CN113155027B - Tunnel rock wall feature identification method - Google Patents
Tunnel rock wall feature identification method Download PDFInfo
- Publication number
- CN113155027B CN113155027B CN202110458537.2A CN202110458537A CN113155027B CN 113155027 B CN113155027 B CN 113155027B CN 202110458537 A CN202110458537 A CN 202110458537A CN 113155027 B CN113155027 B CN 113155027B
- Authority
- CN
- China
- Prior art keywords
- point cloud
- point
- rock wall
- cloud data
- tunnel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000011435 rock Substances 0.000 title claims abstract description 83
- 238000000034 method Methods 0.000 title claims abstract description 67
- 229910000831 Steel Inorganic materials 0.000 claims abstract description 56
- 239000010959 steel Substances 0.000 claims abstract description 56
- 238000005507 spraying Methods 0.000 claims abstract description 31
- 239000007921 spray Substances 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims abstract description 6
- 238000001914 filtration Methods 0.000 claims description 60
- 238000013461 design Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 4
- 238000010845 search algorithm Methods 0.000 claims description 3
- 230000000717 retained effect Effects 0.000 claims description 2
- 239000002002 slurry Substances 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 3
- 238000010276 construction Methods 0.000 description 11
- IAZDPXIOMUYVGZ-UHFFFAOYSA-N Dimethylsulphoxide Chemical compound CS(C)=O IAZDPXIOMUYVGZ-UHFFFAOYSA-N 0.000 description 4
- 238000005422 blasting Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 239000003086 colorant Substances 0.000 description 2
- 238000005553 drilling Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000009412 basement excavation Methods 0.000 description 1
- 238000010009 beating Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009432 framing Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/002—Measuring arrangements characterised by the use of optical techniques for measuring two or more coordinates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention provides a tunnel rock wall feature identification method which is used for solving the problems of strong subjectivity and poor tunnel rock wall spraying effect of the existing artificial identification arch. The method comprises the following steps: carrying out three-dimensional scanning on the tunnel rock wall by using a laser scanner to obtain point cloud data containing tunnel rock wall feature information; processing the point cloud data by using a normal differential algorithm to obtain normal differential values of all point clouds to form a line differential feature vector; dividing the line differential feature vector into different clustering categories by using an European clustering method, so as to realize the identification of tunnel rock wall features; and acquiring three-dimensional coordinates of the identified tunnel rock wall features. The method realizes the classification and the accurate determination of the spatial positions of the tunnel rock wall steel arches, and can guide the subsequent automatic guniting operation. The invention can be fed back to the upper computer of the wet spraying trolley, so that a developer can conveniently formulate a path plan of the wet spraying machine arm support and the spray gun according to the identified three-dimensional coordinates of the steel arch, and the subsequent automatic spraying operation is realized.
Description
Technical neighborhood
The invention relates to the technical field of tunnel engineering construction, in particular to a tunnel rock wall feature identification method.
Background
The tunnel drilling and blasting method is constructed by taking control blasting as an excavation method, and after earth and stones are excavated in the tunnel according to the designed size, the exposed rock wall of the tunnel is initially supported, so that the stability of surrounding rock is kept as much as possible, and a stable cavity is formed. In the tunnel drilling and blasting method construction process, the existing primary support arm section mainly comprises anchor rod beating, reinforcement mesh laying and vertical arch centering. After the initial support is carried out on the exposed rock wall of the tunnel, an operator controls the wet spraying machine through the handle to carry out spraying operation, the position of the arch frame is usually required to be manually observed and determined, the control handle moves the arm support and the spray gun to the area between two adjacent arch frames to carry out concrete spraying operation, the concrete spraying operation is influenced by subjectivity of the operator, and the artificial identification arch frames have larger deviation, so that the concrete covering thickness on the surface of the steel arch frames exceeds the spraying control standard and cannot meet the field construction requirement. On the other hand, the covering thickness of the concrete on the surface of the steel arch is too large, and the concrete needs to be manually shoveled off in the later period, so that the follow-up construction process cannot be carried out on time. Therefore, how to accurately identify the position of the tunnel rock wall steel arch has important significance for guaranteeing the spraying quality, saving the spraying construction time and guaranteeing the good connection of the follow-up construction process.
At present, the three-dimensional laser scanning technology is used as a novel measuring technology, and is widely applied to tunnel construction with high measuring precision and high scanning speed. And carrying out panoramic scanning on the tunnel outline by a laser scanner to obtain three-dimensional point cloud data containing tunnel rock mass characteristics. Based on three-dimensional point cloud data, most of the research is focused on tunnel super-undermining measurement, and less on tunnel rock wall feature identification to assist in the research of spraying operation. Therefore, the invention provides a tunnel rock mass feature identification method utilizing three-dimensional point cloud data, which guides the subsequent automatic guniting operation by accurately identifying the position coordinates of an initial support steel arch.
Disclosure of Invention
Aiming at the technical problems of strong subjectivity and poor tunnel rock wall spraying effect of the existing artificial identification arch, the invention provides a tunnel rock wall feature identification method, which realizes the classification and the accurate determination of the space position of the tunnel rock wall steel arch through point cloud data acquisition, non-target object rejection and steel arch point cloud identification, and can guide the subsequent automatic spraying operation.
In order to achieve the above purpose, the technical scheme of the invention is realized as follows: a tunnel rock wall feature identification method comprises the following steps:
step one: carrying out three-dimensional scanning on the tunnel rock wall by using a laser scanner to obtain point cloud data containing tunnel rock wall feature information;
step two: processing the point cloud data by using a normal differential algorithm to obtain normal differential values of all point clouds to form a line differential feature vector;
step three: dividing the line differential feature vector into different clustering categories by using an European clustering method, so as to realize the identification of tunnel rock wall features;
step four: and (3) acquiring three-dimensional coordinates of the identified tunnel rock wall feature according to the tunnel rock wall feature identified in the step (III).
Further, the tunnel rock wall feature is a steel arch; the three-dimensional coordinates of the identified steel arch are fed back to the upper computer of the wet spraying trolley, so that the path planning of the arm frame and the spray gun of the wet spraying trolley can be formulated, and the automatic spraying operation is realized. The wet spraying trolley can be guided by identifying the steel arch on the tunnel rock wall, so that automatic spraying is realized.
Further, the laser scanner in the first step is arranged at a position 10-15 m away from the tunnel rock wall steel arch, and a spherical prism target is placed at a position 5-10 m away from the laser scanner; and (3) measuring and absolute positioning the laser scanner by erecting a total station in the tunnel, and obtaining the geodetic absolute coordinate of the laser scanner in the tunnel. The point cloud data can be converted into the geodetic coordinate system according to the geodetic absolute coordinates, so that the point cloud data can be conveniently processed.
Further, the type of the point cloud data in the step one includes an X coordinate, a Y coordinate, a Z coordinate, and a reflection intensity value of the point. And filtering processing of the point cloud data can be realized according to the reflection intensity value.
Further, filtering the point cloud data in the first step to remove non-target objects, and then processing by using a normal differential algorithm; the method for filtering the point cloud data comprises the following steps:
step 1, designing a filtering threshold radius according to the central axis coordinate of a tunnel and the design radius of an initial support of the tunnel, and calculating the distance from the central axis coordinate to point cloud dataDistance of passage->Filtering the point cloud data by comparing with a filtering threshold radius to obtain point cloud data after primary filtering; the primary filtering filters invalid point cloud data according to the distance and the filtering threshold radius.
Step 2, counting the reflection intensity values of all points in the point cloud data after primary filtering to obtain the reflection intensity range [ arch ] of the point cloud data corresponding to the surface of the steel arch s ,arch l ];
Step 3, according to the threshold range [ arch ] s ,arch l ]And performing secondary point cloud filtering on the point cloud data after primary filtering: if the reflection intensity value of the point cloud data after primary filtering is within the threshold range [ arch ] s ,arch l ]The point cloud data after primary filtering is reserved; otherwise, eliminating the point cloud data on the concrete surface of the tunnel rock wall, and obtaining the point cloud data after secondary filtering.
Further, the implementation method of the step 1 is as follows:
step 1.1, according to the central axis coordinate P of the tunnel c =(x c ,y c ,z c ) Calculate the distance from the point cloud data p= (x, y, z):
step 1.2, calculating a filtering threshold radius R according to the design radius R of the primary support of the tunnel 1 And radius R 2 The method comprises the following steps of:
wherein, alpha and beta are both filtering scale factors;
step 1.3, comparing the distance between the central axis coordinate of the tunnel and the point cloud data of the tunnel rock wallAnd a filter threshold radius R 1 Radius R 2 If distance->The point cloud data is retained; otherwise, the point cloud data are rejected.
Further, the method for realizing the normal differential algorithm comprises the following steps:
step 4, setting a first neighborhood radius r for estimating normal vector of point cloud in tunnel rock wall 1 And a second neighborhood radius r 2 And r is 1> r 2 ;
Step 5, for each point in the point cloud data after the secondary filtering, utilizing a first neighborhood radius r 1 Calculating the normal vector n of each point 1 ;
Step 6, for each point in the point cloud data after the secondary filtering, utilizing a second neighborhood radius r 2 Calculating the normal vector n of each point 2 ;
Step 7, calculating and normalizing the normal difference value delta n of the same point under different neighborhood radiuses for each point in the point cloud data after the secondary filtering to obtain the normal difference value of each point after normalization;
and 8, filtering the normal differential values of all the normalized point clouds according to a preset normal differential threshold value to obtain a normal differential feature vector of the filtered point clouds. The normal differential algorithm can filter the point cloud data again, so that the point cloud corresponding to the steel arch can be identified by a subsequent European clustering method.
Further, the normal vector n of each point is calculated in the step 5 1 The method of (1) is as follows:
s5.1, searching a first neighborhood radius r according to the coordinates of each point 1 All neighborhood points in the range form a neighborhood point cloud set M 1 ;
S5.2, calculating a neighborhood point cloud set M 1 Center point coordinate center of (a) 1 Coordinates of (c):
wherein, for neighborhood point cloud set M 1 Point of the center point coordinates of (a) i =(x i ,y i ,z i ) For neighborhood point cloud set M 1 Three-dimensional coordinates of the ith point in the region, M is a neighborhood point cloud set M 1 The number of the internal point clouds, i=1, 2,. -%, m;
Further, solving the normal vector n in the step 6 2 The method of (1) is as follows:
s6.1, searching a second neighborhood radius r according to the coordinates of each point 2 All neighborhood points in the range form a neighborhood point cloud set M 2 ;
S6.2, calculating a neighborhood point cloud set M 2 Center point coordinate center of (a) 2 Coordinates of (c):
wherein, for neighborhood point cloud set M 2 Point of the center point coordinates of (a) j =(x j ,y j ,z j ) For neighborhood point cloud set M 2 Three-dimensional coordinates of the jth point in the neighborhood, n is a neighborhood cloud set M 2 Number of internal point clouds, j=1, 2,..n;
Further, the normal difference value Deltan isThe normal differential value is +.>Wherein, the term "represents a norm of the normal difference value Δn.
Further, in the step 8, the method for filtering the normal differential values of all the normalized point clouds includes: according to a preset normal difference threshold delta n threshold If the normal difference value delta n of the normalized point cloud normalize >=Δn threshold Retaining the normal differential value; otherwise, eliminating the normal difference value; and the vector formed by the normal differential values of all the point clouds after filtering is the normal differential feature vector.
Further, the implementation method of the European clustering method comprises the following steps:
step (3.1): according to the filtered normal differential values of the point cloud, randomly selecting the normal differential values corresponding to p points as an initialized Cluster center Cluster t K normal differential values are found out to each initialized Cluster center Cluster through a neighbor search algorithm t The nearest point; wherein t=1, 2,···,p;
Step (3.2): according to a given distance threshold d threshold Calculating k normal differential values to corresponding initialized Cluster centers Cluster t If the distance of the normal difference value is smaller than the set threshold d threshold Then cluster to cluster set Q t ;
Step (3.3): if cluster set Q t The number of the internal point clouds is not increased any more, and the clustering process is ended; otherwise, the initial cluster center point needs to be updated, and a cluster set Q is selected t Is used as the clustering center, and the step S5.5.3 is repeatedly executed until the clustering set Q t The number of point clouds does not increase any more. The European clustering method clusters the point cloud data to obtain different clustering categories.
Further, the method for obtaining the three-dimensional coordinates in the fourth step comprises the following steps: according to the clustered set Q t Acquiring point cloud data sets corresponding to different steel arches, namely, different clustering categories correspond to different steel arches, and each clustering set Q t Including three-dimensional coordinate information of the steel arch.
Compared with the prior art, the invention has the beneficial effects that: according to the three-dimensional coordinates and the reflection intensity of the point cloud, the method utilizes the threshold filter radius and the reflection intensity range of the steel arch to realize rapid filtration of the point cloud data of the tunnel rock wall steel arch; and extracting normal differential characteristics from the filtered tunnel rock wall steel arch point cloud data, so that the steel arch position of the tunnel rock wall can be accurately identified. In addition, the three-dimensional coordinates of the tunnel rock wall steel arch frame identified by the invention can be fed back to the wet spraying trolley upper computer, so that a developer can conveniently formulate a path planning of the wet spraying machine arm frame and the spray gun according to the identified three-dimensional coordinates of the steel arch frame, and the subsequent automatic spraying operation is realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of the present invention.
Fig. 2 is a diagram showing the recognition effect of the tunnel rock wall steel arch in the embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which are obtained by a person of ordinary skill in the art without any inventive effort, are within the scope of the present invention based on the embodiments of the present invention.
As shown in fig. 1, a tunnel rock wall feature identification method comprises the following steps:
s1, placing a laser scanner at a position which is 15m away from a tunnel rock wall steel arch frame, and placing a spherical prism target at a position which is 6m away from the laser scanner.
The laser scanner is used for measuring point cloud data of the tunnel rock wall, takes a spherical prism as a target, and aims at the target through the total station to determine the azimuth angle and the distance between the current station and the target so as to realize positioning.
And S2, before the measurement operation of the laser scanner, erecting a total station in the tunnel to measure and absolutely position the laser scanner, thereby obtaining the geodetic absolute coordinate of the laser scanner in the tunnel.
And (3) aiming the spherical prism nearby the laser scanner through the total station, determining the azimuth angle and the distance between the current measuring point and the target, and carrying out coordinate conversion by combining the known coordinates of the target point and the azimuth angle and the distance to realize absolute positioning, namely obtaining the absolute coordinates of the laser scanner nearby the measuring point. The function of acquiring the geodetic absolute coordinates is to convert the point cloud data acquired by the laser scanner into a geodetic coordinate system.
And S3, carrying out three-dimensional panoramic scanning on the tunnel rock wall through a laser scanner, and obtaining point cloud data containing tunnel rock wall steel arch frame information.
The type of the point cloud data acquired in the step S3 is the reflection intensity value of the point, which is the X coordinate, the Y coordinate, the Z coordinate.
S4, filtering the obtained original point cloud data according to the central axis coordinates of the tunnel and the design radius of the primary support of the tunnel, and removing non-target objects such as an air pipe, a construction trolley and field personnel in the tunnel.
The specific implementation steps of the method for filtering the point cloud data are as follows:
s4.1, calculating a filtering threshold radius R according to the central axis coordinate of the tunnel provided by the tunnel construction party and the design radius R of the primary support of the tunnel 1 Radius R 2 And the distance between the central axis coordinate and the original point cloud dataDistance of passage->Comparing the filtered point cloud data with a filtering threshold radius, and removing non-target objects such as air pipes, construction trolleys and field personnel outside the tunnel rock wall to obtain point cloud data after primary filtering; the method comprises the following specific steps:
s4.1.1, according to the central axis coordinate P of the tunnel c =(x c ,y c ,z c ) And calculating the distance between the central axis coordinate of the tunnel and the original point cloud data of the tunnel rock wall according to the original point cloud data P= (x, y, z):
s4.1.2 calculating a filter threshold radius R according to a certain scale factor according to the design radius R of the primary support of the tunnel 1 And radius R 2 The method comprises the following steps of:
wherein, alpha and beta are both filtering scale factors, alpha is generally set to 0.7-0.9, and beta is generally set to 1.1-1.3.
S4.1.3 by comparing the distance between the central axis coordinate of the tunnel and the original point cloud data of the rock wall of the tunnelAnd a filter threshold radius R 1 Radius R 2 If->The original point cloud data is reserved; otherwise, the invalid point cloud data is regarded as invalid point cloud data, and the invalid point cloud data is rejected.
S4.2, counting the reflection intensity values of all points in the point cloud data after primary filtering, and counting the reflection intensity range [ wall ] of the point cloud data corresponding to the tunnel rock wall, namely the tunnel concrete wall surface according to the principle that the reflection intensities of laser on different object surfaces are different s ,wall l ]Reflection intensity range [ arch ] of point cloud data corresponding to steel arch surface s ,arch l ]。
Statistics of the reflection intensity Range of the Tunnel rock wall s ,wall l ]The method aims at distinguishing the difference of the point cloud intensity between the non-target object and the target object. By drawing the intensity distribution of the point cloud data after primary filtering, framing the point cloud data corresponding to the steel arch in a man-machine interaction mode, and determining the lower intensity limit arch corresponding to the arch point cloud data s Upper limit arch l Thereby obtaining the reflection intensity range [ arch ] s ,arch l ]。
S4.3, according to the threshold value range of the reflection intensity of the steel arch surface [ arch ] s ,arch l ]And (3) performing secondary point cloud filtering: if the reflection intensity value of the point cloud data is between [ arch ] s ,arch l ]If so, retaining point cloud data; otherwise, the point cloud data of the concrete surface of the tunnel rock wall is removed as a non-target object, and the point cloud data of the tunnel rock wall steel arch after secondary filtering is obtained.
S5, according to point cloud data of the tunnel rock wall steel arch after secondary filtering, the tunnel rock wall steel arch is identified by utilizing a normal differential algorithm, and three-dimensional coordinates of the tunnel rock wall steel arch after identification are obtained, wherein the method comprises the following specific steps:
s5.1, setting a larger neighborhood radius r for estimating a point cloud normal vector in a tunnel rock wall 1 Smaller neighborhood radius r 2 Larger neighborhood radius r 1 Typically set to 30-40, a smaller neighborhood radius r 2 Is set to 10 to 20.
S5.2, for each point in the point cloud data after secondary filtering, utilizing a larger neighborhood radius r 1 Calculating the normal vector n of each point 1 The method comprises the following specific steps of:
s5.2.1, according to the coordinates of each point and the larger neighborhood radius r 1 Searching for a neighborhood radius r 1 All neighborhood points in the range form a neighborhood point cloud set M 1 ;
S5.2.2 according to the point cloud set M 1 Computing a neighborhood point cloud set M 1 Center point coordinate center of (a) 1 :
Wherein, for neighborhood point cloud set M 1 Point of the center point coordinates of (a) i =(x i ,y i ,z i ) For neighborhood point cloud set M 1 Three-dimensional coordinates of the ith point in the region, M is a neighborhood point cloud set M 1 Number of internal point clouds, i=1, 2,..m.
S5.2.3 knowing the coordinates center of all points and center points in the point cloud neighborhood 1 Solving for normal vector n 1 Can be converted into seeking a normal vector n 1 The projection point distribution of all the neighborhood points in the direction is most concentrated, namely, the minimization problem is solved:
solving the above minimization problem by using least square method to obtain normal vector n 1 。
S5.3, for each point in the point cloud data after the secondary filtering, utilizing a smaller neighborhood radius r 2 Calculating the normal vector n of each point 2 The method comprises the following specific steps of:
s5.3.1 according to each point coordinate and smaller neighborhood radius r 2 Searching all neighborhood points in the neighborhood radius range to form a neighborhood point cloud set M 2 。
S5.3.2, according to the neighborhood point cloud set M 2 Computing a neighborhood point cloud set M 2 Center point coordinate center of (a) 2 :
Wherein, for neighborhood point cloud set M 2 Point of the center point coordinates of (a) j =(x j ,y j ,z j ) For neighborhood point cloud set M 2 Three-dimensional coordinates of the jth point in the neighborhood, n is a set of neighboring points M 2 Number of internal point clouds, j=1, 2,..n.
S5.3.3 knowing the coordinates center of all points and center points in the point cloud neighborhood 2 Solving for normal vector n 2 Can be converted into seeking a normal n 2 So that all neighborhood points are in the direction n 2 The distribution of the projection points is most concentrated:
s5.4, for each point in the point cloud data after the secondary filtering, calculating the normal difference value delta n of the same point under different neighborhood radiuses, and normalizing to obtain the normal difference value delta n of each point after normalization normalize :
Wherein, the term represents the norm of the normal difference value deltan, i.e. the magnitude of the normal difference value deltan, is used for the normalization process.
S5.5, according to a preset normal difference threshold delta n threshold Filtering the normal differential values of all the normalized point clouds to obtain the normal differential values of the filtered point clouds, and identifying the point clouds corresponding to the steel arch by using the European clustering method, wherein the specific steps are as follows:
s5.5.1 according to a preset normal differential threshold value Deltan threshold If the normal difference value delta n of the normalized point cloud normalize >=Δn threshold The corresponding normal differential value is reserved; otherwise, the corresponding normal difference value is eliminated.
S5.5.2 according to the filtered point cloud normal differential values, randomly selecting the normal differential values corresponding to p points in the space as an initialized Cluster center Cluster t K normal differential values are found out to each initialized Cluster center Cluster through a neighbor search algorithm t The nearest point. Wherein t=1, 2, the sum of the values is p.
S5.5.3, according to a given distance threshold d threshold Calculating k normal differential values to corresponding initialized Cluster centers Cluster t If the distance of the k normal difference values is smaller than the set threshold d threshold Then cluster to cluster set Q t ,t=1,2,···,p;
S5.5.4 if cluster set Q t The number of the internal point clouds is not increased any more, and the clustering process is ended; otherwise, the initial cluster center point needs to be updated, and a cluster set Q is selected t Is used as the clustering center, and the step S5.5.3 is repeatedly executed until the clustering set Q t The number of point clouds does not increase any more.
S5.5.5 according to the clustered set Q of the clustered point cloud t And acquiring point cloud data sets corresponding to different steel arches, namely, different clustering categories correspond to different steel arches, and each clustering set comprises three-dimensional coordinate information of the steel arch.
Calculating the distance d from each filtered normal differential value to the rest normal differential value in the normal differential feature vector; according to a given distance threshold d threshold The normal differential feature vector is divided into different clustering categories, namely different steel arches, with the distance d, and the different clustering categories are distinguished by different colors. And searching and identifying the three-dimensional coordinates of the steel arch according to indexes of different clustering categories. Visualization of steel arch recognition results, and different colors are adopted to distinguish clustered arch information, so that field operators can visually check the information conveniently, as shown in fig. 2.
S6, feeding back the three-dimensional coordinates of the identified tunnel rock wall steel arch to the wet spraying trolley upper computer, and enabling operators to formulate path planning of the wet spraying machine arm support and the spray gun according to the three-dimensional coordinates of the identified steel arch so as to realize subsequent automatic spraying operation.
The method comprises the following steps: placing a laser scanner at a position 10-15 m away from the tunnel rock wall steel arch frame, and placing a spherical prism target at a position 5-10 m away from the scanner; before the measurement operation of the laser scanner, the total station is erected to measure and absolutely position the laser scanner; carrying out three-dimensional panoramic scanning on the tunnel rock wall by using a scanner to obtain point cloud data containing tunnel rock wall steel arch information; filtering the original point cloud data according to the central axis coordinates of the tunnel and the design radius of the primary support of the tunnel, and removing non-target objects such as an air pipe, a construction trolley, field personnel and the like in the tunnel; according to the point cloud data of the filtered tunnel rock wall steel arch, the identification of the tunnel rock wall steel arch is realized by utilizing a normal differential algorithm, and the three-dimensional coordinates of the tunnel rock wall steel arch are obtained; the three-dimensional coordinates of the identified tunnel rock wall steel arch are fed back to the wet spraying trolley upper computer, and a developer can formulate a path planning of the wet spraying machine arm frame and the spray gun according to the three-dimensional coordinates of the identified steel arch, so that subsequent automatic spraying operation is realized.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (12)
1. The tunnel rock wall feature identification method is characterized by comprising the following steps of:
step one: carrying out three-dimensional scanning on the tunnel rock wall by using a laser scanner to obtain point cloud data containing tunnel rock wall feature information;
the tunnel rock wall feature is a steel arch;
filtering the point cloud data in the first step to remove non-target objects, and then processing by using a normal differential algorithm; the method for filtering the point cloud data comprises the following steps:
step 1, designing a filtering threshold radius according to the central axis coordinate of a tunnel and the design radius of an initial support of the tunnel, and calculating the distance from the central axis coordinate to point cloud dataDistance of passage->Filtering the point cloud data by comparing with a filtering threshold radius to obtain point cloud data after primary filtering;
step 2, counting the reflection intensity values of all points in the point cloud data after primary filtering to obtain the reflection intensity range [ arch ] of the point cloud data corresponding to the surface of the steel arch s ,arch l ];
Step 3, according to the threshold range [ arch ] s ,arch l ]And performing secondary point cloud filtering on the point cloud data after primary filtering: if the reflection intensity value of the point cloud data after primary filtering is within the threshold range [ arch ] s ,arch l ]The point cloud data after primary filtering is reserved; otherwise, eliminating point cloud data on the surface of the tunnel rock wall concrete to obtain point cloud data after secondary filtering;
step two: processing the point cloud data by using a normal differential algorithm to obtain normal differential values of all point clouds to form a line differential feature vector;
step three: dividing the line differential feature vector into different clustering categories by using an European clustering method, so as to realize the identification of tunnel rock wall features;
step four: and (3) acquiring three-dimensional coordinates of the identified tunnel rock wall feature according to the tunnel rock wall feature identified in the step (III).
2. The tunnel rock wall feature identification method according to claim 1, wherein the three-dimensional coordinates of the identified steel arch are fed back to a wet spraying trolley upper computer, so that path planning of a wet spraying machine arm frame and a spray gun can be formulated, and automatic slurry spraying operation is achieved.
3. The tunnel rock wall feature recognition method according to claim 2, wherein the laser scanner in the first step is disposed at a position 10-15 m away from the tunnel rock wall steel arch, and a spherical prism target is placed at a position 5-10 m away from the laser scanner; and (3) measuring and absolute positioning the laser scanner by erecting a total station in the tunnel, and obtaining the geodetic absolute coordinate of the laser scanner in the tunnel.
4. A tunnel rock wall feature identification method according to claim 2 or claim 3, wherein the type of point cloud data in the step one includes X-coordinate, Y-coordinate, Z-coordinate and reflected intensity values of points.
5. The tunnel rock wall feature identification method according to claim 4, wherein the implementation method of step 1 is as follows:
step 1.1, according to the central axis coordinate P of the tunnel c =(x c ,y c ,z c ) Calculate the distance from the point cloud data p= (x, y, z):
step 1.2, calculating a filtering threshold radius R according to the design radius R of the primary support of the tunnel 1 And radius R 2 The method comprises the following steps of:
wherein, alpha and beta are both filtering scale factors;
6. The tunnel rock wall feature identification method according to claim 1 or 5, wherein the normal differential algorithm is implemented by:
step 4, setting a first neighborhood radius r for estimating normal vector of point cloud in tunnel rock wall 1 And a second neighborhood radius r 2 And r is 1> r 2 ;
Step 5, for each point in the point cloud data after the secondary filtering, utilizing a first neighborhood radius r 1 Calculating the normal vector n of each point 1 ;
Step 6, for each point in the point cloud data after the secondary filtering, utilizing a second neighborhood radius r 2 Calculating the normal vector n of each point 2 ;
Step 7, calculating and normalizing the normal difference value delta n of the same point under different neighborhood radiuses for each point in the point cloud data after the secondary filtering to obtain the normal difference value of each point after normalization;
and 8, filtering the normal differential values of all the normalized point clouds according to a preset normal differential threshold value to obtain a normal differential feature vector of the filtered point clouds.
7. The method for identifying tunnel rock wall features according to claim 6, wherein the normal vector n of each point is calculated in the step 5 1 The method of (1) is as follows:
s5.1, searching a first neighborhood radius r according to the coordinates of each point 1 All neighborhood points in the range form a neighborhood point cloud set M 1 ;
S5.2, calculating a neighborhood point cloud set M 1 Center point coordinate center of (a) 1 Coordinates of (c):
wherein, for neighborhood point cloud set M 1 Point of the center point coordinates of (a) i =(x i ,y i ,z i ) For neighborhood point cloud set M 1 Three-dimensional coordinates of the ith point in the region, M is a neighborhood point cloud set M 1 The number of the internal point clouds, i=1, 2,. -%, m;
8. The method for identifying tunnel rock wall features according to claim 7, wherein the normal vector n is solved in the step 6 2 The method of (1) is as follows:
s6.1, searching a second neighborhood radius r according to the coordinates of each point 2 All neighborhood points in the range form a neighborhood point cloud set M 2 ;
S6.2, calculating a neighborhood point cloud set M 2 Center point coordinate center of (a) 2 Coordinates of (c):
wherein, for neighborhood point cloud set M 2 Point of the center point coordinates of (a) j =(x j ,y j ,z j ) For neighborhood point cloud set M 2 Three-dimensional coordinates of the jth point in the neighborhood, n is a neighborhood cloud set M 2 Number of internal point clouds, j=1, 2,..n;
10. The method for identifying tunnel rock wall features according to claim 9, wherein the method for filtering the normal differential values of all the normalized point clouds in step 8 comprises the following steps: according to a preset normal difference threshold delta n threshold If the normal difference value delta n of the normalized point cloud normalize >=Δn threshold Retaining the normal differential value; otherwise, eliminating the normal difference value; and the vector formed by the normal differential values of all the point clouds after filtering is the normal differential feature vector.
11. The tunnel rock wall feature identification method according to claim 1 or 10, wherein the implementation method of the euro-clustering method is as follows:
step (3.1): according to the filtered normal differential values of the point cloud, randomly selecting the normal differential values corresponding to p points as an initialized Cluster center Cluster t K normal differential values are found out to each initialized Cluster center Cluster through a neighbor search algorithm t The nearest point; wherein t=1, 2, p;
step (3.2): according to a given distance threshold d threshold Calculating k normal differential values to corresponding initialized Cluster centers Cluster t If the distance of the normal difference value is smaller than the set threshold d threshold Then cluster to cluster set Q t ;
Step (3.3): if cluster set Q t The number of the internal point clouds is not increased any more, and the clustering process is ended; otherwise, the initial cluster center point needs to be updated, and a cluster set Q is selected t Is used as a clustering center, and the step S3.2 is repeatedly executed until the clustering set Q t The number of point clouds does not increase any more.
12. The tunnel rock wall feature identification method according to claim 11, wherein the three-dimensional coordinate acquisition method in the fourth step is as follows: according to the clustered set Q t Acquiring point cloud data sets corresponding to different steel arches, namely, different clustering categories correspond to different steel arches, and each clustering set Q t Including three-dimensional coordinate information of the steel arch.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110458537.2A CN113155027B (en) | 2021-04-27 | 2021-04-27 | Tunnel rock wall feature identification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110458537.2A CN113155027B (en) | 2021-04-27 | 2021-04-27 | Tunnel rock wall feature identification method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113155027A CN113155027A (en) | 2021-07-23 |
CN113155027B true CN113155027B (en) | 2023-05-26 |
Family
ID=76871260
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110458537.2A Active CN113155027B (en) | 2021-04-27 | 2021-04-27 | Tunnel rock wall feature identification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113155027B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117677862A (en) * | 2021-08-09 | 2024-03-08 | 深圳市速腾聚创科技有限公司 | Pseudo image point identification method, terminal equipment and computer readable storage medium |
CN113884010A (en) * | 2021-12-07 | 2022-01-04 | 湖南联智科技股份有限公司 | Tunnel steel frame detection method |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105910556A (en) * | 2016-04-13 | 2016-08-31 | 中国农业大学 | Leaf area vertical distribution information extraction method |
CN111915662B (en) * | 2019-05-07 | 2024-01-12 | 北京京东乾石科技有限公司 | Three-dimensional laser point cloud data preprocessing method and device |
CN110415259B (en) * | 2019-07-30 | 2023-04-28 | 南京林业大学 | Street tree point cloud identification method based on laser reflection intensity |
CN110866449A (en) * | 2019-10-21 | 2020-03-06 | 北京京东尚科信息技术有限公司 | Method and device for identifying target object in road |
CN111027484B (en) * | 2019-12-11 | 2023-04-28 | 中南大学 | Tunnel steel arch identification method based on three-dimensional imaging |
CN111274976B (en) * | 2020-01-22 | 2020-09-18 | 清华大学 | Lane detection method and system based on multi-level fusion of vision and laser radar |
CN111369600B (en) * | 2020-02-12 | 2024-02-09 | 苏州大学 | Laser radar point cloud data registration method based on reference object |
CN111553292B (en) * | 2020-04-30 | 2023-05-02 | 南京工业大学 | Rock mass structural plane identification and occurrence classification method based on point cloud data |
-
2021
- 2021-04-27 CN CN202110458537.2A patent/CN113155027B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN113155027A (en) | 2021-07-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113155027B (en) | Tunnel rock wall feature identification method | |
CN106091972B (en) | A kind of building change detecting method projecting dot density based on moving window | |
CN111811420B (en) | Tunnel three-dimensional contour integral absolute deformation monitoring method and system | |
CN108107444B (en) | Transformer substation foreign matter identification method based on laser data | |
CN108225184B (en) | A kind of shield tunnel Point Cloud Data from Three Dimension Laser Scanning removes the method made an uproar | |
CN109613546B (en) | Three-dimensional measurement method and measurement device for converter furnace chamber based on three-dimensional laser radar auxiliary positioning | |
CN109214994B (en) | Tunnel dense point cloud noise elimination method based on double control points | |
CN110030972A (en) | Tunnel Overbreak & Underbreak detection method based on ExcelVBA | |
CN110008207B (en) | Airborne L iDAR point cloud data vulnerability rapid detection method based on density histogram | |
CN109712190A (en) | The outer ginseng scaling method of three-dimensional laser scanner and three-dimensional laser radar | |
CN109285163B (en) | Laser point cloud based lane line left and right contour line interactive extraction method | |
CN116402866A (en) | Point cloud-based part digital twin geometric modeling and error assessment method and system | |
CN111815611A (en) | Round hole feature extraction method for rivet hole measurement point cloud data | |
CN114234819B (en) | Automatic detection method for quality of steel bar part based on three-dimensional laser scanning | |
CN112325809A (en) | Method for detecting flatness of flange | |
CN110986877B (en) | Railway engineering clearance detection method based on high-precision vehicle-mounted laser mobile measurement system | |
CN115690184A (en) | Tunnel face displacement measurement method based on three-dimensional laser scanning | |
CN114419445A (en) | Tunnel surrounding rock information identification and extraction method based on three-dimensional laser scanning | |
CN112561989B (en) | Recognition method for hoisting object in construction scene | |
CN112785596B (en) | Dot cloud picture bolt segmentation and height measurement method based on DBSCAN clustering | |
CN114067073A (en) | TLS point cloud-based mining area building deformation automatic extraction method | |
CN112907601A (en) | Automatic extraction method and device for tunnel arch point cloud based on feature transformation | |
CN108257222B (en) | The automatic blending algorithm of steel stove converter three-dimensional laser point cloud | |
CN113946982B (en) | Method for obtaining topography profile of dangerous rock mass | |
CN115682970A (en) | Method and system for measuring steel bar spacing based on binocular vision |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |