Displacement detection method of high hanging objects in tunnel based on visual recognition technology

Abstract

The invention discloses a method for detecting displacement of high-hanging objects in a tunnel based on visual recognition technology, and belongs to the field of displacement detection technology. The invention extracts the equipment characteristics of each high-hanging object, and classifies the high-hanging objects according to the equipment characteristics; uses a visual sensor to collect plane images of the high-hanging objects in the tunnel, uses the collected multiple plane images to build a three-dimensional model of each high-hanging object, and lays points on the three-dimensional model of the high-hanging object to obtain a three-dimensional point set of each high-hanging object; analyzes the displacement of each point to find three positioning points for judging the displacement of the high-hanging object; respectively calculates the positions of the three positioning points of each high-hanging object when it is working normally and the displacement threshold of each high-hanging object when it is working in the tunnel; uses the displacement threshold of each high-hanging object to judge the real-time displacement and issue an abnormal warning; calculates the drop threshold and the fault threshold, and uses the two thresholds to classify the abnormal warning.

Description

Technical Field

[0001] The present invention relates to the technical field of displacement detection, and in particular to a method for detecting the displacement of high-hanging objects in a tunnel based on visual recognition technology. Background Art

[0002] Tunnels are an important part of transportation and infrastructure, and their safety is crucial to public safety and operational efficiency. If suspended objects in tunnels, such as cables and pipes, are displaced or deformed, they may cause safety hazards such as power outages, communication failures, or fires. Traditional detection of high-hanging objects in tunnels mainly relies on manual inspection or mechanical sensors. These methods usually have the following problems: time-consuming and dangerous, especially in tunnels with heavy traffic or harsh environments, high installation and maintenance costs, and may affect the tunnel structure; and now with the rapid development of automation technology, sensor technology, computer technology, etc., the displacement detection and maintenance components of high-hanging objects in tunnels have been replaced by manual inspection, and intelligent systems are used to detect the displacement of high-hanging objects in tunnels. However, there are many types of high-hanging objects in tunnels, and each high-hanging object has different shapes and sizes. Therefore, the intelligent system cannot accurately judge the displacement detection of high-hanging objects in tunnels, and there is a large judgment error for irregular high-hanging objects. There is no further judgment on the displacement of high-hanging objects, which makes maintenance personnel unclear about the specific displacement of high-hanging objects. Therefore, how to avoid judgment errors caused by irregular high-hanging objects in tunnels and further judgment of abnormal warnings is crucial. Summary of the invention

[0003] The purpose of the present invention is to provide a method for detecting the displacement of high hanging objects in a tunnel based on visual recognition technology, so as to solve the problems raised in the above-mentioned background technology.

[0004] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0005] A method for detecting displacement of high hanging objects in a tunnel based on visual recognition technology, the method comprising the following steps:

[0006] S100, collecting historical equipment data of high-hanging objects in tunnels, extracting equipment features of each high-hanging object, and classifying the high-hanging objects according to the equipment features;

[0007] Furthermore, the specific steps for classifying high hanging objects according to equipment characteristics are as follows:

[0008] S101, collect the equipment data of each high hanging object in the tunnel in history as {S1, S2, S3, ..., S n}, S1, S2, S3, ..., S nrepresents the 1st, 2nd, 3rd, ..., nth device data of the high hanging objects in the tunnel collected, where n is a positive integer; the values ​​of each device data before and after the high hanging objects in the tunnel are displaced in the collection history are {S_z1, S_z2, S_z3, ..., S_z n}, S_z1, S_z2, S_z3,..., S_z n Indicates the 1st, 2nd, 3rd, ..., nth device data values ​​before the high hanging object is displaced in the collected history; {S_z'1, S_z'2, S_z'3, ..., S_z' n}, S_z'1, S_z'2, S_z'3,..., S_z' n Indicates the 1st, 2nd, 3rd, ..., nth device data values ​​after the high hanging objects are displaced in the collected history;

[0009] S102, collecting the equipment data before and after the high hanging objects are displaced m times, and calculating the average difference of each equipment data before and after the high hanging objects in the tunnel are displaced in history, the formula is:

[0010]

[0011] In the formula, S_c represents the difference between each device data before and after the displacement of the high hanging objects in the tunnel in the calculated history, S_z represents the device data value collected in the history before the displacement of the high hanging objects, and S_z' represents the device data value collected in the history after the displacement of the high hanging objects; the average device data differences obtained by calculating the collected n types of high hanging object data are {S_c1, S_c2, S_c3, ..., S_c n}, S_c1, S_c2, S_c3,..., S_c n Indicates the calculated data difference of the 1st, 2nd, 3rd, ..., nth type of equipment for high hanging objects;

[0012] S103, judging the difference of n types of equipment data after the displacement of the high hanging object, extracting the maximum value of the equipment data difference as the equipment feature of the high hanging object, calculating the equipment of all the high hanging objects in the tunnel, and obtaining the equipment feature of each high hanging object as {T1, T2, T3, ..., T B}, T1, T2, T3, ..., T B represents the equipment characteristics of the 1st, 2nd, 3rd, ..., Bth high hanging objects in the tunnel, where B is a positive integer; high hanging objects with the same equipment characteristics are selected as the same type of high hanging objects, and finally the types of high hanging objects in the tunnel are determined to be {H1, H2, H3, ..., H g}, H1, H2, H3, ..., H g Represents the 1st, 2nd, 3rd, ...gth high hanging objects after classification, where g is a positive integer.

[0013] After classifying the high hanging objects in the tunnel, when conducting various types of monitoring and analysis on the high hanging objects in the tunnel, since the same types of high hanging objects have the same characteristics, it is only necessary to judge the type of high hanging objects, and various types of analysis can be carried out more conveniently according to the type of high hanging objects;

[0014] S200, using a visual sensor to collect plane images of high-hanging objects in the tunnel, using the collected multiple plane images to construct a three-dimensional model of each high-hanging object, analyzing the three-dimensional model of each high-hanging object, laying out points on the three-dimensional model of the high-hanging object, and obtaining a three-dimensional point set of each high-hanging object;

[0015] Furthermore, the specific steps for obtaining the three-dimensional point set of each high hanging object are:

[0016] S201, using visual sensors to photograph and measure the plane images of all high hanging objects in the tunnel, extracting data from the plane images, and selecting the maximum distance of the high hanging objects in the three-dimensional direction in the plane image as (X max , Y max , Z max ), X max Indicates the maximum length of the high hanging object in the plane image, Y max Indicates the maximum height of the high hanging object in the plane image, Z max Indicates the maximum width of the high hanging object in the plane image; the maximum circumscribed rectangle of the high hanging object is drawn using the maximum distance of the high hanging object in the three-dimensional direction in the plane image. (X max , Y max , Z max ) are the length, height and width of the largest circumscribed rectangle respectively;

[0017] S202, collect all the records of displacement of each high hanging object in the tunnel in history, extract the displacement amount of each high hanging object in the records when it is displaced {L1, L2, L3, ..., L k}, L1, L2, L3, ..., L k Indicates the displacement of the high hanging objects collected in history for the 1st, 2nd, 3rd, ..., kth time, where k is a positive integer; compares the displacement of the high hanging objects collected for k times, and selects the minimum displacement as the minimum interval L_min when laying points for the maximum circumscribed rectangle of the high hanging objects; calculates the high hanging objects in B tunnels to obtain the maximum circumscribed rectangle and minimum spacing of each high hanging object;

[0018] S203, using the calculated minimum spacing to lay points on the maximum circumscribed rectangle of each high hanging object, and calculating the number of laid points in different three-dimensional directions, the formula is:

[0019]

[0020] In the formula, Q X , Q Y , Q Z It is expressed as the total number of points in the length, height and width directions when laying points on the maximum circumscribed rectangle of the high hanging object. max , Y max , Z max represents the length, height and width of the maximum circumscribed rectangle of the high hanging object, L_min represents the minimum spacing when laying points for each high hanging object; for each high hanging object in the tunnel, the corresponding maximum circumscribed rectangle is laid using points to form a maximum circumscribed rectangle composed of points; all points constituting the maximum circumscribed rectangle are classified into a unified set as a three-dimensional point set, and finally the three-dimensional point set of all high hanging objects in the tunnel is {G1, G2, G3, ..., G B}, G1, G2, G3, ..., G B Represents the three-dimensional point set of the 1st, 2nd, 3rd, ..., Bth high hanging objects in the tunnel.

[0021] Drawing the maximum circumscribed rectangle of a high hanging object and laying out the points can clearly find the positioning points where the high hanging object reflects excessive displacement when it is displaced, thus avoiding the inability to accurately find the positioning points due to the different irregular shapes of different high hanging objects and the inability to judge the displacement of the high hanging objects;

[0022] S300, collecting the displacement of each point in the three-dimensional point set under different working conditions of the high-hanging objects in the tunnel after the high-hanging objects have been displaced in the history, analyzing the displacement of each point and searching for three positioning points for determining the displacement of the high-hanging objects;

[0023] Furthermore, the specific steps of analyzing the displacement of each point and finding the three positioning points for determining the displacement of the high hanging object are as follows:

[0024] S301, collect P records of displacement of high hanging objects in the tunnel causing abnormality of the high hanging objects in history, extract the displacement of each point in the maximum circumscribed rectangular three-dimensional point set of the high hanging objects in the records: Indicates the 1st, 2nd, 3rd, ..., Qth points in the maximum circumscribed rectangle of the high hanging object in the extracted record X +Q Y +Q Z The displacement of each point, Q X +Q Y +Q ZThe displacement of each point is compared, and the maximum value is selected as the feature point of the collected high-hanging object displacement that causes the high-hanging object to generate abnormal records. The feature points in P records are extracted, and the number of times the same feature point appears in P records is calculated. The feature point with the largest number of occurrences is used as the farthest positioning point A_far among the three high-hanging object positioning points;

[0025] S302, collect the R records in history where the high hanging objects in the tunnel are displaced but the high hanging objects do not produce abnormalities, and similarly extract the displacement of each point in the maximum circumscribed rectangular three-dimensional point set of the high hanging objects in the records: Indicates the 1st, 2nd, 3rd, ..., Qth points in the maximum circumscribed rectangle of the high hanging object in the extracted record X +Q Y +Q Z The displacement of each point, Q X +Q Y +Q Z The displacement of each point is compared, and the minimum value is selected as the feature point of the collected high-hanging object displacement but no abnormal record of the high-hanging object is generated. The feature points in the R records are extracted, and the number of times the same feature point appears in the R records is calculated. The feature point with the largest number of occurrences is used as the nearest positioning point A_re among the three high-hanging object positioning points;

[0026] S303, calculate the median point of the three positioning points of the high hanging object according to the farthest positioning point and the nearest positioning point determined, use the laser radar to measure the distances between the farthest positioning point, the nearest positioning point and the tunnel wall in the maximum circumscribed rectangle of the high hanging object, respectively L_far and L_re, calculate the median value of the distance between the farthest positioning point, the nearest positioning point and the tunnel wall, and determine the median point of the three positioning points, the formula is:

[0027]

[0028] In the formula, M_A represents the distance between each point in the maximum circumscribed rectangle of the high hanging object and the median value of the distance between the farthest positioning point, the nearest positioning point and the tunnel wall, and L_A represents the distance between each point in the maximum circumscribed rectangle of the high hanging object and the tunnel wall. It represents the intermediate value of the distance between the farthest positioning point, the nearest positioning point and the tunnel wall; each point in the three-dimensional point set of the maximum circumscribed rectangle of the high hanging object is calculated to obtain Q X +Q Y +Q Z -The distance between the two points and the middle value, select the point corresponding to the minimum value as the median point A_mid among the three positioning points.

[0029] When the high hanging objects are displaced, the three main positioning points can ensure the accuracy of the displacement judgment of the high hanging objects. The farthest positioning point ensures a judgment when the high hanging objects are rotating and displacing. The combination of the nearest positioning point and the midpoint point can judge the displacement of the high hanging objects in any direction.

[0030] S400, collecting the two equipment positions of each high hanging object in the tunnel when an abnormality occurs and when it is working normally in history, and respectively calculating the positions of three positioning points of each high hanging object when it is working normally and the displacement threshold of each high hanging object when it is working in the tunnel;

[0031] Furthermore, the specific steps for calculating the positions of the three points of each high hanging object during normal operation are as follows:

[0032] S411, establish a spatial rectangular coordinate system in the tunnel, take the laser radar as the origin, the direction perpendicular to the horizontal plane and the tunnel wall as the X-axis, the direction parallel to the horizontal plane and the tunnel wall as the Z-axis, and the vertical direction perpendicular to the ground as the Y-axis to establish the spatial coordinate system;

[0033] S412. Collect historical images of each high-hanging object in the tunnel when it is working normally without displacement, map the collected historical images using the three-dimensional point set of the maximum circumscribed rectangle of each high-hanging object, determine the positions of the three positioning points in the historical images when the high-hanging object is working normally without displacement, and use trigonometric functions in the spatial rectangular coordinate system in the constructed tunnel to calculate the normal coordinates of the three positioning points when each high-hanging object is working normally without displacement.

[0034] The specific steps for the displacement threshold of each type of high hanging object when working in the tunnel are:

[0035] S421. Collect the displacement of three positioning points when each high hanging object in the tunnel is displaced and causes abnormality of the high hanging object in the history. The displacement of the farthest positioning point is

[0036] {J_far1, J_far2, J_far3,...,J_far N}, J_far1, J_far2, J_far3,..., J_far N It indicates the displacement of the farthest positioning point when the high hanging object is displaced for the 1st, 2nd, 3rd, ..., Nth time in the collected history, causing the high hanging object to be abnormal. The displacement of the nearest positioning point is

[0037] {J_re1, J_re2, J_re3,...,J_re N}, J_re1, J_re2, J_re3,..., J_re N It represents the displacement of the nearest positioning point when the high hanging object is displaced for the 1st, 2nd, 3rd, ..., Nth time in the collected history, causing the high hanging object to become abnormal. The displacement of the median point is

[0038] {J_mid1, J_mid2, J_mid3,...,J_mid N},

[0039] J_mid1, J_mid2, J_mid3, ..., J_mid N It indicates the median displacement of the 1st, 2nd, 3rd, ..., Nth time in the collected history when the high hanging object is displaced and causes the high hanging object to be abnormal, and N is a positive integer;

[0040] S422. Calculate the displacement threshold of the high hanging object according to the displacement of the three positioning points when each high hanging object is displaced and causes the high hanging object to be abnormal in the collected history. The formula is:

[0041]

[0042] In the formula, Th_far represents the displacement threshold of the maximum positioning point in the calculated high hanging object. The displacement thresholds of the minimum positioning point and the midpoint are calculated using the same formula to obtain the minimum positioning point displacement threshold Th_re and the midpoint displacement threshold Th_mid.

[0043] S500, using laser radar to locate three points of each high hanging object working in the tunnel based on the displacement distance of the normal working point, and using the displacement threshold of each high hanging object to judge the real-time displacement and issue an abnormal warning;

[0044] Furthermore, the specific steps of using the displacement threshold of each high hanging object to judge the real-time displacement and issue an abnormal warning are as follows:

[0045] S501, using laser radar to measure in real time the distance between the three positioning points on each high hanging object in the tunnel and the normal coordinates calculated in S412: {F(far, re, mid)1, F(far, re, mid)2, F(far, re, mid)3, ..., F(far, re, mid) B}, F(far,re,mid)1,F(far,re,mid)2,F(far,re,mid)3,...,F(far,re,mid) B It indicates the distance between the three positioning points on the 1st, 2nd, 3rd, ..., Bth high hanging objects and the normal coordinates measured in real time by the laser radar;

[0046] S502. Use the displacement threshold of the three positioning points of the high hanging objects to judge the real-time measured distance. When F(far, re, mid) ≥ Th_(far, re, mid) and at least one of the three positioning points satisfies the inequality, it is judged that the corresponding high hanging object has an abnormal risk and an abnormal warning is issued; when F(far, re, mid) < Th_(far, re, mid) and all three positioning points are less than the real-time distance, it is judged that there is no abnormal risk for the high hanging object. Th_(far, re, mid) represents the displacement threshold of the three positioning points of the high hanging object; the displacements of B high hanging objects in the tunnel are judged to obtain a set of abnormal high hanging objects that have abnormal risks and issue warnings.

[0047] S600: After determining the displacement of the high hanging objects in the tunnel and issuing an early warning, collect the displacement numbers of the high hanging objects in history when they fall or fail to work normally, calculate the falling threshold and the failure threshold, and use the two thresholds to classify abnormal early warnings.

[0048] Furthermore, the specific steps for calculating the drop threshold and the failure threshold are as follows:

[0049] S601, collect the median displacement of the high hanging objects in the tunnel when they fall in the history {L_un1, L_un2, L_un3, ..., L_un v}, L_un1, L_un2, L_un3,..., L_un v It represents the median displacement of the high hanging object when it falls for the 1st, 2nd, 3rd, ..., vth time in the collected history;

[0050] S602: Calculate the drop displacement threshold of the high hanging object, the formula is:

[0051]

[0052] In the formula, fall represents the displacement threshold of the high hanging object, the formula represents the calculation of the average value of the displacements less than the middle value in v historical displacements as the displacement threshold, u represents the number of displacements less than the middle value in v historical displacements obtained through judgment, L_un max Indicates the maximum historical displacement, L_un min Indicates the minimum value of historical displacement;

[0053] S602. Perform a secondary judgment on the distances between the three positioning points of the high-hanging objects measured in real time in S500. When F(mid)≥fall, it is judged that the corresponding high-hanging object has a risk of falling, and the warning is judged to be a falling warning; when F(mid)<fall, it is judged that the warning is only a fault threshold; maintenance personnel formulate different maintenance plans for high-hanging objects according to different warning types.

[0054] Classifying abnormal warnings of high-hanging objects and determining whether the warnings are falling or fault warnings can enable maintenance personnel to formulate different maintenance plans when performing abnormal maintenance on high-hanging objects, making it easier for maintenance personnel to complete the maintenance of high-hanging objects.

[0055] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0056] 1. After the present invention extracts and classifies the features of the high-hanging objects in the tunnel, it can reduce the steps of analyzing each high-hanging object according to the same features of different types of high-hanging objects, greatly reduce the amount of data calculation, speed up the calculation process, and make displacement detection more convenient.

[0057] 2. The present invention draws the maximum circumscribed rectangle for all high-hanging objects in the tunnel, and searches for the three positioning points of the high-hanging objects based on the maximum circumscribed rectangle, thereby avoiding the problem that the positioning points cannot be accurately found due to the different irregular shapes of different high-hanging objects, and the displacement of the high-hanging objects cannot be accurately judged.

[0058] 3. The present invention uses the normal coordinates of the three positioning points of each high-hanging object, the displacement threshold and the real-time displacement data to complete the monitoring of the displacement of the high-hanging objects. When the real-time displacement is greater than the threshold, an early warning is issued, and the types of early warnings are distinguished in detail to ensure that maintenance personnel formulate different maintenance plans to complete the maintenance of the high-hanging objects in the tunnel, thereby reducing the occurrence of abnormalities of high-hanging objects. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0060] In the attached picture:

[0061] Figure 1 It is a flow chart of the steps of the method for detecting the displacement of high hanging objects in a tunnel based on the visual recognition technology of the present invention;

[0062] Figure 2 It is a schematic diagram of the maximum circumscribed rectangle of the method for detecting the displacement of high hanging objects in a tunnel based on the visual recognition technology of the present invention. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] See also Figure 1-Figure 2, the present invention provides a technical solution:

[0065] A method for detecting displacement of high hanging objects in a tunnel based on visual recognition technology, the method comprising the following steps:

[0066] S100, collecting historical equipment data of high-hanging objects in tunnels, extracting equipment features of each high-hanging object, and classifying the high-hanging objects according to the equipment features;

[0067] The specific steps to classify high hanging objects according to equipment characteristics are:

[0068] S101. Collect the equipment data of each high hanging object in the tunnel in history.

[0069] {S1, S2, S3, ..., S n}, S1, S2, S3, ..., S n represents the 1st, 2nd, 3rd, ..., nth device data of the high hanging objects in the tunnel collected, where n is a positive integer; the values ​​of each device data before and after the high hanging objects in the tunnel are displaced in the collection history are {S_z1, S_z2, S_z3, ..., S_z n}, S_z1, S_z2, S_z3,..., S_z n Indicates the 1st, 2nd, 3rd, ..., nth device data values ​​before the high hanging object is displaced in the collected history; {S_z'1, S_z'2, S_z'3, ..., S_z' n}, S_z'1, S_z'2, S_z'3,..., S_z' n Indicates the 1st, 2nd, 3rd, ..., nth device data values ​​after the high hanging objects are displaced in the collected history;

[0070] S102, collecting the equipment data before and after the high hanging objects are displaced m times, and calculating the average difference of each equipment data before and after the high hanging objects in the tunnel are displaced in history, the formula is:

[0071]

[0072] In the formula, S_c represents the difference between each device data before and after the displacement of the high hanging objects in the tunnel in the calculated history, S_z represents the device data value collected in the history before the displacement of the high hanging objects, and S_z' represents the device data value collected in the history after the displacement of the high hanging objects; the average device data differences obtained by calculating the collected n types of high hanging object data are {S_c1, S_c2, S_c3, ..., S_c n}, S_c1, S_c2, S_c3,..., S_c n Indicates the calculated data difference of the 1st, 2nd, 3rd, ..., nth type of equipment for high hanging objects;

[0073] S103, judging the difference of n types of equipment data after the displacement of the high hanging object, extracting the maximum value of the equipment data difference as the equipment feature of the high hanging object, calculating the equipment of all the high hanging objects in the tunnel, and obtaining the equipment feature of each high hanging object as {T1, T2, T3, ..., T B}, T1, T2, T3, ..., T B represents the equipment characteristics of the 1st, 2nd, 3rd, ..., Bth high hanging objects in the tunnel, where B is a positive integer; high hanging objects with the same equipment characteristics are selected as the same type of high hanging objects, and finally the types of high hanging objects in the tunnel are determined to be {H1, H2, H3, ..., H g}, H1, H2, H3, ..., H g Represents the 1st, 2nd, 3rd, ...gth high hanging objects after classification, where g is a positive integer.

[0074] After classifying the high hanging objects in the tunnel, when conducting various types of monitoring and analysis on the high hanging objects in the tunnel, since the same types of high hanging objects have the same characteristics, it is only necessary to judge the type of high hanging objects, and various types of analysis can be carried out more conveniently according to the type of high hanging objects;

[0075] S200, using a visual sensor to collect plane images of high-hanging objects in the tunnel, using the collected multiple plane images to construct a three-dimensional model of each high-hanging object, analyzing the three-dimensional model of each high-hanging object, laying out points on the three-dimensional model of the high-hanging object, and obtaining a three-dimensional point set of each high-hanging object;

[0076] The specific steps to obtain the three-dimensional point set of each high hanging object are:

[0077] S201, using visual sensors to photograph and measure the plane images of all high hanging objects in the tunnel, extracting data from the plane images, and selecting the maximum distance of the high hanging objects in the three-dimensional direction in the plane image as (X max , Y max , Z max ), X max Indicates the maximum length of the high hanging object in the plane image, Y max Indicates the maximum height of the high hanging object in the plane image, Z max Indicates the maximum width of the high hanging object in the plane image; the maximum circumscribed rectangle of the high hanging object is drawn using the maximum distance of the high hanging object in the three-dimensional direction in the plane image. (X max , Y max , Z max ) are the length, height and width of the largest circumscribed rectangle respectively;

[0078] S202, collect all the records of displacement of each high hanging object in the tunnel in history, extract the displacement amount of each high hanging object in the records when it is displaced {L1, L2, L3, ..., L k},

[0079] L1, L2, L3, ..., L k Indicates the displacement of the high hanging objects collected in history for the 1st, 2nd, 3rd, ..., kth time, where k is a positive integer; compares the displacement of the high hanging objects collected for k times, and selects the minimum displacement as the minimum interval L_min when laying points for the maximum circumscribed rectangle of the high hanging objects; calculates the high hanging objects in B tunnels to obtain the maximum circumscribed rectangle and minimum spacing of each high hanging object;

[0080] S203, using the calculated minimum spacing to lay points on the maximum circumscribed rectangle of each high hanging object, and calculating the number of laid points in different three-dimensional directions, the formula is:

[0081]

[0082] In the formula, Q X , Q Y , Q Z It is expressed as the total number of points in the length, height and width directions when laying points on the maximum circumscribed rectangle of the high hanging object. max , Y max , Z max represents the length, height and width of the maximum circumscribed rectangle of the high hanging object, L_min represents the minimum spacing when laying points for each high hanging object; for each high hanging object in the tunnel, the corresponding maximum circumscribed rectangle is laid using points to form a maximum circumscribed rectangle composed of points; all points constituting the maximum circumscribed rectangle are classified into a unified set as a three-dimensional point set, and finally the three-dimensional point set of all high hanging objects in the tunnel is {G1, G2, G3, ..., G B}, G1, G2, G3, ..., G B Represents the three-dimensional point set of the 1st, 2nd, 3rd, ..., Bth high hanging objects in the tunnel.

[0083] Drawing the maximum circumscribed rectangle of a high hanging object and laying out the points can clearly find the positioning points where the high hanging object reflects excessive displacement when it is displaced, thus avoiding the inability to accurately find the positioning points due to the different irregular shapes of different high hanging objects and the inability to judge the displacement of the high hanging objects;

[0084] S300, collecting the displacement of each point in the three-dimensional point set under different working conditions of the high-hanging objects in the tunnel after the high-hanging objects have been displaced in the history, analyzing the displacement of each point and searching for three positioning points for determining the displacement of the high-hanging objects;

[0085] The specific steps for analyzing the displacement of each point and finding the three positioning points used to determine the displacement of high hanging objects are:

[0086] S301, collect P records of displacement of high hanging objects in the tunnel causing abnormality of the high hanging objects in history, extract the displacement of each point in the maximum circumscribed rectangular three-dimensional point set of the high hanging objects in the records:

[0087] Indicates the 1st, 2nd, 3rd, ..., Qth points in the maximum circumscribed rectangle of the high hanging object in the extracted record X +Q Y +Q Z The displacement of each point, Q X +Q Y +Q Z The displacement of each point is compared, and the maximum value is selected as the feature point of the collected high-hanging object displacement that causes the high-hanging object to generate abnormal records. The feature points in P records are extracted, and the number of times the same feature point appears in P records is calculated. The feature point with the largest number of occurrences is used as the farthest positioning point A_far among the three high-hanging object positioning points;

[0088] S302, collect the R records in history where the high hanging objects in the tunnel are displaced but the high hanging objects do not produce abnormalities, and similarly extract the displacement of each point in the maximum circumscribed rectangular three-dimensional point set of the high hanging objects in the records:

[0089] Indicates the 1st, 2nd, 3rd, ..., Qth points in the maximum circumscribed rectangle of the high hanging object in the extracted record X +Q Y +Q Z The displacement of each point, Q X +Q Y +Q Z The displacement of each point is compared, and the minimum value is selected as the feature point of the collected high-hanging object displacement but no abnormal record of the high-hanging object is generated. The feature points in the R records are extracted, and the number of times the same feature point appears in the R records is calculated. The feature point with the largest number of occurrences is used as the nearest positioning point A_re among the three high-hanging object positioning points;

[0090] S303, calculate the median point of the three positioning points of the high hanging object according to the farthest positioning point and the nearest positioning point determined, use the laser radar to measure the distances between the farthest positioning point, the nearest positioning point and the tunnel wall in the maximum circumscribed rectangle of the high hanging object, respectively L_far and L_re, calculate the median value of the distance between the farthest positioning point, the nearest positioning point and the tunnel wall, and determine the median point of the three positioning points, the formula is:

[0091]

[0092] In the formula, M_A represents the distance between each point in the maximum circumscribed rectangle of the high hanging object and the median value of the distance between the farthest positioning point, the nearest positioning point and the tunnel wall, and L_A represents the distance between each point in the maximum circumscribed rectangle of the high hanging object and the tunnel wall. It represents the intermediate value of the distance between the farthest positioning point, the nearest positioning point and the tunnel wall; each point in the three-dimensional point set of the maximum circumscribed rectangle of the high hanging object is calculated to obtain Q X +Q Y +Q Z -The distance between the two points and the middle value, select the point corresponding to the minimum value as the median point A_mid among the three positioning points.

[0093] When the high hanging objects are displaced, the three main positioning points can ensure the accuracy of the displacement judgment of the high hanging objects. The farthest positioning point ensures a judgment when the high hanging objects are rotating and displacing. The combination of the nearest positioning point and the midpoint point can judge the displacement of the high hanging objects in any direction.

[0094] S400, collecting the two equipment positions of each high hanging object in the tunnel when an abnormality occurs and when it is working normally in history, and respectively calculating the positions of three positioning points of each high hanging object when it is working normally and the displacement threshold of each high hanging object when it is working in the tunnel;

[0095] The specific steps for calculating the positions of the three points when each type of high hanging object is working normally are:

[0096] S411, establish a spatial rectangular coordinate system in the tunnel, take the laser radar as the origin, the direction perpendicular to the horizontal plane and the tunnel wall as the X-axis, the direction parallel to the horizontal plane and the tunnel wall as the Z-axis, and the vertical direction perpendicular to the ground as the Y-axis to establish the spatial coordinate system;

[0097] S412. Collect historical images of each high-hanging object in the tunnel when it is working normally without displacement, map the collected historical images using the three-dimensional point set of the maximum circumscribed rectangle of each high-hanging object, determine the positions of the three positioning points in the historical images when the high-hanging object is working normally without displacement, and use trigonometric functions in the spatial rectangular coordinate system in the constructed tunnel to calculate the normal coordinates of the three positioning points when each high-hanging object is working normally without displacement.

[0098] The specific steps for the displacement threshold of each type of high hanging object when working in the tunnel are:

[0099] S421. Collect the displacement of three positioning points when each high hanging object in the tunnel is displaced and causes abnormality of the high hanging object in the history. The displacement of the farthest positioning point is

[0100] {J_far1, J_far2, J_far3,...,J_far N}, J_far1, J_far2, J_far3,..., J_far N It indicates the displacement of the farthest positioning point when the high hanging object is displaced for the 1st, 2nd, 3rd, ..., Nth time in the collected history, causing the high hanging object to be abnormal. The displacement of the nearest positioning point is

[0101] {J_re1, J_re2, J_re3,...,J_re N}, J_re1, J_re2, J_re3,..., J_re N It represents the displacement of the nearest positioning point when the high hanging object is displaced for the 1st, 2nd, 3rd, ..., Nth time in the collected history, causing the high hanging object to become abnormal. The displacement of the median point is

[0102] {J_mid1, J_mid2, J_mid3,...,J_mid N},

[0103] J_mid1, J_mid2, J_mid3, ..., J_mid N It indicates the median displacement of the 1st, 2nd, 3rd, ..., Nth time in the collected history when the high hanging object is displaced and causes the high hanging object to be abnormal, and N is a positive integer;

[0104] S422. Calculate the displacement threshold of the high hanging object according to the displacement of the three positioning points when each high hanging object is displaced and causes the high hanging object to be abnormal in the collected history. The formula is:

[0105]

[0106] In the formula, Th_far represents the displacement threshold of the maximum positioning point in the calculated high hanging object. The displacement thresholds of the minimum positioning point and the midpoint are calculated using the same formula to obtain the minimum positioning point displacement threshold Th_re and the midpoint displacement threshold Th_mid.

[0107] S500, using laser radar to locate three points of each high hanging object working in the tunnel based on the displacement distance of the normal working point, and using the displacement threshold of each high hanging object to judge the real-time displacement and issue an abnormal warning;

[0108] The specific steps of using the displacement threshold of each high hanging object to judge the real-time displacement and issue an abnormal warning are as follows:

[0109] S501, using laser radar to measure in real time the distance between the three positioning points on each high hanging object in the tunnel and the normal coordinates calculated in S412: {F(far, re, mid)1, F(far, re, mid)2, F(far, re, mid)3, ..., F(far, re, mid) B}, F(far,re,mid)1,F(far,re,mid)2,F(far,re,mid)3,...,F(far,re,mid) B It indicates the distance between the three positioning points on the 1st, 2nd, 3rd, ..., Bth high hanging objects and the normal coordinates measured in real time by the laser radar;

[0110] S502. Use the displacement threshold of the three positioning points of the high hanging objects to judge the real-time measured distance. When F(far, re, mid) ≥ Th_(far, re, mid) and at least one of the three positioning points satisfies the inequality, it is judged that the corresponding high hanging object has an abnormal risk and an abnormal warning is issued; when F(far, re, mid) < Th_(far, re, mid) and all three positioning points are less than the real-time distance, it is judged that there is no abnormal risk for the high hanging object. Th_(far, re, mid) represents the displacement threshold of the three positioning points of the high hanging object; the displacements of B high hanging objects in the tunnel are judged to obtain a set of abnormal high hanging objects that have abnormal risks and issue warnings.

[0111] S600: After determining the displacement of the high hanging objects in the tunnel and issuing an early warning, collect the displacement numbers of the high hanging objects in history when they fall or fail to work normally, calculate the falling threshold and the failure threshold, and use the two thresholds to classify abnormal early warnings.

[0112] The specific steps for calculating the drop threshold and failure threshold are:

[0113] S601, collect the median displacement of the high hanging objects in the tunnel when they fall in the history {L_un1, L_un2, L_un3, ..., L_un v}, L_un1, L_un2, L_un3,..., L_un v It represents the median displacement of the high hanging object when it falls for the 1st, 2nd, 3rd, ..., vth time in the collected history;

[0114] S602: Calculate the drop displacement threshold of the high hanging object, the formula is:

[0115]

[0116] In the formula, fall represents the displacement threshold of the high hanging object, the formula represents the calculation of the average value of the displacements less than the middle value in v historical displacements as the displacement threshold, u represents the number of displacements less than the middle value in v historical displacements obtained through judgment, L_un max Indicates the maximum historical displacement, L_un min Indicates the minimum value of historical displacement;

[0117] S602. Perform a secondary judgment on the distances between the three positioning points of the high-hanging objects measured in real time in S500. When F(mid)≥fall, it is judged that the corresponding high-hanging object has a risk of falling, and the warning is judged to be a falling warning; when F(mid)<fall, it is judged that the warning is only a fault threshold; maintenance personnel formulate different maintenance plans for high-hanging objects according to different warning types.

[0118] Classifying abnormal warnings of high-hanging objects and determining whether the warnings are falling or fault warnings can enable maintenance personnel to formulate different maintenance plans when performing abnormal maintenance on high-hanging objects, making it easier for maintenance personnel to complete the maintenance of high-hanging objects.

[0119] Embodiment 1: Now, historical equipment data of high-hanging objects in a tunnel is collected, and the high-hanging objects in the tunnel are divided into identification category, structure category, and product category;

[0120] The visual sensor is used to measure the plane image of a camera in the tunnel. The maximum distances in the length, width and height of the camera are measured to be 20, 8 and 13 respectively. The maximum circumscribed rectangle of the camera is drawn, and the minimum displacement of the camera in history is extracted as 2. The minimum distance when laying points on the maximum circumscribed rectangle of the camera is 2. After calculation, the total number of points in the three directions is 40 in the length direction, 16 in the width direction and 24 in the height direction. The points of the maximum circumscribed rectangle are used to form a three-dimensional point set.

[0121] Collect the displacement of the camera when it fails and when it works normally in history, and analyze it to get the three positioning points of the camera: Figure 2 As shown in the figure, the displacement of the camera in normal operation in history is collected, and the coordinate system of the three positioning points is calculated, which are the farthest positioning point (20,10,13), the nearest positioning point (0,18,0), and the median point (10,10,13).

[0122] Collect the displacement of the three positioning points when the camera is abnormal in history, and calculate the displacement threshold. The formula is:

[0123]

[0124] After calculation, the displacement threshold of the farthest positioning point is 25, the displacement threshold of the nearest positioning point is 5, and the displacement threshold of the median point is 10;

[0125] The laser radar is used to measure the distance between the three positioning points and the normal coordinates in real time. The farthest positioning point is displaced by 19, the nearest positioning point is displaced by 8, and the median point is displaced by 12. The real-time camera is judged to have abnormal risks and an abnormal warning is issued;

[0126] Collect the displacement distance of the camera when it falls in history and calculate the drop threshold. The formula is:

[0127]

[0128] The camera drop threshold is calculated to be 15, and the camera warning is determined to be a fault warning. The maintenance personnel formulate different maintenance plans according to different warnings.

[0129] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0130] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for detecting displacement of high hanging objects in a tunnel based on visual recognition technology, characterized in that: The method comprises the following steps: S100, collecting historical equipment data of high-hanging objects in tunnels, extracting equipment features of each high-hanging object, and classifying the high-hanging objects according to the equipment features; S200, using a visual sensor to collect plane images of high-hanging objects in the tunnel, using the collected multiple plane images to construct a three-dimensional model of each high-hanging object, analyzing the three-dimensional model of each high-hanging object, laying out points on the three-dimensional model of the high-hanging object, and obtaining a three-dimensional point set of each high-hanging object; S300, collecting the displacement of each point in the three-dimensional point set under different working conditions of the high-hanging objects in the tunnel after the high-hanging objects have been displaced in the history, analyzing the displacement of each point and searching for three positioning points for determining the displacement of the high-hanging objects; S400, collecting the two equipment positions of each high hanging object in the tunnel when an abnormality occurs and when it is working normally in history, and respectively calculating the positions of three positioning points of each high hanging object when it is working normally and the displacement threshold of each high hanging object when it is working in the tunnel; S500, using laser radar to locate three points of each high hanging object working in the tunnel based on the displacement distance of the normal working point, and using the displacement threshold of each high hanging object to judge the real-time displacement and issue an abnormal warning; S600: After determining the displacement of the high hanging objects in the tunnel and issuing an early warning, collect the displacement numbers of the high hanging objects in history when they fall or fail to work normally, calculate the falling threshold and the failure threshold, and use the two thresholds to classify abnormal early warnings.

2. The method for detecting displacement of high-hanging objects in a tunnel based on visual recognition technology according to claim 1 is characterized in that: The specific steps of classifying high hanging objects according to equipment characteristics in S100 are: S101, collect the equipment data of each high hanging object in the tunnel in history as {S1, S2, S3, ..., S n }, S1, S2, S3, ..., S n Indicates the 1st, 2nd, 3rd, ..., nth type of equipment data collected for high-hanging objects in the tunnel, where n is a positive integer; The data values ​​of each device before and after the high hanging objects in the tunnel are collected are {S_z1, S_z2, S_z3, ..., S_z n }, S_z1, S_z2, S_z3,..., S_z n Indicates the 1st, 2nd, 3rd, ..., nth device data values ​​before the high hanging object is displaced in the collected history; {S_z'1, S_z'2, S_z'3, ..., S_z' n }, S_z'1, S_z'2, S_z'3,..., S_z' n Indicates the 1st, 2nd, 3rd, ..., nth device data values ​​after the high hanging objects are displaced in the collected history; S102, collecting the equipment data before and after the high hanging objects are displaced m times, and calculating the average difference of each equipment data before and after the high hanging objects in the tunnel are displaced in history, the formula is: In the formula, S_c represents the difference between each device data before and after the displacement of the high hanging objects in the tunnel in the calculated history, S_z represents the device data value collected in the history before the displacement of the high hanging objects, and S_z' represents the device data value collected in the history after the displacement of the high hanging objects; the average device data differences obtained by calculating the collected n types of high hanging object data are {S_c1, S_c2, S_c3, ..., S_c n }, S_c1, S_c2, S_c3,..., S_c n Indicates the calculated data difference of the 1st, 2nd, 3rd, ..., nth type of equipment for high hanging objects; S103, judging the difference of n types of equipment data after the displacement of the high hanging object, extracting the maximum value of the equipment data difference as the equipment feature of the high hanging object, calculating the equipment of all the high hanging objects in the tunnel, and obtaining the equipment feature of each high hanging object as {T1, T2, T3, ..., T B }, T1, T2, T3, ..., T B Indicates the equipment characteristics of the 1st, 2nd, 3rd, ..., Bth high hanging objects in the tunnel, where B is a positive integer; The high hanging objects with the same equipment characteristics are selected as the same type of high hanging objects. Finally, after judgment, the types of high hanging objects in the tunnel are {H1, H2, H3, ..., H g }, H1, H2, H3, ..., H g Represents the 1st, 2nd, 3rd, ...gth high hanging objects after classification, where g is a positive integer.

3. The method for detecting displacement of high-hanging objects in a tunnel based on visual recognition technology according to claim 1 is characterized in that: The specific steps of obtaining the three-dimensional point set of each high hanging object in S200 are: S201, using visual sensors to photograph and measure the plane images of all high hanging objects in the tunnel, extracting data from the plane images, and selecting the maximum distance of the high hanging objects in the three-dimensional direction in the plane image as (X max , Y max , Z max ), X max Indicates the maximum length of the high hanging object in the plane image, Y max Indicates the maximum height of the high hanging object in the plane image, Z max Indicates the maximum width of the high hanging object in the plane image; the maximum circumscribed rectangle of the high hanging object is drawn using the maximum distance of the high hanging object in the three-dimensional direction in the plane image. (X max , Y max , Z max ) are the length, height and width of the largest circumscribed rectangle respectively; S202, collect all the records of displacement of each high hanging object in the tunnel in history, extract the displacement amount of each high hanging object in the records when it is displaced {L1, L2, L3, ..., L k }, L1, L2, L3, ..., L k Indicates the displacement of the high hanging objects collected in history for the 1st, 2nd, 3rd, ..., kth time, where k is a positive integer; compares the displacement of the high hanging objects collected for k times, and selects the minimum displacement as the minimum interval L_min when laying points for the maximum circumscribed rectangle of the high hanging objects; calculates the high hanging objects in B tunnels to obtain the maximum circumscribed rectangle and minimum spacing of each high hanging object; S203, using the calculated minimum spacing to lay points on the maximum circumscribed rectangle of each high hanging object, and calculating the number of laid points in different three-dimensional directions, the formula is: In the formula, Q X , Q Y , Q Z It is expressed as the total number of points in the length, height and width directions when laying points on the maximum circumscribed rectangle of the high hanging object. max , Y max , Z max represents the length, height and width of the maximum circumscribed rectangle of the high hanging object, L_min represents the minimum spacing when laying points for each high hanging object; for each high hanging object in the tunnel, the corresponding maximum circumscribed rectangle is laid using points to form a maximum circumscribed rectangle composed of points; all points constituting the maximum circumscribed rectangle are classified into a unified set as a three-dimensional point set, and finally the three-dimensional point set of all high hanging objects in the tunnel is {G1, G2, G3, ..., G B }, G1, G2, G3, ..., G B Represents the three-dimensional point set of the 1st, 2nd, 3rd, ..., Bth high hanging objects in the tunnel.

4. The method for detecting displacement of high-hanging objects in a tunnel based on visual recognition technology according to claim 3 is characterized in that: The specific steps of analyzing the displacement of each point in S300 to find the three positioning points for determining the displacement of the high hanging object are: S301, collect P records of displacement of high hanging objects in the tunnel causing abnormality of the high hanging objects in history, extract the displacement of each point in the maximum circumscribed rectangular three-dimensional point set of the high hanging objects in the records as {L_a1, L_a2, L_a3, ..., L_a QX + QY + QZ }, L_a1, L_a2, L_a3,..., L_a QX+QY+QZ Indicates the 1st, 2nd, 3rd, ..., Qth points in the maximum circumscribed rectangle of the high hanging object in the extracted record X +Q Y +Q Z The displacement of each point, Q X +Q Y +Q Z The displacement of each point is compared, and the maximum value is selected as the feature point of the collected high-hanging object displacement that causes the high-hanging object to generate abnormal records. The feature points in P records are extracted, and the number of times the same feature point appears in P records is calculated. The feature point with the largest number of occurrences is used as the farthest positioning point A_far among the three high-hanging object positioning points; S302, collect the R records in history where the high hanging objects in the tunnel are displaced but the high hanging objects do not produce abnormalities, and similarly extract the displacement of each point in the maximum circumscribed rectangular three-dimensional point set of the high hanging objects in the records: Indicates the 1st, 2nd, 3rd, ..., Qth points in the maximum circumscribed rectangle of the high hanging object in the extracted record X +Q Y +Q Z The displacement of each point, Q X +Q Y +Q Z The displacement of each point is compared, and the minimum value is selected as the feature point of the collected high-hanging object displacement but no abnormal record of the high-hanging object is generated. The feature points in the R records are extracted, and the number of times the same feature point appears in the R records is calculated. The feature point with the largest number of occurrences is used as the nearest positioning point A_re among the three high-hanging object positioning points; S303, calculate the median point of the three positioning points of the high hanging object according to the farthest positioning point and the nearest positioning point determined, use the laser radar to measure the distances between the farthest positioning point, the nearest positioning point and the tunnel wall in the maximum circumscribed rectangle of the high hanging object, respectively L_far and L_re, calculate the median value of the distance between the farthest positioning point, the nearest positioning point and the tunnel wall, and determine the median point of the three positioning points, the formula is: In the formula, M_A represents the distance between each point in the maximum circumscribed rectangle of the high hanging object and the median value of the distance between the farthest positioning point, the nearest positioning point and the tunnel wall, and L_A represents the distance between each point in the maximum circumscribed rectangle of the high hanging object and the tunnel wall. It represents the intermediate value of the distance between the farthest positioning point, the nearest positioning point and the tunnel wall; each point in the three-dimensional point set of the maximum circumscribed rectangle of the high hanging object is calculated to obtain Q X +Q Y +Q Z -The distance between the two points and the middle value, select the point corresponding to the minimum value as the median point A_mid among the three positioning points.

5. The method for detecting displacement of high-hanging objects in a tunnel based on visual recognition technology according to claim 1 is characterized in that: The specific steps of calculating the positions of the three points of each high hanging object during normal operation in S400 are: S411, establish a spatial rectangular coordinate system in the tunnel, take the laser radar as the origin, the direction perpendicular to the horizontal plane and the tunnel wall as the X-axis, the direction parallel to the horizontal plane and the tunnel wall as the Z-axis, and the vertical direction perpendicular to the ground as the Y-axis to establish the spatial coordinate system; S412. Collect historical images of each high-hanging object in the tunnel when it is working normally without displacement, map the collected historical images using the three-dimensional point set of the maximum circumscribed rectangle of each high-hanging object, determine the positions of the three positioning points in the historical images when the high-hanging object is working normally without displacement, and use trigonometric functions in the spatial rectangular coordinate system in the constructed tunnel to calculate the normal coordinates of the three positioning points when each high-hanging object is working normally without displacement.

6. The method for detecting displacement of high-hanging objects in a tunnel based on visual recognition technology according to claim 5 is characterized in that: In S400, the specific steps of the displacement threshold of each high hanging object when working in the tunnel are: S421. Collect the displacement of three positioning points when each high hanging object in the tunnel is displaced and causes abnormality of the high hanging object in the history. The displacement of the farthest positioning point is {J_far1, J_far2, J_far3, ..., J_far N }, J_far1, J_far2, J_far3,..., J_far N It indicates the displacement of the farthest positioning point when the high hanging object is displaced for the 1st, 2nd, 3rd, ..., Nth time in the collected history, resulting in abnormal high hanging objects. The displacement of the nearest positioning point is {J_re1, J_re2, J_re3, ..., J_re N }, J_re1, J_re2, J_re3,..., J_re N It indicates the displacement of the nearest positioning point when the high hanging object is displaced for the 1st, 2nd, 3rd, ..., Nth time in the collected history, resulting in abnormal high hanging objects. The displacement of the median point is {J_mid1, J_mid2, J_mid3, ..., J_mid N }, J_mid1, J_mid2, J_mid3, ..., J_mid N It indicates the median displacement of the 1st, 2nd, 3rd, ..., Nth time in the collected history when the high hanging object is displaced and causes the high hanging object to be abnormal, and N is a positive integer; S422. Calculate the displacement threshold of the high hanging object according to the displacement of the three positioning points when each high hanging object is displaced and causes the high hanging object to be abnormal in the collected history. The formula is: In the formula, Th_far represents the displacement threshold of the maximum positioning point in the calculated high hanging object. The displacement thresholds of the minimum positioning point and the midpoint are calculated using the same formula to obtain the minimum positioning point displacement threshold Th_re and the midpoint displacement threshold Th_mid.

7. The method for detecting displacement of high-hanging objects in a tunnel based on visual recognition technology according to claim 6 is characterized in that: The specific steps of using the displacement threshold of each high hanging object to judge the real-time displacement and issue an abnormal warning in S500 are: S501, using laser radar to measure in real time the distance between the three positioning points on each high hanging object in the tunnel and the normal coordinates calculated in S412: {F(far, re, mid)1, F(far, re, mid)2, F(far, re, mid)3, ..., F(far, re, mid) B }, F(far,re,mid)1,F(far,re,mid)2,F(far,re,mid)3,...,F(far,re,mid) B It indicates the distance between the three positioning points on the 1st, 2nd, 3rd, ..., Bth high hanging objects and the normal coordinates measured in real time by the laser radar; S502. Use the displacement threshold of the three positioning points of the high hanging objects to judge the real-time measured distance. When F(far, re, mid) ≥ Th_(far, re, mid) and at least one of the three positioning points satisfies the inequality, it is judged that the corresponding high hanging object has an abnormal risk and an abnormal warning is issued; when F(far, re, mid) < Th_(far, re, mid) and all three positioning points are less than the real-time distance, it is judged that there is no abnormal risk for the high hanging object. Th_(far, re, mid) represents the displacement threshold of the three positioning points of the high hanging object; the displacements of B high hanging objects in the tunnel are judged to obtain a set of abnormal high hanging objects that have abnormal risks and issue warnings.

8. The method for detecting displacement of high-hanging objects in a tunnel based on visual recognition technology according to claim 1 is characterized in that: The specific steps of calculating the drop threshold and the fault threshold in S600 are: S601, collect the median displacement of the high hanging objects in the tunnel when they fall in the history {L_un1, L_un2, L_un3, ..., L_un v }, L_un1, L_un2, L_un3,..., L_un v It represents the median displacement of the high hanging object when it falls for the 1st, 2nd, 3rd, ..., vth time in the collected history; S602: Calculate the drop displacement threshold of the high hanging object, the formula is: In the formula, fall represents the displacement threshold of the high hanging object, the formula represents the calculation of the average value of the displacements less than the middle value in v historical displacements as the displacement threshold, u represents the number of displacements less than the middle value in v historical displacements obtained through judgment, L_un max Indicates the maximum historical displacement, L_un min Indicates the minimum value of historical displacement; S602, performing a secondary judgment on the distances between the three positioning points of the high hanging object measured in real time in S500, and when F(mid)≥fall, judging that the corresponding high hanging object has a risk of falling, judging that the warning is a falling warning; When F(mid)<fall, the warning is judged to be only the fault threshold; Maintenance personnel formulate different maintenance plans to maintain high-hanging objects according to different warning types.

Images