A construction site safety risk identification alarm method

Abstract

The present application relates to the technical field of safety monitoring, and particularly relates to a construction site safety risk identification and alarm method, which receives video monitoring images of multiple monitors of a construction site; obtains a running area of a construction equipment, and labels the running area on each video monitoring image; reads data of a kind, a stacking position and a stacking amount of combustible materials of the construction site; periodically updates the video monitoring images; respectively performs smoke identification and construction equipment running state identification; if there is smoke on the video monitoring images, a smoke generation position is identified, the smoke generation position is taken as a fire risk position, and a fire risk alarm is sent; if the construction equipment on the video monitoring images is in a running state, it is identified whether there is a non-equipment operator in the running area of the construction equipment, and if there is a non-equipment operator, a device running risk alarm information is sent. The present application has the beneficial technical effects that smoke appearing on a construction site can be quickly identified, and a fire appearing on the construction site can be timely found.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of safety monitoring, in particular to a construction site safety risk identification and alarm method. BACKGROUND

[0002] Construction is a production activity in the implementation phase of engineering construction. Construction includes foundation engineering construction, main structure construction, roof engineering construction, decoration engineering construction, etc. The place of construction operation is called "construction site" or "construction site", also known as the construction site. Construction is a complex technical production process, which needs to be operated by multiple workers in a limited time and space. Hundreds of materials and various mechanical equipment operations must have scientific and advanced organization and management measures. Safety management is an important work. At present, the safety management personnel of the construction site are not in place, the safety inspection is formal, and the safety hidden danger cannot be fully discovered and eliminated in time. The safety of the construction site is affected, and the construction accident hidden danger exists for a long time. Therefore, it is necessary to study the identification and alarm technology of construction site safety risk to ensure the safety of the construction site.

[0003] The prior art discloses a construction safety monitoring and early warning method and system, which includes obtaining real-time state data of a main force bar, the real-time state data including at least one of stress data, strain data, inclination data and settlement data; comparing the real-time state data with the pre-set main force bar attribute data; when the real-time data state deviates from the attribute data, sending a safety warning prompt information to the terminal. When comparing the real-time data with the pre-set main force bar attribute data, the deviation value of the pre-set attribute data is included; the real-time deviation value of the real-time state data deviating from the main force bar attribute data is calculated; when at least one real-time deviation value deviates from the deviation value of the attribute data, a safety warning prompt information is sent to the terminal. The technical scheme can solve the early warning of the main force bar of the construction site when the safety hidden danger occurs, but cannot solve the early warning of other safety risks in the construction site. SUMMARY

[0004] The technical problem to be solved by the present application is the lack of construction site safety risk monitoring scheme. A construction site safety risk identification and alarm method is proposed, which can identify and alarm the safety risks occurring in the construction site, help to dispose the risks in time, and ensure the safety of the construction site.

[0005] To solve the above technical problems, the present application adopts the following technical scheme: a construction site safety risk identification and alarm method, comprising the following steps: receiving video monitoring images of a plurality of monitors in the construction site;

[0006] obtaining a running area of the construction equipment, marking the running area on each video monitoring image;

[0007] reading combustible data including combustible type, storage location and storage amount;

[0008] periodically updating video monitoring images of multiple angles of the construction site;

[0009] performing smoke identification and construction equipment running state identification on each updated video monitoring image respectively;

[0010] if there is smoke on the video monitoring image, identifying a smoke generation location, the smoke generation location being a fire risk location, issuing a fire risk alarm and entering the next step, if there is no smoke on the video monitoring image, directly entering the next step;

[0011] if the construction equipment on the video monitoring image is in a running state, identifying whether there is a non-equipment operator in the running area of the construction equipment, if there is a non-equipment operator, issuing a device running risk alarm information.

[0012] Preferably, the method for smoke identification comprises:

[0013] extracting a color region of a preset color range in the video monitoring image;

[0014] if the pixel area covered by the color region exceeds a preset threshold, it is determined that there is smoke in the video monitoring image, otherwise, it is determined that there is no smoke in the video monitoring image.

[0015] Preferably, the method for obtaining a preset color range comprises:

[0016] reading smoke images of a plurality of historical fires;

[0017] extracting a color range of smoke in each smoke image respectively;

[0018] the union of the color ranges of all smoke images constitutes the preset color range.

[0019] Preferably, the method for obtaining a preset color range comprises:

[0020] reading smoke images of a plurality of historical fires, and associating the smoke images with combustible types;

[0021] eliminating smoke images associated with multiple combustible types;

[0022] extracting a color range of smoke in each remaining smoke image respectively, and associating the color range with a combustible type;

[0023] the union of the color ranges of all smoke images constitutes the preset color range.

[0024] As preferred, the method for determining the smoke generation position comprises:

[0025] According to the color of the smoke on the video monitoring image, the corresponding combustible species is obtained;

[0026] According to the species of the combustible material on the construction site, the stacking position of the corresponding combustible species is obtained as the presumed position of the smoke generation position;

[0027] The pixel distribution area of the smoke on the video monitoring image is obtained, and the bottom of the pixel distribution area is taken as the observation position of the smoke generation position;

[0028] Both the presumed position and the observation position are taken as the fire risk position.

[0029] As preferred, it further comprises a method for identifying the smoke coverage area, which is executed after the smoke generation position is identified, and the method for identifying the smoke coverage area comprises: obtaining the pixel distribution area of the smoke on the video monitoring image, and counting the pixel area S of the pixel distribution area;

[0030] According to the smoke generation position and the position of the monitor, the distance L between the smoke generation position and the monitor is obtained;

[0031] The coverage area C of the smoke is calculated as C=k*S*L, and k is a preset coefficient;

[0032] The coverage area C is compared with a preset coverage area and risk level division scheme to obtain the risk level;

[0033] The fire risk alarm comprises the fire risk position and the risk level.

[0034] As preferred, the risk level division scheme comprises a plurality of coverage area numerical intervals and corresponding risk levels,

[0035] The method for obtaining the risk level division scheme comprises:

[0036] Reading a plurality of smoke images of historical fires marked with fire levels;

[0037] Grouping the smoke images according to the fire levels;

[0038] Identifying the coverage area in each group of smoke images and calculating the mean value as the coverage area corresponding to the fire level;

[0039] Taking the mean value of the coverage area corresponding to adjacent fire levels as the demarcation value to obtain the risk level division scheme.

[0040] As preferred, the method for identifying the construction equipment operation state comprises:

[0041] reading image templates of multiple working postures of the construction equipment, denoted as equipment templates;

[0042] comparing the image in the running area in the video monitoring image with the equipment templates to obtain the working postures of the construction equipment;

[0043] obtaining the working postures of the construction equipment in the video monitoring images of multiple continuous periods;

[0044] if the working postures of the construction equipment in the video monitoring images of multiple continuous periods are not all the same, it is determined that the construction equipment is in a running state, otherwise, it is determined that the construction equipment is in a non-running state.

[0045] As preferred, the method for identifying non-equipment operators comprises:

[0046] comparing the image in the running area in the video monitoring image with the equipment templates to obtain an area occupied by the construction equipment, denoted as an equipment area;

[0047] identifying personnel in the running area, if the personnel are located outside the equipment area, it is determined that they are non-equipment operators, otherwise, if the personnel are located in the equipment area, it is determined that they are equipment operators.

[0048] As preferred, the risk identification and alarm method further comprises a combustible fire prediction method, and the combustible fire prediction method comprises:

[0049] obtaining the wind direction of the construction site;

[0050] judging whether the combustible is located at the downwind of the smoke generation position according to the smoke generation position and the wind direction;

[0051] if the combustible is located at the downwind of the smoke generation position, issuing a combustible fire warning, the combustible fire warning comprising the combustible type, the stacking position and the stacking amount, and if no combustible is located at the downwind of the smoke generation position, no operation is performed.

[0052] The beneficial technical effects of the present application include: the video monitoring image obtained by the monitor, combined with smoke identification, can quickly identify the smoke appearing in the construction site and timely discover the fire in the construction site; combined with the stacking condition of the combustible in the construction site and the identified smoke generation position, the next development of the fire can be predicted, which helps to develop a fire disposal scheme; the type of the combustible is inferred from the smoke, which helps to locate the fire risk occurrence position and is helpful to the effective disposal of the fire; the running state of the construction equipment is identified, non-equipment operators in the working area of the construction equipment are identified, the identification and alarm of the construction equipment running risk are realized, and the safe operation of the construction equipment is ensured.

[0053] Other features and advantages of the present application will be disclosed in the following detailed description of the embodiments, drawings. BRIEF DESCRIPTION OF DRAWINGS

[0054] The application will be further described below with reference to the drawings.

[0055] Figure 1 The full-risk identification alarm method flowchart of the embodiment of the present application.

[0056] Figure 2 The smoke identification method flowchart of the embodiment of the present application.

[0057] Figure 3 The method flowchart of obtaining the preset color range of the embodiment of the present application.

[0058] Figure 4 The method flowchart of associating the combustible species of the embodiment of the present application.

[0059] Figure 5 The method flowchart of judging the smoke generation position of the embodiment of the present application.

[0060] Figure 6 The method flowchart of identifying the smoke coverage area of the embodiment of the present application.

[0061] Figure 7 The method flowchart of obtaining the risk level division scheme of the embodiment of the present application.

[0062] Figure 8 The method flowchart of identifying the construction equipment operation state of the embodiment of the present application.

[0063] Figure 9 The method flowchart of identifying the non-equipment operator of the embodiment of the present application.

[0064] Figure 10 The method flowchart of predicting the combustible material fire of the embodiment of the present application. DETAILED DESCRIPTION

[0065] The technical solutions of the embodiments of the present application will be explained and described below with reference to the drawings of the embodiments of the present application. The following embodiments are preferred embodiments of the present application, and are not all. Based on the embodiments in the embodiments, other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.

[0066] In the following description, the appearance of terms such as "inner", "outer", "upper", "lower", "left", "right", and the like is merely intended to facilitate the description of the embodiments and simplify the description, and does not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the present application.

[0067] Before introducing the technical solutions of the present embodiment, the application background of the present embodiment is introduced. The technical solutions of the present embodiment are used for safety risk identification and alarm of construction sites, and are used for improving the safety degree of construction sites. The construction site mainly refers to the construction site of a building or a special building, especially the construction site of a natural gas station or the decoration stage of a building construction. These construction sites have a high risk of fire, and construction equipment such as large cranes, bulldozers, or tower cranes is also used. The equipment is used for lifting special natural gas storage or transportation equipment or hoisting decoration materials. These equipment also has the safety risk of falling objects or moving structures during operation.

[0068] The management of the construction site is a complex work. The personnel, materials, equipment, and activities of the construction site are constantly changing. Through personnel supervision, it is difficult to effectively manage the construction site. In the industry, a monitor is installed on the site to shoot video images of the site. At the same time, a large number of sensors such as temperature, humidity, wind speed, and noise sensors are arranged on the site, so as to obtain the safety risk situation of the construction site through the sensors, discover the safety risk in time, and obtain the state of the current risk to develop a disposal plan. However, due to the rapid change of materials in the construction site, the position of the risk may also change rapidly. Moreover, the arranged sensors need to be powered and maintained for failure, etc. When the sensor fails, replacing the sensor is a labor-intensive task. Through the fixed sensors arranged, there is a large amount of arrangement work, and there is also a problem of lack of pertinence of risk identification and low identification rate.

[0069] With the development of image recognition technology, through intelligent recognition of on-site monitoring images, not only the identification of objects can be realized, but also the identification of the state of objects and the state of personnel can be realized. Image recognition technology provides a new technical approach for the safety control of construction sites. The image recognition technologies used mainly include feature recognition neural network models and graph convolution neural network models, which belong to the category of neural network models. Through neural network models, various information of the construction site can be obtained, such as the location of goods, the type of goods, the location of personnel, the activity state of personnel, the type of equipment, the working state of equipment, the flame area and the smoke area, etc., which can effectively improve the control level of the construction site and to some extent help to improve the safety level of the construction site. However, there are the following shortcomings in using neural network model technology: neural network model operation requires a large amount of computing power, which puts more requirements on hardware; neural network model needs a large amount of materials for training, and the source of materials is limited; when new risk situations occur, neural network model often cannot effectively identify; there is also the problem of poor portability. These shortcomings result in the use of neural network model recognition for construction site safety monitoring, which has the problems of low recognition efficiency and limited recognition risk types.

[0070] To effectively improve the safety level of the construction site, the embodiment provides a scheme of color and template matching identification of video monitoring images of the construction site by the monitor, and in combination with the related information of the combustible materials stacked in the construction site, the identification and alarm of fire and construction equipment safety risk are realized. The spread of the risk is predicted, the risk of the construction site can be identified and timely alarm can be given, and the related position information of the risk can also be provided to help to correctly formulate and implement the risk disposal scheme.

[0071] The construction site safety risk identification and alarm method provided in the embodiment, please refer to the accompanying Figure 1 , including the following steps:

[0072] Step A01) receiving video monitoring images of multiple monitors of the construction site;

[0073] Step A02) obtaining the running area of the construction equipment, and marking the running area on each video monitoring image;

[0074] Step A03) reading the type, stacking position and stacking amount data of the combustible materials in the construction site;

[0075] Step A04) periodically updating the video monitoring images of multiple angles of the construction site;

[0076] Step A05) respectively performing smoke identification and construction equipment running state identification on each updated video monitoring image;

[0077] Step A06) If there is smoke on the video monitoring image, identify the smoke generation position, the smoke generation position as a fire risk position, issue a fire risk alarm, and enter the next step. If there is no smoke on the video monitoring image, directly enter the next step.

[0078] Step A07) If the construction equipment on the video monitoring image is in a running state, identify whether there is a non-equipment operator in the running area of the construction equipment. If there is a non-equipment operator, issue a device running risk alarm information.

[0079] By identifying the color range of smoke and smoke area, combined with the position of combustible material stacking, the type and stacking position of combustible material where the fire occurs can be basically determined, and the stacking amount of the site can be determined. The color of the smoke generated by the fire has a relatively fixed range, and is obviously different from the color of the background. By identifying the color range, the range of the smoke can be easily extracted. The color of the smoke generated by the fire on the construction site is usually white, black and gray. In rare cases, when the construction site stacks toxic chemical substances, colored smoke will be generated, and yellow-green is the most common. However, the construction site referred to in the embodiment is usually a residential building or an office building, and a large amount of flammable and toxic chemical materials will not usually appear. When applied to a construction site of a chemical plant facility, the color of the smoke generated by the combustion of the chemical materials present on the actual construction site needs to be obtained. The type of combustible material helps to correctly select the fire extinguishing measure, and the stacking position helps to correctly arrange personnel evacuation and implement the fire extinguishing measure. The stacking amount of the combustible material can help determine the development trend of the subsequent fire. That is, if the stacking amount is small, the fire will not be too large, and if the stacking amount is large, the fire will be considered to increase significantly over time. By identifying the working state of the construction equipment and whether there is a non-equipment operator in the running area, if there is a non-equipment operator, an alarm is issued to prompt that there is a safety risk when the construction equipment is running.

[0080] Please refer to the attached Figure 2 The method for identifying smoke includes:

[0081] Step B01) Extract the color region of the preset color range in the video monitoring image;

[0082] Step B02) If the pixel area covered by the color region exceeds a preset threshold, it is determined that there is smoke in the video monitoring image, otherwise it is determined that there is no smoke in the video monitoring image. The color of the smoke generated by the fire has a regular pattern. It is related to the type of combustible material where the fire occurs. Usually, the smoke generated by the fire is from black to grayish white. In the construction site, the color region in the black to grayish white color usually has a small range. The smoke area can be easily extracted by color range. In step B02), only the color region belonging to the smoke is considered for pixel area judgment.

[0083] Furthermore, the smoke will spread upward, so the color region of the smoke will certainly reach the upper boundary of the view of the monitor. If it does not reach the upper boundary, it is not the color region of the smoke. Accordingly, the embodiment specifically provides a method for extracting a color region of a preset color range in a video monitoring image, comprising:

[0084] extracting all patch regions of the color range in the video monitoring image which are consistent with the preset color range;

[0085] judging whether the patch region is in contact with the upper boundary of the video monitoring image, if the patch region is in contact with the upper boundary, it is determined that the patch region is the color region of the smoke, if the patch region is not in contact with the upper boundary, it is determined that the patch region is not the color region of the smoke.

[0086] Please refer to the accompanying Figure 3 The method for obtaining the preset color range comprises:

[0087] Step C01) reading smoke images of historical fire;

[0088] Step C02) extracting the color range of the smoke in each smoke image respectively;

[0089] Step C03) the union of the color ranges of all smoke images constitutes the preset color range.

[0090] By reading the smoke images of historical fire, the color range of the smoke can be obtained for subsequent identification of the smoke. The more the types of combustibles in the historical fire cover, the more accurate the preset color range obtained is for identifying the smoke.

[0091] The types of combustibles are different, and the colors of the smoke generated are also different. For example, there are white smoke, black smoke, black smoke, colored smoke, etc. As shown in Table 1, it is a table of relationship between smoke color and fire intensity.

[0092] Table 1 Table of relationship between smoke color and fire intensity

[0093]

[0094] The color of the smoke can reflect not only the fire, but also the type of combustible material. Specifically, when wood is burning, if the space is sufficient, the color of the smoke is white, and if the space is not sufficient, the wood is not burning sufficiently, and the smoke will be mixed with a large amount of carbon particles, and the color of the smoke will be black. When plastic or other polymer materials are burning, black smoke is usually produced. If white smoke is identified on the video monitoring image, it indicates that the burning is not intense, and the temperature is also relatively low. After a period of time, if the smoke is still white, it indicates that the burning material is wood, or that someone has used water to extinguish the fire. If black smoke is produced, it indicates that the combustible material is organic polymer material or insufficiently burning wood. The smoke produced by the burning of organic polymer materials is highly toxic, and insufficiently burning wood needs to be paid attention to the carbon monoxide contained therein.

[0095] The method of identifying the combustible material only according to the color of the smoke has low accuracy. Only when the wood or organic polymer material is far away from each other on the construction site, the identification can be accurate. Therefore, the embodiment provides a scheme for more accurately identifying the type of combustible material burned in a fire by combining the moving speed of the smoke, which includes:

[0096] Identifying the rising rate of the smoke in the video monitoring image;

[0097] Determining the type of combustible material producing the smoke according to the rising rate of the smoke and the color of the smoke.

[0098] On the other hand, the embodiment provides a method for identifying the rising rate of the smoke in the video monitoring image, which includes:

[0099] Reading a frame of video monitoring image, and intercepting a partial area image of the smoke;

[0100] After waiting for a preset time length, reading a frame of video monitoring image, and finding a corresponding area in the newly read video monitoring image, which is consistent with the partial area image;

[0101] Calculating the moving pixel distance of the corresponding area relative to the partial area image, and taking the quotient of the moving pixel distance and the preset time length as the rising rate of the smoke.

[0102] The moving distance of the smoke area within the preset time length is used to determine the rising rate of the smoke, and the change of the image in the intercepted partial area image of the smoke is small. Otherwise, after waiting for the preset time length, the corresponding area cannot be found. When the fire is large, the smoke rolls quickly, which can reduce the accuracy of the calculation of the rising rate. Further, the embodiment also provides another method for identifying the rising rate of the smoke in the video monitoring image, which is an alternative method, and includes:

[0103] Extracting a color area of a preset color range in the video monitoring image as a smoke area;

[0104] acquire a plurality of pixel values of each pixel point in the smoke area in a preset time length;

[0105] calculate the variance of the plurality of pixel values of each pixel point, and calculate the mean value of the variances of all the pixel points;

[0106] the product of the mean value and a preset coefficient is taken as the rising rate of the smoke.

[0107] The rising rate of the smoke is represented by calculating the speed of the change of the pixels in the smoke area, which has the advantages of fast calculation speed and high calculation accuracy. The rising rate of the smoke is divided into two intervals of slow and fast through a preset division interval.

[0108] When the smoke is white and the rising rate is slow, it indicates that the burning material is a non-high polymer material such as wood, and the burning is not intense. When the smoke is white and the rising rate is fast, it indicates that the burning material is wood, the burning is intense, and the burning is sufficient. When the smoke is black and the rising rate is slow, it indicates that the burning material is a high polymer material, and the burning is not intense. When the smoke is black and the rising rate is fast, it indicates that the burning material is a high polymer material and the burning is intense, or it is wood that burns intensely but not sufficiently. However, when the smoke color is gray, it indicates that the burning material is a mixture of wood and high polymer material.

[0109] In another embodiment, a different preset color range acquisition scheme is adopted. Please refer to the accompanying drawings Figure 4 The method for obtaining the preset color range comprises:

[0110] Step D01) reading a plurality of smoke images of historical fire conditions, and associating the smoke images with combustible material categories;

[0111] Step D02) eliminating the smoke images associated with a plurality of combustible material categories;

[0112] Step D03) extracting the color range of the smoke in each remaining smoke image respectively, and associating the color range with the combustible material category;

[0113] Step D04) the set of color ranges of all the smoke images constitutes the preset color range.

[0114] The smoke images obtained historically not only include the color information of the smoke, but also include the information of the combustible material categories. By associating the smoke images with the combustible material categories, the color range can be associated with the combustible material categories. Thus, the combustible material categories can be identified, and more abundant alarm information is provided for the fire condition. For a construction site, the combustible material categories are not many, and by extracting the smoke images of historical fire conditions, the preset color range with high accuracy of combustible material identification for the construction site can be obtained.

[0115] Please refer to the attached Figure 5 , the method for judging the smoke generation position comprises:

[0116] Step E01) obtaining the combustible species according to the color of the smoke on the video monitoring image;

[0117] Step E02) comparing with the types of combustibles on the construction site, obtaining the stacking position of the combustible species that meets the requirements as the presumed position of the smoke generation position;

[0118] Step E03) obtaining the pixel distribution area of the smoke on the video monitoring image, the bottom of the pixel distribution area as the observation position of the smoke generation position;

[0119] Step E04) both the presumed position and the observation position as the fire risk position.

[0120] By identifying the color range of the smoke, the inference of the combustible species can be realized. Although under actual fire, there may be multiple combustible species burning at the same time, and the color of the smoke generated after mixing changes. But because the combustibles are not always mixed and stacked, the smoke generated will not be completely mixed, and the combustible species can still be identified in the un-mixed area. The smoke spreads rapidly in the upward direction, forming a kind of inverted horn shape. By identifying the two side boundaries of the smoke in the video monitoring image, the intersection of the two side boundaries is the smoke generation position, that is, the accumulation point of the combustibles.

[0121] The embodiment provides a method for identifying the bottom of the smoke as the observation position, combining the stacking position of the combustibles on the site with the color of the smoke, and taking the stacking position of the combustibles whose color meets the requirements as the presumed position of the smoke generation position. Meanwhile, the observation position and the presumed position are provided as the fire risk position, which can better dispose the fire.

[0122] On the other hand, the embodiment also provides a method for identifying the coverage area of the smoke, which is executed after identifying the smoke generation position. Please refer to the attached Figure 6 , the method for identifying the coverage area of the smoke comprises:

[0123] Step F01) obtaining the pixel distribution area of the smoke on the video monitoring image, and counting the pixel area S of the pixel distribution area;

[0124] Step F02) obtaining the distance L between the smoke generation position and the monitor according to the smoke generation position and the position of the monitor;

[0125] Step F03) calculating the coverage area C of the smoke, C=k*S*L, k being a preset coefficient;

[0126] Step F04) comparing the coverage area C with a preset coverage area and preset risk level division scheme to obtain a risk level;

[0127] Step F05) the fire risk alarm includes a fire risk location and a risk level.

[0128] According to the smoke generation location and the location of the monitor, the distance L between the smoke generation location and the monitor is obtained. The greater the distance L is, the greater the smoke coverage area corresponding to the same pixel area is. Therefore, the calculation method provided in step F03) can more accurately calculate the coverage area of the smoke. The coverage area of the smoke calculated in the embodiment is not the real area of the smoke, but only the coverage area in the monitoring field of view of the monitor appearing on the video monitoring image. The size of the fire is represented by the coverage area on the video monitoring image. The greater the fire is, the more smoke it generates, and the faster the smoke rises.

[0129] On the other hand, the embodiment also provides a scheme for dividing the fire risk level according to the rising speed of the smoke. The scheme includes:

[0130] extracting a color region of a preset color range in the video monitoring image as a smoke region;

[0131] periodically collecting pixel values of each pixel point in the smoke region within a preset time length to obtain a plurality of pixel values of each pixel point at different times;

[0132] calculating the variance of the plurality of pixel values of each pixel point, and calculating the mean value of the variances of all pixel points;

[0133] The product of the mean value and a preset coefficient is taken as the rising speed of the smoke. By calculating the speed of the pixel change in the smoke region, the rising speed of the smoke is represented, which has the advantages of fast calculation speed and high calculation accuracy. The mean value of the variances is divided into intervals, and each interval corresponds to a risk level.

[0134] Specifically, the risk level division scheme in the embodiment includes a plurality of numerical intervals of coverage areas and corresponding risk levels. Please refer to the accompanying drawings for details. Figure 7 The method for obtaining the risk level division scheme includes:

[0135] Step G01) reading a plurality of smoke images of historical fires marked with fire levels;

[0136] Step G02) grouping the smoke images according to the fire levels;

[0137] Step G03) identifying the coverage areas in each group of smoke images and calculating the mean value as the coverage area corresponding to the fire level;

[0138] Step G04) taking the mean value of the coverage area corresponding to adjacent fire level as the demarcation value to obtain the risk level division scheme.

[0139] By labeling the smoke image of the historical fire, the smoke image of the historical fire is cut to match the size of the video monitoring image during labeling, which can make the fire level labeling more accurate. By setting the demarcation value, the problem of inaccurate division of fire level at the boundary of fire level is avoided. By taking the mean value of the coverage area corresponding to adjacent fire level as the demarcation value, better demarcation value can be obtained with the help of historical data, and the divided fire level is more scientific.

[0140] Please refer to the attached Figure 8 , the method for identifying the operation state of the construction equipment comprises:

[0141] Step H01) reading image templates of multiple working postures of the construction equipment, denoted as equipment templates;

[0142] Step H02) comparing the image in the running area in the video monitoring image with the equipment templates to obtain the working posture of the construction equipment;

[0143] Step H03) obtaining the working postures of the construction equipment in multiple continuous periods of video monitoring images;

[0144] Step H04) if the working postures of the construction equipment in the multiple continuous periods of video monitoring images are not all the same, it is determined that the construction equipment is in the running state, otherwise, it is determined that the construction equipment is in the non-running state.

[0145] Through the image templates of the working postures of the construction equipment, the state of the construction equipment in the video monitoring image can be matched. If the state of the construction equipment changes in multiple video monitoring images, it indicates that the construction equipment is currently working, otherwise, if the state of the construction equipment does not change in multiple video monitoring images, it indicates that the construction equipment is currently in a non-working state. For example, by identifying multiple postures of the mechanical arm of the excavator, it can be determined whether the excavator is currently in a working state. If the postures of the mechanical arm of the excavator are different in multiple video monitoring images, the excavator is in a working state. If the postures of the mechanical arm of the excavator are the same in multiple video monitoring images, the excavator is in a non-working state. When the excavator is in a non-working state, there is no safety risk in the working area of the excavator.

[0146] Please refer to the attached Figure 9 , the method for identifying the non-equipment operator comprises:

[0147] Step I01) comparing the image in the running area in the video monitoring image with the equipment templates to obtain the area occupied by the construction equipment, denoted as the equipment area;

[0148] Step I02) identifying the person in the operating area, if the person is outside the equipment area, determining that it is a non-equipment operator, otherwise, if the person is in the equipment area, determining that it is an equipment operator.

[0149] When the construction equipment is in working state, the non-equipment operator should keep away. For example, when the excavator is working, there should be no non-equipment operator in the working range of the excavator, otherwise there will be a safety risk. By comparing with the equipment template, the working state and position of the equipment can be obtained, and by the position and working radius of the equipment, the area occupied by the construction equipment, i.e. the equipment area, can be obtained.

[0150] On the other hand, in another embodiment, the risk identification and alarm method also includes a combustible fire prediction method, please refer to the attached Figure 10 The combustible fire prediction method includes:

[0151] Step J01) obtaining the wind direction of the construction site;

[0152] Step J02) judging whether the combustible material is located at the downwind of the smoke generation position according to the smoke generation position and the wind direction;

[0153] Step J03) if the combustible material is located at the downwind of the smoke generation position, issuing a combustible fire warning, the combustible fire warning including the type, stacking position and stacking amount of the combustible material, if there is no combustible material located at the downwind of the smoke generation position, no operation is performed.

[0154] The weather data is obtained through the network, which includes the wind direction of the construction site. The smoke generation position of the construction site is identified, if there is other combustible material stacking at the downwind of the smoke generation position, the stacked combustible material will be ignited with high probability, and the obtained spread will occur. Therefore, timely warning should be issued to prompt the development of targeted disposal measures to further improve the safety of the construction site.

[0155] In summary, the technical scheme provided by the embodiment can achieve the following beneficial technical effects: the video monitoring image obtained by the monitor, combined with smoke identification, can quickly identify the smoke appearing in the construction site and timely discover the fire occurring in the construction site; combined with the stacking condition of the combustible material in the construction site and the identified smoke generation position, the next step of the fire development can be predicted, which helps to develop a fire disposal scheme; by identifying the type of combustible material through smoke, the position of the fire risk occurrence is located, which helps to effectively dispose the fire; by identifying the running state of the construction equipment and the non-equipment operator in the working area of the construction equipment, the construction equipment running risk is identified and alarmed, and the safe operation of the construction equipment is ensured.

[0156] The above merely illustrates the specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and those skilled in the art should understand that the present application includes but is not limited to the contents described in the above specific embodiments and the drawings. Any modification not deviating from the functional and structural principles of the present application will be included in the scope of the claims.

Claims

1. A construction site safety risk identification alarm method, characterized by, The method comprises the following steps: receiving video monitoring images of a plurality of monitors at a construction site; obtaining an operating area of a construction equipment, and marking the operating area on each video monitoring image; reading data of types, stacking positions and stacking amounts of combustibles at the construction site; periodically updating video monitoring images of a plurality of angles at the construction site; respectively performing smoke identification and construction equipment operating state identification on each updated video monitoring image; if there is smoke on the video monitoring image, identifying a smoke generation position, taking the smoke generation position as a fire risk position, issuing a fire risk alarm, and proceeding to the next step; if there is no smoke on the video monitoring image, directly proceeding to the next step; if the construction equipment on the video monitoring image is in an operating state, identifying whether there is a non-equipment operator in the operating area of the construction equipment, and issuing a device operating risk alarm information if there is a non-equipment operator; the method for judging the smoke generation position comprises: obtaining a corresponding combustible type according to the color of the smoke on the video monitoring image; comparing with the types of combustibles at the construction site to obtain a stacking position of the combustible type that meets the requirement as a presumed position of the smoke generation position; obtaining a pixel distribution area of the smoke on the video monitoring image, and taking the bottom of the pixel distribution area as an observed position of the smoke generation position; both the presumed position and the observed position are taken as the fire risk position.

2. The construction site safety risk identification and alarm method according to claim 1, wherein the method for identifying smoke comprises: extracting a color region in a preset color range in the video monitoring image; if the pixel area covered by the color region exceeds a preset threshold, it is determined that there is smoke in the video monitoring image, otherwise, it is determined that there is no smoke in the video monitoring image.

3. The construction site safety risk identification and alarm method according to claim 2, wherein the method for obtaining a preset color range comprises: reading smoke images of a plurality of historical fires; extracting the color range of smoke in each smoke image respectively; the union of the color ranges of all smoke images constitutes the preset color range.

4. The construction site safety risk identification and alarm method according to claim 2, wherein the method for obtaining a preset color range comprises: reading smoke images of a plurality of historical fires, and associating the smoke images with combustible types; eliminating smoke images associated with multiple combustible types; extracting the color range of smoke in each remaining smoke image respectively, and associating the color range with the combustible type; the union of the color ranges of all smoke images constitutes the preset color range.

5. The construction site safety risk identification and alarm method according to any one of claims 1 to 4, wherein it further comprises a method for identifying the smoke coverage area, which is executed after identifying the smoke generation position, and the method for identifying the smoke coverage area comprises: obtaining a pixel distribution area of the smoke on the video monitoring image, and counting the pixel area S of the pixel distribution area; obtaining the distance L between the smoke generation position and the monitor according to the smoke generation position and the position of the monitor; The coverage area C of the smoke is calculated as C=k*S*L, where k is a preset coefficient; The coverage area C is compared with a preset coverage area and risk level division scheme to obtain the risk level; The fire risk alarm includes a fire risk position and a risk level.

6. The construction site safety risk identification and alarm method according to claim 5, wherein The risk level division scheme includes a plurality of coverage area numerical intervals and corresponding risk levels, and the method for obtaining the risk level division scheme includes: reading smoke images of historical fires of a plurality of fire levels; grouping the smoke images according to the fire levels; identifying the coverage areas in each group of smoke images and calculating the average values as the coverage areas corresponding to the fire levels; taking the average values of the coverage areas corresponding to adjacent fire levels as the dividing values to obtain the risk level division scheme.

7. The construction site safety risk identification and alarm method according to any one of claims 1 to 4, wherein The method for identifying the operation state of the construction equipment includes: reading image templates of a plurality of working postures of the construction equipment, denoted as equipment templates; comparing the images in the operation area in the video monitoring images with the equipment templates to obtain the working postures of the construction equipment; obtaining the working postures of the construction equipment in a plurality of continuous periods of video monitoring images; if the working postures of the construction equipment in the plurality of continuous periods of video monitoring images are not all the same, it is determined that the construction equipment is in an operation state, otherwise, it is determined that the construction equipment is in a non-operation state.

8. The construction site safety risk identification and alarm method according to claim 7, wherein The method for identifying non-equipment operators includes: comparing the images in the operation area in the video monitoring images with the equipment templates to obtain the area occupied by the construction equipment, denoted as the equipment area; identifying personnel in the operation area, and if the personnel are located outside the equipment area, it is determined that they are non-equipment operators, otherwise, if the personnel are located in the equipment area, it is determined that they are equipment operators.

9. The construction site safety risk identification and alarm method according to any one of claims 1 to 4, wherein The risk identification and alarm method further includes a combustible material fire prediction method, and the combustible material fire prediction method includes: obtaining the wind direction of the construction site; according to the smoke generation position and the wind direction, sequentially determining whether the combustible material is located at the downwind of the smoke generation position; if the combustible material is located at the downwind of the smoke generation position, issuing a combustible material fire warning, which includes the combustible material type, the stacking position, and the stacking amount, and if there is no combustible material located at the downwind of the smoke generation position, no operation is performed.

Images