Monitoring and alarming method and system for debris flow along traffic and storage medium

By setting up multi-level warning lights and video surveillance systems along the traffic line, and combining the target detection algorithm to calculate the mudslide risk score, the problem of single alarm method and manual judgment dependence in the existing technology is solved, and automated alarm level judgment and multi-source information are realized Fusion improves monitoring efficiency and accuracy.

CN119091587APending Publication Date: 2024-12-06SICHUAN TIBETAN EXPRESSWAY CO LTD +2
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411083662.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-08
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing mudslide geological disaster monitoring and early warning system is single and relies on manual judgment, and cannot achieve automatic hierarchical alarms, which are prone to errors, and cannot effectively integrate multi-source information for risk assessment.

Method used

By setting up multi-level warning lights and video surveillance systems along the traffic line, combining the target detection algorithm to calculate real-time warning scores and monitoring warning scores, weighted decision-making methods are used to calculate the mudslide risk score, and the parameters are corrected by linear regression algorithm to automatically determine the alarm level.

Benefits of technology

It improves the speed, efficiency and accuracy of monitoring alarms, enhances robustness, and improves the detection efficiency and early warning accuracy of mudslide events.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119091587A_ABST
    Figure CN119091587A_ABST
Patent Text Reader

Abstract

The invention discloses a monitoring and alarming method and system for debris flow along a traffic line and a storage medium, a plurality of warning lamps of a first level, a second level and a third level are sequentially arranged on the outer side of the traffic line from inside to outside, the states of the plurality of warning lamps of each level are obtained in real time, and a real-time warning score is obtained through calculation. The method comprises the following steps: respectively acquiring monitoring videos at a first level, a second level and a third level outside a traffic line, then judging a debris flow event based on a target detection algorithm, obtaining a debris flow target confidence coefficient, calculating to obtain a monitoring warning score, and further calculating a debris flow risk score. And determining an alarm grade corresponding to the debris flow risk score s, and correspondingly sending out external alarm information of the grade. The method has the advantages of rapidness in monitoring and alarming, high efficiency, higher precision and higher robustness.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of monitoring and alarming of debris flows along traffic lines, and in particular relates to a monitoring and alarming method, system and storage medium for debris flows along traffic lines. Background Art

[0002] Debris flow is a temporary special torrent that occurs in mountainous areas and contains a large amount of solid materials such as mud, sand, stones and boulders. It is commonly known as "walking dragon", "coming out of dragon", "dragon", etc. It has the characteristics of sudden outbreak, fierce force, rapid movement, and short duration. It often destroys buildings, engineering facilities and people's lives and property in an instant. It has strong destructive power and is a geological disaster that seriously threatens the safety of mountain residents and engineering construction. There is a danger of debris flow along road traffic lines and rail transit lines. Therefore, it is necessary to set up a debris flow geological disaster monitoring and early warning system along important road traffic lines and rail transit lines.

[0003] The existing debris flow geological disaster monitoring and early warning system is mainly composed of on-site monitoring equipment, data communication system, user-built monitoring center (including geological disaster intelligent monitoring and early warning cloud platform), and the geological disaster information is reported through the monitoring and early warning platform. The user terminal information equipment transmits and broadcasts the disaster information, providing timely information for disaster alarm and personnel evacuation. The system is a typical Internet of Things system, which is divided into sensing layer, transmission layer, platform layer and application layer. However, the existing debris flow geological disaster monitoring and early warning methods are relatively simple, with strong dependence on alarm information and prone to errors. Moreover, it is impossible to realize automatic graded alarm, and it is necessary to manually judge the alarm level. Summary of the invention

[0004] The purpose of the present invention is to provide a monitoring and alarm method, system and storage medium for debris flow along traffic lines, aiming to solve the above-mentioned problems.

[0005] The present invention is mainly achieved through the following technical solutions:

[0006] A method for monitoring and alarming debris flow along a traffic line comprises the following steps:

[0007] Step S1: a plurality of warning lights of the first level, the second level and the third level are sequentially arranged on the outside of the traffic line from the inside to the outside, the status of the plurality of warning lights of each level is obtained in real time, and a real-time warning score s1 is calculated;

[0008] S1=αA+βB+θC

[0009] Among them: α, β, θ are the alarm coefficients of the first level, second level and third level respectively;

[0010] A, B, and C are the number of warning light alarms at the first, second, and third levels, respectively;

[0011] Step S2: respectively obtain monitoring videos at the first level, the second level and the third level outside the traffic line, then determine the debris flow event based on the target detection algorithm, obtain the debris flow target confidence, and calculate the monitoring warning score s2;

[0012] S2=λD image1 +λD image2 +ηD image3 Where: D image1 , D image2 , D image3 are the confidence levels of debris flow targets at the first, second and third levels respectively;

[0013] λ, γ, and η are the confidence weights of the first, second, and third levels, respectively;

[0014] Step S3: Based on the real-time warning score s1 obtained in step S1 and the monitoring warning score s2 obtained in step S2, a debris flow risk score s is calculated:

[0015] S=S1+S2+b

[0016] Where: b is the risk compensation value;

[0017] Step S4: Based on the set thresholds of the third-level alarm, the second-level alarm, and the first-level alarm, the alarm level corresponding to the debris flow risk score s in step S3 is determined, and an external alarm information of the corresponding level is issued.

[0018] In order to better implement the present invention, further, the step S4 further includes using a linear regression algorithm to correct the calculation formulas in steps S1 to S3 based on feedback from staff; the step S4 includes the following steps:

[0019] Step S41: Based on the staff's confirmation, the alarm result is corrected to any one of a false alarm, a third-level alarm, a second-level alarm, and a first-level alarm; the comprehensive score s corresponding to the false alarm is set to 0, and the comprehensive scores corresponding to the third-level alarm, the second-level alarm, and the first-level alarm are set to They are 1.0, 1.4 and 1.8 respectively;

[0020] Step S42: using a linear regression algorithm to correct the parameters α, β, θ, λ, γ, η, and b;

[0021] Step S421: Establish a linear regression model of y=wx+b,

[0022] Where: x is A, B, C, D image1 , D image2 , D image3 The vector composed of

[0023] w is a vector composed of α, β, θ, λ, γ, and η;

[0024] The loss function L is defined as:

[0025]

[0026] Where: n is the number of alarm messages after manual confirmation;

[0027] is the comprehensive score s set after the i-th alarm information is manually confirmed;

[0028] s i is the comprehensive score s calculated from the ith alarm information;

[0029] Step S422: Calculate the partial derivatives of w and b respectively, and update the values ​​of w and b by the following formula:

[0030]

[0031] Where: σ is the learning rate;

[0032] Step S423: Iterate step S422 for a set number of times and then end;

[0033] Step S43: Based on the corrected α, β, θ, λ, γ, η, and b parameters, update the calculation formulas of steps S1 to S3.

[0034] In order to better implement the present invention, further, the initial values ​​of α, β, θ are 0.048, 0.096, and 0.190; the initial values ​​of λ, γ, and η are all 0.33; and the initial value of b is 0.

[0035] In order to better implement the present invention, further, in the step S1, first, second and third levels of dams and ditches are sequentially arranged on the outside of the traffic line from the inside to the outside, and a number of warning lights are respectively arranged on the dams and ditches of each level.

[0036] In order to better implement the present invention, further, step S2 includes the following steps:

[0037] Step S21: The monitoring camera device collects continuous video data streams in real time, and analyzes and processes each frame based on the target detection algorithm of YOLOX or YOLOV8;

[0038] Step S22: setting the area where the dam and the ditch are located as the detection area, and filtering the area of ​​the surrounding mountains and forests to obtain the corresponding information of the debris flow target;

[0039] Step S23: adopt a voting mechanism to determine the debris flow event and obtain the debris flow target confidence.

[0040] In order to better implement the present invention, further, in the step S2, cameras are respectively set at the warning light position, 50 meters upstream of the warning light dam, and 50 meters downstream of the warning light.

[0041] In order to better implement the present invention, further, in step S4, the set thresholds of the third-level alarm, the second-level alarm, and the first-level alarm are 0.8, 1.2, and 1.6, respectively.

[0042] The present invention is mainly achieved through the following technical solutions:

[0043] A monitoring and alarm system for debris flow along traffic routes, based on the above method, includes a warning light status detection module, a real-time warning score calculation module, a video monitoring module, a target detection module, a monitoring warning score calculation module, a debris flow risk score calculation module and an alarm level determination module;

[0044] The warning light status detection module is used to detect the status of a number of warning lights at each level and send the status to the real-time warning score calculation module, and the real-time warning score calculation module is used to calculate the real-time warning score;

[0045] The video monitoring module is used to obtain monitoring videos of each level and send them to the target detection module, the target detection module is used to detect debris flow targets in the video stream and calculate the debris flow target confidence, and send it to the monitoring warning score calculation module; the monitoring warning score calculation module is used to calculate the monitoring warning score;

[0046] The debris flow risk score calculation module is used to calculate the debris flow risk score based on the real-time warning score and the monitoring warning score;

[0047] The alarm level determination module is used to obtain the alarm level based on the debris flow risk score and issue external alarm information of the corresponding level.

[0048] A computer-readable storage medium stores a computer program, which implements the above method when executed by a processor.

[0049] The beneficial effects of the present invention are as follows:

[0050] The present invention accurately calculates the real-time warning score through the status of the warning lights at each level; and obtains the monitoring warning score of the corresponding level based on the target detection algorithm of the video monitoring, and finally obtains the debris flow risk score based on the sum of the real-time warning score and the monitoring warning score, and then compares and obtains the final alarm level. The present invention obtains the final alarm level by fusing the detection information in two directions, improves the fusion utilization of information, and improves the detection efficiency, so that the present invention has the advantages of fast and efficient monitoring and alarming, higher accuracy and stronger robustness. Furthermore, the present invention combines manual confirmation of alarm data, and continuously corrects the parameters in the debris flow risk scoring formula through a linear regression algorithm, which further improves the accuracy of the early warning and has good practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 The present invention is a flow chart of the method for monitoring and alarming debris flows along traffic lines. DETAILED DESCRIPTION

[0052] Embodiment 1:

[0053] A method for monitoring and alarming debris flow along a traffic line comprises the following steps:

[0054] Step 1: The warning light gives a graded alarm according to preset rules.

[0055] The ditches and dams beside the highway are important engineering measures to prevent debris flow, and their main purpose is to control and reduce the occurrence of debris flow disasters. Generally, there are one to three levels of dams and ditches around the highway, and guardrails are installed on the side of the road. Three warning lights are installed at each level of dams and ditches, and multiple warning lights are also installed on the guardrails. The alarm platform can obtain the status information of all warning lights in real time (whether it has been hit, whether the light is on, location, etc.), and different levels of warning scores can be obtained according to the preset graded alarm rules.

[0056] Step 2: Perform debris flow detection based on the video stream of the highway monitoring system.

[0057] Step 2.1: The surveillance camera device collects continuous video data streams in real time and analyzes and processes each frame based on the target detection algorithm.

[0058] Step 2.2, construct the detection area, filter the data, and obtain the corresponding information of the debris flow target.

[0059] Step 2.3, use the voting mechanism to determine the debris flow event and obtain the confidence level of the event.

[0060] Step 3: The weighted decision fusion method is used to determine the debris flow event and issue an alarm.

[0061] Through steps 1 and 2, we can obtain the graded alarm data of the warning lights beside the highway and the debris flow detection data of the highway monitoring video stream. Then, we summarize the results, conduct a comprehensive weighted evaluation of the warning light alarm score and the monitoring video stream detection results, and calculate the comprehensive risk score. According to the comprehensive risk score, we determine the final alarm level.

[0062] According to the final alarm level, the platform can trigger multiple warning lights in different colors in front of the debris flow area on the highway to light up, which can warn passing vehicles of the debris flow risk on the road ahead. Through the method of multi-source information fusion, the monitoring and alarm of debris flow events on highways can be realized, which has strong robustness, high timeliness and accuracy.

[0063] Embodiment 2:

[0064] A monitoring and alarm method for debris flow along traffic lines, such as Figure 1 As shown, the following steps are included:

[0065] Step 1: The warning light gives a graded alarm according to preset rules.

[0066] The guardrail is closest to the expressway and is equipped with multiple warning lights. The ditch and dam farthest from the expressway are the third-level ditch and third-level dam, and the closest are the second-level ditch and second-level dam. The dams and ditches are arranged alternately, and each level of dams and ditches is equipped with a group of warning lights, with 3 warning lights in one group.

[0067] When a debris flow occurs, the warning lights of the third-level ditch and the third-level dam are the first to be collided. The triggering rule is: if any one or more of the three warning lights in each level collide, the alarm of that level will be triggered, that is, the warning lights of the third-level ditch and the third-level dam will trigger the third-level alarm, and the warning lights of the second-level ditch and the second-level dam will trigger the second-level alarm. If any warning light installed on the guardrail is collided, the first-level alarm will be triggered.

[0068] The third-level alarm coefficient is α, the number of alarm lights is A, the second-level alarm coefficient is β, the number of alarm lights is B, the first-level alarm coefficient is θ, the number of alarm lights is C. The values ​​of A, B, and C are 0 to 3, and the initial values ​​of α, β, and θ are 0.048, 0.096, and 0.190, respectively. The real-time warning score s1 is calculated by the following formula:

[0069] S1=αA+βB+θC

[0070] Step 2. Debris flow detection from highway monitoring system video stream.

[0071] Step 2.1, construct the detection area and filter irrelevant data.

[0072] In the detection of debris flow on highways, the video stream comes from the monitoring cameras installed along the line. These cameras continuously capture real-time images of the road and surrounding environment, and transmit them to the monitoring system for analysis and processing. The camera has three preset positions, namely the warning light position, 50 meters upstream of the warning light dam, and 50 meters downstream of the warning light. The area where the dam ditch is located is set as the detection area, and the surrounding mountains, woods and other areas are filtered out to reduce false detections caused by the influence of other areas in the image.

[0073] Step 2.2, obtain the corresponding information of the debris flow target.

[0074] Perform target detection on the three preset videos in step 2.1, capture one frame from the video, and independently perform target detection on the image using a target detection algorithm, such as YOLOX or YOLOV8, to obtain information such as the position, category, and confidence of the target detection box, where the confidence value ranges from 0 to 1.

[0075] Step 2.3, using the voting mechanism and combining the confidence level of the debris flow target, the video alarm score is calculated by the following formula, where D image1 , D image2 , D image3 are the confidence levels of the debris flow targets at preset positions 1, 2, and 3 respectively. The initial values ​​of λ, γ, and η are all 0.33

[0076] S2=λD image1 +λD image2 +ηD image3

[0077] Step 3: The weighted decision fusion method is used to determine the debris flow event and issue an alarm.

[0078] The specific working process of step 3 is:

[0079] Step 3.1: According to step 1 and step 2, the debris flow alarm comprehensive score s of the highway monitoring system video stream can be obtained, where the initial value of b is 0:

[0080] S=S1+S2+b

[0081] In step 3.2, the third-level alarm threshold is set to 0.8, the second-level alarm threshold is set to 1.2, and the first-level alarm threshold is set to 1.6. When the comprehensive score of step 3 exceeds a certain level of alarm threshold, it is judged as a debris flow event of that level and the alarm level is determined.

[0082] Step 3.3, trigger a debris flow event and issue an external alarm.

[0083] The platform obtains the final alarm level and triggers the warning lights on both sides of the road a few kilometers ahead of the debris flow area on the highway. Depending on the final alarm level, the color of the warning lights is different, such as red light for the first level alarm, yellow light for the second level alarm, and flashing yellow light for the third level alarm.

[0084] Step 4. System parameter correction

[0085] During the operation of the system, the seven parameters α, β, θ, λ, γ, η, and b in the comprehensive score can be automatically learned and corrected according to actual conditions.

[0086] Step 4.1: After the system alarms, the staff will confirm the alarm event. If it is a false alarm or does not match the actual alarm level, the staff will confirm the alarm result in the system and set the alarm to one of the following: false alarm, level 3 alarm, level 2 alarm, or level 1 alarm.

[0087] Step 4.2: Use linear regression algorithm to correct parameters. The specific steps are as follows:

[0088] Step 4.2.1, establish the model as y=wx+b,

[0089] Where: x is A, B, C, D image1 , D image2 , D image3 The vector composed of

[0090] w is a vector composed of α, β, θ, λ, γ, and η.

[0091] Step 4.2.2: Set the comprehensive score s of the events confirmed by the staff and the false alarm events to 0, and the comprehensive scores of the third-level alarm, the second-level alarm, and the first-level alarm They are set to 1.0, 1.4, and 1.8 respectively, and the loss function is defined as follows:

[0092]

[0093] Where: n is the number of alarm messages after manual confirmation;

[0094] is the comprehensive score s set after the i-th alarm information is manually confirmed;

[0095] s i is the comprehensive score s calculated from the i-th alarm information.

[0096] In step 4.2.3, calculate the partial derivatives of w and b respectively, and update the values ​​of w and b according to the following formula, where σ is the learning rate, and the typical value is 0.01. The iteration ends after m times, and the typical value of m is 500.

[0097]

[0098] In step 4.2.4, the modified parameters are applied to the system, and the subsequent comprehensive alarm score s is calculated using the modified parameters.

[0099] The present invention accurately calculates the real-time warning score through the status of the warning lights at each level; and obtains the monitoring warning score of the corresponding level based on the target detection algorithm of video monitoring, and finally obtains the debris flow risk score based on the sum of the real-time warning score and the monitoring warning score, and then compares and obtains the final alarm level. The present invention obtains the final alarm level by fusing the detection information in two directions, improves the fusion utilization of information, improves the detection efficiency, and makes the result more robust and more accurate. Furthermore, the present invention combines manual confirmation of alarm data, and continuously corrects the parameters in the debris flow risk score formula through a linear regression algorithm, which further improves the accuracy of the early warning and has good practicality.

[0100] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any simple modification or equivalent change made to the above embodiment based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A method for monitoring and alarming debris flow along a traffic line, characterized in that: The following steps are involved: Step S1: a plurality of warning lights of the first level, the second level and the third level are sequentially arranged on the outside of the traffic line from the inside to the outside, the status of the plurality of warning lights of each level is obtained in real time, and a real-time warning score s1 is calculated; S1=αA+βB+θC Among them: α, β, θ are the alarm coefficients of the first level, second level and third level respectively; A, B, and C are the number of warning light alarms at the first, second, and third levels, respectively; Step S2: respectively obtain monitoring videos at the first level, the second level and the third level outside the traffic line, then determine the debris flow event based on the target detection algorithm, obtain the debris flow target confidence, and calculate the monitoring warning score s2; S2=λD image1 +γD image2 +ηD image3 Where: D image1 , D image2 , D image3 are the confidence levels of debris flow targets at the first, second and third levels respectively; λ, γ, and η are the confidence weights of the first, second, and third levels, respectively; Step S3: Based on the real-time warning score s1 obtained in step S1 and the monitoring warning score s2 obtained in step S2, a debris flow risk score s is calculated: S=S1+S2+b Where: b is the risk compensation value; Step S4: Based on the set thresholds of the third-level alarm, the second-level alarm, and the first-level alarm, the alarm level corresponding to the debris flow risk score s in step S3 is determined, and an external alarm information of the corresponding level is issued.

2. The method for monitoring and alarming debris flow along traffic lines according to claim 1, characterized in that: The step S4 further includes using a linear regression algorithm to correct the calculation formulas in steps S1 to S3 based on the feedback from the staff; the step S4 includes the following steps: Step S41: Based on the staff's confirmation, the alarm result is corrected to any one of a false alarm, a third-level alarm, a second-level alarm, and a first-level alarm; the comprehensive score s corresponding to the false alarm is set to 0, and the comprehensive scores corresponding to the third-level alarm, the second-level alarm, and the first-level alarm are set to They are 1.0, 1.4 and 1.8 respectively; Step S42: using a linear regression algorithm to correct the parameters α, β, θ, λ, γ, η, and b; Step S421: Establish a linear regression model of y=wx+b, Where: x is A, B, C, D image1 , D image2 , D image3 The vector composed of w is a vector composed of α, β, θ, λ, γ, and η; The loss function L is defined as: Where: n is the number of alarm messages after manual confirmation; is the comprehensive score s set after the i-th alarm information is manually confirmed; s i is the comprehensive score s calculated from the ith alarm information; Step S422: Calculate the partial derivatives of w and b respectively, and update the values ​​of w and b by the following formula: Where: σ is the learning rate; Step S423: Iterate step S422 for a set number of times and then end; Step S43: Based on the corrected α, β, θ, λ, γ, η, and b parameters, update the calculation formulas of steps S1 to S3.

3. A method for monitoring and alarming debris flow along traffic lines according to claim 1 or 2, characterized in that: The initial values ​​of α, β, and θ are 0.048, 0.096, and 0.190; the initial values ​​of λ, γ, and η are all 0.33; and the initial value of b is 0.

4. The method for monitoring and alarming debris flow along traffic lines according to claim 1, characterized in that: In the step S1, first-level, second-level and third-level dams and ditches are sequentially arranged from the inside to the outside on the outside of the traffic line, and a number of warning lights are respectively arranged on the dams and ditches of each level.

5. The method for monitoring and alarming debris flow along traffic lines according to claim 4, characterized in that: The step S2 comprises the following steps: Step S21: The monitoring camera device collects continuous video data streams in real time, and analyzes and processes each frame based on the target detection algorithm of YOLOX or YOLOV8; Step S22: setting the area where the dam and the ditch are located as the detection area, and filtering the area of ​​the surrounding mountains and forests to obtain the corresponding information of the debris flow target; Step S23: adopt a voting mechanism to determine the debris flow event and obtain the debris flow target confidence.

6. A method for monitoring and alarming debris flow along traffic lines according to claim 5, characterized in that: In step S2, cameras are respectively set at the position of the warning light, 50 meters upstream of the warning light dam, and 50 meters downstream of the warning light.

7. The method for monitoring and alarming debris flow along traffic lines according to claim 1, characterized in that: In step S4, the set thresholds for the third-level alarm, the second-level alarm, and the first-level alarm are 0.8, 1.2, and 1.6, respectively.

8. A monitoring and alarm system for debris flow along traffic lines, based on a monitoring and alarm method for debris flow along traffic lines as claimed in any one of claims 1 to 7, characterized in that: It includes a warning light status detection module, a real-time warning score calculation module, a video monitoring module, a target detection module, a monitoring warning score calculation module, a debris flow risk score calculation module and an alarm level determination module; The warning light status detection module is used to detect the status of a number of warning lights at each level and send the status to the real-time warning score calculation module, and the real-time warning score calculation module is used to calculate the real-time warning score; The video monitoring module is used to obtain monitoring videos of each level and send them to the target detection module, the target detection module is used to detect debris flow targets in the video stream and calculate the confidence of the debris flow targets, and send them to the monitoring warning score calculation module; the monitoring warning score calculation module is used to calculate the monitoring warning score; The debris flow risk score calculation module is used to calculate the debris flow risk score based on the real-time warning score and the monitoring warning score; The alarm level determination module is used to obtain the alarm level based on the debris flow risk score and issue external alarm information of the corresponding level.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a method for monitoring and alarming debris flows along traffic routes as described in any one of claims 1 to 6 is implemented.

Citation Information

Cited By

  • Road side slope collapse monitoring and analyzing method for disaster risk identification

    CN122245139A