is to find a sub-image similar to the template image based on a known template image and use another two-dimensional sliding template for matching. Commonly used is the MAD (Mean Absolute Difference) algorithm.
  • SAD Sud of Absolute Difference
  • NCC Normalized cross correlation
  • Feature-based matching algorithms first extract the features of the image, and then generate feature descriptors. Finally, the features of the two images are matched according to the similarity of the descriptors.
  • SIFT Scale-Invariant Feature Transformation
  • the relation-based matching algorithm is an application of artificial intelligence in image processing. Its breakthrough progress is relatively slow and its application is not widespread. As for which type of matching algorithm the solution of the present invention specifically uses, it is not limited by the embodiments of the present invention.
  • step S12 when the result of the image matching algorithm matches the threshold requirement, it is detected that the captured object in the captured image or video is wearing a helmet, and the remote control center starts a coding recognition device to obtain the helmet. Encoding.
  • the hard hat code in the embodiment of the present invention may be a simple digital code, a one-dimensional barcode, or a two-dimensional code that is widely used now.
  • the hard hat code is the unique identification of all hard hats in this embodiment.
  • the remote control center after performing the above step S13, will use the code as the primary key to read or call a local database.
  • the local database specifically refers to the helmet code and the face of the authorized user in the monitoring area.
  • Feature model relationship mapping library The database stores a one-to-one correspondence between helmets and facial feature models of authorized users in the monitoring area.
  • Each helmet code is bound to a fixed facial area model of authorized users in the monitoring area.
  • obtaining a hard hat code can obtain the facial feature model of the authorized user in the monitoring area bound to it, and obtaining a facial feature model of the authorized user in the monitoring area can also obtain the security bound to it Based on this, step S14 is implemented.
  • the above matching result is determined by the similarity value between the calculated facial feature information of the collected object and the facial feature model of the authorized user in the monitoring area, which is obtained by the remote monitoring center according to the hard hat code.
  • the image matching unit is enabled to scan and match the two, and calculate the similarity value of the collected object matching the facial feature model of the authorized user. Get matching results.
  • whether the matching result is successfully matched is determined mainly based on the relationship between the similarity value of the facial feature information of the collected object and the facial feature model of the authorized user in the monitoring area and the threshold value.
  • the matching Indicates that the collected object has passed the one-person-one-hat condition, and the remote control center sends an instruction to the access control host or the gate to allow the collected object to enter the monitoring area.
  • the embodiment of the present invention combines the biometric identification technology and the image analysis technology, and collects the image information of the captured object when the captured object enters within the preset image collection range of the monitoring area, and Obtain facial feature information of the collected object according to the image information, and then detect whether the collected object is wearing a hard hat according to the collected image information.
  • the code of the hard hat obtains a facial feature model of the authorized user in the monitoring area that has a one-to-one correspondence with the code, and finally matches the facial feature information of the collected object with the facial feature model of the authorized user in the monitored area.
  • the matching result is obtained.
  • the matching result is a successful match, it is determined that the collected object is allowed to enter the monitoring area, so that unauthorized persons often appear on the construction site, and one-man, one-cap control in the monitoring area of the construction site is prevented to avoid accidents.
  • the step of the remote control center obtaining the facial feature information of the collected object through interaction with the infrared sensor and the image acquisition terminal specifically includes:
  • the infrared sensor sends a triggered prompt.
  • the triggered prompt includes information that the captured object enters a preset image collection range of the monitoring area.
  • the remote control center receives the triggered prompt and sends an image information acquisition instruction to the image acquisition terminal.
  • the image acquisition terminal in the embodiment of the present invention includes a terminal that can acquire image information, such as a camera, a camera, a snapper, a camera, and a scanner.
  • the image acquisition terminal receives an image information acquisition instruction and starts acquiring image information of the object to be acquired.
  • the image information of the collected object includes an image or a video, and the face of the collected object must be displayed in the image or video.
  • the image acquisition terminal synchronously sends the acquired image information to the remote control center.
  • the remote control center tracks and extracts facial feature information of the collected object.
  • the remote control center before acquiring facial feature information of an object to be collected through a gate and a gate of a construction site, it is determined whether an object has entered a preset image information collection range through an infrared sensor.
  • the remote control center does not control the image acquisition terminal for image information collection until it receives a prompt that the infrared sensor is triggered.
  • the image acquisition terminal collects image information
  • the remote control center automatically tracks and extracts the facial feature information of each collected object, which can make the entire system operate more efficiently.
  • the image information is collected only when the infrared sensor is triggered, which is conducive to reducing system maintenance costs.
  • an aspect of the present invention further provides another method for monitoring a construction site based on image analysis, which specifically includes the following steps:
  • step S11 has been described in the foregoing embodiment, and is also applicable to this embodiment, and it will not be described here too much.
  • an image acquisition range can be set up between A and B. This range can cover all the open space between construction A and B, or it can be within a certain range of the entrance of construction B. Since the image acquisition terminal of the present invention adopts a distributed design, the image acquisition terminal can be selected according to the needs. Coverage of image information acquisition.
  • the infrared sensor is triggered and a trigger prompt is sent.
  • the remote control center starts the image acquisition terminal to collect the image information of Li Si.
  • the remote control center receives the image acquisition terminal. After sending the image information, the collected image information is tracked, and the facial feature information of Li Si is extracted, and a voice prompt "Successful extraction of facial feature information of the object" can be performed.
  • step S12 has been described in the foregoing embodiment, and is also applicable to this embodiment, and it will not be described too much here.
  • step S12 may include the following implementations:
  • the step of determining whether the collected object is wearing a helmet according to the feature point matching ratio value specifically includes:
  • a first step when the feature point matching ratio value approaches a threshold, it is determined that the collection object is wearing a helmet;
  • a second step when the value of the feature point matching ratio is discrete from a threshold, it is determined that the collection object is not wearing a helmet.
  • a preset comparison image is embedded in the image display window.
  • the comparison image may be a schematic diagram of a staff wearing a standard helmet, or it may simply be a schematic diagram of the uniform use of the entire construction site.
  • the SIFT algorithm can be used to eliminate key points that have no matching relationship due to image occlusion and background chaos.
  • SIFT feature detection is performed to locate key points. Feature points are filtered and accurately located in feature matching.
  • SIFT feature vectors are generated. Take one SIFT keypoint in one image, and find the first two keypoints that are closest to the Euclidean distance in another image.
  • a feature point matching ratio value By determining whether the feature point matching ratio value is less than a preset threshold T, whether to accept the feature point is considered.
  • the above-mentioned feature point matching ratio threshold value recommended by the SIFT algorithm founder is 0.8.
  • the preset The threshold is selected based on countless matching results, and it varies according to the level of matching requirements. At the same time, visualizing the comparison process is conducive to enhancing the user experience.
  • the feature points are accepted and it is determined that Li Si is wearing a safety helmet.
  • the feature point matching ratio value is discrete from the threshold value T, that is, the feature point matching ratio value is greatly different from the threshold value T, and if the feature point is not accepted, it is determined that Li Si is not wearing a helmet, and at this time, a corresponding response can be generated. Broadcast information for voice broadcast "Do not enter without a helmet” or "You do not have a helmet”.
  • step S13 has been described in the above embodiment and is also applicable to this embodiment.
  • the difference is that the acquisition object is the code of the helmet worn on the four heads of Li in this example. Excessive description.
  • step S14 has been described in the foregoing embodiment, and it is also applicable to this embodiment, and it will not be described here too much.
  • the step of obtaining the facial feature model of the authorized user of the monitoring area that has a one-to-one correspondence with the code according to the code of the hard hat specifically includes:
  • the model table records all hard hat codes corresponding to the monitoring area, and facial feature models of authorized users in the monitoring area corresponding to each hard hat code.
  • the remote control center can send a calling instruction to the data calling module, and the data calling module performs traversal and calling to complete).
  • step S15 has been described in the foregoing embodiment, and is also applicable to this embodiment, and it will not be described here too much.
  • step S15 includes:
  • the remote control center compares the facial feature information of Li Si obtained in step S11 with the face of the helmet wearing on Li Si's head obtained in step S14 and has a one-to-one correspondence or binding relationship with the face of the authorized user in the monitoring area. Match the facial feature model and calculate the similarity value between the two. When the matching similarity value between Li Si's facial feature information and the authorized user's facial feature model is greater than or equal to a preset similarity threshold, determine the matching result For matching success, similarly, when the similarity value is less than a preset similarity threshold, it is determined that the matching result is a matching failure.
  • step S16 has been described in the foregoing embodiment, and is also applicable to this embodiment, and it will not be described too much here.
  • the image information of the captured object is collected, and the facial feature information of the collected object is obtained according to the image information, and then according to the collected image
  • obtain the code of the hard hat According to the code of the hard hat, obtain One-to-one correspondence between facial feature models of authorized users in the monitoring area, and finally matching facial feature information of the collected object with facial feature models of authorized users in the monitoring area to obtain a matching result.
  • the facial features of the collected object When the facial features of the collected object When the matching result between the information and the facial feature model of the authorized user in the monitoring area is successful, it is determined that the collected object is allowed to enter the monitoring area, which can realize the safety management of one person and one hat at the construction site, and can also effectively reduce the hardware equipment of the monitoring system. loss.
  • the method further includes:
  • the collected object When the collected object is an authorized user in the monitoring area, obtain basic information of the collected object, add the basic information of the collected object to a preset blacklist library, and determine that the captured object is not allowed
  • the collection object enters the monitoring area; wherein the basic information of the collection object includes the name, position, department, job number and contact information of the collection object;
  • Li Si When Li Si is determined to be an unauthorized user of the construction project or the construction team entering the monitoring area, an alarm message is sent directly to the administrator, and this alarm information can be sent to the administrator's mobile terminal in a voice manner, such as a mobile phone, a walkie-talkie It can also be a short message reminder. It can also be directly broadcasted on the construction site by broadcasting, such as: "Be alert! A stranger breaks in.”
  • Li Si is an authorized user in the monitoring area, it is determined that Li Si is not wearing a safety helmet.
  • the basic information of Li Si is retrieved from the basic information database of authorized users in the preset monitoring area, and displayed in the display window, such as "Name: Li Si; Position: Welder; The said department: Weld a group; work ID: xxyyzz ... ", add the above basic information to the blacklist database, and do not allow Li Si to enter construction B for construction.
  • the method before the step of adding basic information of the collected object to a preset blacklist library, the method further includes:
  • the above method further includes: determining that the collected object is allowed to enter the object when a frequency of occurrence of the work number of the collected object in the preset blacklist database is less than a preset upper limit value. Monitoring area.
  • the monitoring area is conducive to enhancing the awareness of authorized users wearing helmets in the monitoring area.
  • a certain captured object wants to pass through the access control or gate passage of the construction site, or for an unknown identity
  • the monitoring and analysis of the collected objects' activities in the construction site is taken as an example to illustrate the process of the above-mentioned image analysis-based construction site monitoring method. Specifically, referring to FIG. 5, the process in this example is as follows:
  • a distributed image acquisition terminal acquires an image and acquires facial feature information of a collected object
  • step S502. Determine whether to wear a helmet. If yes, go to step S503. If no, go to step S504.
  • step S504. Determine whether it is an authorized user. If so, perform step S510; if not, perform step S509;
  • step S507 Determine whether the matching result is successful. If yes, go to step S508; if no, go to step S504;
  • step S511 Determine whether the frequency of the blacklist library exceeds the upper limit. If yes, go to step S512; if not, go to step S513;
  • the construction site monitoring method based on image analysis in the embodiment of the present invention is more effective and stricter than that of the prior art, and the labor cost of the construction company is relatively reduced. It can solve the problem that unauthorized persons often enter the monitoring area on the construction site, and realize one-man one-hat safety control.
  • the construction site monitoring device 6 based on image analysis includes a memory 61, a processor 62, and a memory 61 stored in the memory 61.
  • a computer program 63 running on a processor 62 which, when executing the computer program 63, implements the steps of the method for monitoring a construction site based on image analysis described above.
  • the processor 62 executes the computer program 63, the following steps are implemented: when a captured object enters within a preset image acquisition range of the monitoring area, receiving image information of the captured object sent by the image acquisition terminal, and Acquiring facial feature information of the collected object according to the image information; detecting whether the collected object is wearing a hard hat according to the image information; and acquiring the detected object when detecting that the collected object is wearing a hard hat Obtaining the facial feature model of the authorized user of the monitoring area that has a one-to-one correspondence with the code according to the encoding of the hard hat; and combining the facial feature information of the collected object with the The facial feature model of the authorized user in the monitoring area is matched to obtain a matching result; when the matching result of the facial feature information of the collected object and the facial feature model of the authorized user in the monitored area is a successful match, it is determined to allow the The collected object enters the monitoring area.
  • the processor 62 executes the computer program 63, it further implements the following steps: when it is detected that the collected object is not wearing a helmet or the facial characteristic information of the collected object and the facial characteristics of the user authorized in the monitoring area When the matching result of the model is that the matching fails, it is determined whether the collected object is an authorized user of the monitoring area according to the facial feature information of the collected object; when the collected object is an authorized user of the monitoring area When acquiring basic information of the collected object, adding the basic information of the collected object to a preset blacklist library, and determining that the collected object is not allowed to enter the monitoring area; wherein the collected object
  • the basic information of the object includes the name, position, department, job number and contact information of the collected object;
  • the processor 62 executes the computer program 63, it further implements the following steps: using an image matching algorithm to perform feature point matching with the preset comparison image to obtain a feature point matching ratio value; wherein, said The comparison image is a standard wearing schematic diagram of the helmet; and it is determined whether the collected object is wearing a helmet according to the feature point matching ratio value.
  • the processor 62 executes the computer program 63, the following steps are further implemented: when the feature point matching ratio value approaches a threshold value, it is determined that the collection object is wearing a hard hat; when the feature point matching ratio value is discrete At the threshold, it is determined that the collection object is not wearing a helmet.
  • the processor 62 executes the computer program 63, the following steps are further implemented: using the hard hat code as a main key, obtaining the monitoring having a one-to-one correspondence with the code from a preset facial feature model table; A facial feature model of an area authorized user; wherein, the facial feature model table records all hard hat codes corresponding to the monitoring area, and a facial feature model of the authorized user of the monitoring area corresponding to each hard hat code.
  • the processor 62 executes the computer program 63, it further implements the following steps: calculating a similarity value of the facial feature information of the collected object and the facial feature model of the authorized user; when the similarity value is greater than or When it is equal to a preset similarity threshold, the matching result is determined to be a successful match; when the similarity value is less than a preset similarity threshold, the matching result is determined to be a matching failure.
  • the processor 62 executes the computer program 63, it also implements the following steps: traversing a preset blacklist library according to the work number of the collected object, and counting the work numbers of the collected object in the preset blacklist library The frequency of occurrence; when the frequency of occurrence of the number of the collected object in the preset blacklist database is equal to or greater than a preset upper limit value, executing adding the basic information of the collected object to the preset Steps to blacklist the library.
  • the processor 62 executes the computer program 63, it further implements the following step: when the frequency of occurrence of the work number of the collected object in the preset blacklist library is less than a preset upper limit value, it is determined to allow the The collection object enters the monitoring area.
  • the steps of implementing the above-mentioned method for monitoring construction site based on image analysis can solve the problem of unauthorized construction on the construction site.
  • one person and one cap can be controlled in the monitoring area of the construction site to avoid accidents.
  • the above-mentioned image analysis-based construction site monitoring device 6 may be a service device such as a computer, a notebook, a remote server, a cloud server, and a remote control center.
  • the construction site monitoring device 6 based on image analysis may include, but is not limited to, a processor 62 and a memory 61.
  • a processor 62 and a memory 61 may be included in the schematic diagram.
  • the schematic diagram is only an example of the construction site monitoring device 6 based on image analysis, and does not constitute a limitation on the construction site monitoring device 6 based on image analysis, and may include more or fewer components than shown in the figure. , Or combine some parts, or different parts.
  • the processor 62 of the construction site monitoring apparatus 6 based on image analysis implements the steps of the above-mentioned construction site monitoring method based on image analysis when the computer program 63 is executed, all embodiments of the foregoing construction site monitoring method based on image analysis are It is applicable to the construction site monitoring device 6 based on image analysis, and can all achieve the same or similar beneficial effects.
  • An aspect of the embodiments of the present invention further provides a computer-readable storage medium.
  • the computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the above-mentioned image analysis-based site monitoring method.
  • the following steps are implemented: when a captured object enters within a preset image collection range of the monitoring area, receiving image information of the captured object sent by the image collection terminal, and according to The image information acquires facial feature information of the collected object; detects whether the collected object is wearing a helmet according to the image information; and acquires the detected object when it is detected that the collected object is wearing a helmet A code of a hard hat; obtaining a facial feature model of the authorized user of the monitoring area having a one-to-one correspondence with the code according to the code of the hard hat; combining the facial feature information of the collected object with the monitoring The facial feature model of the authorized user in the area is matched to obtain a matching result. When the matching result of the facial feature information of the collected object and the facial feature model of the authorized user in the monitored area is a successful match, it is determined that the passive feature is allowed.
  • the collection object enters the monitoring area.
  • the following steps are further implemented: when it is detected that the collected object is not wearing a helmet or the facial characteristic information of the collected object and the facial characteristic model of the user authorized in the monitoring area
  • determine whether the collected object is an authorized user of the monitoring area according to the facial feature information of the collected object when the collected object is an authorized user of the monitoring area Obtaining the basic information of the collected object, adding the basic information of the collected object to a preset blacklist library, and determining that the collected object is not allowed to enter the monitoring area; wherein the collected object
  • the basic information includes the name, position, department, work number and contact information of the collected object; when the collected object is not an authorized user in the monitoring area, an alarm message is sent to the administrator.
  • the following steps are further implemented: using an image matching algorithm to match the acquired image information with a preset comparison image to obtain feature points to obtain a feature point matching ratio value; wherein the ratio
  • the image is a schematic diagram of the standard wearing of the helmet; it is determined whether the collected object is wearing a helmet according to the feature point matching ratio value.
  • the following steps are further implemented: when the feature point matching ratio value approaches a threshold value, it is determined that the collection object is wearing a hard hat; when the feature point matching ratio value is discrete from At the threshold, it is determined that the collection object is not wearing a helmet.
  • the hard hat code is used as a key
  • the monitoring area having a one-to-one correspondence with the code is obtained from a preset facial feature model table.
  • a facial feature model of an authorized user wherein the facial feature model table records all hard hat codes corresponding to the monitoring area, and the facial feature models of the authorized user in the monitoring area corresponding to each hard hat code.
  • the following steps are further implemented: calculating a similarity value of the facial feature information of the collected object and the facial feature model of the authorized user; when the similarity value is greater than or equal to When the similarity threshold is preset, the matching result is determined to be a successful match, and when the similarity value is less than the preset similarity threshold, the matching result is determined to be a matching failure.
  • the following steps are further implemented: traversing a preset blacklist library according to the work number of the collected object, and counting the work numbers of the collected object appearing in the preset blacklist database When the number of occurrences of the ID of the collected object in the preset blacklist database is equal to or greater than a preset upper limit value, executing the adding basic information of the collected object to the preset Steps to blacklist the library.
  • the following steps are further implemented: when the frequency of occurrence of the work number of the collected object in the preset blacklist library is less than a preset upper limit value, determining that the collected data is allowed The subject enters the monitoring area.
  • the computer program of the computer-readable storage medium includes computer program code
  • the computer program code may be in a source code form, an object code form, an executable file, or some intermediate form.
  • the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM, Read-Only Memory) , Random Access Memory (RAM, Random Access Memory), electric carrier signals, telecommunication signals, and software distribution media.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Business, Economics & Management (AREA)
  • General Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Human Computer Interaction (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Artificial Intelligence (AREA)
  • Primary Health Care (AREA)
  • Marketing (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Educational Administration (AREA)
  • Development Economics (AREA)
  • Computer Security & Cryptography (AREA)
  • Strategic Management (AREA)
  • Computing Systems (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Image Analysis (AREA)
  • Alarm Systems (AREA)

Abstract

一种基于图像分析的工地监控方法、装置及可读存储介质，所述方法包括：当监控区域的预设图像采集范围内有被采集对象进入时，采集被采集对象的图像信息，并根据图像信息获取被采集对象的面部特征信息（S11），根据图像信息，检测被采集对象是否佩戴有安全帽（S12），当检测到被采集对象佩戴有安全帽时，获取安全帽的编码（S13），根据安全帽的编码，获取与编码具有一一对应关系的监控区域授权用户的脸部特征模型（S14），将被采集对象的面部特征信息与监控区域授权用户的脸部特征模型进行匹配，得到匹配结果（S15），当二者的匹配结果为匹配成功时，确定允许被采集对象进入所述监控区域（S16），能够实现工地监控区域的一人一帽管控，避免发生意外。

Description

基于图像分析的工地监控方法、装置及可读存储介质
本申请要求于2018年7月25日提交中国专利局，申请号为201810830408.X、发明名称为“基于图像分析的工地监控方法、装置及可读存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本发明涉及工地监控管理领域，尤其涉及一种基于图像分析的工地监控方法、装置及可读存储介质。
背景技术
近年来，随着经济社会的发展，大大小小的施工工地在城乡随处可见，但由于工地各类事故时有发生，施工工地的安全管控也成为工程商为之头疼的问题。
为了提高施工工地的管理效率，预防闲杂人等闯入工地，加强工地安全管控，工程商通常的选择是为工地建立一套安全可靠的视频监控系统。而且生物识别技术早已被广泛应用到采矿、修地铁、房屋建筑等各类施工工地的监控系统中，例如采用人脸识别技术或指纹识别技术进行工地安全管控，当然现有技术中也有采用对安全帽或工作服进行识别的图像分析技术来为工地施工安全保驾护航，不可否认这些都起到了某些效果。但想要最大程度地满足工程商对一人一帽安全管控的要求，这些是远远不够的，现有技术在采用生物识别与图像分析相结合以实现安全管控的技术上仍然存在一些问题。
发明内容
本发明的主要目的在于提供一种基于图像分析的工地监控方法、装置及可读存储介质，以解决施工工地不能实现一人一帽管控，以及因不能实现一人一帽管控而多发意外状况的问题。
为实现上述目的，本发明第一方面提供了一种基于图像分析的工地监控方法，包括：
当监控区域的预设图像采集范围内有被采集对象进入时，采集所述被采集对象的图像信息，并根据所述图像信息获取所述被采集对象的面部特征信息；
根据所述图像信息，检测所述被采集对象是否佩戴有安全帽；
当检测到所述被采集对象佩戴有安全帽时，获取所述安全帽的编码；
根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户 的脸部特征模型；
将所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型进行匹配，得到匹配结果；
当所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配成功时，确定允许所述被采集对象进入所述监控区域。
本发明的第二方面还提供了一种基于图像分析的工地监控装置，包括存储器、处理器以及存储在存储器中并可在处理器上运行的计算机程序，该处理器执行计算机程序时实现上述的基于图像分析的工地监控方法的步骤。
本发明的第三方面还提供了一种计算机可读存储介质，计算机可读存储介质存储有计算机程序，该计算机程序被处理器执行时实现上述的基于图像分析的工地监控方法的步骤。
本发明的上述方案至少包括以下有益效果：
通过当监控区域的预设图像采集范围内有被采集对象进入时，采集被采集对象的图像信息，并根据图像信息获取被采集对象的面部特征信息，然后根据所采集的图像信息，检测被采集对象是否佩戴有安全帽，当检测到被采集对象佩戴有安全帽时，获取安全帽的编码，根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型，最后将被采集对象的面部特征信息与监控区域授权用户的脸部特征模型进行匹配，得到匹配结果，当被采集对象的面部特征信息与监控区域授权用户的脸部特征模型的匹配结果为匹配成功时，确定允许被采集对象进入所述监控区域，从而实现工地监控区域的一人一帽管控，避免发生意外。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本发明实施例提供的一种网络系统架构示意图；
图2为本发明实施例提供的一种基于图像分析的工地监控方法的流程示意图；
图3为本发明实施例提供的一种远程控制中心与红外感应器和图像采集终端的交互流程示意图；
图4为本发明实施例提供的另一种基于图像分析的工地监控方法的流程示意图；
图5为本发明实施例提供的一种基于图像分析的工地监控方法在应用场景中流程示意图；
图6为本发明实施例提供的一种基于图像分析的工地监控装置的结构示意图；
具体实施方式
为了使本技术领域的人员更好地理解本发明方案，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分的实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都应当属于本发明保护的范围。
本发明说明书、权利要求书和附图中出现的术语“包括”和“具有”以及它们任何变形，意图在于覆盖不排他的包含。例如包含了一系列步骤或单元的过程、方法、系统、产品或设备没有限定于已列出的步骤或单元，而是可选地还包括没有列出的步骤或单元，或可选地还包括对于这些过程、方法、产品或设备固有的其它步骤或单元。此外，术语“第一”、“第二”和“第三”等是用于区别不同的对象，而并非用于描述特定的顺序。
下面首先结合相关附图和举例来对本方案的实施例进行说明，但举例并不对本发明的实施例造成限定。
首先结合相关附图来举例介绍下本申请实施例的方案可能应用到的网络系统架构。
具体参见图1，图1所示的网络系统架构涉及到图像采集终端、远程控制中心、显示终端。其中，远程控制中心是整个系统的大脑，可以理解为执行行本方案方法的服务器，控制着系统其它构件的运作；图像采集终端用于在采集范围内采集图像信息，具体可以为摄像头、抓拍机、相机等；显示终端用于显示图像采集终端采集到的图像信息，当然也可以为远程控制中心的一些操作提供可视化服务，各部分通过传输组件或网络进行互通，使整个基于图像分析的监控方法的实施得以有系统架构的支撑。
实施例一
参见图2，图2为本发明提供的一种基于图像分析的工地监控方法的流程示意图，图2所示的方法可以基于图1举例所示的网络系统架构来具体实施。
如图2所示，本发明一方面提供一种基于图像分析的工地监控方法，所述方法包括：
S11，当监控区域的预设图像采集范围内有被采集对象进入时，采集所述被采集对象的图像信息，并根据所述图像信息获取所述被采集对象的面部特征信息。
其中，在本实施例中，监控区域可以是整个工地，预设图像采集范围可以是图像采集 终端能够覆盖到的范围，例如以进出施工工地通道口为中心的几米内、十几米内、几十米内，甚至可以是更宽的范围。此处以被采集对象通过施工工地的大门或通过闸机通道为例，当检测到预设图像采集范围内有被采集(具体可通过红外感应器感应的方式检测)时，启动设于大门或闸机通道处的分布式图像采集终端(摄像机、摄像头或抓拍机等)，采集含有被采集对象脸部或面部的图像信息，包括图像、视频等，并自动对采集到的图像或视频进行人脸跟踪，提取采集对象面部特征信息点之间的几何构成关系，这其中就包括眼睛、鼻子、嘴巴、额头等特征信息点。根据特征信息点间的构成关系确定进入预设图像采集范围内的被采集对象的面部特征信息，需要说明的是，在获取到上述被采集对象的面部特征信息之后，并不着急将其与预先存储的能够进入监控区域授权用户的脸部特征模型进行匹配。
S12，根据所述图像信息，检测所述被采集对象是否佩戴有安全帽。
其中，在本实施例中，采用生物识别技术中的人脸识别技术获取到该被采集对象的面部特征信息之后，首先判断该采集对象是否佩戴有安全帽，在本发明的实施例中，主要采用图像匹配算法来对采集到的图像信息进行检测。
需要说明的是，图像匹配算法分为3类，分别是基于灰度的匹配算法，基于特征的匹配算法和基于关系的匹配算法，基于灰度的匹配算法主要有基于灰度的模板匹配算法，模板匹配(Blocking Matching)是根据已知模板图像到另一幅图像中寻找与模板图像相似的子图像，利用空间二维滑动模板进行匹配，常见的有MAD(Mean Absolute Difference，平均绝对差)算法、SAD(Sum of Absolute Difference，绝对误差和)算法、NCC(Normalized cross correlation，归一化交叉相关)算法等。基于特征的匹配算法首先提取图像的特征，再生成特征描述子，最后根据描述子的相似程度对两幅图像的特征之间进行匹配，常见的有SIFT(Scale-Invariant Feature Transform，尺度不变特征变换)算法、SURF(Speeded-Up Robust Features，加速稳健特征)算法、BRISK(Binary Robust Invariant Scalable Keypoints，二进制鲁棒不变尺度特征)算法等。基于关系的匹配算法是人工智能领域在图像处理中的应用，其突破性进展相对缓慢，应用并不广泛。至于本发明的方案具体采用哪一类匹配算法，并不受本发明实施例的限制。
S13，当检测到所述被采集对象佩戴有安全帽时，获取所述安全帽的编码。
其中，在步骤S12的基础上，图像匹配算法匹配的结果达到阈值要求时，检测出所采集到的图像或视频中的被采集对象佩戴有安全帽，远程控制中心启动编码识别装置，获取该安全帽上的编码。
需要说明的是，本发明实施例中的安全帽编码可以是简单的数字编码、一维条形码，也可以是现在应用广泛的二维码。该安全帽编码是本实施例中所有安全帽的唯一身份标识。
S14，根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型。
其中，在本发明的实施例中，在执行完上述步骤S13后，远程控制中心会以该编码为主键读取或者调用本地数据库，该本地数据库具体指安全帽编码与监控区域授权用户的脸部特征模型关系映射库，该数据库中存储有安全帽与监控区域授权用户的脸部特征模型之间的一一对应关系，每一安全帽编码绑定固定的监控区域授权用户的脸部特征模型。显而易见的，获取到一个安全帽编码就能获取到与之绑定的监控区域授权用户的脸部特征模型，获取到一个监控区域授权用户的脸部特征模型同样能获取到与之绑定的安全帽编码，基于此，步骤S14得以实现。
S15，将所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型进行匹配，得到匹配结果。
其中，在本实施例中，上述匹配结果由计算出来的被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型匹配的相似程度值决定，远程监控中心根据安全帽编码获取到与之一一对应的监控区域授权用户的脸部特征模型后，启用图像匹配单元将上述二者进行扫描、匹配，计算被采集对象与所述授权用户的脸部特征模型匹配的相似程度值，得到匹配结果。
S16，当所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配成功时，确定允许所述被采集对象进入所述监控区域。
其中，在本实施例中，主要根据被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型匹配的相似程度值与阈值的关系来判断匹配结果是否匹配成功，当匹配成功时，表示上述被采集对象达到了一人一帽的通过条件，远程控制中心向门禁控制主机或闸机发送指令，允许该被采集对象进入监控区域。
有益效果：与现有技术相比较，本发明实施例结合生物识别技术与图像分析技术，通过当监控区域的预设图像采集范围内有被采集对象进入时，采集被采集对象的图像信息，并根据图像信息获取被采集对象的面部特征信息，然后根据所采集的图像信息，检测被采集对象是否佩戴有安全帽，当检测到被采集对象佩戴有安全帽时，获取安全帽的编码，根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部 特征模型，最后将被采集对象的面部特征信息与监控区域授权用户的脸部特征模型进行匹配，得到匹配结果，当匹配结果为匹配成功时，确定允许被采集对象进入监控区域，从而解决施工工地上时常有非授权人出现，实现工地监控区域的一人一帽管控，避免发生意外。
进一步的，优选方案，如图3所示，在本发明的具体实施例中，远程控制中心可通过与红外感应器和图像采集终端的交互获得被采集对象的面部特征信息的步骤，具体包括：
S21，红外感应器发送被触发提示。
其中，被触发提示包括有被采集对象进入到监控区域的预设图像采集范围内的信息。
S22，远程控制中心接收到被触发提示，并向图像采集终端发送图像信息采集指令。
其中，本发明实施例中的图像采集终端包括摄像机、摄像头、抓拍机、相机以及扫描仪等可以采集图像信息的终端。
S23，图像采集终端接收图像信息采集指令，开始采集所述被采集对象的图像信息。
其中，上述被采集对象的图像信息包括图像或视频，且该图像或视频中须呈现出被采集对象的面部。
S24，图像采集终端将采集到的图像信息同步发送至远程控制中心。
S25，远程控制中心跟踪、提取所述被采集对象的面部特征信息。
具体的，本发明的实施例在对想要通过施工工地的大门及闸机通道的被采集对象进行面部特征信息获取之前，通过红外感应器判断是否有被采集对象进入到预设图像信息采集范围，远程控制中心在接收到红外感应被触发的提示后才进行控制图像采集终端进行图像信息采集的操作，而且，图像采集终端在采集图像信息时，同步上传采集到的图像信息到远程控制中心，远程控制中心自动跟踪、提取每个被采集对象的面部特征信息，能够使整个系统的运转效率更高，在红外感应器被触发的情况下才进行图像信息采集，有利于系统维护成本的降低。
实施例二
参见图4所示，本发明的一方面还提供了另一种基于图像分析的工地监控方法，具体包括以下步骤：
S11，当监控区域的预设图像采集范围内有被采集对象进入时，采集所述被采集对象的图像信息，并根据所述图像信息获取所述被采集对象的面部特征信息。
步骤S11的具体实施方式在上述实施例中已有相关说明，同样也适用于本实施例，此处便不再作过多的描述。
以施工组从施工工地A处换到B处为例，假设李四现在正打算和小组成员一起进入施工B处，在本实施例中，可以在A处和B处之间设立图像采集范围内，该范围可以覆盖施工A处与B之间所有的空地，也可以是施工B处入口的某一范围内，由于本发明的图像采集终端采用分布式设计，完全可以根据需要选择图像采集终端进行图像信息采集的覆盖范围。当李四进入监控区域的预设图像采集范围内时，红外感应器被触发，并发送被触发提示，远程控制中心启动图像采集终端，采集李四的图像信息，远程控制中心接收到图像采集终端发送的图像信息后，跟踪采集到的图像信息，提取到李四的面部特征信息，并可以进行语音提示“该对象面部特征信息提取成功”。
S12，根据所述图像信息，检测所述被采集对象是否佩戴有安全帽。
步骤S12的具体实施方式在上述实施例中已有相关说明，同样也适用于本实施例，此处便不再作过多的描述。
其中，步骤S12具体可以包括以下实现方式：
S31，采用图像匹配算法将采集到的图像信息与预先设定的比对图像进行特征点匹配，得到特征点匹配比率值；
S32，根据所述特征点匹配比率值判断所述被采集对象是否佩戴有安全帽；
进一步的，优选方案，上述根据所述特征点匹配比率值判断所述被采集对象是否佩戴有安全帽的步骤，具体包括：
第一步，当所述特征点匹配比率值趋近于阈值时，则确定所述采集对象佩戴有安全帽；
第二步，当所述特征点匹配比率值离散于阈值时，则确定所述采集对象未佩戴有安全帽。
在上述步骤得到李四的面部特征信息的基础上，对李四是否佩戴有安全帽进行检测。当监测到有图像采集指令或操作时，在图像显示窗口嵌入预设比对图像，该比对图像可以是员工标准佩戴安全帽的示意图，也可以仅仅是整个工地统一使用的安全帽的示意图。采用SIFT算法可以排除因为图像遮挡和背景混乱而产生的无匹配关系的关键点，首先进行SIFT特征检测，定位关键点，在特征匹配中进行特征点过滤和精确定位，最后生成SIFT特征向量。取一幅图像中的一个SIFT关键点，并找出其与另一幅图像中欧式距离最近的前两个关键点，在这两个关键点中，根据最近的距离除以次近的距离得到一个特征点匹配比率值，通过判断这个特征点匹配比率值是否小于预设阈值T，来考虑是否接受这一特征点，SIFT算法创始人推荐的上述特征点匹配比率阈值为0.8，当然，预设阈值是根据无数次匹配结果来选取的，根据对匹配程度的高低要求不同而有所不一样。同时，可视化显示比对 过程，有利于增强用户体验。
显然，当所述特征点匹配比率值无限趋近于阈值T时，则接受特征点，确定李四佩戴有安全帽。当所述特征点匹配比率值离散于阈值T时，也就是说特征点匹配比率值与阈值T差距较大，不接受特征点，则确定李四未佩戴有安全帽，此时，可以生成相应广播信息进行语音广播“未佩戴安全帽不得进入”或“您未佩戴有安全帽”。
需要说明的是，在前面实施例已经有过描述，在判断采集对象，也就是此例中的李四，是否佩戴有安全帽的步骤上，采用何种图像匹配算法并不受本发明的实施例限制，在此处采用SIFT算法进行特征点匹配并不是本实施例唯一的实施方式。
S13，当检测到所述被采集对象佩戴有安全帽时，获取所述安全帽的编码。
步骤S13的具体实施方式在上述实施例中已有相关说明，同样也适用于本实施例，区别在于，获取对象是本示例中李四头上佩戴的安全帽的编码，此处便不再作过多的描述。
S14，根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型。
步骤S14的具体实施方式在上述实施例中已有相关说明，同样也适用于本实施例，此处便不再作过多的描述。
进一步的，优选方案，上述根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型的步骤，具体包括：
以所述安全帽编码为主键，从预先设定的脸部特征模型表中获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型；其中，所述脸部特征模型表中记录有所述监控区域对应的所有安全帽编码，以及每个安全帽编码对应的监控区域授权用户的脸部特征模型。
在前面获取到李四头上佩戴有安全帽并识别到安全帽编码后，以安全帽编码为主键，遍历预设的脸部特征模型表或库，调取与李四头上的安全帽的编码绑定的上述监控区域授权用户的脸部特征模型(具体可以以远程控制中心向数据调用模块发送调用指令，由数据调用模块进行遍历以及调用来完成)。
S15，将所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型进行匹配，得到匹配结果。
步骤S15的具体实施方式在上述实施例中已有相关说明，同样也适用于本实施例，此处便不再作过多的描述。
其中，步骤S15的具体实现方式包括：
S41，计算所述被采集对象的面部特征信息与所述授权用户的脸部特征模型匹配的相似程度值；
S42，当所述相似程度值大于或等于预设相似程度阈值时，判定所述匹配结果为匹配成功；
S43，当所述相似程度值小于预设相似程度阈值时，判定所述匹配结果为匹配失败。
远程控制中心将步骤S11中获取到的李四的面部特征信息与步骤S14中获取到的与李四头上佩戴的安全帽的编码具有一一对应关系或绑定关系的监控区域授权用户的脸部特征模型进行匹配，计算二者之间的相似程度值，当李四的面部特征信息与授权用户的脸部特征模型间的匹配相似程度值大于或等于预设相似程度阈值时，判定匹配结果为匹配成功，同理，当所述相似程度值小于预设相似程度阈值时，判定匹配结果为匹配失败。
S16，当所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配成功时，确定允许所述被采集对象进入所述监控区域。
步骤S16的具体实施方式在上述实施例中已有相关说明，同样也适用于本实施例，此处便不再作过多的描述。
有益效果：本发明实施例通过当监控区域的预设图像采集范围内有被采集对象进入时，采集被采集对象的图像信息，并根据图像信息获取被采集对象的面部特征信息，然后根据所采集的图像信息，采用图像匹配算法检测被采集对象是否佩戴有安全帽，当检测到被采集对象佩戴有安全帽时，获取安全帽的编码，根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型，最后将被采集对象的面部特征信息与监控区域授权用户的脸部特征模型进行匹配，得到匹配结果，当被采集对象的面部特征信息与监控区域授权用户的脸部特征模型的匹配结果为匹配成功时，确定允许被采集对象进入所述监控区域，能够实现施工工地一人一帽的安全管控，还能有效降低监控系统硬件设备的损耗。
进一步的，优选方案，上述方法还包括：
当检测到所述被采集对象未佩戴有安全帽或者所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配失败时，根据所述被采集对象的面部特征信息，判断所述被采集对象是否为所述监控区域的授权用户；
当所述被采集对象是所述监控区域的授权用户时，获取所述被采集对象的基本信息，将所述被采集对象的基本信息添加到预设黑名单库，并确定不允许所述被采集对象进入所述监控区域；其中，所述被采集对象的基本信息包括被采集对象的姓名、职位、所属部门、 工号及联系方式；
当所述被采集对象不是所述监控区域的授权用户时，向管理员发送报警信息。
在确定李四为本施工项目或本施工组进入监控区域的非授权用户时，直接向管理员发送报警信息，该报警信息可以以语音的方式发送到管理员的移动终端，比如说手机、对讲机等，也可以是短信息提醒，还可以直接以广播的方式在工地播报，如：“警惕！有陌生人闯入。”在李四为监控区域的授权用户时，判定李四未佩戴安全帽或安全帽佩戴不合规，从预先设置的监控区域的授权用户基本信息库中调取李四的基本信息，并显示在显示窗口，如“姓名：李四；职位：焊工；所述部门：焊接一组；工号：xxyyzz……”，将上述基本信息添加到黑名单库，不允许李四进入施工B处进行施工。
有益效果：当检测到被采集对象未佩戴有安全帽或者被采集对象的面部特征信息与监控区域授权用户的脸部特征模型的匹配结果为匹配失败时，根据被采集对象的面部特征信息，判断被采集对象是否为监控区域的授权用户，不是则报警，是监控区域的授权用户则判定未佩戴安全帽或安全帽佩戴不合规，将该采集对象添加到黑名单库，进一步提高了工地一人一帽管控的严密度。
进一步的，优选方案，在所述将所述被采集对象的基本信息添加到预设黑名单库的步骤之前，所述方法还包括：
根据所述被采集对象的工号遍历预设黑名单库，统计所述被采集对象的工号在所述预设黑名单库中出现的频数；
当所述被采集对象的工号在所述预设黑名单库中出现的频数等于或大于预设上限值时，执行所述将所述被采集对象的基本信息添加到预设黑名单库的步骤。
进一步的，优选方案，上述方法还包括：当所述被采集对象的工号在所述预设黑名单库中出现的频数小于预设上限值时，确定允许所述被采集对象进入所述监控区域。
将未佩戴安全帽和安全帽佩戴不合规的监控区域授权用户的基本信息添加到预设的黑名单库，根据这类用户在黑名单中出现的频数确定是否允许这类人进入施工工地的监控区域，有利于提高监控区域授权用户佩戴安全帽的意识。
在此，为便于理解本方案中基于图像分析的工地监控方法的执行过程，在本实施例中，以某一被采集对象想要通过施工工地的门禁或闸机通道，或对某一不明身份的被采集对象在施工场所内活动进行监控分析为例，阐述上述基于图像分析的工地监控方法的流程。具体的，参见图5所示，在该示例中流程如下：
S501，分布式图像采集终端采集图像并获取被采集对象的面部特征信息；
S502，判断是否佩戴安全帽，若是，则执行步骤S503，若否，则执行步骤S504；
S503，获取安全帽编码；
S504，判断是否为授权用户，若是，则执行步骤S510，若否，则执行步骤S509；
S505，获取与安全帽编码一一对应的授权用户的脸部特征模型；
S506，将面部特征信息与授权用户的脸部特征模型进行匹配，得到匹配结果；
S507，判断匹配结果是否匹配成功，若是，则执行步骤S508，若否，则执行步骤S504；
S508，允许被采集对象进入；
S509，向管理员发送报警信息；
S510，获取基本信息，遍历黑名单库；
S511，判断黑名单库频数是否超过上限，若是，则执行步骤S512，若否，则执行步骤S513；
S512，不允许被采集对象进入；
S513，添加到黑名单库；
S514，结束流程。
显而易见的，通过上述示例中的流程演示，本发明实施例中的基于图像分析的工地监控方法较现有技术相比，对工地的监控更有效、更严密，相对降低了工程商的人力成本，能够解决施工工地上时常有非授权人进入监控区域的问题，实现了一人一帽的安全管控。
实施例三
参见图6所示，本发明一方面的具体实施例提供了一种基于图像分析的工地监控装置，该基于图像分析的工地监控装置6包括存储器61、处理器62以及存储在存储器61中并可在处理器62上运行的计算机程序63，该处理器62执行计算机程序63时实现上述的基于图像分析的工地监控方法的步骤。
具体的，处理器62执行计算机程序63时实现如下步骤：当监控区域的预设图像采集范围内有被采集对象进入时，接收所述图像采集终端发送的所述被采集对象的图像信息，并根据所述图像信息获取所述被采集对象的面部特征信息；根据所述图像信息，检测所述被采集对象是否佩戴有安全帽；当检测到所述被采集对象佩戴有安全帽时，获取所述安全帽的编码；根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型；将所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型进行匹配，得到匹配结果；当所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配成功时，确定允许所述被采集对象进入所述监 控区域。
优选的，处理器62执行计算机程序63时还实现如下步骤：当检测到所述被采集对象未佩戴有安全帽或者所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配失败时，根据所述被采集对象的面部特征信息，判断所述被采集对象是否为所述监控区域的授权用户；当所述被采集对象是所述监控区域的授权用户时，获取所述被采集对象的基本信息，将所述被采集对象的基本信息添加到预设黑名单库，并确定不允许所述被采集对象进入所述监控区域；其中，所述被采集对象的基本信息包括被采集对象的姓名、职位、所属部门、工号及联系方式；
优选的，处理器62执行计算机程序63时还实现如下步骤：采用图像匹配算法将采集到的图像信息与预先设定的比对图像进行特征点匹配，得到特征点匹配比率值；其中，所述比对图像为所述安全帽的标准佩戴示意图；根据所述特征点匹配比率值判断所述被采集对象是否佩戴有安全帽。
优选的，处理器62执行计算机程序63时还实现如下步骤：当所述特征点匹配比率值趋近于阈值时，则确定所述采集对象佩戴有安全帽；当所述特征点匹配比率值离散于阈值时，则确定所述采集对象未佩戴有安全帽。
优选的，处理器62执行计算机程序63时还实现如下步骤：以所述安全帽编码为主键，从预先设定的脸部特征模型表中获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型；其中，所述脸部特征模型表中记录有所述监控区域对应的所有安全帽编码，以及每个安全帽编码对应的监控区域授权用户的脸部特征模型。
优选的，处理器62执行计算机程序63时还实现如下步骤：计算所述被采集对象的面部特征信息与所述授权用户的脸部特征模型匹配的相似程度值；当所述相似程度值大于或等于预设相似程度阈值时，判定所述匹配结果为匹配成功；当所述相似程度值小于预设相似程度阈值时，判定所述匹配结果为匹配失败。
优选的，处理器62执行计算机程序63时还实现如下步骤：根据所述被采集对象的工号遍历预设黑名单库，统计所述被采集对象的工号在所述预设黑名单库中出现的频数；当所述被采集对象的工号在所述预设黑名单库中出现的频数等于或大于预设上限值时，执行所述将所述被采集对象的基本信息添加到预设黑名单库的步骤。
优选的，处理器62执行计算机程序63时还实现如下步骤：当所述被采集对象的工号在所述预设黑名单库中出现的频数小于预设上限值时，确定允许所述被采集对象进入所述监控区域。
即，在本发明的具体实施例中，基于图像分析的工地监控装置6的处理器62执行计算机程序63时实现上述的基于图像分析的工地监控方法的步骤，能够解决施工工地上时常有非授权人出现，实现工地监控区域的一人一帽管控，避免发生意外。
示例性的，上述基于图像分析的工地监控装置6可以是计算机、笔记本、远程服务器、云端服务器及远程控制中心等服务设备。基于图像分析的工地监控装置6可包括，但不仅限于处理器62、存储器61。本领域技术人员可以理解，所述示意图仅仅是基于图像分析的工地监控装置6的示例，并不构成对基于图像分析的工地监控装置6的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件。
需要说明的是，由于基于图像分析的工地监控装置6的处理器62执行计算机程序63时实现上述的基于图像分析的工地监控方法的步骤，因此上述基于图像分析的工地监控方法的所有实施例均适用于该基于图像分析的工地监控装置6，且均能达到相同或相似的有益效果。
实施例四
本发明实施例的一方面还提供了一种计算机可读存储介质，计算机可读存储介质存储有计算机程序，该计算机程序被处理器执行时实现上述的基于图像分析的工地监控方法的步骤。
具体的，计算机程序被处理器执行时实现如下步骤：当监控区域的预设图像采集范围内有被采集对象进入时，接收所述图像采集终端发送的所述被采集对象的图像信息，并根据所述图像信息获取所述被采集对象的面部特征信息；根据所述图像信息，检测所述被采集对象是否佩戴有安全帽；当检测到所述被采集对象佩戴有安全帽时，获取所述安全帽的编码；根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型；将所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型进行匹配，得到匹配结果；当所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配成功时，确定允许所述被采集对象进入所述监控区域。
优选的，计算机程序被处理器执行时还实现如下步骤：当检测到所述被采集对象未佩戴有安全帽或者所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配失败时，根据所述被采集对象的面部特征信息，判断所述被采集对象是否为所述监控区域的授权用户；当所述被采集对象是所述监控区域的授权用户时，获取所述被采集对象的基本信息，将所述被采集对象的基本信息添加到预设黑名单库，并确定 不允许所述被采集对象进入所述监控区域；其中，所述被采集对象的基本信息包括被采集对象的姓名、职位、所属部门、工号及联系方式；当所述被采集对象不是所述监控区域的授权用户时，向管理员发送报警信息。
优选的，计算机程序被处理器执行时还实现如下步骤：采用图像匹配算法将采集到的图像信息与预先设定的比对图像进行特征点匹配，得到特征点匹配比率值；其中，所述比对图像为所述安全帽的标准佩戴示意图；根据所述特征点匹配比率值判断所述被采集对象是否佩戴有安全帽。
优选的，计算机程序被处理器执行时还实现如下步骤：当所述特征点匹配比率值趋近于阈值时，则确定所述采集对象佩戴有安全帽；当所述特征点匹配比率值离散于阈值时，则确定所述采集对象未佩戴有安全帽。
优选的，计算机程序被处理器执行时还实现如下步骤：以所述安全帽编码为主键，从预先设定的脸部特征模型表中获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型；其中，所述脸部特征模型表中记录有所述监控区域对应的所有安全帽编码，以及每个安全帽编码对应的监控区域授权用户的脸部特征模型。
优选的，计算机程序被处理器执行时还实现如下步骤：计算所述被采集对象的面部特征信息与所述授权用户的脸部特征模型匹配的相似程度值；当所述相似程度值大于或等于预设相似程度阈值时，判定所述匹配结果为匹配成功，当所述相似程度值小于预设相似程度阈值时，判定所述匹配结果为匹配失败。
优选的，计算机程序被处理器执行时还实现如下步骤：根据所述被采集对象的工号遍历预设黑名单库，统计所述被采集对象的工号在所述预设黑名单库中出现的频数；当所述被采集对象的工号在所述预设黑名单库中出现的频数等于或大于预设上限值时，执行所述将所述被采集对象的基本信息添加到预设黑名单库的步骤。
优选的，计算机程序被处理器执行时还实现如下步骤：当所述被采集对象的工号在所述预设黑名单库中出现的频数小于预设上限值时，确定允许所述被采集对象进入所述监控区域。
即，计算机可读存储介质的计算机程序被处理器执行时实现上述的基于图像分析的工地监控方法的步骤，能够解决施工工地上时常有非授权人出现，实现工地监控区域的一人一帽管控，避免发生意外。
示例性的，计算机可读存储介质的计算机程序包括计算机程序代码，所述计算机程序代码可以为源代码形式、对象代码形式、可执行文件或某些中间形式等。所述计算机可读 介质可以包括：能够携带所述计算机程序代码的任何实体或装置、记录介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、电载波信号、电信信号以及软件分发介质等。
需要说明的是，由于计算机可读存储介质的计算机程序被处理器执行时实现上述的基于图像分析的工地监控方法的步骤，因此上述基于图像分析的工地监控方法的所有实施例均适用于该计算机可读存储介质，且均能达到相同或相似的有益效果。
以上对本发明实施例进行了详细介绍，本文中应用了具体个例对本发明的原理及实施方式进行了阐述，以上实施例的说明只是用于帮助理解本发明的方法及其核心思想；同时，对于本领域的一般技术人员，依据本发明的思想，在具体实施方式及应用范围上均会有改变之处，综上所述，本说明书内容不应理解为对本发明的限制。

Claims (10)

  1. 一种基于图像分析的工地监控方法，其特征在于，所述方法包括：
    当监控区域的预设图像采集范围内有被采集对象进入时，采集所述被采集对象的图像信息，并根据所述图像信息获取所述被采集对象的面部特征信息；
    根据所述图像信息，检测所述被采集对象是否佩戴有安全帽；
    当检测到所述被采集对象佩戴有安全帽时，获取所述安全帽的编码；
    根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型；
    将所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型进行匹配，得到匹配结果；
    当所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配成功时，确定允许所述被采集对象进入所述监控区域。
  2. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    当检测到所述被采集对象未佩戴有安全帽或者所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型的匹配结果为匹配失败时，根据所述被采集对象的面部特征信息，判断所述被采集对象是否为所述监控区域的授权用户；
    当所述被采集对象是所述监控区域的授权用户时，获取所述被采集对象的基本信息，将所述被采集对象的基本信息添加到预设黑名单库，并确定不允许所述被采集对象进入所述监控区域；其中，所述被采集对象的基本信息包括被采集对象的姓名、职位、所属部门、工号及联系方式；
    当所述被采集对象不是所述监控区域的授权用户时，向管理员发送报警信息。
  3. 根据权利要求1所述的方法，其特征在于，所述根据所述图像信息，检测所述被采集对象是否佩戴有安全帽的步骤，具体包括：
    采用图像匹配算法将采集到的图像信息与预先设定的比对图像进行特征点匹配，得到特征点匹配比率值；其中，所述比对图像为所述安全帽的标准佩戴示意图；
    根据所述特征点匹配比率值判断所述被采集对象是否佩戴有安全帽。
  4. 根据权利要求3所述的方法，其特征在于，所述根据所述特征点匹配比率值判断所述被采集对象是否佩戴有安全帽的步骤，具体包括：
    当所述特征点匹配比率值趋近于阈值时，则确定所述采集对象佩戴有安全帽；
    当所述特征点匹配比率值离散于阈值时，则确定所述采集对象未佩戴有安全帽。
  5. 根据权利要求1所述的方法，其特征在于，所述根据所述安全帽的编码，获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型的步骤，具体包括：
    以所述安全帽编码为主键，从预先设定的脸部特征模型表中获取与所述编码具有一一对应关系的所述监控区域授权用户的脸部特征模型；其中，所述脸部特征模型表中记录有所述监控区域对应的所有安全帽编码，以及每个安全帽编码对应的监控区域授权用户的脸部特征模型。
  6. 根据权利要求1所述的方法，其特征在于，所述将所述被采集对象的面部特征信息与所述监控区域授权用户的脸部特征模型进行匹配，得到匹配结果的步骤，具体包括：
    计算所述被采集对象的面部特征信息与所述授权用户的脸部特征模型匹配的相似程度值；
    当所述相似程度值大于或等于预设相似程度阈值时，判定所述匹配结果为匹配成功；
    当所述相似程度值小于预设相似程度阈值时，判定所述匹配结果为匹配失败。
  7. 根据权利要求2所述的方法，其特征在于，在所述将所述被采集对象的基本信息添加到预设黑名单库的步骤之前，所述方法还包括：
    根据所述被采集对象的工号遍历预设黑名单库，统计所述被采集对象的工号在所述预设黑名单库中出现的频数；
    当所述被采集对象的工号在所述预设黑名单库中出现的频数等于或大于预设上限值时，执行所述将所述被采集对象的基本信息添加到预设黑名单库的步骤。
  8. 根据权利要求7所述的方法，其特征在于，所述方法还包括：
    当所述被采集对象的工号在所述预设黑名单库中出现的频数小于预设上限值时，确定允许所述被采集对象进入所述监控区域。
  9. 一种基于图像分析的工地监控装置，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，其特征在于，所述处理器执行所述计算机程序时实现如权利要求1至8任一项所述的基于图像分析的工地监控方法的步骤。
  10. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求1至8任一项所述的基于图像分析的工地监控方法的步骤。
PCT/CN2018/124839 2018-07-25 2018-12-28 基于图像分析的工地监控方法、装置及可读存储介质 WO2020019673A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810830408.X 2018-07-25
CN201810830408.XA CN109344679A (zh) 2018-07-25 2018-07-25 基于图像分析的工地监控方法、装置及可读存储介质

Publications (1)

Publication Number Publication Date
WO2020019673A1 true WO2020019673A1 (zh) 2020-01-30

Family

ID=65291684

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/124839 WO2020019673A1 (zh) 2018-07-25 2018-12-28 基于图像分析的工地监控方法、装置及可读存储介质

Country Status (2)

Country Link
CN (1) CN109344679A (zh)
WO (1) WO2020019673A1 (zh)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111753705A (zh) * 2020-06-19 2020-10-09 神思电子技术股份有限公司 一种基于视频分析的智慧工地安全作业的检测方法
CN111860256A (zh) * 2020-07-10 2020-10-30 云镝智慧科技有限公司 安全检测方法、装置、计算机设备和存储介质
CN112232239A (zh) * 2020-10-20 2021-01-15 华雁智能科技(集团)股份有限公司 变电站施工安全标示牌监控方法、装置和电子设备
CN112347989A (zh) * 2020-11-30 2021-02-09 广联达科技股份有限公司 反光衣识别方法、装置、计算机设备和可读存储介质
CN113343846A (zh) * 2021-06-04 2021-09-03 温州大学 一种基于深浅层特征融合的反光衣检测系统
CN113408365A (zh) * 2021-05-26 2021-09-17 广东能源集团科学技术研究院有限公司 一种复杂场景下的安全帽识别方法及装置
CN113516632A (zh) * 2021-05-18 2021-10-19 浙江大华技术股份有限公司 检测方法、装置、系统、存储介质及电子装置
CN113837030A (zh) * 2021-09-06 2021-12-24 中国福利会国际和平妇幼保健院 用于疫情防控的人员智能管控方法、系统及计算机设备
CN114063489A (zh) * 2021-09-26 2022-02-18 中通服和信科技有限公司 一种基于3d可视化和物联网技术的智慧门控系统
CN114155669A (zh) * 2021-11-30 2022-03-08 安徽富煌钢构股份有限公司 一种基于bim的建筑施工安全预警防护系统
CN115588167A (zh) * 2022-12-08 2023-01-10 山东法尔智能科技有限公司 一种工程施工用安全监测方法及系统
CN116758269A (zh) * 2023-08-22 2023-09-15 雄安雄创数字技术有限公司 一种位置校验方法

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112016777A (zh) * 2019-05-30 2020-12-01 上海若泰工程建设管理技术有限公司 铁路建设工程质量安全管理系统及方法
CN110533811A (zh) * 2019-08-28 2019-12-03 深圳市万睿智能科技有限公司 基于ssd实现安全帽检查的方法及装置与系统和存储介质
CN111178246B (zh) * 2019-12-27 2024-02-09 深圳供电局有限公司 基于电力施工操作的远程监测方法、装置和计算机设备
CN111443633A (zh) * 2020-04-13 2020-07-24 深圳市水务技术服务有限公司 一体化泵站远程监控管理方法、装置、设备以及存储介质
CN113570756B (zh) * 2021-06-18 2023-02-10 广联达科技股份有限公司 一种人员统计方法、装置及电子设备
CN113469043B (zh) * 2021-06-30 2024-10-18 南方科技大学 安全帽佩戴状态检测方法、装置、计算机设备和存储介质
CN113741317A (zh) * 2021-09-16 2021-12-03 镇江绿建工程咨询有限公司 一种太阳能供电的智慧工地用危险区域人员接近探测系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104506562A (zh) * 2015-01-13 2015-04-08 东北大学 融合二维码与人脸识别的会议身份认证装置及方法
CN106778609A (zh) * 2016-12-15 2017-05-31 国网浙江省电力公司杭州供电公司 一种电力施工现场人员制服穿着识别方法
CN107944838A (zh) * 2017-12-18 2018-04-20 四川省大益科技有限公司 一种智能道闸现场监控系统及方法
CN108198281A (zh) * 2017-12-15 2018-06-22 四川省大益科技有限公司 智能道闸现场监控系统及其控制方法

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101340287B1 (ko) * 2012-04-13 2013-12-10 경기대학교 산학협력단 스마트 홈에서 마이닝 기반의 패턴 분석을 이용한 침입탐지 시스템
CN102829788A (zh) * 2012-08-27 2012-12-19 北京百度网讯科技有限公司 一种实景导航方法和实景导航装置
CN104517206A (zh) * 2013-09-27 2015-04-15 中国移动通信集团黑龙江有限公司 一种业务处理的方法及装置
CN108287833A (zh) * 2017-01-09 2018-07-17 北京艺鉴通科技有限公司 一种用于艺术品鉴定的以图搜图方法
CN106993168B (zh) * 2017-04-28 2019-09-06 珠海格力电器股份有限公司 一种监控方法、装置、电子设备及系统
CN108198318B (zh) * 2018-02-09 2024-04-09 厦门路桥信息股份有限公司 一种实名制通道门禁方法及系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104506562A (zh) * 2015-01-13 2015-04-08 东北大学 融合二维码与人脸识别的会议身份认证装置及方法
CN106778609A (zh) * 2016-12-15 2017-05-31 国网浙江省电力公司杭州供电公司 一种电力施工现场人员制服穿着识别方法
CN108198281A (zh) * 2017-12-15 2018-06-22 四川省大益科技有限公司 智能道闸现场监控系统及其控制方法
CN107944838A (zh) * 2017-12-18 2018-04-20 四川省大益科技有限公司 一种智能道闸现场监控系统及方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
XU, JIAN ET AL: "Research on Conference Identity Authentication System Based on Two-Dimensional Bar Code and Face Recognition", NETINFO SECURITY, no. 4, 15 July 2015 (2015-07-15), pages 13 - 18 *

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111753705B (zh) * 2020-06-19 2023-11-03 神思电子技术股份有限公司 一种基于视频分析的智慧工地安全作业的检测方法
CN111753705A (zh) * 2020-06-19 2020-10-09 神思电子技术股份有限公司 一种基于视频分析的智慧工地安全作业的检测方法
CN111860256A (zh) * 2020-07-10 2020-10-30 云镝智慧科技有限公司 安全检测方法、装置、计算机设备和存储介质
CN112232239A (zh) * 2020-10-20 2021-01-15 华雁智能科技(集团)股份有限公司 变电站施工安全标示牌监控方法、装置和电子设备
CN112347989A (zh) * 2020-11-30 2021-02-09 广联达科技股份有限公司 反光衣识别方法、装置、计算机设备和可读存储介质
CN113516632A (zh) * 2021-05-18 2021-10-19 浙江大华技术股份有限公司 检测方法、装置、系统、存储介质及电子装置
CN113408365A (zh) * 2021-05-26 2021-09-17 广东能源集团科学技术研究院有限公司 一种复杂场景下的安全帽识别方法及装置
CN113408365B (zh) * 2021-05-26 2023-09-08 广东能源集团科学技术研究院有限公司 一种复杂场景下的安全帽识别方法及装置
CN113343846A (zh) * 2021-06-04 2021-09-03 温州大学 一种基于深浅层特征融合的反光衣检测系统
CN113343846B (zh) * 2021-06-04 2024-03-15 温州大学 一种基于深浅层特征融合的反光衣检测系统
CN113837030A (zh) * 2021-09-06 2021-12-24 中国福利会国际和平妇幼保健院 用于疫情防控的人员智能管控方法、系统及计算机设备
CN113837030B (zh) * 2021-09-06 2024-03-22 中国福利会国际和平妇幼保健院 用于疫情防控的人员智能管控方法、系统及计算机设备
CN114063489A (zh) * 2021-09-26 2022-02-18 中通服和信科技有限公司 一种基于3d可视化和物联网技术的智慧门控系统
CN114155669A (zh) * 2021-11-30 2022-03-08 安徽富煌钢构股份有限公司 一种基于bim的建筑施工安全预警防护系统
CN115588167A (zh) * 2022-12-08 2023-01-10 山东法尔智能科技有限公司 一种工程施工用安全监测方法及系统
CN116758269A (zh) * 2023-08-22 2023-09-15 雄安雄创数字技术有限公司 一种位置校验方法
CN116758269B (zh) * 2023-08-22 2023-11-14 雄安雄创数字技术有限公司 一种位置校验方法

Also Published As

Publication number Publication date
CN109344679A (zh) 2019-02-15

Similar Documents

Publication Publication Date Title
WO2020019673A1 (zh) 基于图像分析的工地监控方法、装置及可读存储介质
KR101215948B1 (ko) 신체정보 및 얼굴인식에 기반한 감시 시스템의 영상정보 마스킹 방법
KR102012672B1 (ko) 얼굴인식 기반의 사람 특징인식을 이용한 방범 시스템 및 방법
CN206515931U (zh) 一种人脸识别系统
KR102039277B1 (ko) 보행자 얼굴 인식 시스템 및 그 방법
CN101174337A (zh) 基于人脸识别的监控装置及嵌入式门禁监控系统
KR102237086B1 (ko) 안면인식 기능에 기초하여 5g 이동통신망의 이용이 가능한 통신 단말기를 통해 영상감시가 가능한 로비폰 제어 장치 및 방법
CN110807888A (zh) 一种园区智能安防方法、系统及存储介质
JP2010154134A (ja) 映像監視システム
KR101372860B1 (ko) 영상 검색 시스템 및 영상 분석 서버
KR20190030960A (ko) 차량 번호 인식 시스템을 사용한 출입 관리 시스템 및 방법
CN105574477A (zh) 一种安全防盗方法、装置及系统
CN106599879A (zh) 基于云端的安防控制方法及终端
CN109784231A (zh) 安防信息管理方法、装置及存储介质
CN109495727A (zh) 智能监控方法及装置、系统、可读存储介质
CN109800638A (zh) 一种基于人脸识别技术的重点人监控方法
US10783365B2 (en) Image processing device and image processing system
CN111064928A (zh) 一种具有人脸识别功能的视频监控系统
Mun et al. Design for visitor authentication based on face recognition technology Using CCTV
KR20210043960A (ko) IoT와 인공지능 기술을 이용한 행동인식 기반의 안전 감시 시스템 및 방법
CN111385440A (zh) 具有人脸记录与查询功能的监控摄像机
CN205582055U (zh) 智能无障碍通道安全识别系统
KR20100077662A (ko) 지능형 영상감시 시스템 및 영상감시 방법
CN111563174A (zh) 图像处理方法、图像处理装置、电子设备及存储介质
WO2023093241A1 (zh) 行人重识别方法及装置、存储介质

Legal Events

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

Ref document number: 18927798

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18927798

Country of ref document: EP

Kind code of ref document: A1