CN118604790B - A SLAM three-dimensional laser scanner calibration system and calibration method - Google Patents

Abstract

The invention discloses a system and a method for calibrating an SLAM three-dimensional laser scanner, which can rapidly calibrate the SLAM three-dimensional laser scanner with high precision and multiple parameters, and reduce the calibration difficulty; the calibration system comprises a calibration field, a feature, a reference point module, a etalon, an SLAM three-dimensional laser scanner to be calibrated, a scanning point cloud resolving and preprocessing module, a scanning point cloud previewing module and a data analysis module; the calibration method includes planning a path for a calibration scan on a calibration site; determining the positions of datum points on two sides of the path; respectively setting patterns, features and datum points with regular shapes on the ground at two sides of the path; respectively obtaining the measurement results of the patterns, the features and the datum points; the SLAM three-dimensional laser scanner to be calibrated acquires scanning point cloud data; scanning point cloud data to obtain resolved point cloud data; and (3) resolving point cloud data processing to obtain measurement results of corresponding parameters of the SLAM three-dimensional laser scanner to be calibrated, analyzing the measurement results and the like.

Description

A SLAM three-dimensional laser scanner calibration system and calibration method

Technical Field

The present invention relates to the field of metrology and calibration technology, and in particular to a SLAM three-dimensional laser scanner calibration system and a calibration method.

Background Art

SLAM (Simultaneous Localization and Mapping) technology refers to simultaneous positioning and mapping technology, which is mainly used to solve the positioning and map construction problems of robots moving in unknown environments, and is mainly used for large-space three-dimensional scanning systems for composition and navigation. SLAM three-dimensional laser scanner is a high-precision measurement device based on SLAM technology, which is widely used in three-dimensional reconstruction and data acquisition in indoor and outdoor environments. SLAM three-dimensional laser scanner uses laser radar to emit laser pulses and measure their reflection time to calculate the distance of objects, thereby obtaining high-precision point cloud data of the surrounding environment. These point cloud data contain information such as the three-dimensional coordinates, reflectivity and texture of the object. Before using the SLAM three-dimensional laser scanner, it needs to be calibrated to ensure that its accuracy indicators meet the requirements.

The calibration of traditional 3D laser scanners usually uses standard balls and/or standard ball sticks to measure distance. The following problems may occur during the calibration process: First, the stability of the installation of the standard balls and/or standard ball sticks will affect the accuracy of the calibration results; second, due to the small size of the standard balls and/or standard ball sticks (the diameter of the standard balls is not more than 50.8mm), when measuring large spaces, there are problems such as substandard range and high test difficulty; third, only a single indicator (such as distance) can be calibrated, and comprehensive calibration of multiple parameters cannot be achieved.

Summary of the invention

At least one of the purposes of the present invention is to provide a SLAM three-dimensional laser scanner calibration system and calibration method to overcome the problems existing in the above-mentioned prior art, which can quickly perform multi-parameter calibration of the SLAM three-dimensional laser scanner in a large space, improve the calibration accuracy of the SLAM three-dimensional laser scanner, and reduce the difficulty of calibration.

In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention includes the following aspects.

A SLAM three-dimensional laser scanner calibration system includes: a calibration site, a feature object, a reference point module, a standard device, a SLAM three-dimensional laser scanner to be calibrated, a scanning point cloud solution and preprocessing module, a scanning point cloud preview module and a data analysis module. Among them:

A path for calibration scanning is planned in the calibration site, and patterns for characterizing the indication error of the graphic size measurement are respectively arranged on both sides of the path;

The feature objects are respectively arranged on both sides of the path, and the feature objects include: a first feature object for characterizing the indication error of plane angle measurement and point cloud thickness, a second feature object for characterizing the indication error of feature object size measurement, a third feature object for characterizing the plumb bob and horizontal direction angle measurement error, and a fourth feature object for characterizing repeated measurement accuracy;

The reference point module is used to determine the reference point positions on both sides of the path, and after setting the reference point at the later reference point position, measure the coordinates of the reference point;

The standard device includes: a first standard device for obtaining the angle between adjacent first feature objects, a second standard device for obtaining the distance between adjacent second feature objects, and a third standard device for measuring the geometric dimensions of the patterns on both sides of the path;

The SLAM three-dimensional laser scanner to be calibrated is used to collect scanning point cloud data and store the collected scanning point cloud data;

The scanning point cloud solution and preprocessing module is used to perform coordinate transformation processing on the scanning point cloud data, convert the scanning point cloud data into the same coordinate system, and obtain solution point cloud data;

The scanning point cloud preview module is used to preview the solved point cloud data processed by the scanning point cloud solving and preprocessing module, and measure the size of the pattern through the previewed solved point cloud data;

The data analysis module is used to analyze the solved point cloud data processed by the scanning point cloud solution and preprocessing module and to construct a geometric model.

Preferably, the first feature is a flat plate, and a plurality of flat plates are circumferentially arranged on both sides of the path or arranged at different positions in a vertical direction of the path, and the angles between adjacent flat plates increase in steps, and the angle between two adjacent flat plates is greater than 0 degrees and not greater than 180 degrees.

The second feature object is a sphere with a diameter of 30 mm to 300 mm and a sphericity of no more than 0.1 mm; multiple spheres are arranged on both sides of the path along the circumference of the path or multiple spheres are arranged at different positions along the vertical direction of the path.

The third characteristic object is a cylinder, and multiple cylinders are arranged on both sides of the path along the path. The upper end of each cylinder is suspended in the air by a tooling, and the lower end of the cylinder is suspended in the air. The hook of the tooling is on the geometric center line of the cylinder, and the geometric center line of the cylinder is vertically downward. The length of each cylinder is not less than 2m and the coaxiality is not greater than 0.1mm.

The calibration site is a rectangular site, and the fourth characteristic object is not limited to a cube. Four fourth characteristic objects are respectively arranged at the four corners of the calibration site.

The length of the calibration site is not less than 40m and the width is not less than 40m; the path is a circular path, and the path overlaps 2m to 5m at the starting point and the end point.

Preferably, it also includes a fifth feature object for characterizing the indication error of the feature object size measurement, the fifth feature object adopts a standard spacing gauge, and a plurality of the fifth features are respectively arranged horizontally, vertically and obliquely on both sides of the path.

A SLAM three-dimensional laser scanner calibration method comprises the following steps:

Step S1: planning a path for calibration scanning on a calibration site;

Step S2: Determine the positions of the reference points on both sides of the path;

Step S3: setting regular shaped patterns, features and reference points on both sides of the path;

Step S4: obtaining measurement results of the pattern, feature object and reference point respectively;

Step S5: Make the SLAM three-dimensional laser scanner to be calibrated scan along the path to collect scanning point cloud data of the path and both sides of the path;

Step S6: enabling the SLAM three-dimensional laser scanner to be calibrated to separately collect scanning point cloud data of the first feature object;

Step S7: in the absolute coordinate system, respectively solve the scanned point cloud data collected in step S5 and step S6 to obtain corresponding solved point cloud data;

Step S8: Processing the point cloud data solved in step S7 according to the type of parameters to be calibrated, and obtaining the measurement results of the corresponding parameters of the SLAM three-dimensional laser scanner to be calibrated;

Step S9: According to the type of parameters to be calibrated, the measurement results of the corresponding parameters of the SLAM 3D laser scanner to be calibrated are analyzed to determine whether the accuracy index of the corresponding parameters of the SLAM 3D laser scanner to be calibrated meets the requirements.

Preferably, the parameter types to be calibrated include dimension measurement error, plane angle measurement indication error, coordinate measurement error, plumb and horizontal angle measurement error, point cloud thickness and repeated measurement accuracy; the dimension measurement error includes feature object dimension measurement indication error and graphic dimension measurement indication error.

Preferably, the step S4 specifically includes:

The geometric dimensions of each pattern on both sides of the path are obtained by the third standard device; when the pattern is circular, the diameters φ ₀₁ ~φ _0i are measured; when the pattern is rectangular, the lengths L ₀₁ ~ L _0j and the widths K ₀₁ ~ K _0j are measured; wherein i represents the number of circular patterns, i is an integer not less than 1; j represents the number of rectangular patterns, j is an integer not less than 1;

Obtaining distances D ₀₁ to D _0k-1 between adjacent second characteristic objects by a second standard device, wherein k represents the number of the second characteristic objects and k is an integer not less than 2;

Obtaining angles α ₀₁ to α _0m-1 between adjacent first characteristic objects by a first standard device, wherein m represents the number of first characteristic objects, and m is an integer not less than 2;

The coordinates of the reference points (H ₀₁ , S ₀₁ , N ₀₁ )~(H _0p , S _0p , N _0p ) are obtained through static measurement and analysis of the GNSS receiver network, where p represents the number of reference points and p is an integer not less than 1;

The distances d ₀₁ to d _0q-1 between adjacent measurement plates are read through the standard scale lines on the body of the fifth feature, wherein q represents the number of measurement plates and q is an integer not less than 2.

Preferably, in step S5, the process of making the SLAM three-dimensional laser scanner to be calibrated scan along the path and collecting scanning point cloud data of the path and both sides of the path includes:

Step S51: Turn on the SLAM three-dimensional laser scanner to be calibrated and initialize it;

Step S52: Make the SLAM three-dimensional laser scanner to be calibrated walk and scan along the path, and collect scanning point cloud data of the path and both sides of the path;

Step S53: After the SLAM three-dimensional laser scanner to be calibrated walks along the path for one circle, the SLAM three-dimensional laser scanner to be calibrated is turned off;

When more paths and scanning point cloud data on both sides of the paths need to be collected, steps S51 to S53 are repeated.

Preferably, in step S7, the process of solving the scanning point cloud data collected in step S5 includes:

Solve the scan point cloud data n collected for the nth time respectively to obtain the solved point cloud data n, and save the solved point cloud data n; wherein n is an integer not less than 1; when n>1, merge the solved point cloud data 1 to the solved point cloud data n to generate and save a set of solved point cloud data 1-n;

After solving the scanned point cloud data n+1 collected in step S6, solved point cloud data n+1 is obtained, and the solved point cloud data n+1 is saved separately.

Preferably, in step S8, the process of processing the point cloud data solved in step S7 includes:

View the data through the browsing software and measure the size of each pattern; for circular patterns, measure the diameter φ ₁₁ ~φ _1i , for rectangular patterns, measure the length L ₁₁ ~ L _1j and the width K ₁₁ ~ K _1j ; obtain the coordinates of the reference point (H ₁₁ , S ₁₁ , N ₁₁ ) ~ (H _1p , S _1p , N _1p ) through the browsing software;

Import the solved point cloud data 1-n into CAD, cut the solved point cloud data 1-n, check the angles β ₁ ~β _a between the surface or axis point cloud of the same third feature object in the solved point cloud data 1-n and the vertical direction, and the angles γ ₁ ~γ _a between the surface or axis point cloud of the third feature object and the horizontal direction; then check the maximum distances S ₁ ~S _b between the points of the same feature of the same fourth feature object; wherein a represents the number of third feature objects, and a is an integer not less than 1; b represents the number of fourth feature objects, and b is an integer not less than 1;

Import the solved point cloud data n+1 into CAD, cut the solved point cloud data n+1, and check the point cloud thickness T of the first feature object;

The solved point cloud data 1 to the solved point cloud data n are respectively imported into Geomagic, 3DR or PolyWorks software, and fitting processing is performed respectively. According to the fitting processing results, the angles α ₁₁ ~α _1m-1 between adjacent first feature objects, the distances D ₁₁ ~ D _1k-1 between adjacent second feature objects, and the distances d ₁₁ ~ d _1q-1 between adjacent measurement plates of the fifth feature object are respectively solved.

Preferably, the step S9 specifically includes: subtracting the measurement results of each parameter of the SLAM three-dimensional laser scanner to be calibrated obtained in step S8 from the measurement results of the corresponding parameters obtained in step S4 to obtain the analysis value of each parameter, and selecting the maximum value from the analysis value of each parameter as the calibration result.

In summary, due to the adoption of the above technical solution, the present invention has at least the following beneficial effects:

By setting up a calibration site, planning a path on the calibration site, and setting patterns, features and reference points on both sides of the path, the SLAM three-dimensional laser scanner to be calibrated can obtain the scanning point cloud data of the path and both sides of the path after scanning along the path. After the scanning point cloud data is processed, the measurement values of various parameters of the SLAM three-dimensional laser scanner to be calibrated can be obtained. Combined with the measurement results of the corresponding parameters measured by the standard instrument, the different parameters of the SLAM three-dimensional laser scanner to be calibrated can be quickly calibrated to improve the calibration accuracy and calibration efficiency. The SLAM three-dimensional laser scanner calibration method of the present invention can calibrate various parameters of the SLAM three-dimensional laser scanner to be calibrated in a large space, reducing the difficulty of calibration. Compared with the calibration method of a single indicator, the method of the present invention can comprehensively evaluate the accuracy index of the SLAM three-dimensional laser scanner to be calibrated, and improve the reliability of the SLAM three-dimensional laser scanner to be calibrated in the measurement process of different occasions.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of a SLAM three-dimensional laser scanner calibration system according to an exemplary embodiment of the present invention.

FIG. 2 is a schematic diagram of the arrangement of patterns, features, and reference points on a calibration site according to an exemplary embodiment of the present invention.

FIG. 3 is a schematic structural diagram of a fifth feature object of an exemplary embodiment of the present invention.

FIG. 4 is a flow chart of a SLAM three-dimensional laser scanner calibration process according to an exemplary embodiment of the present invention.

Symbols in the figure: 1-calibration site, 2-path, 3-pattern, 4-first feature object, 5-second feature object, 6-third feature object, 7-fourth feature object, 8-fifth feature object, 81-body, 82-slider, 83-measuring plate, 84-linear guide, 9-reference point, 10-starting point, 11-end point.

DETAILED DESCRIPTION

The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments to make the purpose, technical solutions and advantages of the present invention more clearly understood. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

Referring to FIG1 , the SLAM three-dimensional laser scanner calibration system of the exemplary embodiment of the present invention includes: a calibration site, a feature object, a reference point module, a standard, a SLAM three-dimensional laser scanner to be calibrated, a scanning point cloud solution and preprocessing module, a scanning point cloud preview module, and a data analysis module. Among them:

A path for calibration scanning is planned in the calibration site, and patterns (circle, rectangle, triangle, pentagon, etc.) are set on both sides of the path. The patterns are used to represent the indication error of the graphic size measurement;

The feature objects are respectively arranged on both sides of the path, and the feature objects include: a first feature object for characterizing the indication error of plane angle measurement and point cloud thickness, a second feature object for characterizing the indication error of feature object size measurement, a third feature object for characterizing the plumb bob and horizontal direction angle measurement error, and a fourth feature object for characterizing the repeat measurement accuracy; the feature objects also include a fifth feature object for characterizing the indication error of feature object size measurement;

A plurality of first features are respectively arranged on both sides of the path, a plurality of second features are respectively arranged on both sides of the path, a plurality of third features are respectively arranged on both sides of the path, a plurality of fourth features are respectively arranged around the calibration site, and a plurality of fifth features are respectively arranged on both sides of the path;

The benchmark module (including several GNSS receivers or several total stations) is used to determine the positions of the benchmarks on both sides of the path, and after setting the benchmarks at the later benchmark positions, measure the coordinates of the benchmarks (elevation H, longitude S, latitude N);

The standard device includes a first standard device, which is used to measure the angle between adjacent first feature objects and/or the point cloud thickness of the first feature object;

A second standard device, used to obtain the distance between adjacent second characteristic objects;

The third standard device is used to measure the geometric dimensions of the patterns on both sides of the path;

The SLAM 3D laser scanner to be calibrated is integrated with a storage module, and the SLAM 3D laser scanner to be calibrated is used to collect scanning point cloud data and store the collected scanning point cloud data;

The scanning point cloud solution and preprocessing module is used to perform coordinate transformation processing on the scanning point cloud data, convert the scanning point cloud data into the same coordinate system, and obtain the solution point cloud data;

The scanning point cloud preview module is used to preview the solved point cloud data processed by the scanning point cloud solution and pre-processing module, and measure the size of the pattern through the previewed solved point cloud data;

The data analysis module is used to analyze the scanned point cloud solution and the solved point cloud data processed by the preprocessing module and to construct a geometric model.

Referring to FIG2 , the first feature 4 is a flat plate, and a plurality of first features 4 (at least two first features) are respectively arranged on both sides of the path 2 along the path 2, and the plurality of first features 4 are arranged circumferentially along the path 2, and the angles between adjacent first features 4 are increased in steps, and the angle between two adjacent first features 4 is greater than 0 degrees and not greater than 180 degrees (the angle is preferably an obtuse angle). The cross-section of the flat plate is not limited to a rectangular, triangular or other shape. When the flat plate is a rectangular flat plate, the length of each flat plate is not less than 0.2m and the width is not less than 0.2m. The length of the rectangular flat plate is preferably 0.5m and the width is preferably 0.5m.

The edges of two adjacent first features 4 may be in contact or spaced apart, so that an angle exists between the two adjacent first features 4 .

The first feature object 4 can be set on the ground of the calibration site 1. When the area of the calibration site 1 is limited, the first feature object 4 can also be set in the air through tooling; when the first feature object 4 is set in the air, multiple first feature objects 4 are set at different positions along the vertical direction of the path.

The second feature object 5 is a sphere, the diameter of the sphere is 30-300 mm, and the sphericity is not greater than 0.1 mm; multiple second feature objects 5 (at least two second feature objects) are respectively arranged on both sides of the path 2 along the path 2, and multiple second feature objects 5 are arranged circumferentially along the path 2, and the heights of adjacent second feature objects 5 from the ground are different (the heights can be increased or decreased in sequence, and the heights can also be set arbitrarily, as long as the heights of adjacent second feature objects from the ground are different). When the area of the calibration site 1 is limited, the second feature object 5 can also be arranged in the air through tooling, and in the vertical direction of the path 2, the multiple second feature objects 5 are arranged at different positions.

The third characteristic object 6 is a cylinder with a diameter of not less than 0.2m, a length of not less than 2m, and a coaxiality of not more than 0.1mm; a plurality of third characteristic objects 6 (at least two third characteristic objects) are respectively arranged on both sides of path 2 along path 2; the upper end of each third characteristic object 6 is suspended in the air through a tooling, and the lower end of the third characteristic object 6 is suspended in the air, and the hook of the tooling is on the geometric center line of the third characteristic object 6, and the geometric center line of the third characteristic object 6 is vertically downward, and after the third characteristic object 6 is suspended in the air, the distance between the lower end of the third characteristic object 6 and the ground is not less than 3m.

The fourth characteristic object 7 is preferably a cube (it can also be any other three-dimensional structure with no surface defects, such as a triangular prism, a quadrangular prism, etc.), the length of the cube is not less than 0.2m, the width is not less than 0.2m, and the height is not less than 0.2m; four fourth characteristic objects 7 are respectively arranged at the four corners of the calibration site 1 (fourth characteristic objects can also be added between adjacent fourth characteristic objects, and the number of added fourth characteristic objects is set according to calibration needs), and the height of each fourth characteristic object 7 from the ground is not less than 0.5m.

The fifth characteristic object 8 adopts a standard spacing gauge. Referring to FIG3 , the fifth characteristic object 8 includes a main body 81 and a sliding assembly. A linear guide rail 84 is provided on the main body 81. A plurality of sliding assemblies are mounted on the linear guide rail 84. The sliding assembly includes a slider 82 and a measuring plate 83. The measuring plate 83 is mounted on the slider 82. The slider 82 is mounted on the linear guide rail 84. The measuring plates 83 are parallel to each other. Standard scale lines are also provided on the main body 81 to facilitate reading the spacing between the measuring plates 83. Coordinate tables are provided on two opposite surfaces of each measuring plate 83. The edge of the measuring plate 83 is provided with a chamfer (not shown) to facilitate reading the position of the measuring plate 83.

2 , multiple (at least 3) fifth features 8 are arranged on both sides of the path 2, and the multiple fifth features 8 are respectively arranged horizontally, vertically and obliquely on the calibration site 1. When the fifth features 8 are arranged vertically or obliquely, the fifth features 8 can be supported by tooling to ensure the stability of the fifth features 8 on the calibration site 1.

During use of the fifth characteristic object 8, the distance between adjacent measuring plates 83 is read through the standard scale lines on the body 81. When reading the distance, coordinate points with the same reading are selected on two adjacent measuring plates 83 for reading.

When a sphere is used as the second characteristic object, the distance between adjacent second characteristic objects is the distance between the centers of two adjacent spheres. When characterizing the indication error of the characteristic object size measurement, the center of the sphere is a virtual point, and there will be errors when the center of the sphere is obtained by subsequent fitting. When measuring the distance between the measuring plates of the fifth characteristic object, the distance between the surfaces of adjacent measuring plates is measured. The surfaces of the measuring plates are solid points. Compared with the distance between the centers of two adjacent spheres, the distance between two adjacent measuring plates has a smaller error, which is conducive to improving the measurement accuracy.

The reference point 9 is a metal nail, an intersection line or a corner point of the pattern 3 (or a feature object), etc., and multiple reference points 9 (at least four reference points) are arranged on both sides of the path 2 along the path 2; the coordinates of the multiple reference points 9 can be measured by a GNSS receiver or a total station.

During the measurement process, the structure of the tooling is not limited to the bracket. According to the specific setting form of each feature object, a tooling with a corresponding structure can be used to ensure the stability of different feature objects on the calibration site.

Referring to FIG. 4 , a SLAM three-dimensional laser scanner calibration method according to an exemplary embodiment of the present invention includes the following steps:

Step S1: planning a path for calibration scanning on a calibration site;

Step S2: Determine the positions of the reference points on both sides of the path through the reference point module;

Step S3: setting regular shaped patterns, features and reference points on both sides of the path;

Step S4: obtaining measurement results of the pattern, feature object and reference point respectively;

Step S5: Make the SLAM three-dimensional laser scanner to be calibrated scan along the path to collect scanning point cloud data of the path and both sides of the path;

Step S6: enabling the SLAM three-dimensional laser scanner to be calibrated to separately collect scanning point cloud data of the first feature object;

Step S7: in the absolute coordinate system, respectively solve the scanned point cloud data collected in step S5 and step S6 to obtain corresponding solved point cloud data;

Step S8: Processing the point cloud data solved in step S7 according to the type of parameters to be calibrated, and obtaining the measurement results of the corresponding parameters of the SLAM three-dimensional laser scanner to be calibrated;

Step S9: According to the type of parameters to be calibrated, the measurement results of the corresponding parameters of the SLAM 3D laser scanner to be calibrated are analyzed to determine whether the accuracy index of the corresponding parameters of the SLAM 3D laser scanner to be calibrated meets the requirements.

The path in step S1 is a circular closed path so that the SLAM three-dimensional laser scanner can scan continuously. The path overlaps a certain distance at the starting point and the end point (for example, referring to Figure 2, the path overlaps 2m~5m at the starting point 10 and the end point 11), and satellite signals can be received on all sides of the path; the calibration site can be a site of any shape. During the calibration process, a flat rectangular site is preferably used as the calibration site. The length of the calibration site is not less than 40m, the width is not less than 40m, and its length×width size is preferably 50m×50m.

In step S2, when determining the position of the reference point, more than three GNSS receivers or more than three total stations are used to determine the position of the reference point.

In step S3, patterns are set on the ground on both sides of the path. The shape of the pattern is not limited to circles, rectangles, triangles, pentagons and the like, as long as the geometric dimensions of the pattern can be conveniently measured. The pattern shape types on both sides of the path can be set arbitrarily. For example, one type of pattern of the same shape can be set on one side of the path, and another type of pattern of the same shape can be set on the other side of the path. Patterns of the same shape can be set on both sides of the path. Patterns of different shapes can also be set on both sides of the path.

The pattern can be set directly on the ground on both sides of the path, or the pattern can be set on a plate or banner and then the plate or banner can be set on both sides of the path, the plate or banner can be set along the circumference of the path, or the plate or banner can be set in the air in the vertical direction of the path.

The features include a first feature, a second feature, a third feature, and a fourth feature.

When setting the first feature, the angle between adjacent plates or the angle between plates in adjacent plate groups increases in steps, and during the process of increasing the angle of the plate, the distance between the plate and the starting point of the path increases accordingly; the plate can be set on the ground or in the air through a bracket; when the plate is set on the ground, the adjacent plates can be separated or the edges of the adjacent plates can be connected; when the adjacent plates are separated, five plates are preferably used, the angle between the first and second plates is 45 degrees, the angle between the second and third plates is 90 degrees, the angle between the third and fourth plates is 135 degrees, and the angle between the fourth and fifth plates is 180 degrees. When the edges of adjacent plates are connected, eight plates are preferably used, each two plates form a plate group, and four plate groups are set along the path, the angle between the two plates in the first plate group is 45 degrees, the angle between the two plates in the second plate group is 90 degrees, the angle between the two plates in the third plate group is 135 degrees, and the angle between the two plates in the fourth plate group is 180 degrees.

All the plates or groups of plates can be arranged on one side of the path (the side close to the center of the path in the radial direction of the path), or they can be arranged on the other side of the path (the side away from the center of the path in the radial direction of the path); some of the plates can also be arranged on one side of the path and some on the other side of the path, as long as the plates are arranged along the path.

When setting the second feature, multiple spheres (at least two spheres) are arranged along the path; multiple spheres can be arranged on one side of the path, or on the other side of the path, or some spheres can be arranged on one side of the path and some on the other side of the path; the heights of adjacent spheres from the ground are different (the heights can be increased or decreased in sequence, and the heights can also be set arbitrarily, as long as the heights of adjacent spheres from the ground are different). In the present invention, five spheres are preferably used, and the five spheres are arranged on both sides of the path, respectively, and the distances between adjacent spheres increase in sequence. For example, the distance between the first sphere and the second sphere is 10m, the distance between the second sphere and the third sphere is 20m, the distance between the third sphere and the fourth sphere is 30m, and the distance between the fourth sphere and the fifth sphere is 50m.

When setting the third characteristic object, multiple cylinders are arranged along the path. Multiple cylinders can be arranged on one side of the path or on the other side of the path. Some cylinders can be arranged on one side of the path and some on the other side of the path. In the present invention, three cylinders are preferably used.

When setting the fourth feature object, four cubes are arranged at the four corners of the calibration site (cubes can also be added between adjacent cubes, and the number of added cubes is set according to calibration needs), and the height of each cube from the ground is not less than 0.5m.

After the fourth feature object is set, the fifth feature object is set horizontally, vertically and obliquely on both sides of the path; in the present invention, it is preferred to set three fifth feature objects, one horizontally, one vertically and one obliquely.

The reference point is set at the reference point position determined in step S2. When setting the reference point, multiple (at least four) reference points are arranged along the path, and the multiple reference points are arranged on both sides of the path respectively.

In step S4, the process of obtaining the measurement result of the pattern specifically includes: obtaining the geometric dimensions of each pattern on both sides of the path respectively through a third standard instrument; the third standard instrument can be a steel tape measure, or a standard instrument for measuring geometric dimensions with a range and accuracy that meets the standards such as a laser tracker. When the pattern is circular, the diameters φ ₀₁ ~φ _0i are measured; when the pattern is rectangular, L ₀₁ ~ L _0j and widths K ₀₁ ~ K _0j are measured; wherein i represents the number of circular patterns, i is an integer not less than 1; j represents the number of rectangular patterns, j is an integer not less than 1.

In step S4, the process of obtaining the measurement result of the characteristic object specifically includes:

The distances D ₀₁ ~ D _0k-1 between adjacent second feature objects are obtained by the second standard, where k represents the number of second feature objects and is an integer not less than 2; the second standard can be a laser tracker, which collects points on the surfaces of adjacent second feature objects respectively and obtains the distances between adjacent second feature objects by fitting; during the measurement of the laser tracker, when the diameter of the second feature object is the same as the diameter of the target ball of the laser tracker, the second feature object and the target ball can share a common mounting base. The second standard can also be a ground scanner (ground three-dimensional laser scanner), which scans adjacent second feature objects by the ground scanner to obtain point clouds of adjacent second feature objects, and the distances between adjacent second feature objects are obtained by analysis using Geomagic software (or other point cloud processing software);

The angles α ₀₁ ~α _0m-1 between adjacent first feature objects are obtained by the first standard, wherein m represents the number of the first feature objects and is an integer not less than 2; the first standard adopts a ground scanner, and the ground scanner is used to scan the adjacent first feature objects to obtain the point cloud of the adjacent first feature objects, and the angles between the adjacent first feature objects are obtained by analysis using Geomagic software (or other point cloud processing software).

In step S4, the process of obtaining the measurement results of the reference points specifically includes: obtaining the coordinates of the reference points (H ₀₁ , S ₀₁ , N ₀₁ ) to (H _0p , S _0p , N _0p ) through static measurement and analysis of the GNSS receiver network, where p represents the number of reference points and p is an integer not less than 1.

Step S4 further includes: reading the distances d ₀₁ - d _0q-1 between adjacent measurement plates through the standard scale lines on the body of the fifth feature object, wherein q represents the number of measurement plates and q is an integer not less than 2.

In step S5, the process of making the SLAM three-dimensional laser scanner to be calibrated scan along the path and collecting scanning point cloud data of the path and both sides of the path includes:

Step S51: Turn on the SLAM three-dimensional laser scanner to be calibrated and initialize it;

Step S52: Make the SLAM 3D laser scanner to be calibrated walk and scan along the path, and collect scanning point cloud data of the path and both sides of the path; the walking of the SLAM 3D laser scanner along the path can be carried by the measurement personnel, or can be carried out by a robot dog, an AGV trolley, or a guide rail + drive mechanism, as long as the SLAM 3D laser scanner to be calibrated can walk and scan along the path; the SLAM 3D laser scanner to be calibrated can walk at a uniform speed or at a non-uniform speed, and the walking speed can be changed according to the actual measurement requirements; near the feature object, the SLAM 3D laser scanner to be calibrated can be close to the feature object, and can also slow down the speed to increase the density and quantity of scanning point cloud data collection, so as to avoid the scanning dead zone caused by the relative limitation between the feature object and the SLAM 3D laser scanner to be calibrated, which affects the scanning result;

Step S53: After the SLAM three-dimensional laser scanner to be calibrated walks one circle along the path, the SLAM three-dimensional laser scanner to be calibrated is turned off.

When more scanning point cloud data of the path and both sides of the path need to be collected, repeat steps S51 to S53; the scanning point cloud data collected for the nth time is recorded as scanning point cloud data n (n is an integer not less than 1), and after n scans, the scanning point cloud data obtained are scanning point cloud data 1, scanning point cloud data 2, ..., scanning point cloud data n; the scanning point cloud data collected by the SLAM three-dimensional laser scanner to be calibrated each time can be stored through its own storage module.

In the present invention, it is preferred that the SLAM three-dimensional laser scanner to be calibrated walks three circles along the path to collect three scanning point cloud data, namely scanning point cloud data 1, scanning point cloud data 2 and scanning point cloud data 3.

In step S6, the process of making the SLAM three-dimensional laser scanner to be calibrated separately collect the scanning point cloud data of the first feature object includes: before collecting, turning on the SLAM three-dimensional laser scanner to be calibrated and initializing it, making the SLAM three-dimensional laser scanner to be calibrated face the first feature object, and the distance between the SLAM three-dimensional laser scanner to be calibrated and the first feature object is about 1m, and then making the SLAM three-dimensional laser scanner to be calibrated walk and scan from near to far, and when walking to a certain distance (such as 20~30m) from the first feature object, stop walking and scanning, save the collected data through the storage module, and record the collected data as scanning point cloud data n+1. For example, when three scanning point cloud data are collected along the path in step S5, the scanning point cloud data in step S6 is recorded as scanning point cloud data 4; when collecting the scanning point cloud data of the first feature object, the scanning point cloud data of all the first feature objects can be collected separately, or the scanning point cloud data of one of the first feature objects can be collected.

In step S7, in the process of solving the scanning point cloud data collected in step S5, when the scanning point cloud data is collected once in step S5, the collected scanning point cloud data is solved to obtain solved point cloud data 1; when the scanning point cloud data is collected n times (n>1) in step S5, the scanning point cloud data collected each time is solved respectively to obtain corresponding solved point cloud data 1 to solved point cloud data n, and the solved point cloud data 1 to solved point cloud data n are saved respectively; then the solved point cloud data 1 to solved point cloud data n are merged to obtain a group of solved point cloud data 1-n, and the solved point cloud data 1-n are saved; when the scanning point cloud data is collected n times (n>1) in step S5, after solving the scanning point cloud data n+1 collected in step S6, the solved point cloud data n+1 is obtained, and the solved point cloud data n+1 is saved separately.

For example, when three scanning point cloud data collections are performed in step S5, the scanning point cloud data 1, scanning point cloud data 2 and scanning point cloud data 3 collected in step S5 are solved respectively to obtain solved point cloud data 1, solved point cloud data 2 and solved point cloud data 3, and the three solved point cloud data are saved through the storage module respectively; then the solved point cloud data 1, solved point cloud data 2 and solved point cloud data 3 are merged to generate a group of solved point cloud data 1-3, and the solved point cloud data 1-3 are saved through the storage module; for the scanning point cloud data 4 obtained in step S6, after solving, solved point cloud data 4 is obtained, and the solved point cloud data 4 is saved separately in the storage module.

In step S8, the process of processing the point cloud data solved in step S7 includes:

View the data through the browsing software and measure the size of the pattern (for example, view the solved point cloud data n through the CAD software, and measure the size of the pattern through CAD); for a circular pattern, measure the diameter φ ₁₁ ~φ _1i , and for a rectangular pattern, measure the length L ₁₁ ~ L _1j and the width K ₁₁ ~ K _1j ; the browsing software can also obtain the coordinates of the reference point (H ₁₁ , S ₁₁ , N ₁₁ ) ~ (H _1p , S _1p , N _1p );

Import the solved point cloud data 1-3 generated in step S7 into CAD, crop the solved point cloud data 1-3, check the angles β ₁ ~β _a between the surface point cloud of the same third feature object in the solved point cloud data 1-3 and the vertical direction, and the angles γ ₁ ~γ _a between the surface point cloud of the same third feature object and the horizontal direction; then check the maximum distances S ₁ ~S _b between points of the same feature of the same fourth feature object. For example, the solved point cloud data 1-3 is synthesized into a group of data by solving point cloud data 1, solving point cloud data 2 and solving point cloud data 3. In the solved point cloud data 1-3, three identical cubes are superimposed at the position of each fourth feature object (such as a cube), and there is a distance between the same points of the three cubes, and the maximum distance is recorded as S; when the number of cubes is four, in the solved point cloud data 1-3, the maximum distance measured at the position of the first cube is S ₁ , and the maximum distance measured at the position of the second cube is S ₂ , the maximum distance measured at the position of the third cube is S ₃ , the maximum distance measured at the position of the fourth cube is S ₄ ;

Import the solved point cloud data 4 generated in step S7 into CAD, cut the solved point cloud data 4, and check the point cloud thickness T of the first feature object;

The solved point cloud data 1, solved point cloud data 2 and solved point cloud data 3 generated in step S7 are respectively imported into Geomagic (or 3DR, PolyWorks, etc.) software, and fitting processing is performed respectively. According to the fitting processing results, the angles α ₁₁ ~α _1m-1 between adjacent first feature objects, the distances D ₁₁ ~ D _1k-1 between adjacent second feature objects, and the distances d ₁₁ ~ d _{1q-1 between adjacent measurement plates of the fifth feature object are respectively solved} .

When viewing the point clouds of the third and fourth feature objects, you can crop and delete the point clouds other than the third and fourth feature objects, and you can also generate new point cloud data from the cropped point clouds of the third and fourth feature objects to reduce the amount of data and facilitate subsequent analysis.

In step S9, the parameters to be calibrated of the SLAM three-dimensional laser scanner include: size measurement error (feature size measurement indication error and graphic size measurement indication error), plane angle measurement indication error, coordinate measurement error, plumb and horizontal angle measurement error, point cloud thickness and repeated measurement accuracy.

The analysis process of the indication error of the feature size measurement includes: subtracting the distance between the adjacent second feature objects obtained in step S8 from the distance between the corresponding adjacent second feature objects obtained in step S4 to obtain the indication error D _1k-1 -D _0k-1 ; subtracting the distance between the adjacent measurement plates of the fifth feature object obtained in step S8 from the distance between the corresponding adjacent measurement plates of the fifth feature object obtained in step S4 to obtain the indication error d _1q-1 -d _0q-1 .

The analysis process of the indication error of the pattern size measurement includes: subtracting the pattern size obtained in step S8 from the corresponding pattern size obtained in step S4 to obtain the indication error; for example, for a circular pattern, the indication error is φ _1i -φ _0i ; for a rectangular pattern, the indication error is L _1j - L _0j , K _1j -K _0j .

The analysis process of the indication error of the plane angle measurement includes: subtracting the angle between adjacent first characteristic objects obtained in step S8 from the angle between corresponding adjacent first characteristic objects obtained in step S4 to obtain the indication error α _1m-1 -α _0m-1 .

The analysis process of the coordinate measurement error includes: subtracting the coordinates of the reference point obtained in step S8 (H _1p , S _1p , N _1p ) from the coordinates of the corresponding reference point obtained in step S4 (H _0p , S _0p , N _0p ) to obtain the coordinate indication error (H _1p - H _0p , S _1p - S _0p , N _1p - N _0p ).

The analysis process of the plumb and horizontal angle measurement errors includes: the maximum angle β _max obtained from the angles β ₁ ~β _a between the surface or axis point cloud of the third feature object obtained in step S8 and the vertical direction is the plumb direction angle measurement error, and the maximum angle γ _max -90° obtained from the angles γ ₁ ~γ _a between the surface point cloud of the third feature object and the horizontal direction is the horizontal direction angle measurement error.

The point cloud thickness is the point cloud thickness T of the first feature object obtained in step S8.

The analysis process of the repeat measurement accuracy includes: the maximum distances S ₁ to S _b between points of the same feature of the same fourth feature object obtained in step S8 are obtained as S _max , which is the repeat positioning accuracy.

According to the types of parameters to be calibrated, after analyzing the analysis values of the parameters to be calibrated, the maximum value is selected from the analysis values of each parameter as the calibration result; then the maximum value of each parameter is compared with the standard range corresponding to the parameter to determine whether the analysis value of each parameter is within the standard range corresponding to the parameter, thereby determining whether the accuracy indicators of the corresponding parameters of the SLAM three-dimensional laser scanner to be calibrated meet the requirements.

The above is only a detailed description of the specific implementation of the present invention, rather than a limitation of the present invention. Various substitutions, modifications and improvements made by those skilled in the relevant art without departing from the principle and scope of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A SLAM three-dimensional laser scanner calibration system, characterized in that it includes: a calibration site, a feature object, a reference point module, a standard device, a SLAM three-dimensional laser scanner to be calibrated, a scanning point cloud solution and preprocessing module, a scanning point cloud preview module and a data analysis module; Wherein, a path for calibration scanning is planned in the calibration site, and patterns for characterizing the indication error of graphic size measurement are respectively arranged on both sides of the path; The feature objects are respectively arranged on both sides of the path, and the feature objects include: a first feature object for characterizing the indication error of plane angle measurement and point cloud thickness, a second feature object for characterizing the indication error of feature object size measurement, a third feature object for characterizing the plumb bob and horizontal direction angle measurement error, and a fourth feature object for characterizing repeated measurement accuracy; The reference point module is used to determine the reference point positions on both sides of the path, and after setting the reference point at the later reference point position, measure the coordinates of the reference point; The standard device includes: a first standard device for obtaining the angle between adjacent first feature objects, a second standard device for obtaining the distance between adjacent second feature objects, and a third standard device for measuring the geometric dimensions of the patterns on both sides of the path; The SLAM three-dimensional laser scanner to be calibrated is used to collect scanning point cloud data and store the collected scanning point cloud data; The scanning point cloud solution and preprocessing module is used to perform coordinate transformation processing on the scanning point cloud data, convert the scanning point cloud data into the same coordinate system, and obtain solution point cloud data; The scanning point cloud preview module is used to preview the solved point cloud data processed by the scanning point cloud solving and preprocessing module, and measure the size of the pattern through the previewed solved point cloud data; The data analysis module is used to analyze the solved point cloud data processed by the scanning point cloud solution and preprocessing module and to construct a geometric model. 2. The SLAM three-dimensional laser scanner calibration system according to claim 1, characterized in that the first feature object is a flat plate, and a plurality of flat plates are arranged on both sides of the path circumferentially along the path or a plurality of flat plates are arranged at different positions along the vertical direction of the path, and the angle between adjacent flat plates increases in steps, and the angle between two adjacent flat plates is greater than 0 degrees and not greater than 180 degrees; The second feature object is a sphere, the diameter of the sphere is 30 mm to 300 mm, and the sphericity is not greater than 0.1 mm; multiple spheres are arranged on both sides of the path along the circumference of the path or multiple spheres are arranged at different positions along the vertical direction of the path; The third characteristic object is a cylinder, and multiple cylinders are arranged on both sides of the path along the path. The upper end of each cylinder is suspended in the air by a tool, and the lower end of the cylinder is suspended in the air. The hook of the tool is on the geometric center line of the cylinder, and the geometric center line of the cylinder is vertically downward. The length of each cylinder is not less than 2m, and the coaxiality is not greater than 0.1mm; The calibration site is a rectangular site, the fourth feature object is not limited to a cube, and four fourth feature objects are respectively arranged at the four corners of the calibration site; The length of the calibration site is not less than 40m and the width is not less than 40m; the path is a circular path, and the path overlaps 2m to 5m at the starting point and the end point. 3. The SLAM three-dimensional laser scanner calibration system according to claim 1 or 2 is characterized in that it also includes a fifth feature object for characterizing the feature object size measurement indication error, the fifth feature object adopts a standard spacing gauge, and multiple fifth features are arranged horizontally, vertically, and obliquely on both sides of the path. 4. A SLAM three-dimensional laser scanner calibration method, characterized in that it comprises the following steps: Step S1: planning a path for calibration scanning on a calibration site; Step S2: Determine the positions of the reference points on both sides of the path; Step S3: setting regular shaped patterns, features and reference points on both sides of the path; Step S4: obtaining measurement results of the pattern, feature object and reference point respectively; The measurement result of the feature object includes: distances D ₀₁ to D _0k-1 between adjacent second feature objects, wherein k represents the number of second feature objects, and k is an integer not less than 2; The angles between adjacent first features are α ₀₁ ~α _0m-1 , where m represents the number of first features and m is an integer not less than 2; The distances between adjacent measurement plates on the body of the fifth feature object are d ₀₁ to d _0q-1 , where q represents the number of measurement plates and q is an integer not less than 2; The measurement result of the reference point includes the coordinates of the reference point (H ₀₁ , S ₀₁ , N ₀₁ ) to (H _0p , S _0p , N _0p ), wherein p represents the number of the reference point, and p is an integer not less than 1; Step S5: Make the SLAM three-dimensional laser scanner to be calibrated scan along the path to collect scanning point cloud data of the path and both sides of the path; Step S6: enabling the SLAM three-dimensional laser scanner to be calibrated to separately collect scanning point cloud data of the first feature object; Step S7: in the absolute coordinate system, respectively solve the scanned point cloud data collected in step S5 and step S6 to obtain corresponding solved point cloud data; Step S8: Processing the point cloud data solved in step S7 according to the type of parameters to be calibrated, and obtaining the measurement results of the corresponding parameters of the SLAM three-dimensional laser scanner to be calibrated; Step S9: According to the type of parameters to be calibrated, the measurement results of the corresponding parameters of the SLAM 3D laser scanner to be calibrated are analyzed to determine whether the accuracy index of the corresponding parameters of the SLAM 3D laser scanner to be calibrated meets the requirements. 5. The SLAM three-dimensional laser scanner calibration method according to claim 4 is characterized in that the parameter types to be calibrated include size measurement error, plane angle measurement indication error, coordinate measurement error, plumb and horizontal angle measurement error, point cloud thickness and repeated measurement accuracy; the size measurement error includes feature object size measurement indication error and graphic size measurement indication error. 6. The SLAM three-dimensional laser scanner calibration method according to claim 4, characterized in that the step S4 specifically comprises: The geometric dimensions of each pattern on both sides of the path are obtained by the third standard device; when the pattern is circular, the diameters φ ₀₁ ~φ _0i are measured; when the pattern is rectangular, the lengths L ₀₁ ~ L _0j and the widths K ₀₁ ~ K _0j are measured; wherein i represents the number of circular patterns, i is an integer not less than 1; j represents the number of rectangular patterns, j is an integer not less than 1; Obtaining distances D ₀₁ to D _0k-1 between adjacent second characteristic objects by a second standard device, wherein k represents the number of the second characteristic objects and k is an integer not less than 2; Obtaining angles α ₀₁ to α _0m-1 between adjacent first characteristic objects by a first standard device, wherein m represents the number of first characteristic objects, and m is an integer not less than 2; The coordinates of the reference points (H ₀₁ , S ₀₁ , N ₀₁ )~(H _0p , S _0p , N _0p ) are obtained through static measurement and analysis of the GNSS receiver network, where p represents the number of reference points and p is an integer not less than 1; The distances d ₀₁ to d _0q-1 between adjacent measurement plates are read through the standard scale lines on the body of the fifth feature, wherein q represents the number of measurement plates and q is an integer not less than 2. 7. The SLAM three-dimensional laser scanner calibration method according to claim 6, characterized in that in the step S5, the process of making the SLAM three-dimensional laser scanner to be calibrated scan along the path and collecting scanning point cloud data of the path and both sides of the path comprises: Step S51: Turn on the SLAM three-dimensional laser scanner to be calibrated and initialize it; Step S52: Make the SLAM three-dimensional laser scanner to be calibrated walk and scan along the path, and collect scanning point cloud data of the path and both sides of the path; Step S53: After the SLAM three-dimensional laser scanner to be calibrated walks along the path for one circle, the SLAM three-dimensional laser scanner to be calibrated is turned off; When more paths and scanning point cloud data on both sides of the paths need to be collected, steps S51 to S53 are repeated. 8. The SLAM three-dimensional laser scanner calibration method according to claim 7, characterized in that, in the step S7, the process of solving the scanning point cloud data collected in the step S5 comprises: solving the scanning point cloud data n collected for the nth time, obtaining solved point cloud data n, and saving the solved point cloud data n; wherein n is an integer not less than 1; when n>1, merging the solved point cloud data 1 to the solved point cloud data n to generate and save a set of solved point cloud data 1-n; After solving the scanned point cloud data n+1 collected in step S6, solved point cloud data n+1 is obtained, and the solved point cloud data n+1 is saved separately. 9. The SLAM three-dimensional laser scanner calibration method according to claim 8, characterized in that in step S8, the process of processing the point cloud data solved in step S7 comprises: View the data through the browsing software and measure the size of each pattern; for circular patterns, measure the diameter φ ₁₁ ~φ _1i , for rectangular patterns, measure the length L ₁₁ ~ L _1j and the width K ₁₁ ~ K _1j ; obtain the coordinates of the reference point (H ₁₁ , S ₁₁ , N ₁₁ ) ~ (H _1p , S _1p , N _1p ) through the browsing software; Import the solved point cloud data 1-n into CAD, cut the solved point cloud data 1-n, check the angles β ₁ ~β _a between the surface or axis point cloud of the same third feature object in the solved point cloud data 1-n and the vertical direction, and the angles γ ₁ ~γ _a between the surface or axis point cloud of the third feature object and the horizontal direction; then check the maximum distances S ₁ ~S _b between the points of the same feature of the same fourth feature object; wherein a represents the number of the third feature objects, and a is an integer not less than 1; b represents the number of the fourth feature objects, and b is an integer not less than 1; Import the solved point cloud data n+1 into CAD, cut the solved point cloud data n+1, and check the point cloud thickness T of the first feature object; The solved point cloud data 1 to the solved point cloud data n are respectively imported into Geomagic, 3DR or PolyWorks software, and fitting processing is performed respectively. According to the fitting processing results, the angles α ₁₁ ~α _1m-1 between adjacent first feature objects, the distances D ₁₁ ~D _1k-1 between adjacent second feature objects, and the distances d ₁₁ ~d _1q-1 between adjacent measurement plates of the fifth feature object are respectively solved. 10. The SLAM three-dimensional laser scanner calibration method according to any one of claims 4 to 8 is characterized in that step S9 specifically comprises: subtracting the measurement results of each parameter of the SLAM three-dimensional laser scanner to be calibrated obtained in step S8 from the measurement results of the corresponding parameters obtained in step S4 to obtain analysis values of each parameter, and selecting the maximum value from the analysis values of each parameter as the calibration result.