A long tunnel portal water curtain early warning system and method

Abstract

The application relates to the technical field of traffic control, in particular to a long-tunnel-port water curtain early warning system and method. The method comprises the following steps: acquiring an actual water curtain image and wind direction and wind speed, acquiring a wind drift coefficient based on the similarity of different row coding vectors of a standard water curtain image and the actual water curtain image and the change of the coding vector sequence; in a complete water curtain generation, acquiring a total displacement distance vector based on water droplet displacement distance vectors under different wind directions and wind speeds, water droplet duration and drop-out time; then acquiring a predicted water curtain image, acquiring a water droplet displacement amount based on the difference of the wind drift coefficients of the predicted water curtain image and the actual water curtain image to construct a target function; determining a matching nozzle based on the difference between the overall displacement vector of the water droplet in all rows and the position of the water curtain nozzle; spraying water droplets out through the matching nozzle to generate a water curtain to complete early warning; the application improves the water curtain information generation quality, improves the early warning accuracy, and further improves the safe operation of the tunnel.

Description

Technical Field

[0001] This application relates to the field of traffic control technology, specifically to a water curtain early warning system and method for long tunnel entrances. Background Technology

[0002] Long tunnels, as vital passages traversing mountains and connecting regions, occupy a crucial position in road networks. However, due to their enclosed and elongated structural characteristics, tunnels face severe challenges in operational safety. Dramatic changes in lighting conditions within tunnel areas can easily lead to visual adaptation difficulties and distraction for drivers. Simultaneously, the limited space inside tunnels makes it difficult for rescue vehicles to arrive quickly in the event of a traffic accident. If following traffic fails to receive timely information and continues to surge in, secondary collisions are highly likely, escalating the accident and causing severe traffic congestion. To improve tunnel safety and prevent and reduce accidents, various intelligent early warning technologies have emerged. Among them, the water curtain flexible early warning system, with its striking visual impact and physical barrier effect, can generate a water curtain barrier at the tunnel entrance within seconds of an accident, effectively warning following vehicles to slow down or stop, preventing further traffic from entering, and buying valuable time for accident scene isolation and rescue operations. It has become an important technical means in the tunnel safety control system.

[0003] However, existing water curtain warning technologies still have significant shortcomings in practical applications. Currently, the mainstream methods for generating water curtains mainly rely on water curtain projection or digital water curtain technology, both of which use flowing water as the information display medium. This medium is highly susceptible to interference from the complex external environment at tunnel entrances, especially wind conditions. When crosswinds or gusts occur, the airflow can disperse the continuous water curtain or cause the water droplets to deviate from the trajectory of the digital water curtain, resulting in the preset warning text and symbols shifting, distorting, breaking, or even completely disappearing, making key warning information unclear or unrecognizable. This distortion of information transmission due to environmental interference severely weakens the reliability and warning effect of the system. In emergencies, drivers behind may fail to take timely evasive action because they cannot accurately understand the warning content conveyed by the water curtain, thus failing to effectively prevent secondary accidents and failing to fully realize the safety protection value that the water curtain warning system should have. Therefore, there is an urgent need for a new type of water curtain warning system and method for long tunnel entrances that can overcome environmental influences and ensure the stable and accurate transmission of warning information. Summary of the Invention

[0004] To address the technical problem of weak early warning capabilities due to interference and distortion of the water curtain, this application provides a water curtain early warning system and method for long tunnel entrances, the specific technical solution of which is as follows: Firstly, this application proposes a method for early warning of water curtain at the entrance of a long tunnel, the method comprising the following steps: Acquire actual water curtain images at the entrance of the long tunnel, as well as wind direction and speed at each moment; Obtain a standard water curtain image and its ROI region. Obtain the ROI region at the same location in the actual water curtain image and obtain the encoding vector of each row of the ROI region. Determine the wind drift coefficient based on the similarity of the encoding vectors of text contained in the standard water curtain image and the actual water curtain image and the change in the number of rows after reordering the encoding vectors of text contained in the actual water curtain image. The displacement distance vector of water droplets at each moment is obtained by taking the wind direction and wind speed at each moment; the total displacement distance vector is obtained based on the duration and dripping time of the nozzle water droplets and the displacement distance vector at each moment; the encoded vector of the standard water curtain image is moved through the total displacement distance vector to obtain the predicted water curtain image; the difference in wind drift coefficient between the predicted water curtain image and the actual water curtain image is used as the objective function, and the displacement of water droplets is obtained through the particle swarm algorithm. The displacement of water droplets along different axes is used to construct a water droplet displacement vector. The overall displacement vector of the water droplets is determined by combining the time it takes for the water droplets to move to each row. The offset vector of the water droplets is obtained by comparing the overall displacement vector of the water droplets with the position of the water curtain nozzles. The matching nozzle for each nozzle is determined based on the offset vector of the water droplets. The system uses matching nozzles to spray water droplets, creating a water curtain to provide an early warning.

[0005] In the aforementioned scheme, this application proposes a water curtain early warning system and method for long tunnel entrances. First, based on the water curtain display under varying wind conditions, a wind drift coefficient is calculated. This coefficient, combined with the encoding of water curtain information, quantifies the error in the water curtain display, enhancing the accuracy of water curtain quality measurement. This quantification allows the water curtain quality to be incorporated into subsequent error compensation. Then, by combining nozzle and droplet displacement data, an objective function is constructed to quantify droplet displacement under different wind conditions. Based on the quantification results, appropriate nozzles are selected to dispense water droplets to compensate for droplet drift, ensuring the accuracy of the water curtain information. This method facilitates accurate quantification of droplet drift, enabling compensation based on environmental wind changes and maximizing the quality of water curtain information generation. Ultimately, this results in a water curtain early warning system and method for long tunnel entrances with higher stability and accuracy, improving tunnel safety operations.

[0006] In one embodiment, the method for obtaining the encoding vector of each row of the ROI region is as follows: For each row in the ROI region, obtain the encoding sequence. In the encoding sequence, pixels that display text are marked as 1, and pixels that do not display text are marked as 0.

[0007] In one embodiment, the method for determining the wind drift coefficient based on the similarity of the encoded vectors of text contained in the standard water curtain image and the actual water curtain image, and the change in the number of rows after reordering the encoded vectors of text contained in the actual water curtain image, is as follows: The encoding vector containing 1 in the standard water curtain image is denoted as the text encoding vector. The similarity between each text encoding vector and the encoding vector containing 1 in all actual water curtain images is calculated. The encoding vector containing 1 in the actual water curtain image corresponding to the maximum similarity is denoted as the actual similar vector. The actual similar vectors are sorted in descending order of the text encoding vectors to obtain the encoding order sequence. The first difference of the number of rows in the encoding order sequence is calculated to obtain the proportion of positive numbers. The wind drift coefficient is determined based on the proportion of positive values ​​in the first-order difference and the similarity between the text encoding vector and the actual similar vector.

[0008] In one embodiment, the wind drift coefficient is positively correlated with the text encoding vector and the actual similarity vector, and negatively correlated with the proportion of positive values ​​in the first-order difference.

[0009] In one embodiment, the method for obtaining the displacement distance vector of the water droplets at each moment by using the wind direction and wind speed at each moment is as follows: With the center of the nozzle array as the origin, a spatial coordinate system is obtained by defining three coordinate axes: the tunnel direction, the vertical downward direction, and the direction perpendicular to both directions. The speed and direction of each water droplet are adjusted according to the wind direction and speed to obtain the displacement distance vector of the water droplet every second.

[0010] In one embodiment, the method for obtaining the total displacement distance vector based on the duration of the nozzle water droplets, the droplet ejection time, and the displacement distance vector at each moment is as follows: , Indicates the time of a complete water curtain information generation; This represents the dripping time of the water droplet in the nth row within a complete water curtain information generation time; , This indicates the duration of the water droplets in the nth row; , This represents the time elapsed between the moment the nth water droplet is released. This represents the displacement distance vector of the nth water droplet at the moment it is dropped; j represents the j-th moment. This represents the displacement distance vector corresponding to the j-th time moment. This represents the total displacement distance vector of the water droplets in the nth row.

[0011] In one embodiment, the objective function is the absolute value of the difference between the wind drift coefficient of the actual water curtain image and the wind drift coefficient of the predicted water curtain image; the displacement of the water droplets in the predicted water curtain image corresponding to the minimum value of the objective function is taken as the displacement of the water droplets.

[0012] In one embodiment, the expression for the offset vector of the water droplet is: , This represents the displacement vector of the water droplet. This indicates the time when the device moves to the u-th row. This indicates the position of the s-th nozzle. Let represent the offset vector of the water droplet ejected from the s-th nozzle at the u-th row.

[0013] In one embodiment, the method for determining the matching nozzle for each nozzle based on the water droplet offset vector is as follows: Obtain the coordinates of each nozzle in the same row, subtract the coordinates from the offset vector to obtain the accuracy vector of each row, average the accuracy vectors of all rows to obtain the accuracy vector of each nozzle, and select the nozzle corresponding to the minimum value of the magnitude of the accuracy vector and the offset vector of the water droplet as the matching nozzle.

[0014] Secondly, embodiments of this application also provide a long tunnel entrance water curtain early warning system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described long tunnel entrance water curtain early warning methods.

[0015] The beneficial effects of this application are as follows: This application proposes a water curtain early warning system and method for long tunnel entrances. First, based on the water curtain display under varying wind conditions, a wind drift coefficient is calculated. This coefficient, combined with the encoding of water curtain information, quantifies the error in the water curtain display, enhancing the accuracy of water curtain quality measurement. This quantification allows the water curtain quality to be incorporated into subsequent error compensation. Then, by combining nozzle and droplet displacement data, an objective function is constructed to quantify droplet displacement under different wind conditions. Based on the quantification results, appropriate nozzles are selected to deliver water droplets to compensate for droplet drift, ensuring the accuracy of the water curtain information. This method facilitates accurate quantification of droplet drift, enabling compensation based on environmental wind changes and maximizing the quality of water curtain information generation. Ultimately, this results in a water curtain early warning system and method for long tunnel entrances with higher stability and accuracy, improving tunnel safety. Attached Figure Description

[0016] Figure 1 A flowchart illustrating a water curtain early warning method at the entrance of a long tunnel, provided in one embodiment of this application; Figure 2 This is a schematic diagram of a water curtain early warning system at the tunnel entrance. In the diagram, 1 is the tunnel entrance, 2 is a high-definition camera, 3 is the water curtain, 4 is a light warning strip, and 5 is an ultrasonic anemometer. Detailed Implementation

[0017] An example of a water curtain early warning method for long tunnel entrances: The following description, in conjunction with the accompanying drawings, details a specific scheme for a water curtain early warning method for long tunnel entrances provided in this application.

[0018] Please see Figure 1 The diagram illustrates a flowchart of a water curtain early warning method for a long tunnel entrance according to an embodiment of this application. The method includes the following steps: Step S001: Obtain the actual water curtain image, as well as the wind direction and wind speed.

[0019] like Figure 2 The schematic diagram of the tunnel entrance water curtain early warning system shows that the system uses an ultrasonic anemometer installed above the tunnel entrance to collect wind speed and direction data at the tunnel entrance once per second; it uses a high-definition camera to collect actual water curtain images, and uses a denoising algorithm to preprocess the actual water curtain images. Among different processing methods, denoising can include, but is not limited to, median filtering and mean filtering.

[0020] At this point, actual water curtain images, as well as wind speed and direction at the tunnel entrance, were obtained.

[0021] Step S002: Obtain the wind drift coefficient of the actual water curtain image based on the similarity of the encoding vectors of different rows between the standard water curtain image and the actual water curtain image, as well as the change in the order of the encoding vectors.

[0022] Due to the complexity of the tunnel environment, accidents within tunnels can easily trigger secondary accidents. Furthermore, a large influx of vehicles into the tunnel can hinder the arrival of rescue vehicles, escalating the consequences. Therefore, it is crucial to immediately notify and warn vehicles behind the tunnel after an accident. Water curtain flexible warning systems, with their striking visual impact and rapid response, help enhance the warning effect, making them an effective solution for tunnel accident early warning. However, since water curtain warnings use water flow as the information carrier, they are highly susceptible to the influence of wind, causing information distortion or even complete loss of information in strong winds. In such cases, accurate information transmission becomes difficult, compromising the effectiveness of tunnel accident warnings. Therefore, it is necessary to optimize water curtain warning systems to ensure their effectiveness even under wind-affected conditions.

[0023] During the generation of a water curtain at a tunnel entrance, the water curtain is highly susceptible to changes in environmental wind force, which can distort information and lead to inaccurate data transmission. Therefore, it is necessary to compensate for the impact of environmental wind force and correct the water curtain's offset to minimize information distortion. During the use of the water curtain, higher wind speeds cause a greater offset of the water curtain in the opposite direction of the wind, resulting in greater text distortion. Therefore, when correcting the water curtain, it is essential to determine the offset direction and the corresponding amount of offset.

[0024] In this embodiment, the standard environment for acquiring water curtain images is a laboratory environment that is almost unaffected by wind. The water curtain image with the highest clarity under the standard environment is used as the standard water curtain image. In actual applications, the high-definition camera used for acquiring water curtain images needs to be positioned in the same location as in the standard environment.

[0025] In a standard water curtain image, the water curtain area is manually marked as a Region of Interest (ROI). In a standard water curtain image, this ROI is a rectangular region. The position of the ROI in the water curtain image is determined based on its vertex coordinates. Next, the ROI is encoded. Starting from the bottom, the encoding vector for each row is obtained. Pixels displaying text are marked as 1, and pixels without text are marked as 0. The encoding vector for each row of the ROI is obtained in this way.

[0026] For actual water curtain images, the pixels of the corresponding region in the actual water curtain image are determined based on the coordinate position of the ROI region in the standard water curtain image, and the encoding vector of each row of the corresponding region is obtained in the same way.

[0027] The encoded vector containing 1 is denoted as the text encoded vector. The water droplet offset is obtained based on the similarity between the text encoded vector in the standard water curtain image and the text encoded vector in the actual water curtain image, as well as the difference in the number of rows in the actual water curtain image.

[0028] Specifically, for the text encoding vector in the standard water curtain image, the similarity between it and each text encoding vector in the actual water curtain image is calculated, and the maximum similarity and its corresponding text encoding vector in the actual water curtain image are recorded as the actual similar vector.

[0029] When obtaining the text encoding vector, the text encoding vector is sorted from top to bottom according to the row number of the ROI region of the standard water curtain image; since each text encoding vector corresponds one-to-one with the actual similar vector, the actual similar vector is sorted according to the sorting number to obtain the encoding order sequence, and the first difference of the row number in the encoding order sequence is calculated to obtain its positive percentage.

[0030] The wind drift coefficient is determined based on the proportion of positive values ​​in the first-order difference and the similarity between the text encoding vector and the actual similar vector.

[0031] The wind drift coefficient is positively correlated with the text encoding vector and the actual similarity vector, and negatively correlated with the proportion of positive values ​​in the first-order difference.

[0032] It should be noted that positive correlation means that when one variable increases, the other variable also increases, and the two variables change in the same direction. When one variable changes from large to small or from small to large, the other variable also changes from large to small or from small to large. The specific relationship is determined by the actual application, and this application does not impose any special restrictions.

[0033] It should be noted that negative correlation means that when one variable increases, the other variable decreases accordingly, and the two variables change in opposite directions. When one variable changes from large to small or from small to large, the other variable also changes from small to large or from large to small. The specific relationship is determined by practical application, and this application does not impose any special restrictions.

[0034] Preferably, in this embodiment, the expression for the wind drift coefficient is: , This represents the i-th character encoding vector. Let represent the actual similarity vector corresponding to the i-th character encoding vector. Indicates Hamming distance, This indicates the number of text encoding vectors. This represents the proportion of positive numbers in the first-order difference. This represents the wind drift coefficient.

[0035] Furthermore, when encountering extreme environmental wind disturbances, a large number of positive values ​​in the first-order difference sequence are missing, that is, the proportion of positive values ​​in the first-order difference is reduced. When equal to or close to 0, It is forcibly set to a preset minimum normal quantity safety lower limit to ensure that the system can continue to perform subsequent fitness calculations without data interruption.

[0036] It is understandable that water curtain information will experience some distortion under the influence of wind. However, since the information carrier is continuously flowing water, even with distortion due to wind, the distorted information still maintains a strict vertical order due to the temporal nature of the water flow. The wind simply introduces horizontal movement on top of the vertical direction, causing some compression. Therefore, when displaying water curtain information, the smaller the distortion and the less pronounced the distortion, the higher the similarity between the encoded vectors in each row of the actual water curtain image and the standard water curtain image. Furthermore, the vertical distribution characteristics of the encoded vectors are also more similar, meaning that the elements in the encoded sequence gradually increase in size. In this case, the wind drift coefficient is smaller.

[0037] Thus, the wind drift coefficient was obtained.

[0038] Step S003: In a complete water curtain generation, the total displacement distance vector is obtained based on the water droplet displacement distance vector, water droplet duration, and droplet emergence time under different wind directions and speeds; then, a predicted water curtain image is obtained, and an objective function is constructed based on the difference in wind drift coefficient between the predicted water curtain image and the actual water curtain image to obtain the water droplet displacement.

[0039] The wind drift coefficient was analyzed to assess the droplet offset caused by wind. Compensation is then needed to ensure the water curtain accurately displays warning information. Specifically, the nozzles of the water curtain are arranged in an array. First, the droplet offset under different wind conditions is determined. Then, based on the droplet offset, appropriate nozzles in the nozzle array are selected to dispense water droplets, ensuring that the droplets reach their corresponding positions under appropriate wind conditions.

[0040] The generation of a complete water curtain information is considered as a cycle. The wind speed and wind direction within this cycle are used to form wind speed and wind direction sequences according to the acquisition order.

[0041] A spatial coordinate system is obtained by taking the center of the nozzle array as the origin and defining three coordinate axes: the direction along the tunnel, the vertically downward direction, and the direction perpendicular to both directions. Then, a particle swarm optimization algorithm is used to obtain the optimal water droplet drift.

[0042] Specifically, the position of each particle is a displacement distance vector along three coordinate axes under various wind speed and direction sequences during a complete water curtain information generation process. Since the wind direction and speed differ every second, one displacement distance vector is acquired per second. For example, if a complete water curtain information generation process lasts 5 seconds, based on the frequency of environmental wind data collection, there are 5 wind conditions in the entire process. Each condition corresponds to the displacement vector of a water droplet, which can be decomposed into displacement distance vectors along the three coordinate axes. In the particle swarm optimization algorithm, the number of particles is... Maximum number of system iterations Inertial weight and learning factors , In this embodiment, the number of particles m is 50, and the maximum number of iterations is... The value is 100, representing the inertia weight. Initial learning factor The randomly generated initial particles are used as input, and the particle swarm optimization algorithm is employed to output the optimal particles. The particle swarm optimization algorithm is a well-known technique and will not be described in detail in this embodiment.

[0043] The objective function of the particle swarm optimization algorithm is the difference between the wind drift coefficient of the actual water curtain image and the wind drift coefficient of the predicted water curtain image. Therefore, it is necessary to obtain the predicted water curtain image.

[0044] To predict the encoded vector of the water curtain image, it is necessary to calculate the displacement distance vector of the water droplets after being affected by the wind speed. Therefore, the total displacement distance vector is obtained based on the duration of the water droplets and the displacement distance vector per second.

[0045] Preferably, in this embodiment, the expression for the total displacement distance vector is: , , , This indicates the time required to generate a complete water curtain information sequence, i.e., the cycle time. This indicates the time it takes for the water droplet in the nth row to drip within the period, for example, the water droplet in the nth row drips at 2.5 seconds. This indicates the duration of the water droplets in the nth row; This indicates the time within which the nth row of water droplets falls, for example, the nth row of water droplets falls within 0.5 seconds of the 3rd second; This represents the displacement distance vector of the nth water droplet at the moment it is dropped; j represents the j-th moment. This represents the displacement distance vector corresponding to the j-th time moment. This represents the total displacement distance vector of the water droplets in the nth row.

[0046] The total displacement distance vector of each row of water droplets was obtained through the above method. The encoding vector of the nth row corresponding to the ROI region of the standard water curtain image was spatially moved through the total displacement distance vector. After all the encoding vectors were spatially moved, the predicted water curtain image was formed.

[0047] The objective function in the optimization algorithm is: , This represents the wind drift coefficient for predicting water curtain images. This represents the wind drift coefficient of the actual water curtain image. This represents the objective function.

[0048] The purpose of the particle swarm optimization algorithm is to minimize the objective function. It can be understood that, in the process of generating water curtain information, it is necessary to control the dripping time of water droplets. Since the displacement of water droplets varies under different wind conditions, and it is difficult to detect the displacement of water droplets, the wind drift coefficient constructed above can be used to analyze and compare the changes in water curtain information. The difference between the actual water curtain image and the predicted water curtain image obtained by the particle swarm optimization algorithm is used as the objective function, thereby enabling the measurement of water droplet displacement.

[0049] Therefore, the displacement of water droplets in the predicted water curtain image corresponding to the minimum value of the objective function is taken as the displacement of the water droplets.

[0050] Thus, the displacement of the water droplet was obtained.

[0051] Step S004: Determine the offset vector and match the nozzle based on the difference between the overall displacement vector of the water droplets in all rows and the position of the water curtain nozzle.

[0052] Different numbers of samples were collected, each corresponding to one complete water curtain information generation cycle. The water droplet displacement under various wind force changes was obtained using the method described above. In this embodiment, the sample size was 100. Wind speeds in the same wind direction were used as the abscissa, and the corresponding water droplet displacements as the ordinate. Polynomial fitting was performed, and the fitted curve function was output. It should be noted that since the water droplet displacement involves three axial displacements, fitting can be performed separately for each of the three axes. Polynomial fitting is a well-known technique and will not be elaborated further.

[0053] After obtaining the polynomial fitting curve function, the real-time collected ambient wind speed is substituted into the polynomial fitting curve as an input parameter to calculate and obtain the correction amount of the water droplet horizontal drift velocity under the current wind force state, which is denoted as the water droplet displacement vector. .

[0054] First, the offset vector of each row of water droplets is determined based on the difference between the nozzle position and the distance the water droplets travel due to wind force. The nozzle position is represented by coordinates.

[0055] Preferably, in this embodiment, the expression for the water droplet's offset vector is: , This represents the displacement vector of the water droplet. This indicates the time when the device moves to the u-th row. This indicates the position of the s-th nozzle. Let represent the offset vector of the water droplet ejected from the s-th nozzle at the u-th row.

[0056] Obtain the coordinates of each nozzle in the same row, subtract the coordinates from the offset vector to obtain the accuracy vector of each row, average the accuracy vectors of all rows to obtain the accuracy vector of each nozzle, and select the nozzle corresponding to the minimum value of the magnitude of the accuracy vector and the offset vector of the water droplet as the matching nozzle.

[0057] It should be noted that each dimension of the vector represents the displacement along a coordinate axis. The coordinate axes corresponding to each dimension of the data correspond sequentially to the coordinate axes of the nozzle position coordinates. Therefore, the nozzle position coordinates can be adjusted by directly adding or subtracting the corresponding displacement.

[0058] It is understandable that the water droplets in the water droplet array will drift under the influence of the ambient wind, resulting in a large error in the water curtain information. In order to minimize the error, it is necessary to make the water droplets in the water droplet array reach the predetermined position as much as possible under controllable conditions. Therefore, when the water droplets are released, the water droplets can be adjusted to be released from different nozzles according to the wind conditions in the corresponding time period, so that the water droplets can still reach the corresponding position under the influence of the wind after being released.

[0059] At this point, the matching nozzle for each nozzle has been obtained.

[0060] Step S005: Water droplets are sprayed out through the matching nozzle to generate a water curtain and complete the early warning.

[0061] The matching nozzles required for the operation are calculated through the above steps. The water curtain is generated by matching the nozzles, and an early warning is given based on the generated water curtain.

[0062] Based on the same inventive concept as the above method, this embodiment of the invention also provides a long tunnel entrance water curtain early warning system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described long tunnel entrance water curtain early warning methods.

[0063] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for early warning of water curtain at the entrance of a long tunnel, characterized in that, The method includes the following steps: Acquire actual water curtain images at the entrance of the long tunnel, as well as wind direction and speed at each moment; Obtain a standard water curtain image and its ROI region. Obtain the ROI region at the same location in the actual water curtain image and obtain the encoding vector of each row of the ROI region. Determine the wind drift coefficient based on the similarity of the encoding vectors of text contained in the standard water curtain image and the actual water curtain image and the change in the number of rows after reordering the encoding vectors of text contained in the actual water curtain image. The displacement distance vector of water droplets at each moment is obtained by taking the wind direction and wind speed at each moment; the total displacement distance vector is obtained based on the duration and dripping time of the nozzle water droplets and the displacement distance vector at each moment; the encoded vector of the standard water curtain image is moved through the total displacement distance vector to obtain the predicted water curtain image; the difference in wind drift coefficient between the predicted water curtain image and the actual water curtain image is used as the objective function, and the displacement of water droplets is obtained through the particle swarm algorithm. The displacement of water droplets along different axes is used to construct a water droplet displacement vector. The overall displacement vector of the water droplets is determined by combining the time it takes for the water droplets to move to each row. The offset vector of the water droplets is obtained by comparing the overall displacement vector of the water droplets with the position of the water curtain nozzles. The matching nozzle for each nozzle is determined based on the offset vector of the water droplets. The system uses matching nozzles to spray water droplets, creating a water curtain to provide an early warning.

2. The method for early warning of water curtain at the entrance of a long tunnel as described in claim 1, characterized in that, The method for obtaining the encoding vector of each row of the ROI region is as follows: For each row in the ROI region, obtain the encoding sequence. In the encoding sequence, pixels that display text are marked as 1, and pixels that do not display text are marked as 0.

3. The method for early warning of water curtain at the entrance of a long tunnel as described in claim 2, characterized in that, The method for determining the wind drift coefficient based on the similarity of the encoded vectors of text contained in standard water curtain images and actual water curtain images, as well as the change in the number of rows after reordering the encoded vectors of text contained in actual water curtain images, is as follows: The encoding vector containing 1 in the standard water curtain image is denoted as the text encoding vector. The similarity between each text encoding vector and the encoding vector containing 1 in all actual water curtain images is calculated. The encoding vector containing 1 in the actual water curtain image corresponding to the maximum similarity is denoted as the actual similar vector. The actual similar vectors are sorted in descending order of the text encoding vectors to obtain the encoding order sequence. The first difference of the number of rows in the encoding order sequence is calculated to obtain the proportion of positive numbers. The wind drift coefficient is determined based on the proportion of positive values ​​in the first-order difference and the similarity between the text encoding vector and the actual similar vector.

4. The method for early warning of water curtain at the entrance of a long tunnel as described in claim 1, characterized in that, The wind drift coefficient is positively correlated with the text encoding vector and the actual similarity vector, and negatively correlated with the proportion of positive values ​​in the first-order difference.

5. The method for early warning of water curtain at the entrance of a long tunnel as described in claim 1, characterized in that, The method for obtaining the displacement distance vector of water droplets at each moment by using the wind direction and wind speed at each moment is as follows: With the center of the nozzle array as the origin, a spatial coordinate system is obtained by defining three coordinate axes: the tunnel direction, the vertical downward direction, and the direction perpendicular to both directions. The speed and direction of each water droplet are adjusted according to the wind direction and speed to obtain the displacement distance vector of the water droplet every second.

6. The method for early warning of water curtain at the entrance of a long tunnel as described in claim 1, characterized in that, The method for obtaining the total displacement distance vector based on the duration of the nozzle water droplets, the droplet ejection time, and the displacement distance vector at each moment is as follows: , This indicates the time it takes for a complete water curtain information to be generated; This represents the dripping time of the water droplet in the nth row within a complete water curtain information generation time; , This indicates the duration of the water droplets in the nth row; , This represents the time elapsed between the moment the nth water droplet is released. This represents the displacement distance vector corresponding to the moment the nth water droplet is dropped; j represents the j-th moment. This represents the displacement distance vector corresponding to the j-th time moment. This represents the total displacement distance vector of the water droplets in the nth row.

7. The method for early warning of water curtain at the entrance of a long tunnel as described in claim 1, characterized in that, The objective function is the absolute value of the difference between the wind drift coefficient of the actual water curtain image and the wind drift coefficient of the predicted water curtain image. The displacement of water droplets in the predicted water curtain image corresponding to the minimum value of the objective function is taken as the displacement of the water droplets.

8. A method for early warning of water curtain at the entrance of a long tunnel as described in claim 1, characterized in that, The expression for the offset vector of the water droplet is: , This represents the displacement vector of the water droplet. This indicates the time when the device moves to the u-th row. This indicates the position of the s-th nozzle. Let represent the offset vector of the water droplet ejected from the s-th nozzle at the u-th row.

9. A method for early warning of water curtain at the entrance of a long tunnel as described in claim 1, characterized in that, The method for determining the matching nozzle for each nozzle based on the water droplet offset vector is as follows: Obtain the coordinates of each nozzle in the same row, subtract the coordinates from the offset vector to obtain the accuracy vector of each row, average the accuracy vectors of all rows to obtain the accuracy vector of each nozzle, and select the nozzle corresponding to the minimum value of the magnitude of the accuracy vector and the offset vector of the water droplet as the matching nozzle.

10. A water curtain early warning system for a long tunnel entrance, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the long tunnel entrance water curtain early warning method as described in any one of claims 1-9.

Images