A waterlogging water depth calculation method and system and a readable storage medium

Abstract

The application provides a waterlogging depth calculation method, system and readable storage medium. The method comprises the following steps: calibrating a plurality of entity scales at a specific position in a city along a road in advance; acquiring a road monitoring image of the entity scale, and determining position information of the entity scale in the road monitoring image; creating a perspective projection model in combination with the position information, and constructing a virtual scale; determining a virtual scale pixel height of any point on the road surface in combination with the virtual scale; applying a machine vision model to identify a waterlogging area of the road monitoring image, and segmenting to obtain a boundary range of the waterlogging area; constructing a road line in the model in combination with a boundary position of a road area outside the waterlogging boundary range; determining a road surface area according to an extension direction of the road line and boundary information of the waterlogging area; constructing a virtual scale at a waterlogging monitoring point according to the virtual scale pixel height, and determining a target waterlogging depth according to a height difference between intersection points of the virtual scale and a road surface and a waterlogging surface.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of waterlogging depth monitoring, in particular to a waterlogging depth calculation method and system and a readable storage medium. BACKGROUND

[0002] At present, the methods for measuring waterlogging depth mainly include: using water level line image to extract hierarchical scale data for early warning, using water level sensor to monitor water level, and using meteorological hydrological model to simulate runoff process. However, these methods still have the following defects in practical use: (1) The first method requires specific images with water level lines, but only a few places such as rivers, reservoirs or water conservancy facilities have water level lines, and cannot be applied on a large scale. (2) Waterlogging sensor is a good water level monitoring device that can accurately monitor water level; however, waterlogging sensor equipment is complex and costly, and cannot cover the entire city. (3) Meteorological hydrological model is a commonly used method for simulating waterlogging depth, but it is easily limited by data. Therefore, these shortcomings have a great influence on the accuracy and precision of the calculation results.

[0003] In recent years, artificial intelligence has developed rapidly, and now deep learning methods can be used to assist in obtaining waterlogging area information and even depth information from existing data sources such as road monitoring. For example, (1) The invention patent "Urban waterlogging area monitoring method based on deep learning technology" applied by Xi'an University of Technology proposes a road waterlogging area calculation idea that inputs the monitoring image after perspective transformation to obtain waterlogging area recognition results and obtains the total area of waterlogging area through a certain amount of linear operation under the condition of training waterlogging area instance segmentation model, which brings inspiration for calculating various road waterlogging quantity attributes, but the actual size of the waterlogging area needs to be measured on site, which is difficult to ensure real-time and rapid implementation. (2) The invention patent "Urban waterlogging depth monitoring method based on deep learning" applied by Yangzhou University proposes a method of using a deep learning model to detect a car tire in a video, using the tire as a ruler to measure waterlogging depth, and then using a formula to calculate the waterlogging depth in the traffic monitoring video data, which points out a method of using a ruler to indicate the actual depth and avoiding manual measurement. This method has the advantages of economy, efficiency and speed in calculating road waterlogging depth, but it is also subject to the dynamic characteristics of the tire in the monitoring video, and there are problems such as fluctuation of recognition results and need for improvement of accuracy.

[0004] Therefore, considering some existing deep learning-based urban road waterlogging depth monitoring methods, it is necessary to use dynamic information such as tires in monitoring images and other data sources, which leads to a great extent that the calculation of waterlogging depth is associated with the position of the tire, and there are certain limitations in processing, and the calculation accuracy is not high. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide an inland waterlogging water depth calculation method, system and readable storage medium, which can improve the calculation accuracy.

[0006] The embodiments of the present application also provide an inland waterlogging water depth calculation method, comprising the following steps:

[0007] S1, a plurality of entity scales are calibrated along the road at a specific location in the city in advance;

[0008] S2, the road monitoring image of the entity scale is obtained, and the position information of each entity scale in the road monitoring image is determined;

[0009] S3, a corresponding perspective projection model is created in combination with the position information of each entity scale, and a virtual scale is constructed;

[0010] S4, the virtual scale is combined to determine the virtual scale pixel height of any point on the road surface;

[0011] S5, a machine vision model is applied to identify the waterlogging area in the road monitoring image, and the waterlogging area boundary range is segmented;

[0012] S6, in the road area outside the waterlogging boundary range, a corresponding road line is constructed in the model in combination with the boundary position of the road area;

[0013] S7, according to the extension direction of the road line and the boundary information of the waterlogging area, the road surface area corresponding to the waterlogging area is determined;

[0014] S8, according to the virtual scale pixel height, a corresponding virtual scale is constructed at the waterlogging monitoring point, and the target waterlogging depth is determined according to the height difference between the intersection of the virtual scale and the road surface and the water surface.

[0015] In a second aspect, the embodiments of the present application also provide an inland waterlogging water depth calculation system, which comprises an entity calibration module, a scale positioning module, a model creation module, a virtual calibration module, a waterlogging area identification module, a road line construction module, a road surface area identification module, and a waterlogging depth calculation module, wherein:

[0016] The entity calibration module is used for calibrating a plurality of entity scales along the road at a specific location in the city in advance;

[0017] The scale positioning module is used for obtaining the road monitoring image of the entity scale, and determining the position information of each entity scale in the road monitoring image;

[0018] The model creating module is configured to create a corresponding perspective projection model in combination with the position information of each entity scale, and to construct a virtual scale;

[0019] The virtual calibration module is configured to determine a virtual scale pixel height of any point on the road surface in combination with the virtual scale;

[0020] The accumulated water area identification module is configured to apply a machine vision model to identify an accumulated water area in a road monitoring image, and to segment to obtain a boundary range of the accumulated water area;

[0021] The road line construction module is configured to construct a corresponding road line in the model in combination with the boundary position of the road area outside the accumulated water boundary range;

[0022] The road surface area identification module is configured to determine a road surface area corresponding to the accumulated water area according to the extension direction of the road line and the boundary information of the accumulated water area;

[0023] The accumulated water depth calculation module is configured to construct a corresponding virtual scale at an accumulated water monitoring point according to the virtual scale pixel height, and to determine a target accumulated water depth according to a height difference between the virtual scale and the intersection between the road surface and the accumulated water surface.

[0024] In a third aspect, the embodiments of the present application further provide a readable storage medium, wherein the readable storage medium comprises a waterlogging accumulated water depth calculation method program, and the waterlogging accumulated water depth calculation method program is executed by a processor to implement the steps of the waterlogging accumulated water depth calculation method according to any one of the above.

[0025] As can be seen from the above, the waterlogging accumulated water depth calculation method, system and readable storage medium provided by the embodiments of the present application can determine the intersection of the accumulated water area and the virtual scale in real time, and calculate the accumulated water depth at the position based on the position of the intersection, thereby providing first-hand data for subsequent monitoring of the risk level of the accumulated water depth and making a risk warning. The embodiments of the present application take advantage of the fixed characteristics of the monitoring scene, combine the road information of the same scene when there is no accumulated water, construct the virtual scale, and can be applied to the monitoring of the accumulated water depth under different rainfall conditions, thereby effectively expanding the monitoring range and improving the monitoring efficiency and accuracy. The embodiments of the present application use the entity scale which is relatively easy to set up to determine the actual height in the monitoring image, thereby effectively avoiding the problem of large monitoring fluctuation and insufficient accuracy caused by using other dynamic objects as the scale, and improving the calculation accuracy.

[0026] Other features and advantages of the present application will be set forth in the following description, and in part will be apparent from the description, or can be learned by practice of the application. The objects and other advantages of the present application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings. BRIEF DESCRIPTION OF DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly introduced as follows. It should be understood that the following drawings only show some of the embodiments of the present application, and therefore should not be regarded as a limitation on the scope, and for those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0028] Figure 1 A flow chart of a waterlogging water depth calculation method provided by the embodiments of the present application;

[0029] Figure 2 A schematic diagram of a one-point perspective projection model of a road;

[0030] Figure 3 A schematic diagram of a two-point perspective projection model of a road;

[0031] Figure 4 A schematic diagram of determining the height difference of the intersection of the road surface and the water surface on the image;

[0032] Figure 5 A schematic diagram of determining the extension direction of the road according to the lane line obtained by segmentation;

[0033] Figure 6 A structural schematic diagram of a waterlogging water depth calculation system provided by the embodiments of the present application. DETAILED DESCRIPTION

[0034] The technical solutions of the embodiments of the present application will be described clearly and completely in the embodiments of the present application in combination with the drawings. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. The components of the embodiments of the present application described and shown in the drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of the present application.

[0035] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0036] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a method for calculating the depth of urban flooding in some embodiments of this application. The method includes the following steps:

[0037] Step S1: Mark multiple physical rulers along the road at specific locations in the city in advance.

[0038] Step S2: Obtain road monitoring images of physical rulers and determine the position information of each physical ruler in the road monitoring images.

[0039] Step S3: Combine the position information of each physical ruler to create a corresponding perspective projection model and construct a virtual ruler.

[0040] Step S4: Combine the virtual ruler to determine the height of the virtual ruler scale pixels at any point on the road surface.

[0041] Step S5: Apply a machine vision model to identify waterlogged areas in the road monitoring image and segment the boundary range of the waterlogged areas.

[0042] Step S6: In the road area outside the waterlogged boundary, construct the corresponding road lines in the model based on the boundary position of the road area.

[0043] Step S7: Determine the road surface area corresponding to the water accumulation area based on the extension direction of the road line and the boundary information of the water accumulation area.

[0044] Step S8: Based on the height of the virtual ruler's scale pixels, construct a corresponding virtual ruler at the water accumulation monitoring point, and determine the target water accumulation depth based on the height difference between the intersection of the virtual ruler and the road surface and the water accumulation surface.

[0045] From the above, the waterlogging water depth calculation method disclosed in the application can determine the intersection of the waterlogging area and the virtual scale in real time, calculate the waterlogging depth at the position based on the position of the intersection, provide first-hand data for subsequent monitoring of the risk level of the waterlogging depth and making risk warnings, utilize the fixed characteristics of the monitoring scene, combine the road information of the same scene when there is no waterlogging to realize the construction of the virtual scale, and can be applied to waterlogging depth monitoring under different rainfall conditions, can effectively expand the monitoring range, and improve the monitoring efficiency and accuracy. The actual height in the monitoring image is determined by using the entity scale which is relatively easy to set up, effectively avoiding the problem of large monitoring fluctuation and insufficient accuracy caused by using other dynamic objects as a scale, and improving the calculation accuracy.

[0046] In one of the embodiments, in step S2, the position information of each entity scale in the road monitoring image is determined, including:

[0047] Step S21, determine the target road monitoring video corresponding to the specific position of the city and pre-marked with multiple entity scales.

[0048] Step S22, according to the preset image segmentation rule, cut out multiple road monitoring images from the target road monitoring video.

[0049] Specifically, in the current implementation step, the required road monitoring image is cut out from the target road monitoring video according to the image segmentation rule of cutting frame by frame. Of course, different embodiments are not limited to the above-mentioned image segmentation method, which can be dynamically adjusted according to the actual situation.

[0050] Step S23, call the pre-trained target detection model to identify the entity scale from the multiple road monitoring images, and determine the pixel coordinates and pixel height of each entity scale in the road monitoring image based on the upper and lower edge pixel coordinates of the identification frame marked in the identification process.

[0051] Specifically, the training step of the above-mentioned target detection model includes:

[0052] (1) Obtain the target monitoring image to be identified, and mark the scale set in the target monitoring image to obtain the corresponding annotation data.

[0053] It should be noted that in actual application, the annotation data can be added to the preset txt file to facilitate subsequent data calling.

[0054] (2) Construct an initial detection model, and input the target monitoring image as training data into the initial detection model for model training. During the training process, based on the deviation value between the corresponding associated labeled data and the recognition data output by the model, it is judged whether the training end condition is reached.

[0055] It should be noted that in actual application, the training data created at present, the labels set in advance and the verification data can be added to the preset images folder and labels folder respectively. Among them, the two folders are divided into training set train and verification set val, and the pictures and labels maintain a one-to-one correspondence.

[0056] (3) After the training is completed, the target detection model for stably recognizing the human-set ruler from the monitoring image can be obtained.

[0057] The above embodiment, before the model training, the characteristics of the data set are found, or the results of the data processing and data enhancement operation are viewed in advance, and various configurations are continuously adjusted according to the data analysis results for training. At the same time, after combining the data loading debugging function, the influence of the data operation on the data set can be more deeply analyzed, the algorithm model analysis accuracy is further provided, and the execution effect is improved.

[0058] In one of the embodiments, in step S21, the target road monitoring video corresponding to a specific position of a city and having a plurality of entity rulers pre-set along the road signs is determined, comprising:

[0059] Step S211, acquiring an initial road monitoring video corresponding to a specific position of a city, and identifying a road straight section from the initial road monitoring video.

[0060] Step S212, when the corresponding road straight section is identified, a plurality of entity rulers with clear height information are set along the boundary of the road straight section, so that the entity rulers appear in the monitoring video range of the road, and the target road monitoring video required is determined.

[0061] Specifically, after identifying the road straight section in the monitoring range, a plurality of entity rulers (such as traffic barrels, road poles, etc.) with clear height information can be set along the road boundary, so that the entity rulers appear in the road monitoring video range. It should be noted that the aforementioned entity ruler can be understood as a "height-known indicating object", which needs to be set along the road in the actual scene.

[0062] The above embodiment uses entity rulers which are easy to set to determine the actual height of the monitoring image, effectively avoiding the problem of large monitoring fluctuation and insufficient accuracy caused by using other dynamic objects as rulers.

[0063] In one embodiment, there are multiple vanishing points in the created perspective projection model, and the vanishing points are determined according to the parallel lines represented by each of the entity scales, and the aggregated point is aggregated towards the horizon.

[0064] Please refer to Figures 2-3 According to the extension direction of the road, one-point perspective projection model or two-point perspective projection model can be optionally constructed. For example, when the extension direction of the road is determined to be substantially consistent with the screen of the computer device, one-point perspective projection model can be constructed (as shown in Figure 2 On the contrary, two-point perspective projection model can be constructed (as shown in Figure 3 .

[0065] It should be noted that the principle that parallel lines can be aggregated at the vanishing point in reality is also based on the phenomenon observed by the naked eye. For example, two tracks of a railway actually seem to converge at the horizon. There can be one or two vanishing points in the picture, depending on the coordinate position and direction of the composition; all vanishing points can fall on the horizon, or on the extension line outside the picture plane.

[0066] In the above embodiment, the stability of the perspective model is higher for calculating the water depth using dynamic entity scales, and additional errors caused by large fluctuations due to movement of objects in monitoring are avoided.

[0067] In one embodiment, in step S4, the virtual scale pixel height of any point on the road surface is determined by combining the virtual scale, including:

[0068] In step S41, the pixel coordinates of each vanishing point are calculated according to the position information of each entity scale by the following formula:

[0069] (y2-y0)(x1-x0)=(y1-y0)(x2-x0); (1)

[0070] Where (x1, y1), (x2, y2) are the pixel coordinates corresponding to the two entity scales respectively, and (x0, y0) is the pixel coordinates of the corresponding vanishing point.

[0071] Specifically, assuming that there is a vanishing point along the road direction, the pixel coordinates of the vanishing point along the road axis on the image are (x0, y0), and there is an entity scale with a pixel height of h1 at (x1, y1) in the image. When the scale is moved to another position (x2, y2) in the image, the pixel coordinates (x0, y0) of the vanishing point can be calculated according to the position between the two scales by the above formula (1).

[0072] Step S42, constructing a virtual ruler in the road surface where the calibration point (x1, y1) is located.

[0073] Step S43, calculating the virtual ruler scale pixel height h v1 of any point in the road surface according to the longitudinal pixel coordinate y0 of the vanishing point, the longitudinal pixel coordinate y1 of the calibration point, and the display pixel height h v of one scale on the virtual ruler by the following formula:

[0074] h v =h v1 (y-y0) / (y1-y0); (2)

[0075] h v1 =(h 1* h r0 ) / h c ; (3)

[0076] wherein y is the longitudinal pixel coordinate of any point in the road surface; h1 is the pixel height of the physical ruler at the calibration point, h r0 is the actual unit height indicated by the virtual ruler scale, and h c is the actual height of the physical ruler.

[0077] In one of the embodiments, in step S7, the road surface area corresponding to the water accumulation area is determined according to the extension direction of the road line and the boundary information of the water accumulation area.

[0078] Step S71, identifying the water accumulation area from the road monitoring image, and screening the water accumulation monitoring point from the water accumulation area, which needs to monitor the water accumulation depth.

[0079] Specifically, please refer to Figure 4 , the water accumulation area is the gray area shown in the figure, the water accumulation monitoring point is the point P s (x s , y s ) shown in the figure, the vanishing point is the point P0(x0, y0) shown in the figure, and the vertical distance △h of P s relative to the road extension trajectory is the water accumulation depth required.

[0080] Step S72, identifying the line connecting the water accumulation monitoring point and the vanishing point, and determining the intersection point obtained by the intersection of the line and the boundary of the water accumulation area.

[0081] Specifically, the line connecting the water accumulation monitoring point P s and the vanishing point P0 can be further obtained from Figure 4After the boundary position of each lane (i.e., the road boundary) is determined, a corresponding road line is constructed based on the extension direction of the road boundary (for details, refer to Figure 5 ).

[0082] In step S73, a road extension trajectory is made along the intersection and below the water accumulation area based on the extension direction of the road line.

[0083] In step S74, a point on the road extension trajectory directly below the water accumulation monitoring point is taken as a road surface point, and a road surface area corresponding to the water accumulation area is determined based on the area covered by the road surface point.

[0084] Specifically, the method for determining the road surface point is to make a road extension direction trajectory from the intersection point to the point B, and a point directly below the point B on the trajectory is the road surface point C.

[0085] In one embodiment, in step S8, the virtual scale is constructed at the water accumulation monitoring point based on the target virtual scale pixel height, including:

[0086] In step S81, the target virtual scale pixel height of the water accumulation monitoring point is obtained, and a virtual scale with the target virtual scale pixel height is established between the water accumulation monitoring point and the corresponding road surface point.

[0087] In step S8, the target water depth is determined based on the height difference between the intersection of the virtual scale and the road surface and the water surface, including:

[0088] In step S82, the initial water depth is obtained based on the difference between the longitudinal pixel coordinates of the water accumulation monitoring point and the longitudinal pixel coordinates of the corresponding road surface point.

[0089] Specifically, when the longitudinal pixel coordinates of the water accumulation monitoring point are y s , and the longitudinal pixel coordinates of the corresponding road surface point are y f , the initial water depth is y s -y f .

[0090] In step S83, the actual height of the initial water depth is converted based on the target virtual scale pixel height to obtain the target water depth.

[0091] Specifically, y s -y f is taken as h r0After substituting it into formula (2), formula (2) can be converted into formula (4) as follows. Subsequently, according to formula (4), the required target waterlogging depth Δh can be calculated in the case where the relevant parameters are known:

[0092]

[0093] Please refer to Figure 6 The application discloses a waterlogging depth calculation system 600, which comprises an entity calibration module 601, a ruler positioning module 602, a model creation module 603, a virtual calibration module 604, a waterlogging area identification module 605, a road line construction module 606, a road surface area identification module 607 and a waterlogging depth calculation module 608.

[0094] The entity calibration module 601 is used for calibrating a plurality of entity rulers at specific positions in a city along roads.

[0095] The ruler positioning module 602 is used for acquiring road monitoring images of the entity rulers and determining position information of each entity ruler in the road monitoring images.

[0096] The model creation module 603 is used for creating corresponding perspective projection models in combination with the position information of each entity ruler and constructing virtual rulers.

[0097] The virtual calibration module 604 is used for determining virtual ruler scale pixel heights of any points on a road surface in combination with the virtual rulers.

[0098] The waterlogging area identification module 605 is used for identifying waterlogging areas in the road monitoring images by applying a machine vision model and segmenting to obtain waterlogging area boundary ranges.

[0099] The road line construction module 606 is used for constructing corresponding road lines in the models in combination with boundary positions of road regions outside the waterlogging boundary ranges.

[0100] The road surface area identification module 607 is used for determining road surface areas corresponding to the waterlogging areas according to extension directions of the road lines and boundary information of the waterlogging areas.

[0101] The waterlogging depth calculation module 608 is used for constructing corresponding virtual rulers at waterlogging monitoring points according to the virtual ruler scale pixel heights, determining target waterlogging depths according to height differences between intersection points of the virtual rulers and road surfaces and waterlogging surfaces.

[0102] In one embodiment, each module in the system is further used for performing the method in any optional implementation manner of the above-described embodiments.

[0103] From the above, the waterlogging water depth calculation system disclosed in the application can determine the intersection of the waterlogging area and the virtual scale in real time, calculate the waterlogging depth at the position based on the position of the intersection, provide first-hand data for subsequent monitoring of the risk level of the waterlogging depth and making a risk warning, utilize the fixed characteristics of the monitoring scene, combine the road information of the same scene when there is no waterlogging to realize the construction of the virtual scale, and can be applied to the waterlogging depth monitoring under different rainfall conditions, can effectively expand the monitoring range, improve the monitoring efficiency and accuracy, and use the entity scale which is relatively easy to set to determine the actual height in the monitoring image, effectively avoid the problem of large monitoring fluctuation and insufficient accuracy caused by using other dynamic objects as the scale, and improve the calculation accuracy.

[0104] The computer program is executed by the processor to execute the method in any optional implementation manner of the above-mentioned embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage, a flash memory, a magnetic disk or an optical disk.

[0105] The above-mentioned readable storage medium can determine the intersection of the waterlogging area and the virtual scale in real time, calculate the waterlogging depth at the position based on the position of the intersection, provide first-hand data for subsequent monitoring of the risk level of the waterlogging depth and making a risk warning, utilize the fixed characteristics of the monitoring scene, combine the road information of the same scene when there is no waterlogging to realize the construction of the virtual scale, and can be applied to the waterlogging depth monitoring under different rainfall conditions, can effectively expand the monitoring range, improve the monitoring efficiency and accuracy, use the entity scale which is relatively easy to set to determine the actual height in the monitoring image, effectively avoid the problem of large monitoring fluctuation and insufficient accuracy caused by using other dynamic objects as the scale, and improve the calculation accuracy.

[0106] In the embodiments of the present application, it should be understood that the disclosed apparatus and method can be implemented in other manners. The embodiments described above are merely exemplary, for example, the division of the units is only a logical function division, and there can be another division manner in actual implementation; for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0107] In addition, the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments of the present application.

[0108] In addition, the functional modules in each of the embodiments of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0109] In this document, the terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations.

[0110] The above only describes the embodiments of the present application, and is not used to limit the protection scope of the present application. For those skilled in the art, various modifications and changes can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A waterlogging water depth calculation method characterized by, The method comprises the following steps: S1, a plurality of entity scales are calibrated along the road signs at a specific location in a city in advance; S2, road monitoring images of the entity scales are obtained, and position information of each of the entity scales in the road monitoring images is determined; S3, a corresponding perspective projection model is created in combination with the position information of each of the entity scales, and a virtual scale is constructed; S4, a virtual scale pixel height of an arbitrary point on the road surface is determined in combination with the virtual scale; S5, a machine vision model is applied to identify a water accumulation area in the road monitoring image, and a water accumulation area boundary range is segmented; S6, a corresponding road line is constructed in the model in combination with the boundary position of the road area outside the water accumulation boundary range; S7, a road surface area corresponding to the water accumulation area is determined according to the extension direction of the road line and the boundary information of the water accumulation area; S8, a corresponding virtual scale is constructed at a water accumulation monitoring point according to the virtual scale pixel height, and a target water depth is determined according to a height difference between intersection points of the virtual scale and the road surface and the water surface; In the created perspective projection model, there are a plurality of vanishing points, and the vanishing points are determined according to the parallel lines represented by each of the entity scales, and the vanishing points are aggregated to obtain an aggregation point; In step S4, the virtual scale pixel height of an arbitrary point on the road surface is determined in combination with the virtual scale, comprising: S41, pixel coordinates of each vanishing point are calculated according to the position information of each of the entity scales by the following formula: (y2-y0)(x1-x0)=(y1-y0)(x2-x0); (1) Where (x1, y1), (x2, y2) are pixel coordinates corresponding to two preset entity scales respectively, and (x0, y0) is pixel coordinates of the corresponding vanishing point; S42, the construction of the virtual scale is performed in the road surface where the calibration point (x1, y1) is located; S43, according to the longitudinal pixel coordinates y0 of the vanishing point, the longitudinal pixel coordinates y1 of the calibration point, and the display pixel height h corresponding to one scale on the virtual scale v1 , the virtual scale pixel height h of any point in the road surface is calculated by the following formula v : h v = h v1 (y-y0) / (y1-y0); (2) h v1 = (h 1* h r0 ) / h c ; (3) wherein y is the longitudinal pixel coordinate of any point on the road surface; h1 is the pixel height of the physical scale at the calibration point, h r0 is the actual unit height indicated by the virtual scale, h c is the actual height of the physical scale; In step S7, the road surface area corresponding to the water accumulation area is determined according to the extension direction of the road line and the boundary information of the water accumulation area, comprising: S71, the water accumulation area is identified from the road monitoring image, and a water accumulation monitoring point which needs to monitor the water depth is selected from the water accumulation area; S72, a line connecting the water accumulation monitoring point and the vanishing point is identified, and an intersection point obtained by the intersection of the line and the water accumulation area boundary is determined; S73, a corresponding road extension trajectory is made below the water accumulation area along the intersection point according to the extension direction of the road line; S74, a point directly below the water accumulation monitoring point on the road extension trajectory is taken as a road surface point, and a road surface area corresponding to the water accumulation area is determined based on a region covered by the road surface point.

2. The method of claim 1, wherein, In step S2, the position information of each of the entity scales in the road monitoring images is determined, comprising: S21, a target road monitoring video corresponding to a specific location in a city and having a plurality of entity scales calibrated along the road signs in advance is determined; S22, a plurality of road monitoring images are segmented from the target road monitoring video according to a preset image segmentation rule; S23, calling a pre-trained target detection model to identify the entity scale from the plurality of road monitoring images, and determining the pixel coordinates and pixel height of each entity scale in the road monitoring images based on the upper and lower edge pixel coordinates of the identified bounding box in the identification process.

3. The method of claim 2, wherein, In step S21, the target road monitoring video corresponding to a specific location in a city and pre-marked with a plurality of entity scales is determined, including: S211, obtaining an initial road monitoring video corresponding to a specific location in a city, and identifying a road straight section from the initial road monitoring video; S212, when the corresponding road straight section is identified, marking a plurality of entity scales with clear height information along the boundary of the road straight section, so that the entity scales appear in the monitoring video range of the road, and thus determining the target road monitoring video required.

4. The method of claim 1, wherein, In step S8, the virtual scale is constructed at the water accumulation monitoring point according to the virtual scale pixel height, including: S81, obtaining a target virtual scale pixel height of the water accumulation monitoring point, and establishing a virtual scale with the target virtual scale pixel height between the water accumulation monitoring point and the corresponding road surface point; In step S8, the target water depth is determined according to the height difference between the intersection of the virtual scale and the road surface and the water surface, including: S82, obtaining a corresponding initial water depth based on the difference between the longitudinal pixel coordinates of the water accumulation monitoring point and the longitudinal pixel coordinates of the corresponding road surface point; S83, converting the actual height of the initial water depth according to the target virtual scale pixel height to obtain the corresponding target water depth.

5. A waterlogging accumulated water depth calculation system for implementing the method of any one of claims 1 to 4, characterized by, The system comprises an entity calibration module, a scale positioning module, a model creation module, a virtual calibration module, a water accumulation area identification module, a road line construction module, a road surface area identification module, and a water depth calculation module, wherein: The entity calibration module is used to pre-mark a plurality of entity scales at a specific location in a city along a road; The scale positioning module is used to obtain road monitoring images of the entity scales and determine the position information of each entity scale in the road monitoring images; The model creation module is used to create a corresponding perspective projection model in combination with the position information of each entity scale to construct a virtual scale; The virtual calibration module is used to determine the virtual scale pixel height of any point on the road surface in combination with the virtual scale; The water accumulation area identification module is used to apply a machine vision model to identify a water accumulation area in a road monitoring image and segment to obtain a water accumulation area boundary range; The road line construction module is used to construct a corresponding road line in the model in combination with the boundary position of the road area outside the water accumulation boundary range; The road surface area identification module is used to determine the road surface area corresponding to the water accumulation area according to the extension direction of the road line and the boundary information of the water accumulation area. The waterlogging depth calculation module is configured to construct a corresponding virtual scale at a waterlogging monitoring point according to the virtual scale pixel height, and determine a target waterlogging depth according to a height difference between intersections of the virtual scale and a road surface and a waterlogging surface.

6. A readable storage medium characterized by, The readable storage medium comprises a waterlogging depth calculation method program, and the waterlogging depth calculation method program, when executed by the processor, implements the steps of the method according to any one of claims 1 to 4.

Images