Laser radar tunnel point cloud axis height automatic determination method

Abstract

The invention discloses a laser radar tunnel point cloud axis height automatic determination method. According to the laser radar tunnel point cloud axis height automatic determination method, laser radar equipment is used for scanning a tunnel, and point cloud data are collected; pre-processing the point cloud data, wherein the pre-processing comprises the steps of filtering abnormal values and noise, and scaling and rotating; performing contour fitting on the preprocessed point cloud data by using a fitting algorithm; based on the fitting result, the highest point representing the tunnel axis in the point cloud data is determined, and the highest point is a point corresponding to the maximum Y coordinate value on the fitting curve or the highest point determined through an algorithm; and the determined highest points are fitted into the continuous tunnel axis, so that the height of the tunnel axis can be quickly and accurately determined, and the method has good universality and adaptability.

Description

Technical Field

[0001] The present invention relates to the technical field of laser radar point cloud data processing, and in particular to a method for automatically determining the height of a laser radar tunnel point cloud axis. Background Art

[0002] Most existing tunnel point cloud data processing methods rely on manual operations, which are inefficient and their accuracy is limited by the operator's experience. Automated point cloud data processing methods improve efficiency and accuracy, but when processing tunnel point cloud data, existing technologies often have difficulty in accurately and quickly determining the height of the tunnel axis. When processing tunnel point cloud data, existing technologies usually use the following methods to determine the height of the tunnel axis:

[0003] Based on the method of geometric feature extraction, the position and height of the tunnel axis are inferred by analyzing the geometric shape features of the point cloud data, such as finding the symmetrical points of the tunnel wall or fitting the plane. For example, the tunnel wall is first fitted with a plane, and then the height of the tunnel axis is estimated based on the position relationship of the fitted plane. The plane equation of the tunnel wall is obtained by plane fitting the tunnel wall point cloud using the least squares method, and then the approximate direction and position of the tunnel axis are determined by the intersection of multiple wall planes.

[0004] Based on the cluster analysis method, the point cloud data is clustered, points with similar characteristics are classified into one category, and then the point set representing the tunnel axis is identified from the clustering results. For example, clustering is performed based on the spatial distribution and density of the point cloud, and points that may belong to the tunnel axis are divided into a specific cluster. Using the KMeans clustering algorithm, the point cloud data is divided into several clusters, and then by analyzing the characteristics of each cluster, an attempt is made to find the cluster related to the tunnel axis. However, the above method has certain limitations due to the following reasons, which makes it difficult to accurately and quickly determine the height of the tunnel axis.

[0005] The influence of data noise and outliers. Tunnel point cloud data is usually affected by factors such as measurement errors and environmental interference, and there are a lot of noise and outliers. These noise and outliers will interfere with the process of geometric feature extraction and cluster analysis, making it difficult to determine the height of the tunnel axis. For example, in the plane fitting process, noise points may cause inaccurate fitting planes, thereby affecting the estimation of the axis height. The noise in the point cloud data may cause deviations in the clustering results, incorrectly classifying points that do not belong to the tunnel axis into axis-related clusters, or classifying axis points into the wrong clusters, affecting the determination of the axis height.

[0006] Irregular and complex tunnel shapes. The shape of an actual tunnel is often not a completely regular cylinder, and there may be bends, deformations, branches, etc. This irregularity and complexity increases the difficulty of determining the height of the tunnel axis. For example, in a curved tunnel, the traditional method based on symmetry points or plane fitting may no longer be applicable because it is difficult to find accurate symmetry points or suitable planes for fitting. There may be various ancillary facilities inside the tunnel, such as cable trays, ventilation ducts, etc. These facilities will increase the complexity of the point cloud data and interfere with the identification and height determination of the tunnel axis.

[0007] Lack of effective feature extraction methods. Current feature extraction methods often have difficulty accurately extracting features related to the tunnel axis height when processing tunnel point cloud data. For example, some geometric feature extraction methods may be sensitive to local features of the tunnel wall and ignore the overall axis height information. Existing clustering analysis methods may not be able to effectively distinguish tunnel axis points from other points, especially when the point cloud data density is uneven or there are a large number of interference points.

[0008] In summary, when processing tunnel point cloud data, the existing technology is difficult to accurately and quickly determine the height of the tunnel axis due to factors such as data noise, complexity of tunnel shape and lack of effective feature extraction methods. Summary of the invention

[0009] The present invention provides a method for automatically determining the height of a laser radar tunnel point cloud axis, so as to solve the problem that it is often difficult to accurately and quickly determine the height of a tunnel axis when processing tunnel point cloud data in the prior art.

[0010] The present invention provides a method for automatically determining the height of a laser radar tunnel point cloud axis, comprising:

[0011] Step 1: Use laser radar equipment to scan the tunnel and collect point cloud data;

[0012] Step 2: preprocessing the point cloud data, including filtering outliers and noise, scaling and rotating;

[0013] Step 3, using a fitting algorithm to perform contour fitting on the preprocessed point cloud data;

[0014] Step 4: Based on the fitting result, determine the highest point representing the tunnel axis in the point cloud data, where the highest point is the point corresponding to the maximum ordinate value on the fitting curve or the highest point determined by the algorithm;

[0015] Among them, according to the characteristics of the tunnel point cloud data and the fitting requirements, a function is selected to fit the curve; the fitting function of the tunnel top curve is a circular curve; the fitted circular curve is: solve the point where the first-order derivative is zero; let the first-order derivative of the fitting curve be equal to zero, and solve the equation to obtain the extreme point; for the polynomial equation, use a numerical method to solve the root of the equation; substitute the solved extreme point into the second-order derivative for judgment; if the second-order derivative is less than zero, then the point is the maximum point, which is the highest point of the curve; if there are multiple maximum points, compare their function values ​​and determine the maximum point as the highest point; solve the highest point of the multi-section circular curve to form a spatial set of the highest points;

[0016] Step 5, fitting the determined highest point into a continuous tunnel axis;

[0017] Among them, the highest point spatial set in the tunnel point cloud data is determined, and adjacent point cloud data are selected around the highest point;

[0018] For a given set of data points (x i ,y i ), where i = 0, 1, 2, ..., n, the cubic spline interpolation function S(x) satisfies the following conditions:

[0019] In each subinterval [x i ,x i+1 ], S(x) is a cubic polynomial;

[0020] S(x i )=y i , the value of the interpolation function at a known data point is equal to the given data value;

[0021] S(x), S'(x) and S”(x) are continuous over the entire interval;

[0022] Sort the selected highest point and adjacent point cloud data according to the value of the horizontal axis; divide the interval into several sub-intervals with the highest point and adjacent point cloud data as nodes;

[0023] For each subinterval [x i ,x i+1 ], let the cubic spline interpolation function be S i (x) = a i +b i (xx i )+c i (xx i ) 2 +d i (xx i ) 3 ;

[0024] According to the conditions of cubic spline interpolation, the following set of equations is established:

[0025] Interpolation condition: S(x i )=y i and S(x i+1 )=y i+1 ;

[0026] First-order derivative continuity condition: S' i (x i ) = S′ i (x i+1 );

[0027] Second-order derivative continuity condition: S" i (x i )=S” i (x i+1 );

[0028] Select boundary conditions;

[0029] Write the system of equations in matrix form and obtain the coefficient a by solving the linear equations i ,b i ,c i ,d i ;

[0030] Solve systems of linear equations;

[0031] For a given horizontal coordinate value x, determine the subinterval [x i ,x i+1 ];

[0032] Substitute x into the corresponding cubic spline interpolation function Si(x) and calculate the ordinate value y;

[0033] The above process is repeated to obtain corresponding ordinate values ​​for a series of abscissa values, thereby generating a continuous curve of the tunnel axis.

[0034] Furthermore, step 2 also includes:

[0035] Based on the design parameters of the tunnel, the point cloud data is classified to distinguish the point cloud data belonging to the tunnel wall and the tunnel bottom.

[0036] Furthermore, based on the design parameters of the tunnel, the point cloud data is classified to distinguish the point cloud data belonging to the tunnel wall and the tunnel bottom, including:

[0037] Obtain the design dimensions of the tunnel, which include the height, width, and bottom elevation information of the tunnel; obtain the shape characteristics of the tunnel and the designed slope parameters;

[0038] Using the tunnel bottom elevation information, determine the bottom elevation range;

[0039] Traverse the point cloud data, mark the points whose elevation values ​​are within the preset range of the bottom elevation as points that belong to the tunnel bottom determined for the first time, and mark the remaining points as points that belong to the tunnel wall determined for the first time;

[0040] For the points that are first determined to belong to the tunnel wall, the spatial distribution characteristics of the points are calculated, and the spatial distribution characteristics include the distance from the point to the center of the tunnel; for a circular tunnel, the distance from the point to the center of the circle is calculated; for a rectangular tunnel, the distance from the point to each side is calculated; according to the designed width and height of the tunnel, the range of the tunnel wall is determined; if the distance from the point to the center or side of the tunnel is within the designed range of the tunnel wall, the point is determined to be a tunnel wall point;

[0041] For the points that are determined to belong to the bottom of the tunnel for the first time, the flatness and continuity of the points are analyzed; the degree of elevation change of the points in the local area is calculated, and if the elevation change is less than a preset range and the points are continuously distributed in space, the point is determined to be the bottom point of the tunnel determined for the second time; combined with the designed slope of the tunnel, check whether the elevation change of the bottom point of the tunnel determined for the second time meets the slope requirements; if the elevation change of the point is consistent with the designed slope, the point is confirmed to be the bottom point of the tunnel;

[0042] For each point, analyze the characteristics of its neighborhood points; if a point is marked as a tunnel wall point, and most of its surrounding neighborhood points are marked as tunnel bottom points, recheck whether the point is misclassified; if a point is marked as a tunnel bottom point, and most of its surrounding neighborhood points are marked as tunnel wall points, recheck and adjust; perform multiple iterative optimizations on the classification results, and continuously adjust the misclassified points until the classification results are stable.

[0043] Furthermore, step three includes:

[0044] Divide the tunnel into different sections according to the tunnel design drawings, construction records or on-site survey results; divide the tunnel according to the geological conditions, construction methods, curve segments and straight segments; store and mark the point cloud data of each section separately;

[0045] For each section, observe the geometric shape of the tunnel contour presented by the point cloud data; analyze the flatness and curvature change characteristics of the tunnel wall; for sections with straight segments and regular tunnel contours, choose rectangular fitting or circular fitting; rectangular fitting determines the parameters of the rectangle by finding four boundary points, and is suitable for tunnels with rectangular cross-sections; circular fitting uses the least squares algorithm to determine the center and radius of the circle, and is suitable for circular tunnels; for curved sections, circular curves are used to fit the curved sections of the tunnel; the shape of the curve is adjusted by control points so that the fitting result reflects the tunnel contour; polynomial fitting adapts to different degrees of curve changes according to different polynomial orders; for sections with local deformation, unevenness or geological anomalies, multiple algorithms are combined for fitting; the contour is first determined using the fitting algorithm, and then the local deformation area is corrected using the local fitting algorithm;

[0046] When using the fitting algorithm, adjust the algorithm parameters according to the specific characteristics of each paragraph; verify the fitting results of each paragraph, compare the degree of fit between the fitting curve and the actual point cloud data, and calculate the fitting error to evaluate the fitting effect.

[0047] Furthermore, the method further comprises:

[0048] The fitted tunnel axis is verified and the accuracy and reliability of the fitting are evaluated by comparing it with the tunnel design profile.

[0049] Furthermore, the fitted tunnel axis is verified by comparing it with the tunnel design profile to evaluate the accuracy and reliability of the fitting, including:

[0050] Collect the design drawings of the tunnel, including the plan layout and cross-section drawings, to clarify the design outline, dimensions and theoretical position of the axis of the tunnel; record the key parameters of the tunnel design, including tunnel length, width, height and curvature radius;

[0051] Determine key parameters of the fitted tunnel axis, the key parameters of the tunnel axis include the coordinates of the starting point and the end point of the axis, the length of the axis, and the curvature; uniformly select a preset number of points on the fitted axis, the preset number of points covering different sections of the tunnel;

[0052] Project the points selected on the axis to the plane coordinate system of the tunnel design; measure the distance deviation between the points selected on the axis and the design axis on the plane; if the deviation is within the allowable range, it means that the plane position fitting is accurate; if the deviation is not within the allowable range, further analyze the cause; determine the elevation value of the point on the axis and compare it with the elevation of the design contour at the corresponding position; calculate the elevation deviation and evaluate the accuracy of the fitted axis in the elevation direction;

[0053] For curved tunnels, calculate the curvature of the fitted axis and compare it with the designed curvature; measure the difference in curvature by calculating the radius of the curve or the radius of curvature; if the difference is within the allowable range, the fitted curvature is accurate;

[0054] Statistical analysis is performed on the deviations of all comparison points to calculate the average deviation and standard deviation indicators; the degree of closeness between the fitting result and the design profile is judged based on the average deviation and standard deviation;

[0055] Draw the fitted tunnel axis and the designed outline simultaneously in the graphics software for visual comparison; observe the degree of overlap between the fitted axis and the designed outline, as well as the differences in different sections;

[0056] The stability and reliability of the results were evaluated by repeating the measurement and fitting operations.

[0057] The present invention has the following beneficial effects: a method for automatically determining the height of a laser radar tunnel point cloud axis line of the present invention uses a laser radar device to scan the tunnel and collect point cloud data; preprocesses the point cloud data, and the preprocessing includes filtering outliers and noise, scaling and rotating; uses a fitting algorithm to perform contour fitting on the preprocessed point cloud data; based on the fitting result, determines the highest point representing the tunnel axis in the point cloud data, and the highest point is the point corresponding to the maximum Y coordinate value on the fitting curve or the highest point determined by the algorithm; fitting the determined highest point into a continuous tunnel axis can quickly and accurately determine the height of the tunnel axis, and has good versatility and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0059] Figure 1 A flow chart of the method for automatically determining the height of the laser radar tunnel point cloud axis provided by the present invention. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described in conjunction with the specific embodiments of the present invention and the corresponding drawings. 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. The technical solutions provided by the embodiments of the present invention are described in detail below in conjunction with the drawings.

[0061] See also Figure 1 , a method for automatically determining the height of a laser radar tunnel point cloud axis provided by the present invention, comprising:

[0062] Step 1: Use lidar equipment to scan the tunnel and collect point cloud data.

[0063] Step 2: preprocess the point cloud data, including filtering outliers and noise, scaling and rotating.

[0064] Specifically, the fitting algorithm may be a combination of polynomial fitting, B-spline curve fitting or RANSAC algorithm or more.

[0065] Specifically, step two also includes: based on the design parameters of the tunnel, classifying the point cloud data, and distinguishing the point cloud data belonging to the tunnel wall and the tunnel bottom.

[0066] The specific method is as follows: obtain the design dimensions of the tunnel, which include the height, width, and bottom elevation information of the tunnel; obtain the shape characteristics of the tunnel and the designed slope parameters; use the bottom elevation information of the tunnel to determine the bottom elevation range; traverse the point cloud data, mark the points whose elevation values ​​are within the preset range of the bottom elevation as the points that are first determined to belong to the bottom of the tunnel, and mark the remaining points as the points that are first determined to belong to the tunnel wall; for the points that are first determined to belong to the tunnel wall, calculate the spatial distribution characteristics of the points, which include the distance from the point to the center of the tunnel; for a circular tunnel, calculate the distance from the point to the center of the circle; for a rectangular tunnel, calculate the distance from the point to each side; determine the range of the tunnel wall according to the designed width and height of the tunnel; if the distance from the point to the center or side of the tunnel is within the design range of the tunnel wall, determine the point as a tunnel wall point; for the first judgment The points that belong to the bottom of the tunnel are determined, and the flatness and continuity of the points are analyzed; the degree of elevation change of the points in the local area is calculated. If the elevation change is less than the preset range and the points are continuously distributed in space, the point is determined to be the bottom point of the tunnel determined for the second time; combined with the design slope of the tunnel, check whether the elevation change of the bottom point of the tunnel determined for the second time meets the slope requirements; if the elevation change of the point is consistent with the design slope, the point is confirmed to be the bottom point of the tunnel; for each point, analyze the characteristics of its neighboring points; if a point is marked as a tunnel wall point, and most of its surrounding neighboring points are marked as tunnel bottom points, recheck whether the point is misclassified; if a point is marked as a tunnel bottom point, and most of its surrounding neighboring points are marked as tunnel wall points, recheck and adjust; perform multiple iterations on the classification results, and continuously adjust the misclassified points until the classification results are stable. Through the above steps, the point cloud data can be more accurately classified into point cloud data of the tunnel wall and the tunnel bottom based on the tunnel design parameters.

[0067] Step three: Use the fitting algorithm to perform contour fitting on the preprocessed point cloud data.

[0068] Specifically, step three includes: dividing the tunnel into different sections according to the design drawings, construction records or on-site survey results of the tunnel; dividing the tunnel according to the geological conditions of the tunnel, different construction methods, and factors such as curved and straight segments; and storing and marking the point cloud data of each section separately for subsequent separate processing. For each section, observe the geometric shape of the tunnel outline presented by the point cloud data. For example, the straight segment may present a relatively regular rectangular or circular outline, while the curved segment may have certain deformation and irregularity. Analyze the flatness and curvature change characteristics of the tunnel wall. The tunnel wall of some sections may be relatively smooth, while some sections may be uneven or locally deformed. Considering the geological influence, if the tunnel passes through different geological areas, the tunnel outline of different sections may be affected by the geological conditions. For example, in a soft soil geological section, the tunnel may have a large deformation; in a hard rock section, the tunnel outline may be more regular.

[0069] For straight segments with regular tunnel contours, choose rectangular fitting or circular fitting; rectangular fitting determines the parameters of the rectangle by finding four boundary points, which is suitable for tunnels with rectangular cross-sections; circular fitting uses the least squares algorithm to determine the center and radius, which is suitable for circular tunnels; for curved segments, circular curves are used to fit the curved segments of tunnels; spline curve fitting can adapt well to the changes in the shape of the curve, and adjust the shape of the curve by control points so that the fitting results reflect the tunnel contour; polynomial fitting adapts to different degrees of curve changes according to different polynomial orders; for segments with local deformation, unevenness or geological anomalies, multiple algorithms are combined for fitting; first use the fitting algorithm to determine the contour, and then use the local fitting algorithm to correct the local deformation area; when using the fitting algorithm, adjust the algorithm parameters according to the specific characteristics of each segment; for polynomial fitting, select the appropriate polynomial order to avoid overfitting or underfitting. Verify the fitting results of each segment, and compare the degree of fit between the fitting curve and the actual point cloud data, such as the root mean square error and other indicators, to evaluate the fitting effect. Through the above steps, select appropriate fitting algorithms for different sections of the tunnel to better adapt to the changes in the tunnel contour and improve the processing accuracy of the tunnel point cloud data.

[0070] Step 4: Based on the fitting result, determine the highest point representing the tunnel axis in the point cloud data, where the highest point is the point corresponding to the maximum ordinate value on the fitting curve or the highest point determined by the algorithm.

[0071] Specifically, step four includes: determining the extreme point of the curve as the highest point by analyzing the first-order derivative and the second-order derivative of the fitting curve.

[0072] The first-order derivative reflects the rate of change of a function at a certain point. For a fitted curve, the first-order derivative represents the slope of the curve at that point. When the first-order derivative is zero, the slope of the curve at that point is zero, which may be the extreme point or inflection point of the curve. The second-order derivative reflects the change in the curvature of the function. For a fitted curve, the second-order derivative represents the degree of curvature of the curve. When the second-order derivative is greater than zero, the curve is convex at that point, which may be a minimum point; when the second-order derivative is less than zero, the curve is convex at that point, which may be a maximum point; when the second-order derivative is zero, it may be an inflection point.

[0073] The specific method is as follows: According to the characteristics of the tunnel point cloud data and the fitting requirements, select a suitable function to fit the curve. The common tunnel top curve fitting function is the circular curve; for the fitted circular curve; for the fitted circular curve: solve the point where the first-order derivative is zero; let the first-order derivative of the fitting curve equal to zero, solve the equation to get the extreme point; for polynomial equations, use numerical methods such as Newton's method, bisection method, etc. to solve the root of the equation; substitute the solved extreme point into the second-order derivative for judgment; if the second-order derivative is less than zero, then the point is the maximum point, as the highest point of the curve; if there are multiple maximum points, compare their function values ​​and determine the maximum point as the highest point; solve the highest point of the multi-section circular curve to form a spatial set of the highest points.

[0074] Step five, fitting the determined highest point into a continuous tunnel axis.

[0075] Specifically, step five includes: using a cubic spline interpolation method to generate a continuous curve of the tunnel axis according to the highest point space set and adjacent point cloud data. The specific method is as follows:

[0076] 1. Determine the highest point set

[0077] The highest point set in the tunnel point cloud data is determined by the previous method. This highest point set will serve as an important reference point for generating the continuous curve of the tunnel axis.

[0078] 2. Select adjacent point cloud data

[0079] Select a certain number of adjacent point cloud data around the highest point. These points should be able to better reflect the trend of the tunnel axis in this area.

[0080] Determine the selection range and quantity based on actual conditions. Generally speaking, select points within a certain distance from the highest point, or select a certain number of points based on the density of the point cloud.

[0081] Cubic spline interpolation is a method for constructing a smooth curve through a series of known data points. i ,y i), where i = 0, 1, 2, ..., n, the cubic spline interpolation function S(x) satisfies the following conditions:

[0082] In each subinterval [x i ,x i+1 ], S(x) is a cubic polynomial;

[0083] S(x i )=y i , the value of the interpolation function at a known data point is equal to the given data value;

[0084] S(x), S'(x), and S”(x) are continuous over the entire interval.

[0085] Divide the interval: Sort the selected highest point and adjacent point cloud data according to the value of the horizontal coordinate (or other appropriate parameters). Use these points as nodes to divide the interval into several sub-intervals.

[0086] Establish the system of equations: For each subinterval [x i ,x i+1 ], let the cubic spline interpolation function be S i (x) = a i +b i (xx i )+cs(xx i ) 2 +d i (xx i ) 3 .

[0087] According to the conditions of cubic spline interpolation, the following set of equations is established:

[0088] Interpolation condition: S(x i )=y i and S(x i+1 )=y i+1 ;

[0089] First-order derivative continuity condition: S' i (x i ) = S′ i (x i+1 );

[0090] Second-order derivative continuity condition: S" i (x i )=S” i (x i+1 ).

[0091] Boundary conditions: Select appropriate boundary conditions according to actual conditions, such as natural boundary conditions (the second-order derivative of the two end points is zero), fixed boundary conditions, etc.

[0092] Solve the system of equations: Write the system of equations in matrix form and obtain the coefficient a by solving the linear system of equations i ,b i ,c i ,d i . Use numerical methods such as Gaussian elimination and LU decomposition to solve linear equations.

[0093] Generate a continuous curve: For a given horizontal coordinate value x, determine the subinterval [x i ,x i+1 ]. Substitute x into the corresponding cubic spline interpolation function Si(x) and calculate the ordinate value y.

[0094] The above process is repeated to obtain corresponding ordinate values ​​for a series of abscissa values, thereby generating a continuous curve of the tunnel axis.

[0095] For example, suppose there is a maximum point (x 0 ,y 0 ) and the four adjacent points (x 1 ,y 1 ),(x 2 ,y 2 ),(x 3 ,y 3 ),(x 4 ,y 4 ).

[0096] First, divide the interval into [x 0 ,x 1 ],[x 1 ,x 2 ],[x 2 ,x 3 ],[x 3 ,x 4 ]Four sub-intervals.

[0097] Then, a set of equations is established according to the interpolation conditions, derivative continuity conditions and boundary conditions, and the coefficients of the cubic spline interpolation function of each subinterval are solved.

[0098] Finally, for a given range of abscissas, a continuous curve of the tunnel axis is generated.

[0099] Through the above steps, the cubic spline interpolation method is used to generate a continuous curve of the tunnel axis according to the highest point and adjacent point cloud data. The curve has high smoothness and accuracy and can better reflect the actual shape of the tunnel axis.

[0100] In this embodiment, the method also includes: verifying the fitted tunnel axis, and evaluating the accuracy and reliability of the fit by comparing it with the tunnel design profile. The specific method is as follows: collecting the design drawings of the tunnel, which include plan layout drawings and cross-sectional drawings, to clarify the design profile, dimensions and theoretical position of the tunnel axis; recording the key parameters of the tunnel design, which include tunnel length, width, height, and radius of curvature; determining the key parameters of the fitted tunnel axis, which include the starting and ending coordinates of the axis, the length of the axis, and the curvature; uniformly selecting a preset number of points on the fitted axis, and the preset number of points covers different sections of the tunnel; projecting the points selected on the axis into the plane coordinate system of the tunnel design; measuring the distance deviation between the points selected on the axis and the design axis on the plane; if the deviation is within the allowable range, it means that the plane position is fitted accurately; if the deviation is not within the allowable range, further analyze the cause ; Determine the elevation value of the point on the axis and compare it with the elevation of the design contour at the corresponding position; calculate the elevation deviation and evaluate the accuracy of the fitted axis in the elevation direction; for curved tunnels, calculate the curvature of the fitted axis and compare it with the design curvature; measure the difference in curvature by calculating the radius of the curve or the radius of curvature; if the difference is within the allowable range, it means that the fitted curvature is accurate; perform statistical analysis on the deviations of all compared points and calculate the average deviation and standard deviation indicators; judge the degree of closeness of the fitting result to the design contour based on the average deviation and standard deviation; draw the fitted tunnel axis and the design contour in the graphics software at the same time for visual comparison; observe the degree of overlap between the fitted axis and the design contour, as well as the differences in different sections; evaluate the stability and reliability of the results by repeating the measurement and fitting operations.

[0101] For example, suppose there is a circular tunnel with a design radius of R. Select n points on the fitted axis, measure the distance from these points to the design center, and calculate the average deviation and standard deviation. If the average deviation and standard deviation are small, for example, less than 5mm, it means that the accuracy of the fitting is high. At the same time, the deviation between the axis and the design contour can be intuitively seen through visual inspection to further evaluate the reliability of the fitting. Through the above steps, the fitted tunnel axis is fully verified, the accuracy and reliability of the fitting are evaluated, and a reliable basis is provided for the design, construction and monitoring of tunnel engineering.

[0102] The above-described embodiments of the present invention do not limit the protection scope of the present invention.

Claims

1. A method for automatically determining the height of the axis of a laser radar tunnel point cloud, characterized in that: include: Step 1: Use laser radar equipment to scan the tunnel and collect point cloud data; Step 2: preprocessing the point cloud data, including filtering outliers and noise, scaling and rotating; Step 3, using a fitting algorithm to perform contour fitting on the preprocessed point cloud data; Step 4: Based on the fitting result, determine the highest point representing the tunnel axis in the point cloud data, where the highest point is the point corresponding to the maximum ordinate value on the fitting curve or the highest point determined by the algorithm; Among them, according to the characteristics of the tunnel point cloud data and the fitting requirements, a function is selected to fit the curve; the fitting function of the tunnel top curve is a circular curve; the fitted circular curve is: solve the point where the first-order derivative is zero; let the first-order derivative of the fitting curve be equal to zero, and solve the equation to obtain the extreme point; for the polynomial equation, use a numerical method to solve the root of the equation; substitute the solved extreme point into the second-order derivative for judgment; if the second-order derivative is less than zero, then the point is the maximum point, which is the highest point of the curve; if there are multiple maximum points, compare their function values ​​and determine the maximum point as the highest point; solve the highest point of the multi-section circular curve to form a spatial set of the highest points; Step 5, fitting the determined highest point into a continuous tunnel axis; Among them, the highest point spatial set in the tunnel point cloud data is determined, and adjacent point cloud data are selected around the highest point; For a given set of data points (x i ,y i ), where i = 0, 1, 2, ..., n, the cubic spline interpolation function S(x) satisfies the following conditions: In each subinterval [x i ,x i+1 ], S(x) is a cubic polynomial; S(x i )=y i , the value of the interpolation function at a known data point is equal to the given data value; S(x), S'(x) and S”(x) are continuous over the entire interval; Sort the selected highest point and adjacent point cloud data according to the value of the horizontal axis; divide the interval into several sub-intervals with the highest point and adjacent point cloud data as nodes; For each subinterval [x i ,x i+1 ], let the cubic spline interpolation function be S i (x) = a i +b i (xx i )+c i (xx i ) 2 +d i (xx i ) 3 ; According to the conditions of cubic spline interpolation, the following set of equations is established: Interpolation condition: S(x i )=y i and S(x i+1 )=y i+1 ; First-order derivative continuity condition: S′ i (x i ) = S′ i (x i+1 ); Second-order derivative continuity condition: S" i (x i )=S” i (x i+1 ); Select boundary conditions; Write the system of equations in matrix form and obtain the coefficient a by solving the linear equations i ,b i ,c i ,d i ; Solve systems of linear equations; For a given horizontal coordinate value x, determine the subinterval [x i ,x i+1 ]; Substitute x into the corresponding cubic spline interpolation function Si(x) and calculate the ordinate value y; The above process is repeated to obtain corresponding ordinate values ​​for a series of abscissa values, thereby generating a continuous curve of the tunnel axis.

2. The method for automatically determining the height of the laser radar tunnel point cloud axis according to claim 1, characterized in that: Step 2 also includes: Based on the design parameters of the tunnel, the point cloud data is classified to distinguish the point cloud data belonging to the tunnel wall and the tunnel bottom.

3. The method for automatically determining the height of the laser radar tunnel point cloud axis according to claim 2, characterized in that: Based on the design parameters of the tunnel, the point cloud data is classified to distinguish the point cloud data belonging to the tunnel wall and the tunnel bottom, including: Obtain the design dimensions of the tunnel, which include the height, width, and bottom elevation information of the tunnel; obtain the shape characteristics of the tunnel and the designed slope parameters; Using the tunnel bottom elevation information, determine the bottom elevation range; Traverse the point cloud data, mark the points whose elevation values ​​are within the preset range of the bottom elevation as points that belong to the tunnel bottom determined for the first time, and mark the remaining points as points that belong to the tunnel wall determined for the first time; For the points that are first determined to belong to the tunnel wall, the spatial distribution characteristics of the points are calculated, and the spatial distribution characteristics include the distance from the point to the center of the tunnel; for a circular tunnel, the distance from the point to the center of the circle is calculated; for a rectangular tunnel, the distance from the point to each side is calculated; according to the designed width and height of the tunnel, the range of the tunnel wall is determined; if the distance from the point to the center or side of the tunnel is within the designed range of the tunnel wall, the point is determined to be a tunnel wall point; For the points that are determined to belong to the bottom of the tunnel for the first time, the flatness and continuity of the points are analyzed; the degree of elevation change of the points in the local area is calculated, and if the elevation change is less than a preset range and the points are continuously distributed in space, the point is determined to be the bottom point of the tunnel determined for the second time; combined with the designed slope of the tunnel, check whether the elevation change of the bottom point of the tunnel determined for the second time meets the slope requirements; if the elevation change of the point is consistent with the designed slope, the point is confirmed to be the bottom point of the tunnel; For each point, analyze the characteristics of its neighborhood points; if a point is marked as a tunnel wall point, and most of its surrounding neighborhood points are marked as tunnel bottom points, recheck whether the point is misclassified; if a point is marked as a tunnel bottom point, and most of its surrounding neighborhood points are marked as tunnel wall points, recheck and adjust; perform multiple iterative optimizations on the classification results, and continuously adjust the misclassified points until the classification results are stable.

4. The method for automatically determining the height of the laser radar tunnel point cloud axis according to claim 1, characterized in that: Step three includes: Divide the tunnel into different sections according to the tunnel design drawings, construction records or on-site survey results; divide the tunnel according to the geological conditions, construction methods, curve segments and straight segments; store and mark the point cloud data of each section separately; For each section, observe the geometric shape of the tunnel contour presented by the point cloud data; analyze the flatness and curvature change characteristics of the tunnel wall; for sections with straight segments and regular tunnel contours, choose rectangular fitting or circular fitting; rectangular fitting determines the parameters of the rectangle by finding four boundary points, and is suitable for tunnels with rectangular cross-sections; circular fitting uses the least squares algorithm to determine the center and radius of the circle, and is suitable for circular tunnels; for curved sections, circular curves are used to fit the curved sections of the tunnel; the shape of the curve is adjusted by control points so that the fitting result reflects the tunnel contour; polynomial fitting adapts to different degrees of curve changes according to different polynomial orders; for sections with local deformation, unevenness or geological anomalies, multiple algorithms are combined for fitting; the contour is first determined using the fitting algorithm, and then the local deformation area is corrected using the local fitting algorithm; When using the fitting algorithm, adjust the algorithm parameters according to the specific characteristics of each paragraph; verify the fitting results of each paragraph, compare the degree of fit between the fitting curve and the actual point cloud data, and calculate the fitting error to evaluate the fitting effect.

5. The method for automatically determining the height of the laser radar tunnel point cloud axis according to claim 1, characterized in that: The method further comprises: The fitted tunnel axis is verified and the accuracy and reliability of the fitting are evaluated by comparing it with the tunnel design profile.

6. The method for automatically determining the height of the laser radar tunnel point cloud axis according to claim 5, characterized in that: Verify the fitted tunnel axis and evaluate the accuracy and reliability of the fitting by comparing it with the tunnel design profile, including: Collect the design drawings of the tunnel, including the plan layout and cross-section drawings, to clarify the design outline, dimensions and theoretical position of the axis of the tunnel; record the key parameters of the tunnel design, including tunnel length, width, height and curvature radius; Determine key parameters of the fitted tunnel axis, the key parameters of the tunnel axis include the coordinates of the starting point and the end point of the axis, the length of the axis, and the curvature; uniformly select a preset number of points on the fitted axis, the preset number of points covering different sections of the tunnel; Project the points selected on the axis to the plane coordinate system of the tunnel design; measure the distance deviation between the points selected on the axis and the design axis on the plane; if the deviation is within the allowable range, it means that the plane position fitting is accurate; if the deviation is not within the allowable range, further analyze the cause; determine the elevation value of the point on the axis and compare it with the elevation of the design contour at the corresponding position; calculate the elevation deviation and evaluate the accuracy of the fitted axis in the elevation direction; For curved tunnels, calculate the curvature of the fitted axis and compare it with the designed curvature; measure the difference in curvature by calculating the radius of the curve or the radius of curvature; if the difference is within the allowable range, the fitted curvature is accurate; Statistical analysis is performed on the deviations of all comparison points to calculate the average deviation and standard deviation indicators; the degree of closeness between the fitting result and the design profile is judged based on the average deviation and standard deviation; Draw the fitted tunnel axis and the designed outline simultaneously in the graphics software for visual comparison; observe the degree of overlap between the fitted axis and the designed outline, as well as the differences in different sections; The stability and reliability of the results were evaluated by repeating the measurement and fitting operations.

Images