[go: up one dir, main page]

CN117671887B - Intelligent security early warning management method and system based on big data - Google Patents

Intelligent security early warning management method and system based on big data Download PDF

Info

Publication number
CN117671887B
CN117671887B CN202410062491.6A CN202410062491A CN117671887B CN 117671887 B CN117671887 B CN 117671887B CN 202410062491 A CN202410062491 A CN 202410062491A CN 117671887 B CN117671887 B CN 117671887B
Authority
CN
China
Prior art keywords
early warning
intelligent security
feature
security early
image data
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.)
Active
Application number
CN202410062491.6A
Other languages
Chinese (zh)
Other versions
CN117671887A (en
Inventor
杨瑞士
胡汉生
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Qiaohua Industries Ltd
Original Assignee
Shenzhen Qiaohua Industries Ltd
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 Shenzhen Qiaohua Industries Ltd filed Critical Shenzhen Qiaohua Industries Ltd
Priority to CN202410062491.6A priority Critical patent/CN117671887B/en
Publication of CN117671887A publication Critical patent/CN117671887A/en
Application granted granted Critical
Publication of CN117671887B publication Critical patent/CN117671887B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B19/00—Alarms responsive to two or more different undesired or abnormal conditions, e.g. burglary and fire, abnormal temperature and abnormal rate of flow
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B17/00—Fire alarms; Alarms responsive to explosion
    • G08B17/10—Actuation by presence of smoke or gases, e.g. automatic alarm devices for analysing flowing fluid materials by the use of optical means
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B17/00—Fire alarms; Alarms responsive to explosion
    • G08B17/12—Actuation by presence of radiation or particles, e.g. of infrared radiation or of ions

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Multimedia (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Human Computer Interaction (AREA)
  • Business, Economics & Management (AREA)
  • Emergency Management (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Image Analysis (AREA)

Abstract

The invention relates to the technical field of security early warning and discloses an intelligent security early warning management method and system based on big data. The method comprises the following steps: whether an emergency occurs in the intelligent security early warning area is monitored in real time by installing a series of monitoring equipment in the intelligent security early warning area, and real-time alarming is carried out on the emergent condition through the installed monitoring equipment; meanwhile, the method further comprises the steps of identifying each person in the intelligent security early warning area through the installation of the face identification terminal, improving the storage efficiency of face identification through the index establishment mode, determining face identification information through the multi-frame shooting mode, judging whether the person is dangerous or not through the identification of the face image data through the convolutional neural network, and immediately giving an alarm when the person is detected to be dangerous. The intelligent security early warning management method and system based on big data can effectively manage and monitor personnel entering the intelligent security early warning area, and provide guarantee for the intelligent security early warning area.

Description

Intelligent security early warning management method and system based on big data
Technical Field
The invention relates to the technical field of security early warning, in particular to an intelligent security early warning management method and system based on big data.
Background
With the acceleration of the urban process of China, more and more people enter the urban settlement from the village, and the urban population grows. Urban improvement can improve the living standard of people to a certain extent, but the urban population density is increased, so that the social complexity in the city is increased: the space structure is complex, population density is high, mobility is high, social relations are numerous and complex, traffic jam is serious, and accidents and malignant events are frequent. Moreover, due to the increase of urban population and scale, urban managers are also challenged to manage cities and protect public safety.
Therefore, when the city manager protects, the state of the protected object needs to be considered, and if the obtained object state is lagged or has a large error, the protection by the city manager is often lagged or erroneous.
In the prior art CN109064698a, whether abnormal behavior exists is determined only by analyzing the advancing track and behavior pattern of the person in the monitoring area, so that the state of the monitored object is not monitored accurately and in real time, and the method has a great limitation.
Disclosure of Invention
(One) solving the technical problems
Aiming at the defects of the prior art, the invention provides the intelligent security early warning management method and system based on big data, which have the advantages of accurate identification, real-time alarm and the like, and solve the problems of lag and larger error of public security states.
(II) technical scheme
In order to solve the technical problems of lag and larger error of the public security state, the invention provides the following technical scheme:
the embodiment discloses an intelligent security early warning management method based on big data, which specifically comprises the following steps:
S1, installing a series of monitoring devices and face recognition terminals around an intelligent security early warning area;
s2, monitoring the surrounding environment of the intelligent security early warning area in real time through installed monitoring equipment;
S3, collecting face image data of each person in the intelligent security early warning area, identifying the face image data through convolution, and processing and storing the identified face image data;
face image data of each dangerous person in the alarm information are collected, the face image data are identified through convolution, and the face image data after identification are processed and stored;
s4, monitoring the intelligent security early warning area in real time through a series of installed monitoring equipment and a face recognition terminal;
S5, the face recognition terminal determines face image information according to the current shot image data and processes the face image information;
s6, after the face target is determined, comparing the extracted feature array with the feature arrays in the stored indexes by using a convolutional neural network;
s7, judging whether an alarm needs to be given or not in real time according to the comparison result;
preferably, the installed monitoring device monitors the real-time environment of the intelligent security early warning area, and comprises the following steps:
the installed monitoring device includes: the device comprises a smoke alarm, infrared detection equipment, a camera and an alarm device;
S21, setting a protection area in the intelligent security early warning area;
S22, opening a camera through infrared detection equipment when detecting that people come around a protection area, monitoring nearby people in real time through the camera, and generating an alarm signal and waking up an alarm device when detecting that illegal behaviors exist in the people;
s23, monitoring whether flame exists in the protection area in real time through infrared detection equipment; when detecting that flame exists in the protection area, the infrared detection equipment can generate an alarm signal and wake up the alarm device;
S24, monitoring whether smoke exists in the protection area in real time through an installed smoke alarm, and when the smoke exists in the protection area, generating an alarm signal and waking up an alarm device by the smoke alarm.
Preferably, the collecting face image data of each person in the intelligent security early warning area, identifying the face image data through convolution, and processing and storing the identified face image data includes the following steps:
the convolutional neural network recognition of the face image data comprises the following steps:
s31, dividing face image data into n small-area image data blocks;
s32, transmitting the divided image data blocks to a convolutional neural network for feature extraction;
S33, unfolding and combining the extracted characteristic data to obtain a characteristic array;
S34, optimizing the obtained feature array according to a clustering algorithm, and establishing a corresponding index according to a clustering result;
s35, constructing independent indexes for each clustered class center and storing all the indexes together.
Preferably, the optimizing the obtained feature array according to the clustering algorithm includes the following steps:
Initializing k category centers according to the obtained feature array, and k 1,k2,k3,...,kn;
initially setting each feature array as a category center of each category;
establishing a feature vector based on the obtained feature data and the feature array;
LinkWa=<w1,w2,...,wL>
LinkWb=<w1,w2,...,wN>
Link Wa represents a value of the feature vector w a,w1 with the largest occurrence number in the feature array, and the feature vector is mapped into a directed graph based on the occurrence number and the occurrence position of the feature data;
judging the similarity of the two feature arrays by calculating the maximum cosine similarity of the two feature vectors;
Wherein Len （Wa）、Len（Wb） represents breadth-first traversal information of the feature data in the directed graph; dep （Wpab） represents the most recent common root node for which feature data is mapped in the directed graph; dep （Wb）、Dep（Wa） represents depth information of the feature data in the mapped directed graph;
When the similarity Sim is [0,1], the two feature arrays are similar, and when the similarity Sim is [0, -1], the two feature arrays are dissimilar;
when the two feature arrays are similar, the similar feature arrays are arranged in a descending order according to the size of the similarity; selecting arrays with similarity between [0,1] to merge two by two to form a set; and selecting a set center point as a new category center based on the combined set size;
Clustering is carried out between the new class center and the unmixed feature arrays by carrying out similarity again until the new class center does not appear any more, and iteration is stopped.
Preferably, the transmitting the divided image data blocks to the convolutional neural network for feature extraction includes:
after receiving an input image data block, the convolution layer moves on the input image data block according to a set step length through a convolution kernel and performs multiply accumulation on a corresponding area of each step and a characteristic value of the area, so that characteristic extraction of each image data block is realized;
The convolution calculation formula is as follows:
Wherein, The input characteristic value is represented by a value of the input characteristic,Representing the weight of the corresponding convolution kernel, b representing the offset value;
for the extracted characteristics of the image data block, the characteristics of the extracted image data block are subjected to characteristic dimension reduction through a pooling layer, the important information in most input characteristics is reserved, the resolution of the input characteristics is reduced to the greatest extent for the subsequent convolution layer and the full connection layer, and the calculated amount of the model is reduced;
in the neural network, the output of the upper layer is used as the input of the lower layer, and the convolutional neural network is formed by continuously stacking;
the data must be subjected to the process of activating the function during the process of inputting the data to the lower layer;
assume that the input value is (I=1, 2,) n), each output valueThe corresponding input weight isB is an offset, and output results obtained after the input values are input into the convolutional neural network are as follows:
Wherein, Is a corresponding activation function;
Data is transmitted to the full connection layer after convolution and pooling which are continuously stacked;
And unfolding and combining the characteristic data through the full connection layer to obtain the characteristic data.
Preferably, the real-time monitoring of the intelligent security early warning area through the installed series of monitoring devices and the face recognition terminal comprises the following steps:
s41, numbering q monitoring points q 1,q2,...qq based on the set monitoring points;
S42, taking a monitoring point q 1 as a coordinate origin, establishing a three-dimensional control coordinate system according to the q 1 coordinate, and determining the coordinates of other monitoring points based on the monitoring point q 1;
s43, storing the coordinates of each monitoring point in a database of the system;
S44, the face recognition device installed at each monitoring point shoots face images of the person at the current position in real time.
Preferably, the face recognition terminal determines face image information according to the currently photographed image data, and processes the face image information, including the following steps:
S51, setting the number of frames shot by a face recognition terminal, and shooting a plurality of frames of face target images;
S52, transmitting the shot face target image to a trained convolutional neural network, identifying the face target, storing characteristic points, and obtaining the approximate coordinate position of the face target in one frame of image;
And S53, extracting the characteristics of another frame of image, obtaining a new coordinate position, comparing the new coordinate position with the characteristic points of the previous frame, and judging whether the image is the same face target or not according to the fact that the coincidence degree is higher than 80%.
Preferably, the step of comparing the extracted feature array with the feature arrays in the stored indexes by the convolutional neural network after the face target is determined includes the following steps:
S61, calling each stored index into a convolutional neural network;
the step of calling the stored indexes into the convolutional neural network comprises the following steps:
s611, performing preliminary matching by calculating the similarity between the extracted feature array and each index;
S612, after the matching is successful, calling each characteristic data in the corresponding index into a convolutional neural network based on the matching result;
s62, comparing the extracted feature array with the feature arrays of all people in the intelligent security early warning area in all the stored indexes;
S63, comparing the extracted feature array with the feature arrays of all dangerous persons in the alarm information in all the stored indexes;
Preferably, the determining whether the alarm is needed in real time according to the comparison result includes:
If the matching is successful, verifying that the intelligent security early warning area is indicated as a worker, and performing accurate matching based on the index passing the matching verification; if the intelligent security early warning areas are not matched, verification fails, and the intelligent security early warning areas are not workers;
step S63 is carried out, secondary matching is carried out, if the matching is carried out, verification is passed, dangerous personnel in the alarm information is indicated, and the face recognition terminal can report the alarm information; if the information is not matched, the verification failure indicates that the information is not dangerous personnel in the alarm information.
The embodiment also discloses an intelligent security early warning management system based on big data, which specifically comprises: the system comprises a smoke alarm, a face recognition terminal, infrared detection equipment, a camera and an alarm device;
The infrared detection equipment is used for monitoring surrounding conditions in real time, and starting a camera according to the surrounding conditions;
The camera is used for shooting surrounding condition information in real time;
The alarm equipment is used for receiving the alarm signal in real time, generating alarm information after receiving the alarm signal and reporting the alarm information through the Internet;
The smoke alarm is used for monitoring whether smoke exists in the protection area in real time and generating an alarm signal according to the smoke condition.
(III) beneficial effects
Compared with the prior art, the invention provides a smart security early warning management method and system based on big data, which have the following beneficial effects:
1. According to the intelligent security early warning system, the intelligent security early warning area is monitored and early warned in a mode that different monitoring points are arranged in the intelligent security early warning area, and a series of monitoring equipment and face recognition terminals are arranged in the monitoring; meanwhile, the system has the functions of early warning, processing and preventing flames, smoke and dangerous personnel through the cooperation of a series of installed monitoring equipment and the face recognition terminal, and the intelligent level of monitoring in an intelligent security early warning area is improved.
2. The invention provides a basic reference for the recognition of the face recognition terminal by collecting face image information data of people in the intelligent security early warning area and carrying out feature extraction and recognition through the convolutional neural network, and simultaneously realizes the detection of the people coming and going in the intelligent security early warning area by extracting the face image data shot in real time and comparing the face image data with the recognized face image features, and arranges corresponding precautionary measures based on the detection condition, thereby reducing the potential safety hazards in the intelligent security early warning area.
3. According to the invention, through a multi-frame shooting mode, the characteristics of a plurality of face images shot continuously are extracted, and whether the face images are the same face target is judged through a characteristic point comparison mode, so that the accuracy and the reliability of face recognition are improved.
4. According to the invention, the face characteristic data after identification is clustered by using a clustering algorithm, and the collected face characteristic data is classified by using the clustering algorithm, so that the storage efficiency of the face characteristic data is improved.
5. According to the invention, the storage efficiency of the face feature data is improved by establishing an index mode for the identified feature data, and meanwhile, the frequency of feature data comparison is reduced and the face recognition speed is improved by comparing the face feature extracted by real-time shooting with the feature data stored in the index.
Drawings
FIG. 1 is a schematic flow chart of a smart security early warning management method based on big data;
Fig. 2 is a schematic diagram of a face recognition terminal judging flow structure according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The embodiment discloses an intelligent security early warning management method based on big data, which specifically comprises the following steps:
S1, installing a series of monitoring devices and face recognition terminals around an intelligent security early warning area;
s2, monitoring the surrounding environment of the intelligent security early warning area in real time through installed monitoring equipment;
the installed monitoring device includes: the device comprises a smoke alarm, infrared detection equipment, a camera and an alarm device;
further, the monitoring equipment installed carries out real-time environment monitoring on the intelligent security early warning area, and the method comprises the following steps:
S21, setting a protection area in the intelligent security early warning area;
S22, opening a camera through infrared detection equipment when detecting that people come around a protection area, monitoring nearby people in real time through the camera, and generating an alarm signal and waking up an alarm device when detecting that illegal behaviors exist in the people;
s23, monitoring whether flame exists in the protection area in real time through infrared detection equipment; when detecting that flame exists in the protection area, the infrared detection equipment can generate an alarm signal and wake up the alarm device;
S24, monitoring whether smoke exists in the protection area in real time through an installed smoke alarm, and when the smoke exists in the protection area, generating an alarm signal and waking up an alarm device by the smoke alarm;
S3, collecting face image data of each person in the intelligent security early warning area, identifying the face image data through convolution, and processing and storing the identified face image data;
the convolutional neural network recognition of the face image data comprises the following steps:
s31, dividing face image data into n small-area image data blocks;
s32, transmitting the divided image data blocks to a convolutional neural network for feature extraction;
Further, transmitting the divided image data blocks to a convolutional neural network for feature extraction includes:
after receiving an input image data block, the convolution layer moves on the input image data block according to a set step length through a convolution kernel and performs multiply accumulation on a corresponding area of each step and a characteristic value of the area, so that characteristic extraction of each image data block is realized;
The convolution calculation formula is as follows:
Wherein, The input characteristic value is represented by a value of the input characteristic,Representing the weight of the corresponding convolution kernel, b representing the offset value;
for the extracted characteristics of the image data block, the characteristics of the extracted image data block are subjected to characteristic dimension reduction through a pooling layer, the important information in most input characteristics is reserved, the resolution of the input characteristics is reduced to the greatest extent for the subsequent convolution layer and the full connection layer, and the calculated amount of the model is reduced;
in the neural network, the output of the upper layer is used as the input of the lower layer, and the convolutional neural network is formed by continuously stacking;
the data must be subjected to the process of activating the function during the process of inputting the data to the lower layer;
assume that the input value is (I=1, 2,) n), each output valueThe corresponding input weight isB is an offset, and output results obtained after the input values are input into the convolutional neural network are as follows:
Wherein, Is a corresponding activation function;
Data is transmitted to the full connection layer after convolution and pooling which are continuously stacked;
unfolding and combining the characteristic data through the full connection layer to obtain one characteristic data;
S33, unfolding and combining the extracted characteristic data to obtain a characteristic array;
S34, optimizing the obtained feature array according to a clustering algorithm, and establishing a corresponding index according to a clustering result;
Clustering the obtained feature array comprises the following steps:
Initializing k category centers according to the obtained feature array, and k 1,k2,k3,...,kn;
initially setting each feature array as a category center of each category;
further, establishing a feature vector based on the obtained feature data and the feature array;
LinkWa=<w1,w2,...,wL>
LinkWb=<w1,w2,...,wN>
Link Wa represents a value of the feature vector w a,w1 with the largest occurrence number in the feature array, and the feature vector is mapped into a directed graph based on the occurrence number and the occurrence position of the feature data;
judging the similarity of the two feature arrays by calculating the maximum cosine similarity of the two feature vectors;
Wherein Len （Wa）、Len（Wb） represents breadth-first traversal information of the feature data in the directed graph; dep （Wpab） represents the most recent common root node for which feature data is mapped in the directed graph; dep （Wb）、Dep（Wa） represents depth information of the feature data in the mapped directed graph;
When the similarity Sim is [0,1], the two feature arrays are similar, and when the similarity Sim is [0, -1], the two feature arrays are dissimilar;
further, when the two feature arrays are similar, the similar feature arrays are arranged in a descending order according to the size of the similarity; selecting arrays with similarity between [0,1] to merge two by two to form a set; and selecting a set center point as a new category center based on the combined set size;
further, clustering is carried out between the new class center and the unmixed feature arrays in a mode of carrying out similarity again until the new class center does not appear any more, and iteration is stopped;
s35, constructing independent indexes for each clustered class center and summarizing and storing all indexes;
Further, face image data of each dangerous person in the alarm information are collected, the face image data are identified through convolution, and the face image data after identification are processed and stored;
s4, monitoring the intelligent security early warning area in real time through a series of installed monitoring equipment and a face recognition terminal;
the real-time monitoring of the intelligent security early warning area through the installed series of monitoring equipment and the face recognition terminal comprises the following steps:
the steps of setting q monitoring points around the intelligent security early warning area and installing a face recognition device on each monitoring point comprise the following steps:
s41, numbering q monitoring points q 1,q2,...qq based on the set monitoring points;
S42, taking a monitoring point q 1 as a coordinate origin, establishing a three-dimensional control coordinate system according to the q 1 coordinate, and determining the coordinates of other monitoring points based on the monitoring point q 1;
s43, storing the coordinates of each monitoring point in a database of the system;
S44, the face recognition device installed at each monitoring point shoots face images of the person at the current position in real time.
S5, the face recognition terminal determines face image information according to the current shot image data and processes the face image information;
The face recognition terminal can determine face image information according to the current shot image data and process the face image information, and the face recognition terminal comprises the following steps:
S51, setting the number of frames shot by a face recognition terminal, and shooting a plurality of frames of face target images;
S52, transmitting the shot face target image to a trained convolutional neural network, identifying the face target, storing characteristic points, and obtaining the approximate coordinate position of the face target in one frame of image;
S53, extracting features of another frame of image, obtaining a new coordinate position, comparing the new coordinate position with the feature point of the previous frame, and judging whether the new coordinate position is the same face target or not according to the fact that the coincidence degree is higher than 80%;
S6, after the face target is determined, comparing the extracted feature array with the feature arrays in the stored indexes by a convolutional neural network, wherein the method specifically comprises the following steps of:
S61, calling each stored index into a convolutional neural network;
the step of calling the stored indexes into the convolutional neural network comprises the following steps:
s611, performing preliminary matching by calculating the similarity between the extracted feature array and each index;
S612, after the matching is successful, calling each characteristic data in the corresponding index into a convolutional neural network based on the matching result;
s62, comparing the extracted feature array with the feature arrays of all people in the intelligent security early warning area in all the stored indexes;
S63, comparing the extracted feature array with the feature arrays of all dangerous persons in the alarm information in all the stored indexes;
Further, determining whether an alarm is needed in real time according to the comparison result includes:
If the matching is successful, verifying that the intelligent security early warning area is indicated as a worker, and performing accurate matching based on the index passing the matching verification; if the intelligent security early warning areas are not matched, verification fails, and the intelligent security early warning areas are not workers;
Step S73, performing secondary matching, if the matching is successful, verifying that the person is a dangerous person in the alarm information, and reporting the alarm information by the face recognition terminal; if the information is not matched, the verification failure indicates that the information is not dangerous personnel in the alarm information;
The embodiment also discloses an intelligent security early warning management system based on big data, which specifically comprises: the system comprises a smoke alarm, a face recognition terminal, infrared detection equipment, a camera and an alarm device;
The infrared detection equipment is used for monitoring surrounding conditions in real time, and starting a camera according to the surrounding conditions;
The camera is used for shooting surrounding condition information in real time;
The alarm equipment is used for receiving the alarm signal in real time, generating alarm information after receiving the alarm signal and reporting the alarm information through the Internet;
the smoke alarm is used for monitoring whether smoke exists in the protection area in real time and generating an alarm signal according to the smoke condition;
The face recognition terminal is used for scanning and comparing the face image information and displaying the compared result.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (9)

1. The intelligent security early warning management method based on big data is characterized by comprising the following steps of:
S1, installing a series of monitoring devices and face recognition terminals around an intelligent security early warning area;
s2, monitoring the surrounding environment of the intelligent security early warning area in real time through installed monitoring equipment;
S3, collecting face image data of each person in the intelligent security early warning area, identifying the face image data through convolution, and processing and storing the identified face image data;
face image data of each dangerous person in the alarm information are collected, the face image data are identified through convolution, and the face image data after identification are processed and stored;
The method for collecting the face image data of each person in the intelligent security early warning area, identifying the face image data through a convolutional neural network, and processing and storing the identified face image data comprises the following steps:
the convolutional neural network recognition of the face image data comprises the following steps:
s31, dividing face image data into n small-area image data blocks;
s32, transmitting the divided image data blocks to a convolutional neural network for feature extraction;
S33, unfolding and combining the extracted characteristic data to obtain a characteristic array;
S34, optimizing the obtained feature array according to a clustering algorithm, and establishing a corresponding index according to a clustering result; the optimizing of the obtained feature array according to the clustering algorithm comprises the following steps:
Initializing k category centers according to the obtained feature array, and k 1,k2,k3,...,kn;
initially setting each feature array as a category center of each category;
establishing a feature vector based on the obtained feature data and the feature array;
LinkWa＝<w1,w2,...,wL>
LinkWb＝<w1,w2,...,wN>
Link Wa represents a value of the feature vector w a,w1 with the largest occurrence number in the feature array, and the feature vector is mapped into a directed graph based on the occurrence number and the occurrence position of the feature data;
judging the similarity of the two feature arrays by calculating the maximum cosine similarity of the two feature vectors;
Wherein Len (Wa)、Len(Wb) represents breadth-first traversal information of the feature data in the directed graph; dep (Wpab) represents the most recent common root node for which feature data is mapped in the directed graph; dep (Wb)、Dep(Wa) represents depth information of the feature data in the mapped directed graph;
When the similarity Sim is [0,1], the two feature arrays are similar, and when the similarity Sim is [0, -1], the two feature arrays are dissimilar;
when the two feature arrays are similar, the similar feature arrays are arranged in a descending order according to the size of the similarity; selecting arrays with similarity between [0,1] to merge two by two to form a set; and selecting a set center point as a new category center based on the combined set size;
Clustering is carried out between the new class center and the feature arrays which are not combined in a mode of carrying out similarity again until the new class center does not appear any more, and iteration is stopped;
s35, constructing independent indexes for each clustered class center and summarizing and storing all indexes;
s4, monitoring the intelligent security early warning area in real time through a series of installed monitoring equipment and a face recognition terminal;
S5, the face recognition terminal determines face image information according to the current shot image data and processes the face image information;
s6, after the face target is determined, comparing the extracted feature array with the feature arrays in the stored indexes by using a convolutional neural network;
and S7, judging whether an alarm needs to be given or not in real time according to the comparison result.
2. The intelligent security early warning management method based on big data according to claim 1, wherein the installed monitoring equipment monitors the real-time environment of the intelligent security early warning area, and comprises the following steps:
the installed monitoring device includes: the device comprises a smoke alarm, infrared detection equipment, a camera and an alarm device;
S21, setting a protection area in the intelligent security early warning area;
S22, opening a camera through infrared detection equipment when detecting that people come around a protection area, monitoring nearby people in real time through the camera, and generating an alarm signal and waking up an alarm device when detecting that illegal behaviors exist in the people;
s23, monitoring whether flame exists in the protection area in real time through infrared detection equipment; when detecting that flame exists in the protection area, the infrared detection equipment can generate an alarm signal and wake up the alarm device;
S24, monitoring whether smoke exists in the protection area in real time through an installed smoke alarm, and when the smoke exists in the protection area, generating an alarm signal and waking up an alarm device through the smoke alarm device to alarm.
3. The intelligent security early warning management method based on big data according to claim 1, wherein the transmitting the divided image data blocks to a convolutional neural network for feature extraction comprises:
after receiving an input image data block, the convolution layer moves on the input image data block according to a set step length through a convolution kernel and performs multiply accumulation on a corresponding area of each step and a characteristic value of the area, so that characteristic extraction of each image data block is realized;
The convolution calculation formula is as follows:
f(x)＝I(x)·w+b
wherein, I (x) represents an input characteristic value, w represents a weight value of a corresponding convolution kernel, and b represents a bias value;
for the extracted characteristics of the image data block, the characteristics of the extracted image data block are subjected to characteristic dimension reduction through a pooling layer, the important information in most input characteristics is reserved, the resolution of the input characteristics is reduced to the greatest extent for the subsequent convolution layer and the full connection layer, and the calculated amount of the model is reduced;
in the neural network, the output of the upper layer is used as the input of the lower layer, and the convolutional neural network is formed by continuously stacking;
the data must be subjected to the process of activating the function during the process of inputting the data to the lower layer;
Assuming that the input values are x i (i=1, 2,.. N), each input value x i corresponds to an input weight value w i, b is an offset, and the output result obtained after the input values are input to the convolutional neural network is:
Wherein g is a corresponding activation function;
Data is transmitted to the full connection layer after convolution and pooling which are continuously stacked;
And unfolding and combining the characteristic data through the full connection layer to obtain the characteristic data.
4. The intelligent security early warning management method based on big data according to claim 1, wherein the real-time monitoring of the intelligent security early warning area through the installed series of monitoring devices and face recognition terminals comprises the following steps:
s41, numbering q monitoring points q 1,q2,...qq based on the set monitoring points;
S42, taking a monitoring point q 1 as a coordinate origin, establishing a three-dimensional control coordinate system according to the q 1 coordinate, and determining the coordinates of other monitoring points based on the monitoring point q 1;
s43, storing the coordinates of each monitoring point in a database of the system;
S44, the face recognition device installed at each monitoring point shoots face images of the person at the current position in real time.
5. The intelligent security early warning management method based on big data according to claim 1, wherein the face recognition terminal determines face image information according to the currently photographed image data, and processes the face image information, comprising the following steps:
S51, setting the number of frames shot by a face recognition terminal, and shooting a plurality of frames of face target images;
S52, transmitting the shot face target image to a trained convolutional neural network, identifying the face target, storing characteristic points, and obtaining the approximate coordinate position of the face target in one frame of image;
And S53, extracting the characteristics of another frame of image, obtaining a new coordinate position, comparing the new coordinate position with the characteristic points of the previous frame, and judging whether the image is the same face target or not according to the fact that the coincidence degree is higher than 80%.
6. The intelligent security early warning management method based on big data according to claim 1, wherein the comparison of the extracted feature array with the feature arrays in the stored indexes by the convolutional neural network after the face target is determined comprises the following steps:
S61, calling each stored index into a convolutional neural network;
the step of calling the stored indexes into the convolutional neural network comprises the following steps:
s611, performing preliminary matching by calculating the similarity between the extracted feature array and each index;
S612, after the matching is successful, calling each characteristic data in the corresponding index into a convolutional neural network based on the matching result;
s62, comparing the extracted feature array with the feature arrays of all people in the intelligent security early warning area in all the stored indexes;
S63, comparing the extracted feature array with the feature arrays of all dangerous persons in the alarm information in all the stored indexes.
7. The intelligent security early warning management method based on big data according to claim 6, wherein the real-time judging whether the warning is needed according to the comparison result comprises:
If the matching is successful, verifying that the intelligent security early warning area is indicated to be a worker, performing accurate matching based on the index passing the matching verification, and opening a gate to release; if the intelligent security early warning areas are not matched, verification fails, and the intelligent security early warning areas are not workers; step S63 is carried out, secondary matching is carried out, if the matching is carried out, verification is passed, dangerous personnel in the alarm information is indicated, and the face recognition terminal can report the alarm information; if the information is not matched with the alarm information, the verification failure indicates that the information is not a dangerous person in the alarm information, the gate is opened to release, and the management personnel is reminded.
8. A big data based intelligent security early warning management system for implementing the big data based intelligent security early warning management method of any one of claims 1-7, characterized by comprising a smoke alarm, an infrared detection device, a camera and an alarm device;
The infrared detection equipment is used for monitoring surrounding conditions in real time, and starting a camera according to the surrounding conditions;
The camera is used for shooting surrounding condition information in real time;
The alarm device is used for receiving the alarm signal in real time, generating alarm information after receiving the alarm signal and reporting the alarm information through the Internet;
The smoke alarm is used for monitoring whether smoke exists in the protection area in real time and generating an alarm signal according to the smoke condition.
9. The intelligent security early warning management system based on big data according to claim 8, further comprising a face recognition terminal, wherein the face recognition terminal is used for scanning and comparing face image information and displaying the compared result.
CN202410062491.6A 2024-01-16 2024-01-16 Intelligent security early warning management method and system based on big data Active CN117671887B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410062491.6A CN117671887B (en) 2024-01-16 2024-01-16 Intelligent security early warning management method and system based on big data

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410062491.6A CN117671887B (en) 2024-01-16 2024-01-16 Intelligent security early warning management method and system based on big data

Publications (2)

Publication Number Publication Date
CN117671887A CN117671887A (en) 2024-03-08
CN117671887B true CN117671887B (en) 2024-09-17

Family

ID=90073458

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410062491.6A Active CN117671887B (en) 2024-01-16 2024-01-16 Intelligent security early warning management method and system based on big data

Country Status (1)

Country Link
CN (1) CN117671887B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118230399B (en) * 2024-05-24 2024-08-02 湖南警察学院 A face recognition system
CN118447582B (en) * 2024-07-08 2024-09-24 深圳市迪沃视讯数字技术有限公司 Security protection early warning system based on behavior pattern recognition

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109376596A (en) * 2018-09-14 2019-02-22 广州杰赛科技股份有限公司 Face matching process, device, equipment and storage medium
CN110263703A (en) * 2019-06-18 2019-09-20 腾讯科技（深圳）有限公司 Personnel's flow statistical method, device and computer equipment

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8036425B2 (en) * 2008-06-26 2011-10-11 Billy Hou Neural network-controlled automatic tracking and recognizing system and method
CN107679078B (en) * 2017-08-29 2020-01-10 银江股份有限公司 Bayonet image vehicle rapid retrieval method and system based on deep learning
CN110276320A (en) * 2019-06-26 2019-09-24 杭州创匠信息科技有限公司 Face recognition based access control method, device, equipment and storage medium
CN113673345B (en) * 2021-07-20 2024-04-02 中国铁道科学研究院集团有限公司电子计算技术研究所 Face recognition method and system
CN113673495B (en) * 2021-10-25 2022-02-01 北京通建泰利特智能系统工程技术有限公司 Intelligent security method and system based on neural network and readable storage medium

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109376596A (en) * 2018-09-14 2019-02-22 广州杰赛科技股份有限公司 Face matching process, device, equipment and storage medium
CN110263703A (en) * 2019-06-18 2019-09-20 腾讯科技（深圳）有限公司 Personnel's flow statistical method, device and computer equipment

Also Published As

Publication number Publication date
CN117671887A (en) 2024-03-08

Similar Documents

Publication Publication Date Title
CN112347916B (en) Video image analysis-based power field operation safety monitoring method and device
CN117671887B (en) Intelligent security early warning management method and system based on big data
CN112183265A (en) Video monitoring and alarm method and system for power construction based on image recognition
CN112560745B (en) Method for discriminating personnel on electric power operation site and related device
CN114677640A (en) Intelligent construction site safety monitoring system and method based on machine vision
CN110766894A (en) Community fence crossing early warning method, system, server and computer storage medium
CN111476979A (en) Intelligent security and stability maintenance method and system based on multi-model analysis
CN117523437A (en) Real-time risk identification method for substation near-electricity operation site
CN118736743A (en) Smart manhole cover anti-theft monitoring system
CN119808594A (en) A dynamic digital emergency plan simulation and analysis system for multiple scenarios
CN108269021A (en) Risk monitoring system and method are merged based on NB-IoT electric network informations physics
CN115049875A (en) Detection method for wearing insulating gloves in transformer substation based on deep learning
CN116311729A (en) AI-based on-site safety management system for power grid infrastructure
CN112435240B (en) Deep vision mobile phone detection system for workers to illegally use mobile phones
CN119131691A (en) Campus smart security method based on multi-source data analysis
CN116682162B (en) A robot personnel detection method based on real-time video stream
CN115346170B (en) Intelligent monitoring method and device for gas facility area
CN116052035A (en) Power plant personnel perimeter intrusion detection method based on convolutional neural network
CN118450087B (en) A smart security monitoring method and system
CN114821473A (en) A method and system for monitoring loitering behavior of people
Karnati et al. Data Mining Algorithms for Traffic Interruption Detection.
CN119854343B (en) Battery antitheft system of battery changing cabinet based on artificial intelligence
CN119091195B (en) Protection network abnormality detection system and method
CN116950029B (en) Early warning system based on foundation pit detection
Lin et al. An integrated application platform for safety management and control of smart power plants based on video image technology

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