阅读说明：本技术 基于计算机视觉的加油站危险事故预警方法、装置及设备 (Gas station dangerous accident early warning method, device and equipment based on computer vision ) 是由 黄红星 付少新 陈少红 于 2021-06-22 设计创作，主要内容包括：本发明涉及人工智能技术领域,具体涉及一种基于计算机视觉的加油站危险事故预警方法、装置及设备。该方法包括：获取初始图像；根据初始图像识别存在危险行为的人员数量作为第一因素,根据传感器检测未熄火的车辆数量作为第二因素；根据第一因素和第二因素得到加油站环境的丰富度；获取环境中的空气湿度和油密度；根据空气湿度和油密度获得加油站环境的敏感度；根据丰富度和敏感度对加油站环境中的风险类型等级进行评估,进一步预测出当前加油站环境中可能发生的事故类型。利用本发明,可以有效帮助加油站工作人员对站点内实时情况的监控,检测可能发生的事故类型原因,有针对的进行排查,有效避免重大危险事故产生更大的影响。(The invention relates to the technical field of artificial intelligence, in particular to a gas station dangerous accident early warning method, a gas station dangerous accident early warning device and gas station dangerous accident early warning equipment based on computer vision. The method comprises the following steps: acquiring an initial image; identifying the number of people with dangerous behaviors as a first factor according to the initial image, and detecting the number of vehicles without flameout as a second factor according to the sensor; obtaining the richness of the gas station environment according to the first factor and the second factor; acquiring air humidity and oil density in the environment; obtaining the sensitivity of the gas station environment according to the air humidity and the oil density; and evaluating the risk type grade in the gas station environment according to the richness and the sensitivity, and further predicting the accident type possibly occurring in the current gas station environment. The invention can effectively help the staff of the gas station to monitor the real-time condition in the station, detect the possible accident type reason, carry out the specific investigation, and effectively avoid the major dangerous accident from generating larger influence.)

1. A gas station dangerous accident early warning method based on computer vision is characterized by comprising the following steps:

acquiring an initial image, wherein the initial image comprises a region around a gas station;

identifying the number of people with dangerous behaviors as the first factor according to the initial image, and detecting the number of vehicles without flameout as the second factor according to a sensor; obtaining the abundance of the current gas station environment according to the first factor and the second factor;

acquiring the air humidity and the oil density in the current environment; obtaining the sensitivity of the current gas station environment according to the air humidity and the oil density;

and evaluating the risk type grade in the current gas station environment according to the richness and the sensitivity, and predicting the accident type possibly occurring in the current gas station environment according to the risk type grade.

2. The method of claim 1, wherein the richness of the current gas station environment is positively correlated to the first factor and the second factor.

3. The method according to claim 1, characterized in that the sensitivity of the current gasoline station environment is positively correlated to the air humidity and the oil density.

4. The method of claim 1, wherein said assessing a risk type rating in said current gas station environment based on said richness and said sensitivity comprises:

and acquiring risk type grades corresponding to the evaluation matrix according to the richness and the sensitivity, wherein the risk type grades are divided by analyzing historical richness and historical sensitivity data.

5. The method of claim 4, wherein the step of constructing the evaluation matrix further comprises:

and constructing a two-dimensional data distribution model based on the historical abundance and the historical sensitivity, clustering the data of the historical abundance and the historical sensitivity in the two-dimensional data distribution model to obtain a plurality of categories, wherein different categories correspond to different risk type grades, and determining the evaluation matrix according to the risk type grades.

6. The method of claim 5, wherein the step of determining the evaluation matrix based on the risk type ratings further comprises:

and according to the number of the risk type grades, dividing the historical abundance and the historical sensitivity into a plurality of corresponding threshold ranges respectively, and determining corresponding risk type grades according to the abundance, the sensitivity and the corresponding threshold ranges so as to determine an evaluation matrix.

7. The method of claim 1, wherein predicting the type of accident that may occur in the current gas station environment based on the risk type rating comprises:

and performing analog simulation on the data of the historical abundance and the historical sensitivity to acquire the different accident types, wherein the risk type grades correspond to the different accident types.

8. A dangerous accident early warning device of filling station based on computer vision, its characterized in that, the device includes:

the system comprises an image acquisition unit, a service station acquisition unit and a service management unit, wherein the image acquisition unit is used for acquiring an initial image which comprises a region around the service station;

the image detection and richness acquisition unit is used for identifying the number of people with dangerous behaviors as a first factor according to the initial image and detecting the number of vehicles without flameout as a second factor according to a sensor; obtaining the abundance of the current gas station environment according to the first factor and the second factor;

the sensitivity acquisition unit is used for acquiring the air humidity and the oil density in the current environment; obtaining the sensitivity of the current gas station environment according to the air humidity and the oil density;

and the accident type prediction unit is used for evaluating the risk type grade in the current gas station environment according to the richness and the sensitivity and predicting the accident type which possibly occurs in the current gas station environment according to the risk type grade.

9. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 7 when executing the computer program.

Technical Field

The invention relates to the field of artificial intelligence, in particular to a gas station dangerous accident early warning method and system based on computer vision.

Background

At present, the number of refueling and gas filling stations in various places is increased sharply, but the refueling and gas filling stations have higher risks due to improper management or other factors; under the environment of the refueling station, the influence brought by different vehicle states and personnel behaviors is different.

In a gas station site, the probability of various accidents of the gas station can be greatly increased due to interference of dangerous behaviors of personnel or improper operation in the oil discharge operation; normally, the spark effect of metal collision is not large, but after oil leakage, the density of oil in the air is increased, so that the effect of metal spark is large, and explosion is easily caused.

The dangerous accident early warning in the prior art cannot effectively predict the source of different accident types, and potential safety hazards cannot be timely and effectively discovered when workers at the station of the gas station conduct investigation, so that the prediction of the dangerous accident types occurring in the station of the gas station is very important.

Disclosure of Invention

In order to solve the technical problems, the invention aims to provide a gas station dangerous accident early warning method, a gas station dangerous accident early warning device and gas station dangerous accident early warning equipment based on computer vision, and the adopted technical scheme is as follows:

in a first aspect, an embodiment of the present invention provides a gas station dangerous accident early warning method based on computer vision, which includes the following specific steps:

acquiring an initial image, wherein the initial image comprises a region around a gas station; identifying the number of people with dangerous behaviors as a first factor according to the initial image, and detecting the number of vehicles without flameout as a second factor according to a sensor; obtaining the abundance of the current gas station environment according to the first factor and the second factor; acquiring the air humidity and the oil density in the current environment; obtaining the sensitivity of the current gas station environment according to the air humidity and the oil density; and evaluating the risk type grade in the current gas station environment according to the richness and the sensitivity, and predicting the accident type possibly occurring in the current gas station environment according to the risk type grade.

Preferably, the richness of the current gas station environment is in positive correlation with the first factor and the second factor.

Preferably, the sensitivity of the current filling station environment is positively correlated with the air humidity and the oil density.

Preferably, the step of assessing a risk type rating in the current gas station environment based on the richness and the sensitivity comprises:

and acquiring risk type grades corresponding to the evaluation matrix according to the richness and the sensitivity, wherein the risk type grades are divided by analyzing historical richness and historical sensitivity data.

Preferably, the step of constructing the evaluation matrix further includes:

and constructing a two-dimensional data distribution model based on the historical abundance and the historical sensitivity, clustering the data of the historical abundance and the historical sensitivity in the two-dimensional data distribution model to obtain a plurality of categories, wherein different categories correspond to different risk type grades, and determining the evaluation matrix according to the risk type grades.

Preferably, the step of determining the evaluation matrix according to the risk type grade further includes:

and according to the number of the risk type grades, dividing the historical abundance and the historical sensitivity into a plurality of corresponding threshold ranges respectively, and determining corresponding risk type grades according to the abundance, the sensitivity and the corresponding threshold ranges so as to determine an evaluation matrix.

Preferably, the predicting the type of accident that may occur in the current gas station environment according to the risk type level includes:

and performing analog simulation on the data of the historical abundance and the historical sensitivity to acquire the different accident types, wherein the risk type grades correspond to the different accident types.

In a second aspect, another embodiment of the present invention provides a dangerous accident early warning device for a gas station based on computer vision, comprising:

the system comprises an image acquisition unit, a service station and a display unit, wherein the image acquisition unit acquires an initial image which comprises a region around the service station;

the image detection and richness acquisition unit is used for identifying the number of people with dangerous behaviors as a first factor according to the initial image and detecting the number of vehicles without flameout as a second factor according to the sensor; obtaining the abundance of the current gas station environment according to the first factor and the second factor;

a sensitivity acquisition unit that acquires air humidity and oil density in the current environment; obtaining the sensitivity of the current gas station environment according to the air humidity and the oil density;

and the accident type prediction unit is used for evaluating the risk type grade in the current gas station environment according to the richness and the sensitivity and predicting the accident type which possibly occurs in the current gas station environment according to the risk type grade.

Preferably, an embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

The embodiment of the invention has the beneficial effects that: the method comprises the steps of calculating the abundance and the sensitivity of the current gas station site environment through the dangerous behaviors of personnel in the gas station site environment, the vehicle state, the air humidity and the density of gasoline and diesel oil in the air, predicting the risk type grade according to the data of historical abundance and historical sensitivity, and further predicting the potential accident type which possibly occurs in the current gas station environment, so that the staff can timely perform targeted investigation, the real-time prediction and monitoring effects in the gas station site environment are improved, and the occurrence of dangerous accidents is prevented from generating larger influence.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.

FIG. 1 is a flow chart of a method for providing a dangerous accident warning for a gas station based on computer vision according to an embodiment of the present invention;

FIG. 2 is a logic flow diagram of a gas station dangerous accident warning based on computer vision according to an embodiment of the present invention;

FIG. 3 is a flowchart of a method for obtaining richness of a gas station environment according to an embodiment of the present invention;

FIG. 4 is a graph of cluster segmentation for simulating accident types via historical richness and historical sensitivity data, as provided by an embodiment of the present invention;

FIG. 5 is a block diagram of a dangerous accident early warning device for a gasoline station based on computer vision according to another embodiment of the present invention;

fig. 6 is a schematic diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

To further illustrate the technical means and effects of the present invention for achieving the predetermined objects, the following detailed description of the method and system for early warning of dangerous accidents of gas stations based on computer vision according to the present invention with reference to the accompanying drawings and preferred embodiments shows the following detailed descriptions. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

The embodiment of the invention aims at the following specific scenes: the method and the device are applied to the environment of a gas station in real life, and in order to solve the problem that serious dangerous accidents occur in the gas station due to personnel dangerous behaviors or increased oil leakage density in the loading and unloading process of a fuel truck, the type of accident danger possibly occurring in the current state of the gas station is predicted by analyzing and evaluating the personnel dangerous behaviors, the state of the vehicle during refueling and the current environment state of the gas station.

The invention provides a gas station dangerous accident early warning method, a gas station dangerous accident early warning device and a gas station dangerous accident early warning equipment based on computer vision.

Referring to fig. 1 and fig. 2, fig. 1 is a flowchart illustrating a method for providing a dangerous accident warning for a gas station based on computer vision according to an embodiment of the present invention; FIG. 2 is a logic flow diagram of a computer vision based early warning of dangerous accidents at gas stations according to an embodiment of the present invention.

The method comprises the following steps:

step S100, acquiring an initial image, wherein the initial image comprises the surrounding area of the gas station.

And deploying a camera at the gas station, wherein the camera is positioned above the gas station site and can shoot the complete information of the area of the gas station site.

Step S200, identifying the number of people with dangerous behaviors as a first factor according to the initial image, and detecting the number of vehicles without flameout as a second factor according to the sensor; and obtaining the abundance of the current gas station environment according to the first factor and the second factor.

The gasoline stations are all placed in the external environment, most of the positions of the gasoline stations are located at two sides of a road, the environment is complex and changeable, therefore, the initial image is firstly processed before being analyzed so as to obtain clearer image data,

referring to fig. 3, the specific process is as follows:

step S201, processing the collected initial image by adopting filtering denoising and sharpening processing to obtain an optimized image.

Preferably, the embodiment of the invention selects a self-adaptive median filtering algorithm, automatically selects an applicable filtering window, and performs denoising processing on the initial image.

In order to improve the visual effect of the initial image, make the outline of the initial image clear and facilitate subsequent accurate information extraction, the embodiment of the invention adopts image differential sharpening to process the initial image.

In step S202, a first factor is obtained from the optimized image in step S201.

Preferably, the embodiment of the invention adopts the key point detection network to identify the dangerous behaviors of the personnel and obtain the number of the dangerous behaviors in the current gas station site environment, and the method effectively reduces the calculated amount of the system.

The training process of the key point detection network comprises the following steps:

and carrying out artificial key point labeling on dangerous behaviors existing in the optimized image, taking sample images and label image data collected by a camera as the input of a key point detection network, carrying out feature extraction through a key point detection encoder, and carrying out up-sampling through a key point detection decoder so as to generate a thermodynamic diagram of the key points.

The network adopts a mean square error loss function to carry out iterative training during training, and network parameters are continuously updated.

Furthermore, the personnel danger marked in the embodiment of the invention mainly comprises five categories of clothes beating, alarming, mobile phone using, smoking, open fire and the like.

In other embodiments, methods such as gesture analysis and distance analysis are often used to identify dangerous behaviors.

Step S203, acquiring a second factor according to the detection of the sensor.

Preferably, in the embodiment of the present invention, an engine sensor of a vehicle is used to detect an engine state of the vehicle in real time, and the specific process includes:

the starting and stopping states of the vehicles are obtained according to the engine sensors of the vehicles which are refueling, and the obtained vehicle state information is fed back through a vehicle network, so that the number of the vehicles which are not flameout is obtained.

Methods for detecting whether the vehicle is flameout include voice detection network recognition, sensor detection, and the like.

And step S204, obtaining the richness of the gas station in the current environment according to the first factor and the second factor.

The richness in the embodiment of the invention refers to the complexity of vehicle state information and personnel information in the current gas station environment, and is in positive correlation with a first factor and a second factor;

the function model expression of the environment richness is as follows:

F＝exp(ωK)*ln(C+e)+τ

wherein K is a first factor, omega is a model adjustable factor larger than zero, C is a second factor, and tau is a model offset term.

Preferably, in the embodiment of the invention, the value of ω is 0.2; τ is 5; since the richness of the environment does not increase indefinitely, the threshold range of the richness of the environment is set to (6,50) in the embodiment of the present invention.

When the richness exceeds the threshold range, the system can directly send out buzzing early warning, and the station managers of the gas station need to comprehensively check the stations.

And step S300, acquiring the air humidity and the oil density in the current environment, and acquiring the sensitivity of the current gas station environment according to the air humidity and the oil density.

Sensors are installed at the gasoline station site for measuring air humidity, gasoline density and diesel density in the environment of the gasoline station site.

Preferably, in the embodiment of the invention, an SHT30 temperature and humidity air pressure sensor is selected to detect the humidity of the ambient air, and an online gasoline and gas detector HNAG1000-EXQ is adopted to detect the density rho of gasoline in the air_{Steam generator}Acquiring air diesel density rho by adopting a Honeyeger HAGG1000 diesel gas density detection device_Firewood。

The environmental sensitivity in the embodiment of the invention refers to the state of the substance in the ambient air of the current gas station, and has a positive correlation with the air humidity and the oil density.

The functional model expression of the environmental sensitivity is as follows:

E＝σ*(exp[(10²ρ_{steam generator}+10²ρ_Firewood)-1])

Where σ is the humidity influence factor, ρ_{Steam generator}For the gasoline density, rho, in the station environment of a gasoline station_FirewoodIs the density of diesel in the environment of the station site.

When the current environment humidity of the station of the gas station is higher than the preset standard humidity, the current station of the gas station has overhigh humidity and slow air flow, the gas in the environment is not easy to circulate, and the value of a humidity influence factor is more than 0 and less than 1; when the current environment humidity of the station of the gas station is lower than the preset standard humidity, the current station humidity of the gas station is lower, the air circulation is faster, and the influence factor value 1 is less than sigma.

Preferably, in the embodiment of the present invention, the value of the humidity affecting factor is set to be 0.8 when the ambient humidity is higher than the preset standard humidity, and is set to be 1.2 when the ambient humidity is lower than the preset standard humidity.

In order to facilitate analysis of the influence of the environment richness and the sensitivity on the whole environment, the environment sensitivity model and the environment richness model are normalized in the embodiment of the invention, so that the function value ranges of the two models are both in the (0,1) interval.

And step S400, evaluating the risk type grade in the current gas station environment according to the richness and the sensitivity, and predicting the accident type possibly occurring in the current gas station environment according to the risk type grade.

The specific process is as follows:

and acquiring risk type grades corresponding to the evaluation matrix according to the richness and the sensitivity, wherein the risk type grades are divided by analyzing historical richness and historical sensitivity data.

Referring to fig. 4, simulation is performed on data of history abundance and history sensitivity to obtain different accident types, and the risk type levels correspond to different accident types.

The abscissa of each point in the graph represents the environmental sensitivity of the gas station site, and the ordinate represents the environmental richness of the gas station site;

in order to distinguish the simulated accident types, in the embodiment of the invention, four levels are divided by adopting different gray values to represent different accident types, wherein the four levels are respectively a no-danger accident 01, an operation accident 02, a fire accident 03 and an explosion accident 04.

The specific process for constructing the evaluation matrix comprises the following steps:

and constructing a two-dimensional data distribution model based on the history abundance and the history sensitivity, clustering the data of the history abundance and the history sensitivity in the two-dimensional data distribution model to obtain a plurality of categories, wherein different categories correspond to different risk type grades, and determining an evaluation matrix according to the risk type grades.

Preferably, the data are clustered by adopting a K-means algorithm in the embodiment of the invention, and the data are divided into four categories corresponding to four accident types with different risk degrees.

The specific mode for determining the evaluation matrix according to the risk type grade is as follows:

according to the number of the risk type grades, the history abundance and the history sensitivity are divided into a plurality of corresponding threshold ranges respectively, and the corresponding risk type grades are determined according to the abundance and the sensitivity and the corresponding threshold ranges, so that the evaluation matrix is determined.

As a preferred scheme, four threshold ranges are set in the embodiment of the present invention, respectively: (0,0.25), (0.25,0.5), (0.5,0.75), (0.75, 1.0).

Please refer to the following matrix:

TABLE 1 evaluation matrix for risk type ratings

And F in the matrix is the environment richness, E is the site environment sensitivity, and Z is the site potential risk type grade, and the four accident types are respectively corresponding to the F in the matrix.

When the risk type grade is grade 1, the station of the gas station has no accident risk;

when the risk type grade is grade 2, the accident type corresponding to the station of the gas station is an operation accident;

when the risk type grade is 3 grade, the accident type corresponding to the station of the gas station is a fire accident;

when the risk type grade is 4 grade, the accident type corresponding to the station of the gas station is an explosion accident;

furthermore, staff at a subsequent station can correspondingly evaluate the risk type grade in the matrix according to the numerical ranges of the richness and the sensitivity in the current gas station environment, so that the accident type which possibly occurs can be predicted quickly and efficiently.

According to the embodiment of the invention, an intuitive method is adopted to predict and judge the potential dangerous accident type of the gas station site, and the judgment matrix of the accident type is obtained through data analysis, so that site workers can quickly and accurately master the real-time situation of the gas station site; when the risk type grade of the station of the gas station is greater than 1, the system can send out an early warning prompt to timely warn workers to carry out targeted investigation on the station environment of the gas station, so that the occurrence of major accident types is avoided.

In summary, the embodiment of the present invention provides a gas station dangerous accident early warning method based on computer vision, the method takes into account the richness of the computing environment by taking into account the dangerous behavior of personnel and the vehicle status in the current gasoline station environment, the sensitivity of the environment is calculated by the air humidity in the ambient air of the current gasoline station and the density of the gasoline diesel, accident type and risk type level are simulated through a large amount of data with historical richness and sensitivity, the risk types of the gasoline station sites are accurately predicted according to the evaluation matrix, the gasoline station site accident types in real-time environment are further obtained, therefore, the staff of the gas station can more intuitively know the dangerous source of the station of the gas station, and can timely and purposefully investigate the possible reasons of the occurrence of the dangerous accident, thereby avoiding the dangerous accident from generating larger influence.

Based on the same inventive concept as the method embodiment, the embodiment of the invention also provides a gas station dangerous accident early warning device based on computer vision.

Referring to fig. 5, an embodiment of the present invention provides a gas station dangerous accident early warning device based on computer vision, the device includes: an image acquisition unit 100, an image detection and richness acquisition unit 200, a sensitivity acquisition unit 300, and an accident type prediction unit 400.

An image acquisition unit 100 for acquiring an initial image, the initial image comprising an area around the gasoline station.

The image detection and richness acquisition unit 200 is used for identifying the number of people with dangerous behaviors as a first factor according to the initial image and detecting the number of vehicles without flameout as a second factor according to the sensor; and obtaining the abundance of the current gas station environment according to the first factor and the second factor.

A sensitivity acquisition unit 300 for acquiring air humidity and oil density in the current environment; the sensitivity of the current gas station environment is obtained from the air humidity and the oil density.

And the accident type prediction unit 400 is used for evaluating the risk type grade in the current gas station environment according to the richness and the sensitivity and predicting the accident type which possibly occurs in the current gas station environment according to the risk type grade.

Further, please refer to fig. 6, which illustrates a schematic diagram of an electronic device according to an embodiment of the present invention. The electronic device in this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. The processor, when executing the computer program, implements the steps of one of the above-mentioned embodiments of the method for early warning of dangerous accidents at a gas station based on computer vision, such as the steps shown in fig. 1. Or the processor executes the computer program to realize the functions of the units in the embodiment of the gas station dangerous accident early warning device based on computer vision.

Illustratively, a computer program may be partitioned into one or more units, where one or more units are stored in the memory and executed by the processor to implement the invention. One or more of the elements may be a sequence of computer program instruction segments for describing the execution of the computer program in the electronic device, which can perform certain functions.

The electronic device can be a desktop computer, a notebook, a palm computer, a cloud server and other computing devices. The electronic device may include, but is not limited to, a processor, a memory. It will be appreciated by those skilled in the art that the schematic diagrams are merely examples of the electronic device and do not constitute a limitation of the electronic device, and may include more or less components than those shown, or some components in combination, or different components, e.g. the electronic device may also include input-output devices, buses, etc.

The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like that is the control center for an electronic device and that connects the various parts of the overall electronic device using various interfaces and wires.

It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.