CN118247915A - Monitoring image processing method and system based on computer vision - Google Patents

Abstract

The invention is suitable for the technical field of monitoring image processing, and provides a monitoring image processing method and system based on computer vision, wherein the method comprises the following steps: acquiring monitoring data in real time, wherein the monitoring data comprises a monitoring picture, monitoring time and a monitoring number; comparing the first monitoring data with a preset data set, obtaining first key monitoring data through comparison, and marking the first key monitoring data; acquiring monitoring peripheral monitoring data of the observed first key monitoring data, comparing the peripheral monitoring data with a preset data set, and comparing to obtain second key monitoring data; the first important monitoring data and the second important monitoring data are preprocessed, and warning operation is carried out according to the first important monitoring data and the second important monitoring data, so that the problems that the expansion of a fire range and the moving track of dangerous molecules cannot be obtained through analysis, quick and accurate response cannot be realized, and the safety management efficiency and the safety management effectiveness are low are solved.

Description

Monitoring image processing method and system based on computer vision

Technical Field

The invention belongs to the technical field of monitoring image processing, and particularly relates to a monitoring image processing method and system based on computer vision.

Background

In the current security monitoring field, one of the core functions of the monitoring system is to monitor public or private areas in real time to prevent and respond to various security accidents and emergency situations, and the traditional monitoring system mainly relies on manual monitoring or simple action detection technology, and these methods often cannot provide efficient and accurate response under complex situations. With the rapid development of computer vision and machine learning technology, the field of monitoring image processing technology introduces computer vision to improve the automation level and response capability of a monitoring system.

Existing computer vision techniques have been applied to monitoring systems to improve the ability to automatically detect abnormal events. These systems automatically identify specific events or abnormal behavior in the video using image recognition and pattern analysis techniques, such as fire and smoke detection algorithms, can identify flame and smoke patterns in the image.

However, the prior art has the problems that the expansion of the fire range and the moving track of dangerous molecules cannot be analyzed and obtained, quick and accurate response cannot be realized, and the safety management efficiency and the effectiveness are not high.

Disclosure of Invention

An object of an embodiment of the present invention is to provide a method for processing a monitoring image based on computer vision, which aims to solve the problem set forth in the third section of the background art.

The embodiment of the invention is realized in such a way that a monitoring image processing method and a system based on computer vision are provided, wherein the method comprises the following steps:

Acquiring monitoring data in real time, wherein the monitoring data comprises a monitoring picture, monitoring time and a monitoring number;

comparing the first monitoring data with a preset data set, wherein the preset data set comprises smoke, flame and dangerous molecular face data, obtaining first important monitoring data through comparison, and marking the first important monitoring data;

acquiring first important monitoring data, monitoring peripheral monitoring data, comparing the peripheral monitoring data with a preset data set, and comparing to obtain second important monitoring data;

And preprocessing the first important monitoring data and the second important monitoring data, wherein the preprocessing comprises the steps of adjusting the size of an image, adjusting the definition of the image and removing noise, and executing warning operation according to the first important monitoring data and the second important monitoring data.

Preferably, the step of comparing the first monitored data with a preset data set, where the preset data set includes smoke, flame and dangerous molecule face data, and obtaining first key monitored data through comparison, and performing marking processing on the first key monitored data specifically includes:

comparing the first monitoring data with a preset data set, comparing smoke and flame through an image processing technology, and counting and comparing dangerous molecular faces through an image segmentation and object recognition technology;

comparing the first similarity with a threshold value, and judging the first key monitoring data if the comparison result exceeds the threshold value;

and marking the first key monitoring data, sending the first key monitoring data to the monitoring terminal, and receiving the confirmation information of the monitoring terminal.

Preferably, the step of obtaining the first important monitoring data to monitor the peripheral monitoring data, comparing the peripheral monitoring data with a preset data set, and comparing to obtain the second important monitoring data specifically includes:

Acquiring a first monitoring number of first key monitoring data, and calling a peripheral monitoring number by adopting a breadth-first search algorithm according to the first monitoring number, wherein the breadth-first search algorithm accesses all neighbors of a starting node and then moves to the neighbors of the neighbors;

Obtaining second monitoring data corresponding to the peripheral monitoring number, comparing the second monitoring data with a preset data set, obtaining second similarity through comparison, comparing the second similarity with a threshold value, and judging the second monitored data as second point monitoring data if the comparison result exceeds the threshold value;

And marking the second multipoint monitoring data, sending the second multipoint monitoring data to the monitoring terminal, and receiving the confirmation information of the monitoring terminal.

Preferably, the preprocessing the first key monitoring data and the second key monitoring data includes adjusting an image size, adjusting an image definition, removing noise, and executing a warning operation according to the first key monitoring data and the second key monitoring data, and specifically includes:

Preprocessing the first important monitoring data and the second important monitoring data to obtain preprocessed third monitoring data;

Performing remote storage backup on the third monitoring data, generating an authorization code and sending the authorization code to a monitoring terminal;

and executing warning operation according to the first important monitoring data and the second important monitoring data.

Preferably, the monitoring terminal adopts a computer display screen.

It is another object of an embodiment of the present invention to provide a monitoring image processing system based on computer vision, the system including:

the monitoring acquisition module acquires monitoring data in real time, wherein the monitoring data comprises a monitoring picture, monitoring time and a monitoring number;

The first key monitoring data module is used for comparing the first monitoring data with a preset data set, wherein the preset data set comprises smog, flame and dangerous molecular face data, the first key monitoring data is obtained through comparison, and the first key monitoring data is marked;

the second heavy point monitoring data module is used for acquiring first heavy point monitoring data, monitoring peripheral monitoring data, comparing the peripheral monitoring data with a preset data set and obtaining second heavy point monitoring data;

The warning module is used for preprocessing the first important monitoring data and the second important monitoring data, wherein the preprocessing comprises the steps of adjusting the size of an image, adjusting the definition of the image and removing noise, and warning operation is executed according to the first important monitoring data and the second important monitoring data.

Preferably, the first key monitoring data module includes:

The comparison unit is used for comparing the first monitoring data with a preset data set, comparing smoke and flame through an image processing technology, and counting and comparing dangerous molecular faces through an image segmentation and object recognition technology;

the first similarity unit is used for obtaining first similarity through comparison, comparing the first similarity with a threshold value, and judging the first important monitoring data if the comparison result exceeds the threshold value;

The first marking unit is used for marking the first key monitoring data, sending the first key monitoring data to the monitoring terminal and receiving the confirmation information of the monitoring terminal.

Preferably, the second multipoint monitoring data module includes:

The method comprises the steps of calling a peripheral monitoring unit, obtaining a first monitoring number of first key monitoring data, calling the peripheral monitoring number according to the first monitoring number by adopting a breadth-first search algorithm, accessing all neighbors of a starting node by the breadth-first search algorithm, and moving to the neighbors of the neighbors;

the second similarity unit is used for obtaining second monitoring data corresponding to the peripheral monitoring number, comparing the second monitoring data with a preset data set, obtaining second similarity through comparison, comparing the second similarity with a threshold value, and judging the second monitored data as second point monitoring data if the comparison result exceeds the threshold value;

And the second marking unit is used for marking the second multipoint monitoring data, sending the second multipoint monitoring data to the monitoring terminal and receiving the confirmation information of the monitoring terminal.

Preferably, the warning module includes:

The third monitoring data unit is used for preprocessing the first important monitoring data and the second important monitoring data to obtain preprocessed third monitoring data;

The backup unit is used for carrying out remote storage backup on the third monitoring data, generating an authorization code and sending the authorization code to the monitoring terminal;

and the warning unit is used for executing warning operation according to the first important monitoring data and the second important monitoring data.

Preferably, the monitoring terminal adopts a computer display screen.

According to the monitoring image processing method based on computer vision, monitoring data are obtained in real time, monitoring pictures, monitoring time and monitoring numbers are known through the monitoring data, occurrence can be judged according to the monitoring pictures, occurrence time is judged according to the monitoring time, occurrence places of events are judged according to the monitoring numbers, whether fire, smoke and dangerous molecules occur in an office building can be identified by comparing the first monitoring data with a preset data set, the monitoring data for finding the fire, the smoke and the dangerous molecules are marked, a manager can conveniently inquire in time, meanwhile, when the fire, the smoke and the dangerous molecules are found, the first important monitoring data are obtained for monitoring the peripheral monitoring data, the second important monitoring data are obtained through inquiring the peripheral monitoring data, the spreading condition of the fire and the smoke can be known, the moving track of the dangerous molecules is conveniently inquired and processed in time after warning, the sizes, the image definition and the noise removal of the first important monitoring data and the second important monitoring data are adjusted, the problem that the moving track of the fire range and the dangerous molecules cannot be obtained through analysis is solved, and the quick response, the safety and the safety management cannot be realized effectively is achieved.

Drawings

Fig. 1 is a flowchart of a method for processing a monitoring image based on computer vision according to an embodiment of the present invention;

FIG. 2 is a flowchart showing steps of comparing first monitoring data with a preset data set, wherein the preset data set includes smoke, flame and dangerous molecular face data, obtaining first important monitoring data through comparison, and performing marking processing on the first important monitoring data;

FIG. 3 is a flowchart illustrating steps for obtaining first important monitoring data, monitoring peripheral monitoring data, comparing the peripheral monitoring data with a preset data set, and comparing the peripheral monitoring data with a preset data set to obtain second important monitoring data according to an embodiment of the present invention;

FIG. 4 is a flowchart illustrating steps for preprocessing first and second important monitoring data, including adjusting an image size, adjusting an image sharpness, and removing noise, according to the first and second important monitoring data, and executing a warning operation according to the first and second important monitoring data;

FIG. 5 is a block diagram of a monitoring image processing system based on computer vision according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of a first data module for monitoring data according to an embodiment of the present invention;

FIG. 7 is a block diagram of a second embodiment of the present invention;

fig. 8 is a schematic diagram of an alarm module according to an embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

It will be understood that the terms "first," "second," and the like, as used herein, may be used to describe various elements, but these elements are not limited by these terms unless otherwise specified. These terms are only used to distinguish one element from another element. For example, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script, without departing from the scope of this disclosure.

As shown in fig. 1, a method for processing a monitoring image based on computer vision according to an embodiment of the present invention includes:

And S100, acquiring monitoring data in real time, wherein the monitoring data comprises a monitoring picture, monitoring time and a monitoring number.

In this step, the monitoring data is obtained in real time, and the monitoring picture refers to a real-time image or video captured by the monitoring camera. Its function is to provide visual information to the monitored area, enabling a monitoring person or system to observe and analyze the monitored environment. The monitoring picture can be used for detecting abnormal activities, confirming event occurrence, supervising safety conditions and the like;

The monitoring time refers to a time stamp or a time range in which monitoring data is recorded. Its functions are determining the time of event occurrence, building event timelines, data analysis and backtracking. Monitoring time can help determine the time period that a particular event occurs in order to quickly locate problems or trace back the event's occurrence;

The monitoring number refers to a unique identifier of the monitoring device or monitoring location. Its function is to distinguish between different monitoring sources or locations and to classify and manage the monitoring data. The monitoring number can help a manager or a system to archive, index and retrieve the monitoring data so as to facilitate subsequent searching and analysis, and can timely locate the position when the situation occurs.

Taking a meeting room in an office building as an example, assuming that a monitoring camera is installed in the meeting room, and recording monitoring data in a period of time;

monitoring a picture: the monitoring picture refers to a real-time image or video captured by the camera. In the conference room, the monitoring screen may include: the layout of conference tables and chairs, the activities and communication of participants, and the entrance and exit conditions of conference rooms, such as the entrance and exit conditions of people.

Monitoring time: the monitoring time refers to a time range in which monitoring data is recorded. In conference room monitoring, the monitoring time may include: day 19, 4, 2024, 9:00 am to 10:00 am, all conference activities during this time, including conference start, participant entry and exit, conference in progress, etc.

Monitoring number: the monitoring number refers to a unique identifier of the monitoring device or monitoring location. In conference room monitoring, the monitoring number may be: CONF-ROOM1-CAM1 (conference ROOM1 camera 1), this number can help a manager or system to quickly locate a monitor screen in the conference ROOM.

S200, comparing the first monitoring data with a preset data set, wherein the preset data set comprises smoke, flame and dangerous molecular face data, obtaining first key monitoring data through comparison, and marking the first key monitoring data.

In this step, the first monitoring data is compared with a preset data set, which comprises smoke, flame and dangerous molecular face data, which are usually of a specific color, so that possible smoke or flame areas can be detected by color analysis. Color thresholding techniques or color space transformations may be employed to extract potential fire areas, such as the HSV color space, which is a model for describing colors consisting of three parameters, hue (Hue), saturation (Saturation) and brightness (Value), which can intuitively describe the properties of the colors, and by adjusting the values of the three parameters, the kinds, purity and darkness of the colors can be easily changed;

the texture and shape of the flame and smoke also have certain characteristics, and the fire area can be further identified through texture analysis and shape detection; the image captured by the monitoring camera is subjected to face detection processing, so that a face region appearing in the monitoring image can be identified, and the identification system can identify the detected face. The system extracts the characteristics of each face and compares the characteristics with the face data of staff recorded in advance, meanwhile, dangerous molecules of police office can be accessed, and if the dangerous molecules are found, warning processing can be carried out;

The fire, smoke and dangerous molecules are found through comparison, the monitoring data for finding the fire, smoke and dangerous molecules are defined as first important monitoring data, the first important monitoring data are marked, and the monitoring time and the monitoring number corresponding to the first important monitoring data are obtained, so that quick and accurate response can be realized, and the efficiency and the effectiveness of safety management are improved.

S300, acquiring first important monitoring data, monitoring peripheral monitoring data, comparing the peripheral monitoring data with a preset data set, and comparing to obtain second important monitoring data.

In this step, the monitoring peripheral monitoring data of the first key monitoring data is obtained, for example, when the point A monitors an emergency, monitoring of the adjacent two sides B and C is mobilized through a program, the scope of the emergency is queried, if the emergency is queried at the point B, the monitoring of the adjacent two sides of the point B is mobilized again if the emergency is not queried at the point C, the monitoring data of the point B is compared with a preset data set, and the like;

The fire, the smoke and the dangerous molecules are found through comparison, the monitoring data for finding the fire, the smoke and the dangerous molecules are defined as second point monitoring data, and the monitoring time and the monitoring number corresponding to the second point monitoring data are obtained, so that quick and accurate response can be realized, and the efficiency and the effectiveness of safety management are improved.

S400, preprocessing the first important monitoring data and the second important monitoring data, wherein the preprocessing comprises the steps of adjusting the size of an image, adjusting the definition of the image and removing noise, and executing warning operation according to the first important monitoring data and the second important monitoring data.

In this step, the first key monitoring data and the second key monitoring data are preprocessed, the size of the image is adjusted, the definition of the image is adjusted, noise is removed, the visual effect of the image is improved, the image is more suitable for display or further analysis, the monitoring content and the definition of sound in monitoring can be more clearly observed, the warning operation is executed according to the first key monitoring data and the second key monitoring data, the display can be performed at the monitoring terminal, meanwhile, direct warning can be performed, the safety is improved, the requirement of manual monitoring can be reduced through automation, and the efficiency and the response speed are improved.

As shown in fig. 2, as a preferred embodiment of the present invention, the step of comparing the first monitored data with a preset data set, where the preset data set includes smoke, flame and dangerous molecule face data, obtaining first important monitored data through comparison, and performing marking processing on the first important monitored data specifically includes:

S201, comparing the first monitoring data with a preset data set, comparing smoke and flame through an image processing technology, and counting and comparing dangerous molecular faces through an image segmentation and object recognition technology.

In this step, the first monitoring data is compared with a preset data set, which comprises smoke, flame and dangerous molecular face data, which are usually of a specific color, so that possible smoke or flame areas can be detected by color analysis. Color thresholding techniques or color space transformations may be employed to extract potential fire areas, such as the HSV color space, which is a model for describing colors consisting of three parameters, hue (Hue), saturation (Saturation) and brightness (Value), which can intuitively describe the properties of the colors, and by adjusting the values of the three parameters, the kinds, purity and darkness of the colors can be easily changed;

The texture and shape of the flame and smoke also have certain characteristics, and the fire area can be further identified through texture analysis and shape detection; the image captured by the monitoring camera is subjected to face detection processing, so that a face region appearing in the monitoring image can be identified, and the identification system can identify the detected face. The system extracts the characteristics of each face and compares the characteristics with the face data of staff recorded in advance, meanwhile, dangerous molecules of police authorities can be accessed, and if the dangerous molecules are found, warning processing can be carried out.

S202, comparing the first similarity with a threshold value, and judging the first key monitoring data if the comparison result exceeds the threshold value.

In the step, the first similarity is obtained through comparison, the first similarity is compared with a threshold value, the threshold value of smoke and flame can be set to be 30%, the threshold value of dangerous molecular face data is set to be 50%, if the comparison result of the smoke and the flame exceeds the threshold value, the dangerous molecular face data is judged to be first important monitoring data, and if the comparison result of the smoke and the flame is lower than the threshold value, the dangerous molecular face data is higher than the threshold value, the dangerous molecular face data is not judged to be the first important monitoring data;

For example, in a monitoring system of a office building, a monitoring camera captures a scene of smoke rolling in a hall. The system analyzes pictures in the video stream, detects a smoke area in the hall, and finds that the similarity is 70% after comparing the smoke area with a preset threshold value, and the smoke characteristics are consistent with the fire judgment standard after exceeding the threshold value. The system then triggers an alarm informing the firefighter to go to dispose of the fire. Meanwhile, the system also detects a stranger in the hall, and recognizes that the person is not in the face data of the staff through the face recognition technology, and the similarity is lower than a set threshold value, so that the person can be determined to be a dangerous molecule, and the system can correspondingly trigger an alarm to inform security personnel to take measures.

And S203, marking the first key monitoring data, sending the first key monitoring data to the monitoring terminal, and receiving the confirmation information of the monitoring terminal.

In the step, the first key monitoring data are marked, fire, smoke and dangerous molecules are found through comparison, the monitoring data for finding fire, smoke and dangerous molecules are defined as the first key monitoring data, the first key monitoring data are marked, and the monitoring time and the monitoring number corresponding to the first key monitoring data are obtained, so that quick and accurate response can be realized, and the efficiency and the effectiveness of safety management are improved.

As shown in fig. 3, as a preferred embodiment of the present invention, the step of obtaining the first important monitored data to monitor the peripheral monitored data, comparing the peripheral monitored data with a preset data set, and obtaining the second important monitored data specifically includes:

S301, acquiring a first monitoring number of first key monitoring data, and calling a peripheral monitoring number by adopting a breadth-first search algorithm according to the first monitoring number, wherein the breadth-first search algorithm accesses all neighbors of the starting node and then moves to the neighbors of the neighbors.

In the step, a first monitoring number of first key monitoring data is obtained, and a breadth-first search algorithm is adopted to call a peripheral monitoring number according to the first monitoring number, for example, the peripheral monitoring number starts from a point A where an emergency occurs;

Check the neighbors of point a (B and C): if the side B detects an emergency and the side C does not, continuing to check adjacent points on the side B;

if side C is detected and side B is not, side C is treated similarly.

If both sides are detected or neither are detected, the strategy is stopped or adjusted according to the requirements.

Repeating step 2 for each new point at which an emergency is detected until no new emergency is found;

Assuming that we have a floor monitoring network, the cameras are arranged at a series of points (for example 1 to 10) and each point is connected only to its immediate neighbors;

def find_incidents(start, adjacency_list):

Use of queues to support breadth-first search

from collections import deque

queue = deque([start])

Incidents _found=set () # stores the point at which the emergency was found

while queue:

current = queue.popleft()

if current in incidents_found:

continue

Suppose we can call a function to detect whether there is an emergency

if has_incident(current):

incidents_found.add(current)

# Add neighbor to queue

for neighbor in adjacency_list[current]:

queue.append(neighbor)

return incidents_found

def has_incident(point):

# This is a simulation function, which in practice requires access to the API of the monitoring system

# Assume points 3, 4, 5 have an emergency

return point in {3, 4, 5}

# Build a simple neighbor list representation

adjacency_list = {

1: [2],

2: [1, 3],

3: [2, 4],

4: [3, 5],

5: [4, 6],

6: [5, 7],

7: [6, 8],

8: [7, 9],

9: [8, 10],

10: [9]

}

# Detection starting from Point 3

incident_points = find_incidents(3, adjacency_list)

Print ('monitored emergency points:', incident _points)

S302, second monitoring data corresponding to the peripheral monitoring number are obtained, the second monitoring data and a preset data set are compared, second similarity is obtained through comparison, the second similarity is compared with a threshold value, and if the comparison result exceeds the threshold value, the second multipoint monitoring data is judged.

In this step, the second monitoring data corresponding to the peripheral monitoring number is obtained, the first monitoring number corresponding to the first monitoring data is obtained, the second monitoring data corresponding to the second monitoring number is obtained according to the first monitoring number, the second monitoring data corresponding to the second monitoring number is compared with a preset data set, the second similarity is obtained through comparison, the threshold value of smoke and flame can be set to be 30%, the threshold value of dangerous molecular face data is set to be 50%, if the comparison result of smoke and flame exceeds the threshold value, the dangerous molecular face data is judged to be the second heavy point monitoring data, and if the comparison result of smoke and flame is lower than the threshold value, the dangerous molecular face data is higher than the threshold value, the dangerous molecular face data is not judged to be the second heavy point monitoring data;

And (3) the monitoring numbers inquired by the peripheral monitoring numbers are called through the comparison priority search algorithm, when the comparison result of the smoke and the flame exceeds the threshold value, the dangerous molecular face data is inquired, if the comparison result of the smoke and the flame is lower than the threshold value, the inquiry is carried out, and if the comparison result of the smoke and the flame is higher than the threshold value, the inquiry is stopped.

S303, marking the second multipoint monitoring data, sending the second multipoint monitoring data to the monitoring terminal, and receiving the confirmation information of the monitoring terminal.

In the step, the second point monitoring data is marked, fire, smoke and dangerous molecules are found through comparison, the monitoring data for finding fire, smoke and dangerous molecules are defined as the second point monitoring data, the second point monitoring data is marked, and the monitoring time and the monitoring number corresponding to the second point monitoring data are obtained, so that quick and accurate response can be realized, the efficiency and the effectiveness of safety management are improved, and the occurrence and spreading range of the fire and the moving track of the dangerous molecules can be judged.

As shown in fig. 4, as a preferred embodiment of the present invention, the preprocessing of the first critical monitoring data and the second critical monitoring data includes adjusting the size of an image, adjusting the sharpness of the image, and removing noise, and executing the warning operation according to the first critical monitoring data and the second critical monitoring data, which specifically includes:

s401, preprocessing the first important monitoring data and the second important monitoring data to obtain preprocessed third monitoring data.

In the step, the first important monitoring data and the second important monitoring data are preprocessed, the size of the image is adjusted, the definition of the image is adjusted, noise is removed, the visual effect of the image is improved, the image is more suitable for display or further analysis, the monitoring content and the definition of sound in monitoring can be more clearly observed, and the third monitoring data are obtained after the processing is completed.

S402, carrying out remote storage backup on the third monitoring data, generating an authorization code and sending the authorization code to the monitoring terminal.

In this step, the third monitoring data is remotely stored and backed up, and the video data is encrypted and securely transmitted to the cloud server using a secure transmission protocol (such as TLS). The safety and the integrity of the data in the transmission process are ensured, the situation that the monitoring data is lost due to fire and the like can be prevented, the fire cause can not be inquired, and the workload of inquiring the fire starting point is reduced;

After the authorization code is uploaded to the cloud server, the authorization code is generated and sent to the monitoring terminal, and a manager can access the cloud server through the authorization code, so that the safety and privacy of monitoring data are guaranteed.

S403, executing warning operation according to the first important monitoring data and the second important monitoring data.

In the step, the warning operation is executed according to the first important monitoring data and the second important monitoring data, and the warning operation is executed according to the first important monitoring data and the second important monitoring data, so that the warning device can display at the monitoring terminal, can directly give an alarm, improves the safety, and can reduce the requirement of manual monitoring through automation, and improve the efficiency and the response speed.

As shown in fig. 5, a data processing system based on a server architecture according to an embodiment of the present invention,

The monitoring acquisition module 100 is configured to acquire monitoring data in real time, where the monitoring data includes a monitoring screen, a monitoring time, and a monitoring number.

In the system, the monitoring acquisition module 100 acquires monitoring data in real time, and the monitoring picture refers to a real-time image or video captured by a monitoring camera. Its function is to provide visual information to the monitored area, enabling a monitoring person or system to observe and analyze the monitored environment. The monitoring picture can be used for detecting abnormal activities, confirming event occurrence, supervising safety conditions and the like;

The monitoring time refers to a time stamp or a time range in which monitoring data is recorded. Its functions are determining the time of event occurrence, building event timelines, data analysis and backtracking. Monitoring time can help determine the time period that a particular event occurs in order to quickly locate problems or trace back the event's occurrence;

The monitoring number refers to a unique identifier of the monitoring device or monitoring location. Its function is to distinguish between different monitoring sources or locations and to classify and manage the monitoring data. The monitoring number can help a manager or a system to archive, index and retrieve the monitoring data so as to facilitate subsequent searching and analysis, and can timely locate the position when the situation occurs.

Taking a meeting room in an office building as an example, assuming that a monitoring camera is installed in the meeting room, and recording monitoring data in a period of time;

monitoring a picture: the monitoring picture refers to a real-time image or video captured by the camera. In the conference room, the monitoring screen may include: the layout of conference tables and chairs, the activities and communication of participants, and the entrance and exit conditions of conference rooms, such as the entrance and exit conditions of people.

Monitoring time: the monitoring time refers to a time range in which monitoring data is recorded. In conference room monitoring, the monitoring time may include: day 19, 4, 2024, 9:00 am to 10:00 am, all conference activities during this time, including conference start, participant entry and exit, conference in progress, etc.

Monitoring number: the monitoring number refers to a unique identifier of the monitoring device or monitoring location. In conference room monitoring, the monitoring number may be: CONF-ROOM1-CAM1 (conference ROOM1 camera 1), this number can help a manager or system to quickly locate a monitor screen in the conference ROOM.

The first key monitoring data module 200 is configured to compare the first monitoring data with a preset data set, where the preset data set includes smoke, flame, and dangerous molecular face data, obtain the first key monitoring data through comparison, and perform a marking process on the first key monitoring data.

In the present system, the first critical monitoring data module 200 compares the first monitoring data with a preset data set, and creates a preset data set including smoke, flame and dangerous molecular facial data, which are typically of a specific color, so that possible smoke or flame areas can be detected by color analysis. Color thresholding techniques or color space transformations may be employed to extract potential fire areas, such as the HSV color space, which is a model for describing colors consisting of three parameters, hue (Hue), saturation (Saturation) and brightness (Value), which can intuitively describe the properties of the colors, and by adjusting the values of the three parameters, the kinds, purity and darkness of the colors can be easily changed;

the texture and shape of the flame and smoke also have certain characteristics, and the fire area can be further identified through texture analysis and shape detection; the image captured by the monitoring camera is subjected to face detection processing, so that a face region appearing in the monitoring image can be identified, and the identification system can identify the detected face. The system extracts the characteristics of each face and compares the characteristics with the face data of staff recorded in advance, meanwhile, dangerous molecules of police office can be accessed, and if the dangerous molecules are found, warning processing can be carried out;

The fire, smoke and dangerous molecules are found through comparison, the monitoring data for finding the fire, smoke and dangerous molecules are defined as first important monitoring data, the first important monitoring data are marked, and the monitoring time and the monitoring number corresponding to the first important monitoring data are obtained, so that quick and accurate response can be realized, and the efficiency and the effectiveness of safety management are improved.

The second multipoint monitoring data module 300 is configured to obtain first point monitoring data, monitor surrounding monitoring data, compare the surrounding monitoring data with a preset data set, and obtain second point monitoring data.

In the system, the second point monitoring data module 300 acquires the monitoring peripheral monitoring data of the first point monitoring data, for example, when the point A monitors an emergency, monitors the adjacent two sides B and C through a program, inquires the scope of the emergency, if the point B inquires the emergency, the point C does not inquire the emergency, then monitors the adjacent two sides of the point B, compares the point B monitoring data with a preset data set, and the like;

The fire, the smoke and the dangerous molecules are found through comparison, the monitoring data for finding the fire, the smoke and the dangerous molecules are defined as second point monitoring data, and the monitoring time and the monitoring number corresponding to the second point monitoring data are obtained, so that quick and accurate response can be realized, and the efficiency and the effectiveness of safety management are improved.

The warning module 400 is configured to perform preprocessing on the first important monitoring data and the second important monitoring data, where the preprocessing includes adjusting an image size, adjusting an image definition, and removing noise, and perform a warning operation according to the first important monitoring data and the second important monitoring data.

In this system, warning module 400 carries out preliminary treatment to first important monitored data and second heavy point monitored data, adjust image size, adjust image definition and remove noise, improve the visual effect of image, make it more suitable for the demonstration or further analysis, can more clear observation control content and the interior sound definition of control, carry out the warning operation according to first important monitored data and second heavy point monitored data, can show at the monitor terminal, can directly report to the police simultaneously, the security has been improved, still can reduce the demand of manual monitoring through the automation, improve efficiency and response speed.

As shown in fig. 6, as a preferred embodiment of the present invention, the first critical monitoring data module 200 includes:

the comparison unit 201 is configured to compare the first monitoring data with a preset data set, compare smoke and flame through an image processing technology, and perform counting and dangerous molecule face comparison through an image segmentation and object recognition technology.

In this module, the comparison unit 201 compares the first monitoring data with a preset data set, creating a preset data set comprising smoke, flame and dangerous molecular face data, which are usually of a specific color, so that possible smoke or flame areas can be detected by color analysis. Color thresholding techniques or color space transformations may be employed to extract potential fire areas, such as the HSV color space, which is a model for describing colors consisting of three parameters, hue (Hue), saturation (Saturation) and brightness (Value), which can intuitively describe the properties of the colors, and by adjusting the values of the three parameters, the kinds, purity and darkness of the colors can be easily changed;

The texture and shape of the flame and smoke also have certain characteristics, and the fire area can be further identified through texture analysis and shape detection; the image captured by the monitoring camera is subjected to face detection processing, so that a face region appearing in the monitoring image can be identified, and the identification system can identify the detected face. The system extracts the characteristics of each face and compares the characteristics with the face data of staff recorded in advance, meanwhile, dangerous molecules of police authorities can be accessed, and if the dangerous molecules are found, warning processing can be carried out.

The first similarity unit 202 is configured to compare the first similarity with a threshold, and determine that the first monitored data is the first monitored data if the comparison result exceeds the threshold.

In this module, the first similarity unit 202 obtains a first similarity by comparing the first similarity with a threshold, and may set the threshold of smoke and flame to 30%, set the threshold of dangerous molecular face data to 50%, if the comparison result of smoke and flame exceeds the threshold, the dangerous molecular face data is determined to be the first important monitoring data, and if the comparison result of smoke and flame is below the threshold, the dangerous molecular face data is not determined to be the first important monitoring data;

For example, in a monitoring system of a office building, a monitoring camera captures a scene of smoke rolling in a hall. The system analyzes pictures in the video stream, detects a smoke area in the hall, and finds that the similarity is 70% after comparing the smoke area with a preset threshold value, and the smoke characteristics are consistent with the fire judgment standard after exceeding the threshold value. The system then triggers an alarm informing the firefighter to go to dispose of the fire. Meanwhile, the system also detects a stranger in the hall, and recognizes that the person is not in the face data of the staff through the face recognition technology, and the similarity is lower than a set threshold value, so that the person can be determined to be a dangerous molecule, and the system can correspondingly trigger an alarm to inform security personnel to take measures.

The first marking unit 203 is configured to perform marking processing on the first key monitoring data, send the first key monitoring data to the monitoring terminal, and receive acknowledgement information of the monitoring terminal.

In this module, the first marking unit 203 performs marking processing on the first key monitoring data, and compares the first key monitoring data to find fire, smoke and dangerous molecules, defines the monitoring data for finding fire, smoke and dangerous molecules as first key monitoring data, and performs marking processing on the first key monitoring data to obtain a monitoring time and a monitoring number corresponding to the first key monitoring data, thereby realizing rapid and accurate response and improving efficiency and effectiveness of safety management.

As shown in fig. 7, as a preferred embodiment of the present invention, the second re-point monitoring data module 300 includes:

the invoking peripheral monitoring unit 301 is configured to obtain a first monitoring number of the first key monitoring data, invoke the peripheral monitoring number according to the first monitoring number by using a breadth-first search algorithm, where the breadth-first search algorithm accesses all neighbors of the starting node and moves to the neighbors of the neighbors.

In this module, the peripheral monitoring unit 301 obtains a first monitoring number of the first key monitoring data, and a breadth-first search algorithm is adopted to invoke the peripheral monitoring number according to the first monitoring number, for example, from a point a where an emergency occurs;

Check the neighbors of point a (B and C): if the side B detects an emergency and the side C does not, continuing to check adjacent points on the side B;

if side C is detected and side B is not, side C is treated similarly.

If both sides are detected or neither are detected, the strategy is stopped or adjusted according to the requirements.

Repeating step 2 for each new point at which an emergency is detected until no new emergency is found;

Assuming that we have a floor monitoring network, the cameras are arranged at a series of points (for example 1 to 10) and each point is connected only to its immediate neighbors;

def find_incidents(start, adjacency_list):

Use of queues to support breadth-first search

from collections import deque

queue = deque([start])

Incidents _found=set () # stores the point at which the emergency was found

while queue:

current = queue.popleft()

if current in incidents_found:

continue

Suppose we can call a function to detect whether there is an emergency

if has_incident(current):

incidents_found.add(current)

# Add neighbor to queue

for neighbor in adjacency_list[current]:

queue.append(neighbor)

return incidents_found

def has_incident(point):

# This is a simulation function, which in practice requires access to the API of the monitoring system

# Assume points 3, 4, 5 have an emergency

return point in {3, 4, 5}

# Build a simple neighbor list representation

adjacency_list = {

1: [2],

2: [1, 3],

3: [2, 4],

4: [3, 5],

5: [4, 6],

6: [5, 7],

7: [6, 8],

8: [7, 9],

9: [8, 10],

10: [9]

}

# Detection starting from Point 3

incident_points = find_incidents(3, adjacency_list)

Print ('monitored emergency points:', incident _points)

The second similarity unit 302 is configured to obtain second monitoring data corresponding to the peripheral monitoring number, compare the second monitoring data with a preset data set, obtain a second similarity through comparison, compare the second similarity with a threshold, and determine that the second monitored data is the second point monitoring data if the comparison result exceeds the threshold.

In the module, the second similarity unit 302 obtains second monitoring data corresponding to the peripheral monitoring number, obtains a first monitoring number corresponding to the first monitoring data, obtains second monitoring data corresponding to the second monitoring number according to the first monitoring number, compares the second monitoring data with a preset data set, obtains second similarity through comparison, compares the second similarity with a threshold value, can set the threshold value of smoke and flame to be 30%, sets the threshold value of dangerous molecular face data to be 50%, judges that the second heavy point monitoring data is obtained if the smoke and flame comparison result exceeds the threshold value, judges that the dangerous molecular face data is lower than the threshold value, and does not judge that the second heavy point monitoring data is obtained if the smoke and flame comparison result is higher than the threshold value;

And (3) the monitoring numbers inquired by the peripheral monitoring numbers are called through the comparison priority search algorithm, when the comparison result of the smoke and the flame exceeds the threshold value, the dangerous molecular face data is inquired, if the comparison result of the smoke and the flame is lower than the threshold value, the inquiry is carried out, and if the comparison result of the smoke and the flame is higher than the threshold value, the inquiry is stopped.

And the second marking unit 303 is configured to perform marking processing on the second endpoint monitoring data, send the second endpoint monitoring data to the monitoring terminal, and receive the acknowledgement information of the monitoring terminal.

In this module, the second marking unit 303 performs marking processing on the second multipoint monitoring data, finds fire, smoke and dangerous molecules through comparison, defines the monitoring data for finding fire, smoke and dangerous molecules as second multipoint monitoring data, performs marking processing on the second multipoint monitoring data, and obtains the monitoring time and the monitoring number corresponding to the second multipoint monitoring data, thereby realizing quick and accurate response, improving the efficiency and the effectiveness of safety management, and judging the occurrence and spreading range of the fire and the movement track of the dangerous molecules.

As shown in fig. 8, as a preferred embodiment of the present invention, the alert module 400 includes:

And the third monitoring data unit 401 is configured to pre-process the first important monitoring data and the second important monitoring data, and obtain pre-processed third monitoring data.

In this module, the third monitoring data unit 401 performs preprocessing on the first important monitoring data and the second important monitoring data, adjusts the image size, adjusts the image definition, removes noise, improves the visual effect of the image, makes the image more suitable for display or further analysis, can more clearly observe the monitoring content and the sound definition in the monitoring, and obtains the third monitoring data after the processing is completed.

And the backup unit 402 is configured to perform remote storage backup on the third monitoring data, generate an authorization code, and send the authorization code to the monitoring terminal.

In this module, the backup unit 402 performs remote storage backup on the third monitoring data, encrypts and securely transmits the video data to the cloud server using a secure transmission protocol (such as TLS). The safety and the integrity of the data in the transmission process are ensured, the situation that the monitoring data is lost due to fire and the like can be prevented, the fire cause can not be inquired, and the workload of inquiring the fire starting point is reduced;

After the authorization code is uploaded to the cloud server, the authorization code is generated and sent to the monitoring terminal, and a manager can access the cloud server through the authorization code, so that the safety and privacy of monitoring data are guaranteed.

The warning unit 403 is configured to perform a warning operation according to the first important monitoring data and the second important monitoring data.

In this module, the warning unit 403 performs a warning operation according to the first important monitoring data and the second important monitoring data, can display on the monitoring terminal, and can directly give an alarm at the same time, thereby improving safety, and also improving efficiency and response speed by reducing the requirement of manual monitoring through automation.

In one embodiment, a computer device is presented, the computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the following steps when executing the computer program:

Acquiring monitoring data in real time, wherein the monitoring data comprises a monitoring picture, monitoring time and a monitoring number;

comparing the first monitoring data with a preset data set, wherein the preset data set comprises smoke, flame and dangerous molecular face data, obtaining first important monitoring data through comparison, and marking the first important monitoring data;

acquiring first important monitoring data, monitoring peripheral monitoring data, comparing the peripheral monitoring data with a preset data set, and comparing to obtain second important monitoring data;

And preprocessing the first important monitoring data and the second important monitoring data, wherein the preprocessing comprises the steps of adjusting the size of an image, adjusting the definition of the image and removing noise, and executing warning operation according to the first important monitoring data and the second important monitoring data.

In one embodiment, a computer readable storage medium is provided, having a computer program stored thereon, which when executed by a processor causes the processor to perform the steps of:

Acquiring monitoring data in real time, wherein the monitoring data comprises a monitoring picture, monitoring time and a monitoring number;

comparing the first monitoring data with a preset data set, wherein the preset data set comprises smoke, flame and dangerous molecular face data, obtaining first important monitoring data through comparison, and marking the first important monitoring data;

acquiring first important monitoring data, monitoring peripheral monitoring data, comparing the peripheral monitoring data with a preset data set, and comparing to obtain second important monitoring data;

And preprocessing the first important monitoring data and the second important monitoring data, wherein the preprocessing comprises the steps of adjusting the size of an image, adjusting the definition of the image and removing noise, and executing warning operation according to the first important monitoring data and the second important monitoring data.

It should be understood that, although the steps in the flowcharts of the embodiments of the present invention are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in various embodiments may include multiple sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor do the order in which the sub-steps or stages are performed necessarily performed in sequence, but may be performed alternately or alternately with at least a portion of the sub-steps or stages of other steps or other steps.

Those skilled in the art will appreciate that all or part of the processes in the methods of the above embodiments may be implemented by a computer program for instructing relevant hardware, where the program may be stored in a non-volatile computer readable storage medium, and where the program, when executed, may include processes in the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.

The technical features of the above-described embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above-described embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The foregoing examples illustrate only a few embodiments of the invention and are described in detail herein without thereby limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the invention, which are all within the scope of the invention. Accordingly, the scope of protection of the present invention is to be determined by the appended claims.

The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.

Claims (10)

1. A method for processing a monitored image based on computer vision, the method comprising:

Acquiring monitoring data in real time, wherein the monitoring data comprises a monitoring picture, monitoring time and a monitoring number;

comparing the first monitoring data with a preset data set, wherein the preset data set comprises smoke, flame and dangerous molecular face data, obtaining first important monitoring data through comparison, and marking the first important monitoring data;

acquiring first important monitoring data, monitoring peripheral monitoring data, comparing the peripheral monitoring data with a preset data set, and comparing to obtain second important monitoring data;

And preprocessing the first important monitoring data and the second important monitoring data, wherein the preprocessing comprises the steps of adjusting the size of an image, adjusting the definition of the image and removing noise, and executing warning operation according to the first important monitoring data and the second important monitoring data.

2. The method for processing a monitored image based on computer vision according to claim 1, wherein the step of comparing the first monitored data with a preset data set, wherein the preset data set includes smoke, flame and dangerous molecular face data, obtaining first important monitored data through comparison, and performing marking processing on the first important monitored data specifically includes:

comparing the first monitoring data with a preset data set, comparing smoke and flame through an image processing technology, and counting and comparing dangerous molecular faces through an image segmentation and object recognition technology;

comparing the first similarity with a threshold value, and judging the first key monitoring data if the comparison result exceeds the threshold value;

and marking the first key monitoring data, sending the first key monitoring data to the monitoring terminal, and receiving the confirmation information of the monitoring terminal.

3. The method for processing a monitored image based on computer vision according to claim 1, wherein the step of obtaining the first important monitored data to monitor the peripheral monitored data, comparing the peripheral monitored data with a preset data set, and comparing the peripheral monitored data with the preset data set to obtain the second important monitored data comprises the following steps:

Acquiring a first monitoring number of first key monitoring data, and calling a peripheral monitoring number by adopting a breadth-first search algorithm according to the first monitoring number, wherein the breadth-first search algorithm accesses all neighbors of a starting node and then moves to the neighbors of the neighbors;

Obtaining second monitoring data corresponding to the peripheral monitoring number, comparing the second monitoring data with a preset data set, obtaining second similarity through comparison, comparing the second similarity with a threshold value, and judging the second monitored data as second point monitoring data if the comparison result exceeds the threshold value;

And marking the second multipoint monitoring data, sending the second multipoint monitoring data to the monitoring terminal, and receiving the confirmation information of the monitoring terminal.

4. The method for processing a monitored image based on computer vision according to claim 1, wherein the preprocessing of the first important monitored data and the second important monitored data includes adjusting an image size, adjusting an image sharpness, and removing noise, and executing a warning operation according to the first important monitored data and the second important monitored data, specifically comprising:

Preprocessing the first important monitoring data and the second important monitoring data to obtain preprocessed third monitoring data;

Performing remote storage backup on the third monitoring data, generating an authorization code and sending the authorization code to a monitoring terminal;

and executing warning operation according to the first important monitoring data and the second important monitoring data.

5. The method for processing the monitoring image based on the computer vision according to claim 2, wherein the monitoring terminal adopts a computer display screen.

6. A computer vision-based monitored image processing system, the system comprising:

the monitoring acquisition module acquires monitoring data in real time, wherein the monitoring data comprises a monitoring picture, monitoring time and a monitoring number;

The first key monitoring data module is used for comparing the first monitoring data with a preset data set, wherein the preset data set comprises smog, flame and dangerous molecular face data, the first key monitoring data is obtained through comparison, and the first key monitoring data is marked;

the second heavy point monitoring data module is used for acquiring first heavy point monitoring data, monitoring peripheral monitoring data, comparing the peripheral monitoring data with a preset data set and obtaining second heavy point monitoring data;

The warning module is used for preprocessing the first important monitoring data and the second important monitoring data, wherein the preprocessing comprises the steps of adjusting the size of an image, adjusting the definition of the image and removing noise, and warning operation is executed according to the first important monitoring data and the second important monitoring data.

7. The computer vision-based surveillance image processing system of claim 6 wherein the first accent surveillance data module comprises:

The comparison unit is used for comparing the first monitoring data with a preset data set, comparing smoke and flame through an image processing technology, and counting and comparing dangerous molecular faces through an image segmentation and object recognition technology;

the first similarity unit is used for obtaining first similarity through comparison, comparing the first similarity with a threshold value, and judging the first important monitoring data if the comparison result exceeds the threshold value;

The first marking unit is used for marking the first key monitoring data, sending the first key monitoring data to the monitoring terminal and receiving the confirmation information of the monitoring terminal.

8. The computer vision based surveillance image processing system of claim 7 wherein the second point surveillance data module comprises:

The method comprises the steps of calling a peripheral monitoring unit, obtaining a first monitoring number of first key monitoring data, calling the peripheral monitoring number according to the first monitoring number by adopting a breadth-first search algorithm, accessing all neighbors of a starting node by the breadth-first search algorithm, and moving to the neighbors of the neighbors;

the second similarity unit is used for obtaining second monitoring data corresponding to the peripheral monitoring number, comparing the second monitoring data with a preset data set, obtaining second similarity through comparison, comparing the second similarity with a threshold value, and judging the second monitored data as second point monitoring data if the comparison result exceeds the threshold value;

And the second marking unit is used for marking the second multipoint monitoring data, sending the second multipoint monitoring data to the monitoring terminal and receiving the confirmation information of the monitoring terminal.

9. The computer vision-based surveillance image processing system of claim 8 wherein the alert module comprises:

The third monitoring data unit is used for preprocessing the first important monitoring data and the second important monitoring data to obtain preprocessed third monitoring data;

The backup unit is used for carrying out remote storage backup on the third monitoring data, generating an authorization code and sending the authorization code to the monitoring terminal;

and the warning unit is used for executing warning operation according to the first important monitoring data and the second important monitoring data.

10. The computer vision-based monitoring image processing system of claim 9, wherein the monitoring terminal employs a computer display screen.