Rail seam detection method and system for rail conveyor and medium

Abstract

The invention belongs to the technical field of conveyor rail seam detection, and discloses a rail type conveyor rail seam detection method and system and a medium, and the method comprises the steps: collecting rail type conveyor rail laser three-dimensional contour data, and extracting a rail laser three-dimensional contour data seam feature region; determining feature points a and b and a feature angle beta, and constructing a feature triangle; and judging whether the track seam is a seam or not by utilizing a multi-point linkage verification mechanism, if so, substituting the feature triangle to calculate the track seam detection width bc, and calculating the actual width of the track seam according to the track seam angle. Through depth information analysis, the problem of false detection caused by breakpoints can be effectively solved, and the detection precision of the rail gap width is remarkably improved; a rail seam feature triangle is combined with an angular point algorithm, so that the robustness of the method in a complex environment is improved, and meanwhile, the method is suitable for rail type conveyors of different materials and structures; and through parameter adjustment, application scenes of various types of rail conveyors can be adapted, and the application range of the technology is expanded.

Description

Technical Field

[0001] The present invention belongs to the technical field of conveyor rail gap detection, and particularly relates to a method, a system and a medium for detecting rail gaps of a rail conveyor. Background Art

[0002] For traditional rail gap detection of rail conveyors, manual detection is carried out using tools such as feeler gauges. The rail gap detection method based on laser scanning is to use a laser scanning device to collect three-dimensional contour data of the track surface, and identify the specific position and width of the rail gap by analyzing the contour break points on both sides of the rail gap. This method depends on the clarity of the contour at the edge of the rail gap, and usually calculates the width of the rail gap by detecting the positions of the contour break points on both sides of the rail gap. In some implementation solutions, two-dimensional laser projection and contour comparison may be combined to improve the detection accuracy.

[0003] For traditional manual detection of rail gaps of rail conveyors, the manual detection efficiency is low, and affected by the on-site environment, it is difficult for manual detection at some rail joint positions. In the present invention, a laser three-dimensional contour scanner is carried by an inspection robot, and the rail joint information at positions where manual detection is difficult can be collected.

[0004] Although the rail gap detection method based on laser scanning has certain accuracy and efficiency, the width of the rail gap is calculated by detecting the distance between the contour break points of the rail gap. Due to the influence of the environment, materials, etc. during the laser scanning process, non-joint break points will appear on the track surface contour, and the inability to distinguish whether the break point is a joint will lead to false detection. And affected by the sensor accuracy, the edge of the rail gap contour is not a vertical edge, which will also cause a large detection error.

[0005] Disadvantage 1: Strong dependence on the clarity of the rail gap edge contour, and vulnerable to the state of the track surface.

[0006] The laser scanning method relies on the contour break points on both sides of the rail gap to locate the rail gap area and width. However, in the actual environment, the track surface may be affected by dust, oil stains, rust or wear, resulting in unclear contour break points, which will affect the accuracy of rail gap detection.

[0007] The present invention can use other parameters (such as laser intensity change or regional curvature) to replace the traditional contour break point detection when the track surface state is poor by combining multi-mode data fusion (such as depth information and reflection intensity information) and an intelligent verification algorithm, significantly improving the robustness.

[0008] Disadvantage 2: Limited ability to describe the shape of the rail gap, and difficult to handle complex rail gaps.

[0009] The current method based on laser scanning usually uses a simplified geometric model (such as a straight line or a single break point) to characterize the shape of the rail gap, but for rail gaps with variable width, bending or irregular shapes, this method has insufficient accuracy and may lead to measurement errors.

[0010] The present invention adopts a more refined polygon fitting or curvature analysis technology to perform high-precision modeling on the rail gap area, which can accurately characterize the complex geometric shape of the rail gap and solve the problem that traditional methods cannot handle complex rail gaps.

[0011] Through the above analysis, the problems and defects of the existing technology are as follows: In the rail gap detection of the on-rail conveyor, the traditional laser-based detection method relies on the contour information of the laser scanning the track surface, and calculates the rail gap width by identifying the distance between the contour breakpoints on both sides of the rail gap. However, this method has the following problems:

[0012] (1) False detection of contour breakpoints: The laser scanning process is affected by factors such as ambient light interference, reflectivity of the track surface material, and damage to the track surface, resulting in breakpoints in the contour data, thus causing problems of false rail gaps or missed detection of real rail gaps.

[0013] (2) Poor data integrity: When some rail gap contour information is missing during the scanning process, it is difficult for the traditional method to accurately restore the rail gap width, resulting in a decrease in detection accuracy.

[0014] (3) Insufficient adaptability: The algorithm of the traditional method has poor adaptability to rail gaps with small widths or complex track structures, and cannot meet the application requirements of different types of on-rail conveyor tracks. Summary of the Invention

[0015] To overcome the problems existing in the related technology, the disclosed embodiments of the present invention provide a method, system and medium for detecting rail gaps of an on-rail conveyor, and the technical solutions are as follows:

[0016] The present invention is implemented as follows. The method for detecting rail gaps of an on-rail conveyor includes the following steps:

[0017] S1. Collect the three-dimensional laser contour data of the on-rail conveyor track;

[0018] S2. Extract the seam feature area of the three-dimensional laser contour data of the track;

[0019] S3. Determine the feature points a, b and the feature angle β;

[0020] S4. Construct a triangular feature of the track seam;

[0021] S5. Use a multi-point linkage verification mechanism to determine whether it is a seam. If so, substitute it into the feature triangle to calculate the detected width bc of the track seam, and calculate the actual width of the rail gap according to the track seam angle.

[0022] In step S1, collect the laser three-dimensional profile data of the rail conveyor track, including: preprocess the laser-scanned three-dimensional profile data using a preprocessing algorithm, which includes data denoising and compensation strategies, perform outlier processing using the Z-score method, and then intercept the rail gap feature region according to the characteristics of the collected data;

[0023] (1) Outlier processing using the Z-score method: Use the Z-score to map the data to a distribution with a mean defined as 0 and a standard deviation defined as 1, eliminating the influence of data position and scale. The Z-score calculation formula is:

[0024]

[0025] In the formula, Z i is the Z-score of the i-th data point, x i is the value of the i-th data point on one dimension of the original data, μ is the mean of the original data under this dimension, and σ is the standard deviation of the original data under this dimension;

[0026] (2) Interception of the rail gap feature region: According to the characteristics of the collected data, intercept the points on the tread of the rail head for rail gap detection.

[0027] In step S2, extract the joint feature region of the rail laser three-dimensional profile data, including: through in-depth analysis of the three-dimensional profile data of the rail obtained by laser scanning, extract the depth information inside the rail gap. The depth information is determined by the coordinates of the deepest point a inside the rail gap in the three-dimensional profile of the rail joint. Through the corner detection algorithm, determine the coordinates of the rail gap feature points a and b; obtain the rail gap depth information through the projection of the feature edge ab in the direction perpendicular to the rail tread, and obtain the rail gap detection width information through the projection of the feature edge ab in the direction perpendicular to the rail tread.

[0028] In step S4, construct the rail joint feature triangle, including: based on the rail gap feature triangle, introduce the slope difference algorithm for corner detection, and extract the feature point and feature edge information of the rail gap.

[0029] Furthermore, by constructing the rail gap feature triangle, represent the rail gap feature region in a geometric shape, define the spatial distribution characteristics of the rail gap based on the side lengths and corner positions of the triangle, and calculate the rail gap width according to the characteristics; through the corner detection algorithm, determine the coordinates of the rail gap feature points a and b.

[0030] Furthermore, the slope difference algorithm for corner detection includes:

[0031] Sequentially select the points on the rail joint feature segment in the rail profile data as the center point P = (x 0 , y 0 ). The center point P is the selected point to be detected, and the neighborhood points of the center point P are P i = (x i,y i ), with the neighborhood radius being r and the index of the neighborhood point being i;

[0032] For each neighborhood point P i , calculate the slope k of this point to the center point i , and the expression is:

[0033]

[0034] Sort all the slopes k of the neighborhood points i in ascending order, denoted as k (i) , calculate the difference between adjacent slopes, and the expression is:

[0035] Δk i = k (i+1) - k (i)

[0036] In the formula, Δk i is the difference between adjacent slopes, and k (i+1) is the previous slope;

[0037] Calculate the sum of all slope differences as the corner point score of this point, and the expression is:

[0038]

[0039] In the formula, S is the corner point score of the center point, and n is the number of neighborhood points.

[0040] In step S5, use the multi-point linkage verification mechanism to determine whether it is a seam, including:

[0041] Through the verification algorithm of multi-point linkage, verify the consistency of the rail gap feature point information;

[0042] Let the coordinates of corner points a and b be (x a , y a ), (x b , y b ), select multiple points between points a and b, and let the coordinates be (x j , y j ), calculate the difference Δk between the slopes of this point and a, b j , and the expression is:

[0043]

[0044] If Δk j is close to 0, then verify that a and b are rail gap feature points, otherwise it is a false detection caused by data noise.

[0045] Furthermore, calculating the rail joint based on the characteristic triangle includes:

[0046] According to the installation angle of the laser scanner, given the inclination angle θ between the laser scanner and the horizontal plane, let the angle between side ba and side bc be β, and calculate β according to the equal alternate interior angles rule of parallel lines;

[0047]

[0048] The length of side ab is:

[0049]

[0050] The length of side bc of the characteristic triangle is:

[0051] bc = abcosβ

[0052] Let the track joint angle be The actual track joint width l is:

[0053]

[0054] Another object of the present invention is to provide a rail conveyor track joint detection system, which is used to regulate the rail conveyor track joint detection method, and the system includes:

[0055] A track profile data acquisition module, which is used to acquire the laser three-dimensional profile data of the rail conveyor track;

[0056] A joint feature area extraction module, which is used to extract the joint feature area of the laser three-dimensional profile data of the track;

[0057] A track joint feature triangle construction module, which is used to determine the feature points a, b and the feature angle β;

[0058] A track joint width calculation module, which is used to construct the track joint feature triangle;

[0059] A track joint detection module, which is used to use the multi-point linkage verification mechanism to determine whether it is a joint. If so, substitute it into the feature triangle to calculate the detected width bc of the track joint, and calculate the actual width of the track joint according to the track joint angle.

[0060] Another object of the present invention is to provide a computer-readable storage medium, storing instructions, which when the instructions run on a computer, enable the computer to apply the rail conveyor track joint detection system.

[0061] The present invention proposes a method for detecting rail gaps based on the depth information of feature triangles (DIFT). By innovatively introducing depth information analysis, construction of rail gap feature triangles, and calculation of corner point detection algorithms, higher-precision rail gap detection is achieved. The technical solution of the present invention provides a more efficient detection method for rail inspection and safe operation and maintenance of rail conveyors, strongly promoting the intelligent development of rail conveyor inspection. It is specifically reflected in the following aspects:

[0062] (1) Improve detection accuracy: Depth information analysis can effectively solve the problem of false detection caused by breakpoints, significantly improving the detection accuracy of rail gap width.

[0063] (2) Enhance robustness and adaptability: The combination of rail gap feature triangles and corner point algorithms improves the robustness of the method in complex environments (such as uneven light and rail surface damage), and at the same time adapts to rail conveyors of different materials and structures.

[0064] (3) Reduce the false detection rate and missed detection rate: Multi-point linkage verification and depth information extraction further reduce the probability of false detection and missed detection in the detection, ensuring the accuracy and reliability of the detection results.

[0065] (4) Improve versatility: By adjusting parameters, it can adapt to the application scenarios of various types of rail conveyors, expanding the scope of use of the technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure;

[0067] Figure 1 is a flowchart of the method for detecting rail gaps of a rail conveyor provided by an embodiment of the present invention;

[0068] Figure 2 is a diagram of the rail joint feature triangle provided by an embodiment of the present invention;

[0069] Figure 3 is a diagram of the rail gap feature area provided by an embodiment of the present invention;

[0070] Figure 4 is a diagram of the position of the rail gap feature points provided by an embodiment of the present invention;

[0071] Figure 5 is a diagram of the position of the rail gap feature points provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0072] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following will provide a detailed description of the specific embodiments of the present invention with reference to the accompanying drawings. Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the spirit of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0073] The core technical points of the present invention:

[0074] 1. A method for constructing a rail gap feature region based on laser scanning depth information; the present invention analyzes the three-dimensional contour data of laser scanning, extracts the depth information within the rail gap to determine the rail gap feature region, and solves the problem of misdetection in the traditional method of detecting the rail gap width based on breakpoint detection when contour breakpoints occur due to environmental or material factors. This method is the core innovation of the entire detection technology and directly determines the accuracy of rail gap width calculation.

[0075] 2. Rail gap feature triangle construction technology; based on the rail gap feature region extracted from the depth information, the present invention further constructs a rail gap feature triangle, and based on the sides and corner points of the triangle, accurately determines the feature points and feature edges of the rail gap, thereby avoiding width calculation errors caused by breakpoints. This feature greatly improves the robustness and reliability of rail gap detection in complex environments.

[0076] 3. A high-precision method for calculating the rail gap width based on the corner point algorithm; the present invention uses the corner point algorithm to accurately extract the vertex and edge information of the rail gap feature triangle, and calculates the width of the rail joint based on the positional relationship of the rail gap feature points. Compared with traditional methods, this method can significantly reduce the misdetection rate and improve the accuracy of width calculation at the same time.

[0077] 4. Preprocessing technology for laser scanning three-dimensional contour data; aiming at the problem that the data may be incomplete due to environmental interference (such as light, dust, etc.) during laser scanning, the present invention designs a dedicated preprocessing algorithm, including data denoising and compensation strategies, to improve the stability of subsequent rail gap feature region detection.

[0078] The preprocessing algorithm includes: first, using the Z-score method for outlier processing; then intercepting the rail gap feature region according to the characteristics of the collected data.

[0079] (1) Outlier processing using the Z-score method

[0080] The Z-score (i.e., the standard score) describes a data point based on the relationship between the data point and the mean and standard deviation of a set of points. The Z-score is used to map the data to a distribution with a mean defined as 0 and a standard deviation defined as 1. The effects of data location and scale are eliminated to allow direct comparison of different data sets. In the Z-score outlier detection method, once the data is centralized and rescaled, any value that differs too much from zero (the threshold is often the Z-score 3 or -3) should be regarded as an outlier. The Z-score calculation formula:

[0081]

[0082] In the formula, Z i is the Z-score of the i-th data point, x i is the value of the i-th data point on one dimension of the original data (i.e., the coordinate value on the x or y dimension), μ is the mean of the original data under this dimension, and σ is the standard deviation of the original data under this dimension;

[0083] Set the Z-score threshold to 3, and screen out the points whose absolute value of the Z-score exceeds the threshold.

[0084] (2) Interception of the rail gap feature area

[0085] According to the characteristics of the collected data, the resolution of the single contour data collected by the laser three-dimensional profiler is about 950 points, but the rail gap feature area is only related to the first about 200 points. To ensure the detection accuracy while improving the detection speed, the first 200 points of the rail head tread are intercepted for rail gap detection. The rail gap feature area is as Figure 3 shown, and the positions of the rail gap feature points are as Figure 4 shown.

[0086] Multi-point linkage feature point information verification mechanism;

[0087] To further improve the accuracy of rail gap feature point extraction, the present invention introduces a multi-point linkage verification mechanism in the triangle construction process, and enhances the applicability of the system in the actual scenario by cross-verifying the information consistency between the feature points and the feature edges.

[0088] Algorithm scalability for adapting to multiple types of rail conveyors;

[0089] The method and system of the present invention can adapt to the rail gap characteristics of different rail conveyors, support scenarios with diverse rail widths and materials, and expand its versatility by flexibly adjusting the depth information threshold and feature extraction parameters.

[0090] Example 1, as Figure 1 shown, the rail gap detection method for rail conveyors provided by the embodiments of the present invention includes:

[0091] Depth information detection of the rail gap feature region: The present invention analyzes the depth of the three-dimensional rail profile data obtained by laser scanning to extract the depth information inside the rail gap for identifying the feature region of the rail gap. Compared with the traditional method relying on contour breakpoints, this depth-information-based method can effectively avoid the failure of rail gap identification caused by misdetection or missed detection of breakpoints.

[0092] Construction of the rail gap feature triangle: Based on the extraction of depth information, the present invention constructs a rail gap feature triangle to represent the rail gap feature region in a geometric shape and defines the spatial distribution characteristics of the rail gap based on the side lengths and corner positions of the triangle. Compared with traditional methods, this geometric processing method significantly improves the stability and robustness of rail gap detection, especially in accurately restoring the rail gap characteristics in the breakpoint region.

[0093] The rail joint feature triangle is as Figure 2 shown. Side ac is the depth information side of the rail joint contour, point a is the deepest point of the rail joint that the laser emitted by the laser scanner can reach, and side bc is the tread of the rail head of the rail contour.

[0094] Calculation of the rail gap width based on the corner point algorithm: Based on the rail gap feature triangle, the present invention introduces the slope difference algorithm for corner detection to accurately extract the feature point and feature edge information of the rail gap and calculates the rail gap width according to these features. Compared with traditional methods, this method can avoid misjudging the false breakpoints on both sides of the rail gap and improve the accuracy of width calculation.

[0095] The basic steps of the slope difference algorithm for corner detection are as follows: Select the points on the rail joint feature segment in the rail profile data in sequence as the center point P=(x 0 ,y 0 ), its neighborhood point is P i =(x i ,y i ), the neighborhood radius is r, and i is the index of the neighborhood point.

[0096] For each neighborhood point P i , calculate the slope k i of this point to the center point, and the formula is:

[0097]

[0098] Sort all the slopes k i of the neighborhood points from small to large, and denote them as k (i) .

[0099] Calculate the difference between adjacent slopes, and the formula is:

[0100] Δk i =k (i+1) -k (i)

[0101] Calculate the sum of all slope differences as the corner point score for this point. The formula is:

[0102]

[0103] In the formula, n is the number of neighborhood points. The higher the S score, the greater the possibility of a corner point. Determine corner points a and b based on the scores S of all points in the track joint feature segment of the track profile data.

[0104] Select several points with the highest S scores as possible corner points according to the scores S of all points in the track joint feature segment of the track profile data.

[0105] Since the points between corner points a and b are sparser, the distance differences between the possible corner points and their two nearest neighborhood points on the left and right can be calculated. The two points with the largest absolute value of the distance difference are corner point a or b, where the point with a negative distance difference is point a and the point with a positive distance difference is point b.

[0106] Record the coordinates of the possible corner point as Q = (x 0 , y 0 ), and the neighborhood points are Q -1 = (x -1 , y -1 ) and Q 1 = (x 1 , y 1 ).

[0107]

[0108] Multi-point linkage verification mechanism: During the detection process of the present invention, through a multi-point linkage verification algorithm, the consistency of the rail gap feature point information is verified to ensure the reliability of the detection results. This multi-point linkage mechanism can further reduce false detections caused by data noise. The specific steps are as follows:

[0109] Let the coordinates of corner points a and b be (x a , y a ), (x b , y b ) respectively. Select multiple points between points a and b, and let their coordinates be (x j , y j ). Calculate the difference between the slope of this point and the slopes of a and b. The formula is:

[0110]

[0111] If Δk j is close to 0, then verify that a and b are rail gap feature points; otherwise, it may be a false detection caused by data noise.

[0112] The specific steps for calculating the track joint according to the characteristic triangle are as follows:

[0113] According to the installation angle of the laser scanner, given the inclination angle θ between the laser scanner and the horizontal plane, let the angle between side ba and side bc be β. Calculate β according to the equal alternate interior angles of the parallel line rule. The formula is:

[0114]

[0115] The length of side ab is:

[0116]

[0117] Then the length of side bc of the characteristic triangle is:

[0118] bc = abcosβ

[0119] Let the track joint angle be Then the actual track joint width l is:

[0120]

[0121] The present invention uses the slope difference algorithm of corner detection to detect the depth information of the three-dimensional contour of the track joint; the multi-point linkage verification mechanism is used to determine whether it is a false detection caused by data noise; the formula for calculating the track joint according to the characteristic triangle is used to calculate the actual track joint width according to the characteristic information detected by the previous algorithm.

[0122] Example 2, alternative solutions for depth information detection

[0123] Alternative solution 1: Detection of the joint feature area based on the change of laser intensity

[0124] In addition to depth information, the intensity echo data of laser scanning can also be used to detect the joint feature area. By analyzing the difference in laser reflection intensity on both sides of the joint and combining the position of the intensity break point, the position and feature area of the joint are inferred.

[0125] Advantages: Avoid the dependence on the laser angle for depth information acquisition and have better adaptability under certain specific track materials (such as low-reflectivity surfaces).

[0126] Alternative solution 2: Joint detection method based on image processing

[0127] Laser scanning can be combined with a high-resolution industrial camera to collect image data of the track surface, and the joint feature area is located by image feature extraction (such as texture change, brightness contrast).

[0128] Advantages: Improve the adaptability under changing ambient light, and at the same time can be fused with depth information to form multi-modal detection.

[0129] Example 3, Alternative Solutions for Constructing the Rail Gap Feature Triangle

[0130] Alternative Solution 1: Rail Gap Feature Modeling Method Based on Polygon Fitting

[0131] In addition to triangle modeling, a more complex polygon fitting method (such as rectangle or trapezoid) can also be used to geometrically model the rail gap feature area, and the boundary shape of the rail gap area is characterized by more feature points.

[0132] Advantages: Enhance the ability to describe complex rail gap shapes, especially the adaptability to irregular rail gap structures (such as widening or narrowing rail gaps).

[0133] Example 4, Alternative Solutions for the Corner Point Algorithm:

[0134] Alternative Solution 1: Calculation of Rail Gap Width Based on Edge Detection Algorithm

[0135] Use an edge detection algorithm (such as Canny edge detection) to extract the rail gap edge information, and calculate the rail gap width in combination with the spatial position of the edge.

[0136] Advantages: Suitable for scenarios where it is difficult to extract rail gap feature points or the corner point calculation accuracy is insufficient, and provides a more direct width calculation method.

[0137] Alternative Solution 2: Identification of Rail Gap Feature Points Based on Curvature Analysis

[0138] By calculating the curvature distribution of the inner contour in the rail gap area, determine the curvature extreme points of the rail gap edge as feature points, and further calculate the rail gap width.

[0139] Advantages: Have higher detection accuracy in complex curve tracks, and at the same time enhance the ability to describe the rail gap edge shape.

[0140] Example 5, Alternative Solutions for Multi-Point Linkage Verification:

[0141] Alternative Solution 1: Single-Point Feature Verification Method Based on Weighted Average

[0142] For each detected feature point, calculate the weighted average feature value of its adjacent points to correct the detection error of a single point.

[0143] Advantages: Simplify the calculation complexity of multi-point linkage verification, and at the same time improve the efficiency in real-time detection.

[0144] Alternative Solution 2: Feature Point Consistency Verification Method Based on Machine Learning

[0145] By training a machine learning model (such as support vector machine or neural network), use the spatial position and surrounding information of the rail gap feature points to judge the consistency of the feature points.

[0146] Advantages: It can adapt to more complex detection environments, especially in cases where the characteristics of rail gaps vary greatly.

[0147] Example 6, alternative to the complete technical solution:

[0148] Alternative 1: A fully vision-based rail gap detection system

[0149] Adopt a high-resolution industrial camera and an optical sensor to detect the rail gap feature area and width through machine vision algorithms (such as morphological analysis, deep learning object detection models).

[0150] Advantages: It can completely replace laser detection in an environment with good optical conditions, reduce equipment costs, and improve the detection speed at the same time.

[0151] Alternative 2: A rail gap detection method based on magnetic induction

[0152] Use a magnetic induction sensor to detect the magnetic field change at the rail joint, and calculate the rail gap width by analyzing the characteristics of the magnetic induction signal.

[0153] Advantages: It can adapt to lightless conditions or strongly polluted environments (such as dust, oil stains), expanding the application scenarios of the detection system.

[0154] Determine the coordinates of the rail gap feature points a and b through a corner detection algorithm.

[0155] Obtain the rail gap depth information feature edge through the projection of the feature edge ab in the direction perpendicular to the rail tread; obtain the rail gap detection width information feature edge through the projection of the feature edge ab in the direction perpendicular to the rail tread.

[0156] Denote the right-angle point of the projection as c. Then, the depth information feature edge ac, the width information feature edge bc, and the feature edge ab together form the rail gap feature triangle.

[0157] The width information feature edge does not reflect the real rail gap width. It is necessary to calculate the real rail gap width according to the angle of the rail joint itself. Therefore, it is used as the width information feature edge.

[0158] Feature triangle, as Figure 2 shown. The feature triangle on the physical object, as Figure 5 shown. The width information feature edge is bc, the real rail gap width is bd, and cd is the projection of the side bc on the rail cross-section. Calculate the real rail gap width bd according to the known angle of the rail joint.

[0159] The above is only a relatively preferred specific implementation manner of the present invention. However, the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention shall be covered by the protection scope of the present invention.

Claims

1. A method for detecting rail gaps of a rail conveyor, characterized in that: The method comprises the following steps: S1, collecting the laser 3D profile data of the rail conveyor track; S2, extracting the seam feature area of ​​the track laser 3D profile data; S3, determining feature points a, b and feature angle β; S4, constructing track joint feature triangles; S5. Use the multi-point linkage verification mechanism to determine whether it is a joint. If so, substitute the characteristic triangle to calculate the track joint detection width bc, and calculate the actual width of the track joint based on the track joint angle.

2. The rail conveyor rail gap detection method according to claim 1, characterized in that: In step S1, the laser three-dimensional profile data of the rail conveyor track is collected, including: preprocessing the laser scanning three-dimensional profile data using a preprocessing algorithm, the preprocessing includes data denoising and compensation strategy, outlier processing using the Z score method, and then intercepting the rail joint feature area according to the characteristics of the collected data; (1) Z-score method for outlier processing: The Z-score is used to map the data to a distribution with a mean value defined as 0 and a standard deviation defined as 1, eliminating the influence of data location and scale. The Z-score calculation formula is: In the formula, Z i is the Z score of the ith data point, x i is the value of the i-th data point at a latitude of the original data, μ is the average value of the original data at this latitude, and σ is the standard deviation of the original data at this latitude; (2) Track gap feature area capture: Based on the characteristics of the collected data, points on the rail head tread are captured for track gap detection.

3. The rail conveyor rail gap detection method according to claim 1, characterized in that: In step S2, the joint feature area of ​​the track laser three-dimensional contour data is extracted, including: extracting the depth information inside the track joint by performing depth analysis on the track three-dimensional contour data obtained by laser scanning, the depth information is determined by the coordinates of the deepest point a inside the track joint in the three-dimensional contour of the track joint, and the coordinates of the track joint feature points a and b are determined by a corner point detection algorithm; the track joint depth information is obtained by projecting the feature edge ab in a direction perpendicular to the track tread, and the track joint detection width information is obtained by projecting the feature edge ab in a direction perpendicular to the track tread.

4. The rail conveyor rail gap detection method according to claim 1, characterized in that: In step S4, a track joint feature triangle is constructed, including: based on the track joint feature triangle, a slope difference algorithm for corner point detection is introduced to extract feature points and feature edge information of the track joint.

5. The rail conveyor rail gap detection method according to claim 4, characterized in that: The track joint feature triangle is constructed and the track joint feature area is represented by a geometric shape. The spatial distribution characteristics of the track joint are defined based on the side length and corner point position of the triangle. The track joint width is calculated based on the characteristics. The coordinates of the track joint feature points a and b are determined through the corner point detection algorithm.

6. The rail conveyor rail gap detection method according to claim 5, characterized in that: The slope difference algorithm for corner detection includes: The points of the track joint feature segment in the track profile data are selected in sequence as the center point P = (x0, y0), the center point P is the selected point to be detected, and the neighborhood point of the center point P is P i =(x i ,y i ), the neighborhood radius is r, and the index of the neighborhood point is i; For each neighborhood point P i , calculate the slope k from this point to the center point i , the expression is: The slope k of all neighboring points i Sort from small to large, denoted as k (i) , calculate the difference between adjacent slopes, the expression is: Δk i =k (i+1) -k (i) In the formula, Δk i is the difference between adjacent slopes, k (i+1) is the previous slope; Calculate the sum of all slope differences as the corner score of the point, the expression is: Where S is the corner point score of the center point, and n is the number of neighborhood points.

7. The rail conveyor rail gap detection method according to claim 1, characterized in that: In step S5, a multi-point linkage verification mechanism is used to determine whether it is a seam, including: Through the multi-point linkage verification algorithm, the consistency of the rail joint feature point information is verified; Assume that the coordinates of corner points a and b are (x a ,y a ),(x b ,y b ), select multiple points between points a and b, and set the coordinates to (x j ,y j ), calculate the difference Δk between this point and the slopes of a and b j , the expression is: If Δk j If it is close to 0, it is verified that a and b are track joint feature points, otherwise it is a false detection caused by data noise.

8. The rail conveyor rail gap detection method according to claim 7, characterized in that: Calculation of track joints based on characteristic triangles includes: According to the installation angle of the laser scanner, the inclination angle θ between the laser scanner and the horizontal plane is known. Let the angle between the ba side and the bc side be β. According to the parallel line law, the interior alternate angles are equal and β is calculated. The length of side ab is: The length of the characteristic triangle side bc is: bc=abcosβ Assume the track joint angle is The actual track joint width l is:

9. A rail conveyor rail gap detection system, characterized in that: The system is used to control the rail conveyor rail gap detection method according to any one of claims 1 to 8, and the system comprises: Track profile data acquisition module, used to collect the laser 3D profile data of the track of the rail conveyor; A seam feature region extraction module is used to extract the seam feature region of the track laser 3D profile data; The rail joint feature triangle construction module is used to determine the feature points a, b and the feature angle β; Track joint width calculation module, used to construct track joint characteristic triangle; The rail joint detection module is used to use the multi-point linkage verification mechanism to determine whether it is a joint. If so, the characteristic triangle is substituted into the track joint detection width bc, and the actual width of the track joint is calculated according to the track joint angle.

10. A computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to apply the rail conveyor rail gap detection system according to claim 9.

Images