A road dust dynamic monitoring method and system based on image
Obtaining dynamic monitoring data of road dust through image recognition technology solves the problem of time-consuming, labor-intensive and accurate traditional detection methods, realizing low-cost and high-precision road environment monitoring, and providing fast and reliable data support.
Patent Information
- Application Number
- CN202111388723.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-11-22
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2041-11-22
AI Technical Summary
The pollution source detection method in the traditional environmental protection industry requires a lot of manpower to conduct on-site inspections, which consumes time and effort, and the real-time detection accuracy is not high. The principle of equipment monitoring is far from the national standard, resulting in large data errors and high economic costs, and the inability to effectively evaluate the pollution source to air. Effects of quality.
By using image-based dynamic monitoring method of road dust, the road dust accumulation samples and multi-frame images, the dry/wet conditions of the road surface, the vehicle type and traffic flow, the hourly and daily emissions of road dust are calculated, and the daily emissions of road dust are obtained in segments. Particle concentration of linear sources.
It realizes low-cost and high-precision road environment monitoring, provides fast and reliable data support, reduces the demand for on-site monitoring, and improves the real-time monitoring capabilities of pollution sources.
Smart Images

Figure CN113936260B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of environmental monitoring, and in particular to an image-based road dust dynamic monitoring method and system. Background Art
[0002] At present, the complex and huge pollution sources require a lot of manpower for on-site inspection. Problems are not discovered in time and the problem is not handled in a closed loop. The traditional environmental protection industry uses multiple monitoring devices to obtain the concentration of pollution sources. Because the monitoring principle of the equipment is far from the national standard, the data error is large, the economic cost is high, and it cannot evaluate the impact of pollution sources on air quality. Summary of the invention
[0003] In view of the above problems existing in the prior art, the present invention proposes an image-based road dust dynamic monitoring method and system, which mainly solves the problem that traditional detection methods require time-consuming and labor-intensive on-site inspections and have low real-time detection accuracy.
[0004] In order to achieve the above purpose and other purposes, the technical solution adopted by the present invention is as follows.
[0005] A road dust dynamic monitoring method based on images, comprising:
[0006] Obtain dust samples of the road section to be monitored, and obtain the dust load of the road to be monitored according to the dust samples;
[0007] Acquire multiple frames of images of the road section to be detected, identify the dry / wet conditions of the road surface corresponding to the multiple frames of images, the types of vehicles included, and the traffic flow corresponding to each of the vehicle types, and determine the weight of vehicles passing through the road section to be detected within a corresponding time period according to the vehicle types and the corresponding traffic flow;
[0008] Calculate the hourly emission and / or daily emission of road dust of the road section to be monitored according to the road dust load, the dry / wet condition of the corresponding road surface and the weight of the vehicle;
[0009] The road section to be monitored is segmented to obtain one or more linear sources, and the particle concentration of each linear source at a specified point is obtained according to the hourly emission and / or daily emission.
[0010] Optionally, acquiring multiple frames of images of the road section to be detected, and identifying the vehicle types contained in the multiple frames of images and the traffic volume corresponding to each of the vehicle types, includes:
[0011] Acquire video images collected by each image collection point of the road section to be detected, obtain the multiple frames of images from the video images, detect vehicles contained in the multiple frames of images, and create a target chain;
[0012] Predicting the position of the target vehicle in the target chain, and obtaining the model and travel trajectory of the target vehicle;
[0013] The traffic volume of the corresponding vehicle type is obtained according to the intersection of the travel trajectory and the statistical area pre-demarcated in the road section to be detected.
[0014] Optionally, detecting vehicles contained in the multiple frames of images and creating a target chain includes:
[0015] The target vehicle contained in the current frame image is obtained through the detection algorithm, and an initial target chain is created according to the obtained target vehicle;
[0016] A tracking chain is created with the target vehicle in the initial target chain as the tracking target, and whether the next frame image contains the target vehicle in the tracking chain is detected, and the newly added target vehicle is incorporated into the tracking chain until the detection of all frame images is completed to obtain the target chain.
[0017] Optionally, predicting the position of the target vehicle in the target chain and obtaining the model and travel trajectory of the target vehicle includes:
[0018] Predicting the position information of each target vehicle in the target chain in the next frame of image, and connecting the position information of the target vehicle corresponding to the predicted position information reaching a set threshold with the position information of the target vehicle used for position prediction according to the similarity between the target vehicle corresponding to the predicted position information and the target vehicle used for position prediction, so as to form a moving track of the target vehicle;
[0019] The target chain is modified according to the similarity calculation result.
[0020] Optionally, obtaining the vehicle flow of the corresponding vehicle type according to the intersection of the travel trajectory and a statistical area pre-demarcated in the road section to be detected includes:
[0021] It is determined whether the moving trajectory of the target vehicle intersects with the statistical line of the statistical area. If so, the number of the corresponding target vehicles is increased by one, so as to obtain the traffic flow of the road section to be detected.
[0022] Optionally, the similarity calculation method includes: Hungarian algorithm.
[0023] Optionally, determining the weight of vehicles passing through the road section to be detected within a corresponding time period according to the vehicle type and the corresponding traffic flow includes:
[0024] Obtaining a preset vehicle weight of a corresponding vehicle type according to the vehicle type of the target vehicle;
[0025] The vehicle weight of the road section to be monitored is calculated based on the vehicle weight of each vehicle type.
[0026] Optionally, calculating the hourly emission and / or daily emission of road dust of the road section to be monitored according to the road dust load and the vehicle weight includes:
[0027] Calculating the road dust coefficient of the road section to be monitored according to the vehicle weight of the road section to be monitored and the road dust load;
[0028] The hourly emission and / or daily emission is obtained according to the road dust coefficient and the traffic volume of the road section to be monitored.
[0029] An image-based road dust dynamic monitoring system, comprising:
[0030] A dust load acquisition module is used to acquire dust samples of the road section to be monitored, and acquire the dust load of the road to be monitored according to the dust samples;
[0031] A vehicle flow recognition module, used to obtain multiple frames of images of the road section to be detected, identify the dry / wet conditions of the road surface corresponding to the multiple frames of images, the types of vehicles included, and the traffic flow corresponding to each of the vehicle types, and determine the weight of vehicles passing through the road section to be detected within a corresponding time period according to the vehicle types and the corresponding traffic flow;
[0032] An emission calculation module, used to calculate the hourly emission and / or daily emission of road dust of the road section to be monitored according to the road dust load, the dryness / wetness of the corresponding road surface and the weight of the vehicle;
[0033] The line source impact assessment module is used to segment the road section to be monitored to obtain one or more straight line sources, and obtain the particle concentration of each straight line source at a specified point according to the hourly emission and / or daily emission.
[0034] As described above, the image-based road dust dynamic monitoring method and system of the present invention have the following beneficial effects.
[0035] Through image recognition, information such as vehicle model and traffic flow is obtained, and the emissions of the corresponding road section to be tested are calculated and updated in real time based on the recognition data, achieving low-cost, high-precision, and fast supervision, providing fast and reliable data support for road environment monitoring. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 The figure is a flow chart of a method for dynamic monitoring of road dust based on images in one embodiment of the present invention.
[0037] Figure 2 The figure is a flow chart of vehicle flow detection in one embodiment of the present invention.
[0038] Figure 3FIG. 4 is a schematic diagram of line source cutting in one embodiment of the present invention.
[0039] Figure 4 Schematic diagram of the structure of an image-based road dust dynamic monitoring system in one embodiment of the present invention. DETAILED DESCRIPTION
[0040] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.
[0041] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.
[0042] See also Figure 1 The present invention provides an image-based road dust dynamic monitoring method, which includes the following steps.
[0043] Step S01, obtaining dust samples of a road section to be monitored, and obtaining the dust load of the road to be monitored according to the dust samples;
[0044] Step S02, acquiring multiple frames of images of the road section to be detected, identifying the dry / wet conditions of the road surface corresponding to the multiple frames of images, the types of vehicles included, and the traffic flow corresponding to each of the vehicle types, and determining the weight of vehicles passing through the road section to be detected within a corresponding time period according to the vehicle types and the corresponding traffic flow;
[0045] Step S03, calculating the hourly emission and / or daily emission of road dust of the road section to be monitored according to the road dust load, the dryness / wetness of the corresponding road surface and the weight of the vehicle;
[0046] Step S04, segmenting the road section to be monitored to obtain one or more linear sources, and obtaining the particle concentration of each linear source at a designated point according to the hourly emission and / or daily emission.
[0047] In one embodiment, when obtaining the road dust load, on-site sampling is required, and the specific process is as follows:
[0048] First, mark out a sampling area with a certain width and calculate the sampling area. The cumulative sampling area of each dust sample is recorded as S. Use a vacuum cleaner to sweep the dust on the road for sampling. After sampling, put the dust bag of the vacuum cleaner into a sealed bag.
[0049] At 105°C, a single road dust sample was dried to constant weight, cooled in a desiccator, and recorded as W0 (unit: g);
[0050] Use a soft brush or high-pressure air flow to clean the sieve to prevent it from sticking with debris;
[0051] Place 20-mesh and 200-mesh sieves stacked from top to bottom in a shaking table, weigh 30-100 g of road dust, put it into the 20-mesh sieve, and record the weight as W (unit: g);
[0052] Close the shaker and start the shaking operation for 10 minutes. Take out and weigh the material on the 20-mesh sieve, which is recorded as W20 (unit: g).
[0053] The 200-mesh sieve material was shaken in a shaker until the weight difference between the two weights of the 200-mesh sieve material was less than 3%, and the total time was no more than 40 minutes. The 200-mesh sieve material was taken out and weighed, recorded as W200 (unit: g).
[0054] The calculation method of road dust load SL can be expressed as:
[0055]
[0056] The above only shows one way of obtaining the road dust load, and other obtaining ways may also be used, which are not limited here.
[0057] In one embodiment, acquiring multiple frames of images of the road section to be detected, and identifying the vehicle types contained in the multiple frames of images and the traffic volume corresponding to each of the vehicle types, includes:
[0058] Acquire video images collected by each image collection point of the road section to be detected, obtain the multiple frames of images from the video images, detect vehicles contained in the multiple frames of images, and create a target chain;
[0059] Predicting the position of the target vehicle in the target chain, and obtaining the model and travel trajectory of the target vehicle;
[0060] The traffic volume of the corresponding vehicle type is obtained according to the intersection of the travel trajectory and the statistical area pre-demarcated in the road section to be detected.
[0061] In one embodiment, detecting vehicles contained in the multiple frames of images and creating a target chain includes:
[0062] The target vehicle contained in the current frame image is obtained through the detection algorithm, and an initial target chain is created according to the obtained target vehicle;
[0063] A tracking chain is created with the target vehicle in the initial target chain as the tracking target, and whether the next frame image contains the target vehicle in the tracking chain is detected, and the newly added target vehicle is incorporated into the tracking chain until the detection of all frame images is completed to obtain the target chain.
[0064] In one embodiment, predicting the position of the target vehicle in the target chain and obtaining the model and travel trajectory of the target vehicle include:
[0065] Predicting the position information of each target vehicle in the target chain in the next frame of image, and connecting the position information of the target vehicle corresponding to the predicted position information reaching a set threshold with the position information of the target vehicle used for position prediction according to the similarity between the target vehicle corresponding to the predicted position information and the target vehicle used for position prediction, so as to form a moving track of the target vehicle;
[0066] The target chain is modified according to the similarity calculation result.
[0067] Specifically, see Figure 2 ,The specific process of detecting the traffic flow of the road section to be monitored is as follows:
[0068] Step 1, select a road video / camera, and obtain a video image of the corresponding road section through one or more monitoring devices set up on the road section to be monitored;
[0069] Step 2, target detection, can be performed on the collected video image through a detection algorithm such as yolov5 to obtain the target vehicle in the video image. Specifically, multiple frames of images can be obtained from the collected video for target detection. When the detection algorithm detects multiple target vehicles from the current frame, the detected target vehicles are used to create an initial target chain, and the target vehicles in the initial target chain are used as tracking targets to build a tracking chain. Continue to detect the target vehicle in the next frame of the image through the detection algorithm, and determine whether the next frame of the image contains the target vehicle in the tracking chain. If it does, the tracking target is not empty, and if it does not, the tracking target is empty.
[0070] Step 3, when the tracking target in the target chain is not empty, the position of the target vehicle in the target chain in the next frame image can be predicted by Kalman filtering. The specific prediction process is prior art and will not be repeated here.
[0071] Step 4, obtain the target vehicle corresponding to the predicted position in the next frame image, and further calculate the similarity between the target vehicles in the two frames of images. Specifically, the Hungarian algorithm can be used to calculate the similarity of the target objects in the two frames of images. If there is no match, the unmatched tracking target is deleted; if there is a match, the matching target vehicle information is updated to the tracking chain, and the position information of the matching target vehicle is connected to form the target and the route of the quantity. At the same time, the vehicle type of the tracking target can be identified, and the route of the target vehicle of the corresponding vehicle type can be recorded. The specific vehicle type identification process is a prior art and will not be repeated here.
[0072] In step 5, the target vehicle's trajectory can be further corrected by Kalman filtering to obtain the connection line of the vehicle's front and rear two-frame tracking trajectories.
[0073] Step 6, the statistical area of the road section to be monitored can be pre-demarcated, and the area boundary can be marked in the image by the statistical line. It is further determined whether the trajectory line of the target vehicle intersects with the statistical line. Specifically, the intersection judgment can be performed by the cross multiplication method. The specific process is not repeated here. If it intersects with the statistical line, the number of target vehicles of the corresponding model is increased by 1, and the statistical number of various models of the road section to be monitored is output in sequence to obtain the traffic flow in the corresponding time period.
[0074] In one embodiment, the dry / wet condition of the road surface of the road section to be monitored in the collected real-time image can also be identified by a pre-trained recognition model. The recognition model can use a component model framework such as a recurrent neural network and a convolutional neural network, collect images containing various dry and wet road scenes to construct a training sample set, and annotate the dry and wet conditions of the corresponding images in the training sample set, further construct a logistic regression loss function, and perform model and training based on the annotated training samples to obtain a dry / wet condition recognition model. The logistic regression loss function can be expressed as:
[0075]
[0076] Among them, θ is the parameter to be estimated, i is the training sample, and y i Represents the construction vector, and the training objective can be set to make the predicted probability distribution As close as possible to the true label probability distribution y i The specific model training process is an existing technology and will not be described here.
[0077] For paved roads, the road dust emission coefficient can be expressed as:
[0078] E Ri =K i ×(sL) 0.91 ×(W) 1.02 (1-η)
[0079] 1) E RiIt is the PMi emission coefficient of dust on paved roads, and its unit is g / VKT (indicating the mass of particulate matter generated by a motor vehicle traveling 1 km).
[0080] 2) K i It is the particle size multiplier of PMi in the generated dust.
[0081] 3) SL is the road dust load.
[0082] 4) W is the average vehicle weight in tons. The average vehicle weight represents the average weight of all vehicles passing through a certain level of road. By identifying the vehicle weight information of each vehicle type, the average vehicle weight of the monitored section can be calculated based on the traffic volume of the monitored section and the number of different vehicle types counted in the previous steps.
[0083] (1) Hourly emission estimation
[0084] W Ri =E Ri ×L R ×N R ×(1-n r ) Formula (4)
[0085] Where W Ri PM is particulate matter in road dust sources i Total emissions, g / h.
[0086] E Ri PM in road dust sources i Average emission coefficient, g / (km·vehicle).
[0087] L R is the road length, km.
[0088] N R It is the average traffic volume on this road section during a certain period, vehicles / h.
[0089] n r is the dust-free coefficient. When the rainfall is greater than 1 mm / h, n r Take 1; when the rainfall is less than or equal to 1mm / h, n r Take 0. Obtain the corresponding rainfall according to the dry / wet conditions of the road surface and calculate the real-time emission.
[0090] (2) Daily emission estimation
[0091]
[0092] Where W Ri PM is particulate matter in road dust sources i Total emissions, unit: g / d.
[0093] ERi PM in road dust sources i Average emission coefficient, unit: g / (km·vehicle).
[0094] L R is the road length, in km.
[0095] N R It is the average traffic volume on the corresponding road over a period of time, with the unit of vehicles / day.
[0096] n r To express the number of dust-free hours, the number of hours when the daily precipitation is greater than 1 mm / h can be used.
[0097] According to the above formula, real-time output of vehicle data is used as the key factor, and real-time emissions are output every hour.
[0098] Furthermore, the line source algorithm is optimized based on the principle of point source diffusion.
[0099] like Figure 3 As shown, the curved line source is cut into a long straight line source at the inflection point. Assume that the angle between the wind direction and the line source AB is θ, the line source length is L (m), and the line source strength is Q l (mg / m·s). Take the midpoint O of the line source as the origin and establish a rectangular coordinate system with the x-axis consistent with the wind direction. Choose the y-axis direction appropriately so that the line source is in the first and third quadrants, then θ is strictly limited to between 0° and 90°. The coordinates of all points on OB are positive, and the coordinates of all points on OA are negative.
[0100] The basic formula is as follows:
[0101]
[0102] Q l : Line source strength, mg / m·s.
[0103] u: average wind speed, m / s.
[0104] b: vertical diffusion parameter σ z The coefficient of .
[0105] q: vertical diffusion parameter σ z The index of .
[0106] p: lateral diffusion parameter σ y The index of .
[0107] θ: The angle between wind direction and line source, 0°≤θ≤90°.
[0108] x0: horizontal coordinate of the affected point, m.
[0109] |d|: The vertical distance between the affected point and the line source, m.
[0110] k: constant.
[0111] Φ(t): normal distribution function
[0112] With real-time output of vehicle data as the key factor, real-time emissions are output every hour. The real-time emissions can be calibrated in the collected images and updated dynamically. At the same time, the impact of pollution source emissions at the location of the monitoring equipment on the ambient air automatic monitoring station (environmental monitoring concentration C) is output in combination with information such as wind speed.
[0113] See also Figure 4 This embodiment provides an image-based road dust dynamic monitoring system, which is used to execute the image-based road dust dynamic monitoring method described in the aforementioned method embodiment. Since the technical principle of the system embodiment is similar to the technical principle of the aforementioned method embodiment, the same technical details will not be repeated.
[0114] In one embodiment, an image-based dynamic monitoring system for road dust is characterized in that it includes: a dust load acquisition module 10, which is used to obtain dust samples of a road section to be monitored, and obtain the road dust load to be monitored based on the dust samples; a vehicle flow recognition module 11, which is used to obtain multiple frames of images of the road section to be detected, identify the vehicle types contained in the multiple frames of images and the vehicle flow corresponding to each of the vehicle types, and determine the weight of vehicles passing through the road section to be detected within a corresponding time period based on the vehicle types and the corresponding vehicle flow; an emission calculation module 12, which is used to calculate the hourly emission and / or daily emission of road dust of the road section to be monitored based on the road dust load and the vehicle weight; a line source impact assessment module 13, which is used to determine one or more straight line sources contained in the road section to be monitored, and obtain the dust concentration of each straight line source at a specified point based on the hourly emission and / or daily emission.
[0115] In summary, the present invention provides an image-based road dust dynamic monitoring method and system. Through real-time image recognition, the association between pollution source emissions and images can be established to form a dynamic, spatially distributed atmospheric pollution source hourly automatic emission system, providing support for precise forecast consultation and precise pollution control; AI image and dynamic pollution source emission identification algorithm can replace the traditional environmental instrument monitoring concentration with the iterative algorithm of AI image, quantitatively output the transmission amount of the pollution source on the atmospheric environment air index (PM10, PM2.5), and can show the difference in the contribution of pollutants at different times and in different scenes at the same spatial scale, providing scientific and technological support for the application of pollution source dynamic control technology, effect evaluation, etc.; according to the image detection, whether the corresponding road section to be detected is a dry scene or a wet scene, and further according to the dry and wet conditions of the scene, the real-time dynamic emission source calculation is performed, and the pollution emission generated at any time can be obtained, realizing dynamic emission monitoring based on images; at the same time, the particle concentration (PM10, PM2.5) of the line source is obtained according to the dynamic emission source calculation, and the off-site remote dynamic emission monitoring is realized by combining image recognition, meteorological data and ambient air quality. Therefore, the present invention effectively overcomes various shortcomings in the prior art and has high industrial utilization value.
[0116] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.
Claims
1. A dynamic monitoring method for road dust based on images, It is characterized in that include: Obtain dust samples of the road section to be monitored, and obtain the dust load of the road to be monitored according to the dust samples; Acquire multiple frames of images of the road section to be monitored, identify the dry / wet conditions of the road surface corresponding to the multiple frames of images, the types of vehicles included, and the traffic flow corresponding to each of the vehicle types, and determine the weight of vehicles passing through the road section to be monitored within a corresponding time period according to the vehicle types and the corresponding traffic flow; Acquire multiple frames of images of the road section to be monitored, identify the vehicle types contained in the multiple frames of images and the traffic flow corresponding to each vehicle type, including: acquire video images collected by each image acquisition point of the road section to be monitored, acquire the multiple frames of images from the video images, detect the vehicles contained in the multiple frames of images, and create a target chain; predict the position of the target vehicle in the target chain, acquire the model and travel trajectory of the target vehicle; acquire the traffic flow of the corresponding model according to the intersection of the travel trajectory and the statistical area pre-demarcated in the road section to be monitored; Calculate the hourly emission and / or daily emission of road dust of the road section to be monitored according to the road dust load, the dry / wet condition of the corresponding road surface and the weight of the vehicle; The road section to be monitored is segmented to obtain one or more linear sources, and the particle concentration of each linear source at a designated point is obtained according to the hourly emission and / or daily emission.
2. The method for dynamic monitoring of road dust based on images according to claim 1, It is characterized in that Detecting vehicles contained in the multiple frames of images and creating a target chain includes: The target vehicle contained in the current frame image is obtained through a detection algorithm, and an initial target chain is created based on the obtained target vehicle; A tracking chain is created with the target vehicle in the initial target chain as the tracking target, and whether the next frame image contains the target vehicle in the tracking chain is detected, and the newly added target vehicle is incorporated into the tracking chain until the detection of all frame images is completed to obtain the target chain.
3. The method for dynamic monitoring of road dust based on images according to claim 1, It is characterized in that Predicting the position of the target vehicle in the target chain and obtaining the model and travel trajectory of the target vehicle include: Predicting the position information of each target vehicle in the target chain in the next frame of image, and connecting the position information of the target vehicle corresponding to the predicted position information reaching a set threshold with the position information of the target vehicle used for position prediction according to the similarity between the target vehicle corresponding to the predicted position information and the target vehicle used for position prediction, so as to form a moving track of the target vehicle; The target chain is modified according to the similarity calculation result.
4. The method for dynamic monitoring of road dust based on images according to claim 1, It is characterized in that According to the intersection of the travel trajectory and the statistical area pre-demarcated in the road section to be monitored, the vehicle flow of the corresponding vehicle model is obtained, including: It is determined whether the moving trajectory of the target vehicle intersects with the statistical line of the statistical area. If so, the number of the corresponding target vehicles is increased by one, so as to obtain the traffic flow of the road section to be monitored.
5. The method for dynamic monitoring of road dust based on images according to claim 3, It is characterized in that The similarity calculation method includes: Hungarian algorithm.
6. The method for dynamic monitoring of road dust based on images according to claim 1, It is characterized in that And determining the weight of vehicles passing through the road section to be monitored within a corresponding time period according to the vehicle type and the corresponding traffic flow, including: Obtaining a preset vehicle weight of a corresponding vehicle type according to the vehicle type of the target vehicle; The vehicle weight of the road section to be monitored is calculated based on the vehicle weight of each vehicle type.
7. The method for dynamic monitoring of road dust based on images according to claim 1, It is characterized in that Calculating the hourly emission and / or daily emission of road dust on the road section to be monitored according to the road dust load and the vehicle weight includes: Calculating the road dust coefficient of the road section to be monitored according to the vehicle weight of the road section to be monitored and the road dust load; The hourly emission and / or daily emission is obtained according to the road dust coefficient and the traffic volume of the road section to be monitored.
8. A road dust dynamic monitoring system based on images, It is characterized in that include: A dust load acquisition module is used to acquire dust samples of the road section to be monitored, and acquire the dust load of the road to be monitored according to the dust samples; A vehicle flow recognition module, used to obtain multiple frames of images of the road section to be monitored, identify the dry / wet conditions of the road surface corresponding to the multiple frames of images, the types of vehicles included, and the traffic flow corresponding to each of the vehicle types, and determine the weight of vehicles passing through the road section to be monitored within a corresponding time period according to the vehicle types and the corresponding traffic flow; Acquire multiple frames of images of the road section to be monitored, identify the vehicle types contained in the multiple frames of images and the traffic flow corresponding to each vehicle type, including: acquire video images collected by each image acquisition point of the road section to be monitored, acquire the multiple frames of images from the video images, detect the vehicles contained in the multiple frames of images, and create a target chain; predict the position of the target vehicle in the target chain, acquire the model and travel trajectory of the target vehicle; acquire the traffic flow of the corresponding model according to the intersection of the travel trajectory and the statistical area pre-demarcated in the road section to be monitored; An emission calculation module, used to calculate the hourly emission and / or daily emission of road dust of the road section to be monitored according to the road dust load, the dryness / wetness of the corresponding road surface and the weight of the vehicle; The line source impact assessment module is used to segment the road section to be monitored to obtain one or more straight line sources, and obtain the particle concentration of each straight line source at a specified point according to the hourly emission and / or daily emission.
