CN115661421A - Method for removing abnormal point cloud values, method and device for processing point cloud and related equipment - Google Patents
Method for removing abnormal point cloud values, method and device for processing point cloud and related equipment Download PDFInfo
- Publication number
- CN115661421A CN115661421A CN202210983097.7A CN202210983097A CN115661421A CN 115661421 A CN115661421 A CN 115661421A CN 202210983097 A CN202210983097 A CN 202210983097A CN 115661421 A CN115661421 A CN 115661421A
- Authority
- CN
- China
- Prior art keywords
- point
- point cloud
- neighboring
- cloud
- euclidean distance
- 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.)
- Pending
Links
Images
Landscapes
- Length Measuring Devices With Unspecified Measuring Means (AREA)
Abstract
The application discloses a method for removing a point cloud abnormal value, a point cloud processing method, a device and related equipment, wherein the method comprises the following steps: uniformly dividing the point cloud to be processed into a plurality of layers along a reference coordinate axis; for each point in each layer of the point cloud, determining whether to remove the point based on the bounding box of the point to obtain a second point cloud; acquiring the average neighbor Euclidean distance of the second point cloud, namely the neighbor Euclidean distance of each point; for each point in the second point cloud, determining whether to remove the point based on a difference between a neighboring euclidean distance of the point and the average neighboring euclidean distance, and a standard deviation of the neighboring euclidean distance of each point in the second point cloud and the average neighboring euclidean distance. Through the steps, the method and the device can intuitively and conveniently remove the abnormal value in the point cloud on the premise of keeping the detail characteristics.
Description
Technical Field
The present application relates to the field of computer graphics processing technologies, and in particular, to a method for removing a point cloud outlier, a method for processing a point cloud outlier, an apparatus for processing a point cloud outlier, and related devices.
Background
Three-dimensional point cloud is a representative format for displaying the three-dimensional appearance of an object or a scene in the real world at present, and has been widely used in various fields such as pose estimation, target tracking, remote sensing, cultural relic restoration and the like. Wherein the three-dimensional point cloud is mainly acquired by a depth camera or a laser scanning device. However, due to the existence of factors such as ambient light interference, target reflectivity change or instability of the sensor itself, a large number of abnormal values often appear in the three-dimensional point cloud. Therefore, before the acquired three-dimensional point cloud is delivered to subsequent high-level applications such as reconstruction, segmentation and detection, the abnormal values in the three-dimensional point cloud need to be removed.
The traditional method for processing the abnormal value of the point cloud mainly utilizes the difference of a true value and the abnormal value in time sequence and directly detects and removes the abnormal value of the original point cloud by using characteristic descriptors such as density, projection, clustering and normal vector. However, due to the algorithm architecture, the point cloud processed by the conventional method cannot effectively balance the outlier removal rate and the detail feature retention rate. Although the point cloud processed by the deep learning-based method can solve the problems, the training of the network requires massive data and time support, the generalization capability is poor, and when the difference between the feature distribution of the input point cloud and the training data is large, the deep learning-based method is completely ineffective.
Disclosure of Invention
In view of the above, the present application provides a method for removing a point cloud abnormal value, a method for processing a point cloud, an apparatus and a related device, so as to remove the abnormal value of the point cloud while preserving detailed features.
In order to achieve the above object, a first aspect of the present application provides a method for removing a point cloud outlier, including:
uniformly dividing the point cloud to be processed into a plurality of layers along a reference coordinate axis;
setting an enclosing frame taking the point as a center for each point in each layer of the point cloud, and removing the point if the enclosing frame only contains the point to obtain a second point cloud;
determining a first adjacent point set of the points for each point in the second point cloud, wherein the number of first adjacent points contained in the first adjacent point set of each point is the same, and the Euclidean distance from the first adjacent points to the points is smaller than that from any point outside the first adjacent point set to the points;
obtaining neighbor Euclidean distances of each point in the second point cloud, and obtaining an average neighbor Euclidean distance of the second point cloud, wherein the neighbor Euclidean distance is an average value of the Euclidean distances of each first neighbor point of the point and the point, and the average neighbor Euclidean distance is an average value of the neighbor Euclidean distances of each point in the second point cloud;
for each point in the second point cloud, determining whether to remove the point based on a difference of the neighboring euclidean distance of the point and the average neighboring euclidean distance, and a standard deviation of the neighboring euclidean distance of each point in the second point cloud and the average neighboring euclidean distance.
Preferably, the process of uniformly dividing the point cloud to be processed into a plurality of layers along the reference coordinate axis includes:
the thickness value δ is calculated using the following equation:
wherein n is 1 As the total number of points of the point cloud, S xoy 、S yoz And S xoz Respectively the projection areas of the point cloud on the xoy plane, the yoz plane and the xoz plane;
based on the thickness value δ, the point cloud to be processed is uniformly divided into l layers along the Z-axis:
l＝|z max -z min |/δ
wherein z is max Is the maximum value of the point cloud in the Z axis, Z min And the minimum value of the point cloud in the Z axis is obtained.
Preferably, the process of setting the bounding box centered at the point comprises:
setting a square by taking the point as a center, determining the square as the surrounding frame, and calculating the side length xi of the square by the following equation:
wherein alpha is a preset bounding box threshold value, n 1 As the total number of points of the point cloud, S xoy 、S yoz And S xoz The projection areas of the point cloud on the xoy plane, the yoz plane and the xoz plane are respectively.
Preferably, for each point in the second point cloud, the process of determining a first set of neighboring points of the point comprises:
for each point in the second point cloud, calculating the Euclidean distance between the point and each other point in the second point cloud to obtain (n) 2 -1) distance values;
from the (n) 2 -1) the smaller of the determined values of distance k 1 A distance value of k 1 The distance values constitute a first set of neighbors of the point;
wherein n is 2 Is the total number of points, k, of the second point cloud 1 Is a preset natural number.
Preferably, for each point in the second point cloud, determining whether to remove the point based on a difference of the neighbor euclidean distance of the point and the average neighbor euclidean distance, and a standard deviation of the neighbor euclidean distance of each point in the second point cloud and the average neighbor euclidean distance, comprises:
judging the neighbor Euclidean distance D (i) of any point i in the second point cloud and the average neighbor Euclidean distance
Whether the difference satisfies the following equation:
if not, removing the point i, wherein sigma is a preset standard deviation threshold value, and n 2 And the total point number of the second point cloud is obtained.
A second aspect of the present application provides an apparatus for removing a point cloud abnormal value, including:
the point cloud layering unit is used for uniformly dividing the point cloud to be processed into a plurality of layers along a reference coordinate axis;
a first removing unit, configured to set, for each point in each layer of the point cloud, an enclosure frame centered around the point, and if only the point is included in the enclosure frame, remove the point to obtain a second point cloud;
a neighboring point determining unit, configured to determine, for each point in the second point cloud, a first neighboring point set of the points, where the number of first neighboring points included in the first neighboring point set of each point is the same, and an euclidean distance from the first neighboring point to the point is smaller than an euclidean distance from any point outside the first neighboring point set to the point;
a distance determining unit, configured to obtain neighboring euclidean distances of each point in the second point cloud, and obtain an average neighboring euclidean distance of the second point cloud, where the neighboring euclidean distance is an average of the euclidean distances of each first neighboring point to the point, and the average neighboring euclidean distance is an average of the neighboring euclidean distances of each point in the second point cloud;
a second removing unit configured to determine, for each point in the second point cloud, whether to remove the point based on a difference value of the neighboring euclidean distance of the point and the average neighboring euclidean distance, and a standard deviation of the neighboring euclidean distance of each point in the second point cloud and the average neighboring euclidean distance.
A third aspect of the present application provides a device for removing a point cloud outlier, including: a memory and a processor;
the memory is used for storing programs;
the processor is used for executing the program and realizing the steps of the method for removing the point cloud abnormal value.
A fourth aspect of the present application provides a storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method for removing a point cloud outlier as described above.
A fifth aspect of the present application provides a point cloud processing method, including:
removing the abnormal value of the point cloud to be processed by using the method for removing the abnormal value of the point cloud to obtain a roughly processed point cloud;
removing abnormal values of the roughly processed point cloud by using the method for removing the abnormal values of the point cloud to obtain incomplete point cloud;
calculating the difference value between the roughly processed point cloud and the incomplete point cloud by using a frame difference method to obtain a plurality of filling points;
acquiring a second adjacent point set of each filling point in the incomplete point cloud, wherein the number of second adjacent points contained in the second adjacent point set of each filling point is the same, and the Euclidean distance from the second adjacent points to the filling point is smaller than that from any point outside the second adjacent point set to the filling point;
acquiring a mean value and a covariance of each second neighbor point set, wherein the mean value of each second neighbor point set is a coordinate mean value of each second neighbor point in the second neighbor point set, and the covariance of each second neighbor point set is a covariance of each second neighbor point in the second neighbor point set;
and determining whether the filling points are merged into the incomplete point cloud or not based on the mean and covariance of the second neighboring point set of each filling point to obtain a processed incomplete point cloud, and outputting the processed incomplete point cloud.
Preferably, the process of determining whether to incorporate the filling point into the incomplete point cloud based on the mean and covariance of the second set of neighboring points for each filling point comprises:
the fill point f is calculated using the following equation i To curved surface Q i Mahalanobis distance M (i):
wherein, c i Is a filling point f i The average value of the second neighboring point set is calculated as follows:
cov(Q i ) Is a filling point f i The covariance of the second neighboring point set is calculated as follows:
wherein, the curved surface Q i Is a filling point f i Of the second set of neighboring points, k 2 Is the number of second neighbors in the second set of neighbors, (x) i,j ,y i,j ,z i,j ) Is a curved surface Q i J (th) point q i,j Three-dimensional coordinates of (a);
if M (i) is less than the preset threshold, filling the point f i Incorporated into the incomplete point cloud.
The sixth aspect of the present application provides a point cloud processing apparatus, including:
a first removing unit, configured to remove an abnormal value of the point cloud to be processed by using the method for removing an abnormal value of a point cloud according to any one of claims 1 to 5, so as to obtain a roughly processed point cloud;
a second removing unit, configured to remove the abnormal value of the roughly processed point cloud by using the method for removing the abnormal value of the point cloud according to any one of claims 1 to 5, so as to obtain an incomplete point cloud;
the first acquisition unit is used for calculating the difference value between the roughly processed point cloud and the incomplete point cloud by using a frame difference method to obtain a plurality of filling points;
a second obtaining unit, configured to obtain a second neighboring point set of each filling point in the incomplete point cloud, where the number of second neighboring points included in the second neighboring point set of each filling point is the same, and an euclidean distance between the second neighboring point and the filling point is smaller than an euclidean distance between any point outside the second neighboring point set and the filling point;
a third obtaining unit, configured to obtain a mean value and a covariance of each second neighboring point set, where the mean value of the second neighboring point set is a coordinate mean value of each second neighboring point in the second neighboring point set, and the covariance of the second neighboring point set is a covariance of each second neighboring point in the second neighboring point set;
and the filling unit is used for determining whether the filling points are merged into the incomplete point cloud or not based on the mean value and the covariance of the second adjacent point set of each filling point to obtain the processed incomplete point cloud and outputting the processed incomplete point cloud.
A seventh aspect of the present application provides a point cloud processing apparatus, including: a memory and a processor;
the memory is used for storing programs;
the processor is used for executing the program and realizing the steps of the point cloud processing method.
An eighth aspect of the present application provides a storage medium, on which a computer program is stored, which, when being executed by a processor, realizes the steps of the point cloud processing method as described above.
According to the technical scheme, the point cloud to be processed is uniformly divided into a plurality of layers along the reference coordinate axis. Then, for each point in each layer of the point cloud, an enclosing frame taking the point as a center is arranged, and if the enclosing frame only comprises the point, the point is removed, and a second point cloud is obtained. It is understood that each point in the object point cloud should exist continuously in a certain range, and when the enclosure box is set to be large enough, only discrete points exist in the enclosure box, so that the discrete points can be confirmed to be abnormal points, and the removing operation should be performed on the abnormal points. Then, for each point in the second point cloud, a first set of neighboring points for the point is determined. The number of the first neighbor points contained in the first neighbor point set of each point is the same, and the Euclidean distance from the first neighbor point to the point is smaller than the Euclidean distance from any point outside the first neighbor point set to the point. Next, neighbor euclidean distances of each point in the second point cloud are obtained, and an average neighbor euclidean distance of the second point cloud is obtained. Wherein the neighbor euclidean distance is an average of the euclidean distances between the point and each of the first neighbor points of the point, which reflects the average distance between a certain point in the point cloud and its first neighbor point; the average neighbor euclidean distance is the average of the neighbor euclidean distances of each point in the second point cloud, reflecting the overall average distance of each point in the point cloud to its first neighbor. Finally, for each point in the second point cloud, determining whether to remove the point based on a difference of the neighbor euclidean distance of the point and the average neighbor euclidean distance, and a standard deviation of the neighbor euclidean distance of each point in the second point cloud and the average neighbor euclidean distance. Through the steps, the method and the device can intuitively and conveniently remove the abnormal value in the point cloud on the premise of keeping the detail characteristics.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only the embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a schematic diagram of a method for removing a point cloud outlier disclosed in an embodiment of the present application;
FIG. 2 is a schematic diagram of a point cloud processing method disclosed in an embodiment of the present application;
FIG. 3 illustrates a schematic diagram of an original point cloud containing 30% outliers;
FIG. 4 illustrates an effect diagram of processing an original point cloud by an SOR method;
FIG. 5 illustrates an effect of processing an original point cloud by using a local density method;
FIG. 6 is a schematic diagram illustrating the effect of processing an original point cloud by the method of the present application;
FIG. 7 illustrates a schematic diagram of an original point cloud containing 50% outliers;
FIG. 8 illustrates an effect diagram of processing an original point cloud by an SOR method;
FIG. 9 illustrates an effect of processing an original point cloud by using a local density method;
FIG. 10 is a schematic diagram illustrating the effect of processing an original point cloud using the method of the present application;
FIG. 11 is a schematic diagram of an apparatus for removing a point cloud outlier disclosed in an embodiment of the present application;
FIG. 12 is a schematic diagram of a point cloud processing apparatus disclosed in an embodiment of the present application;
fig. 13 is a schematic diagram of a point cloud outlier removing apparatus and a point cloud processing apparatus disclosed in the embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The method for removing the point cloud abnormal value provided by the embodiment of the application is described below. The method for removing the point cloud Outlier is actually a method for removing a hierarchical Statistical Outlier (LSOR). Referring to fig. 1, a method for removing a point cloud outlier provided in an embodiment of the present application may include the following steps:
step S101, the point cloud to be processed is uniformly divided into a plurality of layers along a reference coordinate axis.
It is to be understood that the point cloud processed by the present application is a three-dimensional point cloud, and assuming that the coordinate system of the three-dimensional point cloud is XYZ coordinate system, and the reference coordinate axis is Z axis, then the three-dimensional point cloud can be divided into a plurality of layers parallel to the xoy plane along the Z axis, where o is the origin of the coordinate system.
Step S102, for each point in each layer of the point cloud, an enclosing frame taking the point as a center is arranged, if the enclosing frame only comprises the point, the point is removed, and a second point cloud is obtained.
The size of the bounding box can be set according to specific parameters and specific characteristics of the point cloud, if the bounding box is set to be large enough, when only one point is contained in the bounding box, the point is easy to know to be an abnormal point, and therefore the point can be removed. When all the outliers of the point cloud to be processed are removed, the resulting point cloud is called the second point cloud.
Step S103, for each point in the second point cloud, determining a first adjacent point set of the point.
The number of the first neighbor points included in the first neighbor point set of each point is the same. And for each point, the Euclidean distance from the first neighbor point to the point is smaller than the Euclidean distance from any point outside the first neighbor point set to the point.
Step S104, acquiring the neighbor Euclidean distance of each point in the second point cloud, and acquiring the average neighbor Euclidean distance of the second point cloud.
The neighbor euclidean distance is an average of euclidean distances of a certain point and first neighbor points of the point, that is, the neighbor euclidean distance of a certain point depends on the point and the first neighbor points of the point.
The average neighbor euclidean distance is the average of the neighbor euclidean distances of each point in the second point cloud.
Step S105, for each point in the second point cloud, whether the point is removed or not is determined based on the difference value between the neighbor Euclidean distance of the point and the average neighbor Euclidean distance and the standard deviation between the neighbor Euclidean distance of each point in the second point cloud and the average neighbor Euclidean distance.
The above-described embodiment of the present application first uniformly divides a point cloud to be processed into a plurality of layers along a reference coordinate axis. Then, for each point in each layer of the point cloud, an enclosing frame with the point as the center is arranged, and if the enclosing frame only contains the point, the point is removed to obtain a second point cloud. It is understood that each point in the object point cloud should exist continuously in a certain range, and when the bounding box is set to be large enough, only discrete points exist in the bounding box, so that it can be confirmed that the discrete points are abnormal points, and the removing operation should be performed on the abnormal points. Then, for each point in the second point cloud, a first set of neighboring points for the point is determined. The number of the first neighbor points included in the first neighbor point set of each point is the same, and the Euclidean distance from the first neighbor point to the point is smaller than that from any point outside the first neighbor point set to the point. Next, neighboring euclidean distances of each point in the second point cloud are obtained, and an average neighboring euclidean distance of the second point cloud is obtained. Wherein, the neighboring Euclidean distance is the average value of the Euclidean distance between the point and each first neighboring point of the point, and reflects the average distance from a certain point in the point cloud to the first neighboring point; the average neighbor euclidean distance is the average of the neighbor euclidean distances of each point in the second point cloud, which reflects the overall average distance of each point in the point cloud to its first neighbor. And finally, for each point in the second point cloud, determining whether to remove the point or not based on the difference value of the neighbor Euclidean distance of the point and the average neighbor Euclidean distance and the standard deviation of the neighbor Euclidean distance of each point in the second point cloud and the average neighbor Euclidean distance. Through the steps, the method and the device can intuitively and conveniently remove the abnormal value in the point cloud on the premise of keeping the detail characteristics.
In some embodiments of the present application, the step S101 of uniformly dividing the point cloud to be processed into a plurality of layers along the reference coordinate axis may include:
s1, calculating a thickness value delta by using the following equation:
wherein n is 1 As the total number of points of the point cloud, S xoy 、S yoz And S xoz The projection areas of the point cloud on the xoy plane, the yoz plane and the xoz plane can be calculated by the following equations:
wherein x is max Is the maximum value of the point cloud on the X axis, X min The minimum value of the point cloud on the X axis is obtained; y is max Is the maximum value of the point cloud in the Y axis, Y min The minimum value of the point cloud on the Y axis is obtained; z is a radical of formula max Is the maximum value of the point cloud in the Z axis, Z min The minimum value of the point cloud on the Z axis.
S2, uniformly dividing the point cloud to be processed into l layers along the Z axis based on the thickness value delta:
l＝|z max -z min |/δ
wherein z is max Is the maximum value of the point cloud in the Z axis, Z min The minimum value of the point cloud on the Z axis.
In some embodiments of the present application, the process of setting the bounding box centered on the point in step S102 may include:
a square is set with the point as the center, and the square is determined as a surrounding frame.
Wherein, the side length xi of the square is calculated by the following equation:
wherein alpha is a preset bounding box threshold value, n 1 Is the total number of points, s, of the point cloud xoy 、s yoz And s xoz The projection areas of the point cloud on the xoy plane, the yoz plane and the xoz plane are respectively.
It can be understood that the α value directly affects the determination of the abnormal point, and if the α value is set too large, there is a possibility that the determination is missed, that is, the point to be determined as the abnormal point is determined as the normal point; if the value α is too small, there is a possibility that a point which should not be determined as an abnormal point is determined as an abnormal point, which is a misjudgment.
Since the bounding box is a two-dimensional surface, and each layer in the point cloud is still a three-dimensional body even if the thickness value δ is smaller, when the bounding box is set as a square, the square is a square parallel to the xoy plane, and the step S102 may include, for each point in each layer of the point cloud, a determination process of whether only the point is included in the bounding box:
s1, acquiring a target point of the x value and the y value in the bounding box from the layer where the point is located.
S2, judging whether only one target point exists; if yes, determining that the bounding box only contains the point; if not, determining that the point is not contained in the bounding box.
In some embodiments of the present application, the step S103 of determining, for each point in the second point cloud, a first neighboring point set of the point may include:
s1, calculating the Euclidean distance between each point in the second point cloud and each other point in the second point cloud to obtain (n) 2 -1) distance values.
Wherein n is 2 The total number of points of the second point cloud.
S2, from (n) 2 -1) k of the determined distance values which is smaller 1 A distance value of from 1 The distance values constitute a first set of neighboring points of the point.
Wherein k is 1 Is a preset natural number. For example, the (n) can be adjusted 2 -1) sorting the distance values from small to large, and then sorting the k front 1 The distance values are determined as first neighbors of the point, from which first neighbors a first set of neighbors is formed.
In some embodiments of the present application, the step S105 of determining, for each point in the second point cloud, whether to remove the point based on a difference between a neighboring euclidean distance of the point and an average neighboring euclidean distance, and a standard deviation between the neighboring euclidean distance of each point in the second point cloud and the average neighboring euclidean distance, may include:
judging the neighbor Euclidean distance D (i) and the average neighbor Euclidean distance of any point i in the second point cloud
Whether the difference satisfies the following equation:
if not, removing the point i.
Wherein, sigma is a preset standard deviation threshold value, n 2 The average nearest neighbor Euclidean distance is the total point number of the second point cloud
Is a neighbor euclidean distance D (i) (i =1,2,. N 2 ) Is measured.
The neighbor euclidean distance D (i) of any point i can be calculated by using the following equation:
wherein, d i (j) The euclidean distance between the ith point and the jth first neighboring point in the second point cloud can be calculated by the following equation:
wherein (x) i ,y i ,z i ) (x) as the coordinates of the ith point in the second point cloud i ,y i ,z i ) The three-dimensional coordinates of the jth first neighbor point of the ith point in the second point cloud.
Based on the method for removing the point cloud abnormal value provided by the embodiments, the application also provides a point cloud processing method. The point cloud processing method provided by the embodiment of the application is described below. Referring to fig. 2, a point cloud processing method provided in an embodiment of the present application may include the following steps:
step S201, removing abnormal values of the point cloud to be processed by using a method for removing abnormal values of the point cloud to obtain a roughly processed point cloud.
And S202, removing abnormal values of the roughly processed point cloud by using a point cloud abnormal value removing method to obtain incomplete point cloud.
The method for removing the point cloud abnormal value mentioned in steps S201 and S202 is the method for removing the point cloud abnormal value provided in the above embodiments. It is understood that steps S202 and S201 use different parameters (α, k) when removing abnormal values from the point cloud 1 σ) or (α, k) 2 σ), and the parameters used in step S202 are more strict, so that the once processed roughly processed point cloud can be processed secondarily, and the outliers can be removed secondarily based on the roughly processed point cloud. Therefore, the parameters (α, k) used in step S201 may be set 1 σ) is determined as a relaxation parameter, which can typically take the value (5, 30, 2); the parameters (α, k) used in step S202 may be set 2 σ) is determined as a severity parameter, which can typically take the value (3,30,1).
Step S203, calculating a difference between the roughly processed point cloud and the incomplete point cloud by using a frame difference method to obtain a plurality of filling points.
The process of calculating the difference value between the roughly processed point cloud and the incomplete point cloud by using a frame difference method comprises the following steps:
s1, comparing the roughly processed point cloud with the incomplete point cloud to obtain points with the same positions;
s2, acquiring points with different positions based on the points with the same positions;
and S3, confirming the points at different positions as difference values.
Exemplarily, it is assumed that the point cloud to be processed in step S201 is P = [ P = 1 ,p 2 ,…,p N ]∈R N×3 The point cloud obtained after the processing in step S201 is P' = [ P = 1 ,p 2 ,…,p N′ ]∈R N′×3 The incomplete point cloud obtained after the processing of step S202 is P = [ P ″ ] 1 ,p 2 ,…,p N″ ]∈R N″×3 . Then, the filling point obtained after the processing in step S203 is F = [ F = 1 ,f 2 ,…,f m ]∈R m×3 Wherein m = N' -N ".
Step S204, a second neighboring point set of each filling point in the incomplete point cloud is obtained.
The number of the second neighboring points contained in the second neighboring point set of each filling point is the same, and the euclidean distance from the second neighboring point to the filling point is smaller than the euclidean distance from any point outside the second neighboring point set to the filling point.
Step S205, obtaining a mean and a covariance of each second neighboring point set.
The mean value of the second neighboring point set is the coordinate mean value of each second neighboring point in the second neighboring point set, and the covariance of the second neighboring point set is the covariance of each second neighboring point in the second neighboring point set.
For example, for fill point f i The second neighbor set is Q i ＝{q ij ,j＝1,2,…,k 2 Then, the second set of neighbors Q i Mean value of c i Can be calculated from the following equation:
second oneSet of neighbors Q i Covariance of (Q) i ) Can be calculated from the following equation:
wherein k is 2 Is the number of second neighbors in the second set of neighbors, (x) i,j ,y i,j ,z i,j ) Is a curved surface Q i J (th) point q i,j Three-dimensional coordinates of (a).
Step S206, determining whether the filling point is merged into the incomplete point cloud or not based on the mean and covariance of the second neighboring point set of each filling point to obtain a processed incomplete point cloud, and outputting the processed incomplete point cloud.
In some embodiments of the present application, the step S206 of determining whether to incorporate the filling point into the incomplete point cloud based on the mean and covariance of the second neighboring point set of each filling point may include:
s1, calculating a filling point f by using the following equation i To curved surface Q i Mahalanobis distance M (i):
wherein, c i Is a filling point f i Of the second set of neighbors, cov (Q) i ) Is a filling point f i Of the second set of neighbors.
S2, if the M (i) is smaller than a preset threshold value, filling the point f i Incorporated into the incomplete point cloud.
In order to verify the accuracy of the point cloud processing method provided by the embodiment of the application, the accuracy is proved through a public data set test experiment. The simulation results of the present invention are shown in fig. 3 to 10, respectively. Therein, fig. 3 shows a visualization of an original point cloud containing 30% outliers. Fig. 4 to 6 show the processing results of the original point cloud by three outlier processing methods. Fig. 4 and 5 are processing results of the SOR method and the local density method, respectively, and it can be seen that a part of the abnormal values around the real point is not removed. Fig. 6 is a processing result of the method of the present invention, and it can be seen that for the point cloud containing 30% of abnormal values, the processing result of the method of the present invention contains almost no abnormal values, and the detail features of the point cloud are completely retained.
Fig. 7 shows the visualization of the original point cloud containing 50% outliers. Fig. 8 to 10 show the processing results of the original point cloud by three outlier processing methods. Fig. 8 shows the processing result of the SOR method, and it can be seen that the SOR method cannot complete the outlier removal operation when the outlier proportion is increased to 50%. Fig. 9 shows the processing result of the local density method, and it can be seen that the local density method can effectively process the point cloud with a high outlier proportion, but some detail features of the target (such as rabbit ear portions) have a missing problem. Fig. 10 is a processing result of the method of the present invention, and it can be seen that for a point cloud containing 50% outliers, the method of the present invention can still obtain excellent outlier removal and detail feature retention results.
The device for removing the point cloud abnormal value provided by the embodiment of the present application is described below, and the device for removing the point cloud abnormal value described below and the method for removing the point cloud abnormal value described above may be referred to in correspondence with each other.
Referring to fig. 11, the apparatus for removing a point cloud outlier provided in the embodiment of the present application may include:
a point cloud layering unit 21, configured to uniformly divide a point cloud to be processed into multiple layers along a reference coordinate axis;
a first removing unit 22, configured to set, for each point in each layer of the point cloud, an enclosure frame centered on the point, and if only the point is included in the enclosure frame, remove the point to obtain a second point cloud;
a neighboring point determining unit 23, configured to determine, for each point in the second point cloud, a first neighboring point set of the points, where the number of first neighboring points included in the first neighboring point set of each point is the same, and an euclidean distance between the first neighboring point and the point is smaller than an euclidean distance between any point outside the first neighboring point set and the point;
a distance determining unit 24, configured to obtain neighboring euclidean distances of each point in the second point cloud, and obtain an average neighboring euclidean distance of the second point cloud, where the neighboring euclidean distance is an average of the euclidean distances of each first neighboring point to the point, and the average neighboring euclidean distance is an average of the neighboring euclidean distances of each point in the second point cloud;
a second removing unit 25 for determining, for each point in the second point cloud, whether to remove the point based on a difference value of the neighboring euclidean distance of the point and the average neighboring euclidean distance, and a standard deviation of the neighboring euclidean distance of each point in the second point cloud and the average neighboring euclidean distance.
In some embodiments of the present application, the process of the point cloud layering unit 21 uniformly dividing the point cloud to be processed into a plurality of layers along the reference coordinate axis may include:
the thickness value δ is calculated using the following equation:
wherein n is 1 As the total number of points of the point cloud, S xoy 、S yoz And S xoz Respectively the projection areas of the point cloud on the xoy plane, the yoz plane and the xoz plane;
based on the thickness value δ, the point cloud to be processed is uniformly divided into l layers along the Z-axis:
l＝|z max -z min |/δ
wherein z is max Is the maximum value of the point cloud in the Z axis, Z min And the minimum value of the point cloud in the Z axis is obtained.
In some embodiments of the present application, the process of the first removing unit 22 setting the bounding box centered on the point may include:
setting a square by taking the point as a center, determining the square as the surrounding frame, and calculating the side length ξ of the square by the following equation:
wherein alpha is a preset bounding box threshold value, n 1 As the total number of points of the point cloud, S xoy 、S yoz And S xoz The projection areas of the point cloud on the xoy plane, the yoz plane and the xoz plane are respectively.
In some embodiments of the present application, the process of determining, for each point in the second point cloud, a first neighboring point set of the point by the neighboring point determining unit 23 may include:
for each point in the second point cloud, calculating the Euclidean distance between the point and each other point in the second point cloud to obtain (n) 2 -1) distance values;
from the (n) 2 -1) the smaller of the determined values of distance k 1 A distance value of k 1 The distance values constitute a first set of neighboring points of said point;
wherein n is 2 Is the total number of points, k, of the second point cloud 1 Is a preset natural number.
In some embodiments of the present application, the determining whether to remove the point based on the difference between the neighbor euclidean distance of the point and the average neighbor euclidean distance and the standard deviation between the neighbor euclidean distance of each point in the second point cloud and the average neighbor euclidean distance by the second removing unit 25 may include, for each point in the second point cloud:
determining the neighbor Euclidean distance D (i) of any point i in the second point cloud and the average neighbor Euclidean distance
Whether the difference satisfies the following equation:
if not, removing the point i, wherein sigma is a preset standard deviation threshold value, and n 2 The total point number of the second point cloud is obtained.
The device for removing the point cloud abnormal value provided by the embodiment of the application can be applied to equipment, such as a computer, for removing the point cloud abnormal value. Alternatively, fig. 13 is a block diagram showing a hardware configuration of a point cloud abnormal value removing apparatus, and referring to fig. 13, the hardware configuration of the point cloud abnormal value removing apparatus may include: at least one processor 31, at least one communication interface 32, at least one memory 33 and at least one communication bus 34.
In the embodiment of the present application, the number of the processor 31, the communication interface 32, the memory 33 and the communication bus 34 is at least one, and the processor 31, the communication interface 32 and the memory 33 complete the communication with each other through the communication bus 34;
the processor 31 may be a central processing unit CPU, or an Application Specific Integrated Circuit ASIC (Application Specific Integrated Circuit), or one or more Integrated circuits configured to implement the embodiments of the present Application, etc.;
the memory 32 may comprise a high-speed RAM memory, and may further comprise a non-volatile memory (non-volatile memory) or the like, such as at least one disk memory;
wherein the memory 33 stores a program and the processor 31 may invoke the program stored in the memory 33, the program being for:
uniformly dividing a point cloud to be processed into a plurality of layers along a reference coordinate axis;
for each point in each layer of the point cloud, setting a surrounding frame taking the point as a center, and if only the point is contained in the surrounding frame, removing the point to obtain a second point cloud;
for each point in the second point cloud, determining a first neighbor point set of the point, wherein the number of first neighbor points contained in the first neighbor point set of each point is the same, and the Euclidean distance from the first neighbor point to the point is smaller than that from any point outside the first neighbor point set to the point;
obtaining neighbor Euclidean distances of each point in the second point cloud, and obtaining an average neighbor Euclidean distance of the second point cloud, wherein the neighbor Euclidean distance is an average value of the Euclidean distances of each first neighbor point of the point and the point, and the average neighbor Euclidean distance is an average value of the neighbor Euclidean distances of each point in the second point cloud;
for each point in the second point cloud, determining whether to remove the point based on a difference of the neighboring euclidean distance of the point and the average neighboring euclidean distance, and a standard deviation of the neighboring euclidean distance of each point in the second point cloud and the average neighboring euclidean distance.
Alternatively, the detailed function and the extended function of the program may refer to the above description.
An embodiment of the present application further provides a storage medium, where the storage medium may store a program adapted to be executed by a processor, where the program is configured to:
uniformly dividing the point cloud to be processed into a plurality of layers along a reference coordinate axis;
setting an enclosing frame taking the point as a center for each point in each layer of the point cloud, and removing the point if the enclosing frame only contains the point to obtain a second point cloud;
determining a first adjacent point set of the points for each point in the second point cloud, wherein the number of first adjacent points contained in the first adjacent point set of each point is the same, and the Euclidean distance from the first adjacent points to the points is smaller than that from any point outside the first adjacent point set to the points;
obtaining neighboring euclidean distances of each point in the second point cloud, and obtaining an average neighboring euclidean distance of the second point cloud, wherein the neighboring euclidean distances are an average of the euclidean distances of each first neighboring point to the point, and the average neighboring euclidean distance is an average of the neighboring euclidean distances of each point in the second point cloud;
for each point in the second point cloud, determining whether to remove the point based on a difference of the neighboring euclidean distance of the point and the average neighboring euclidean distance, and a standard deviation of the neighboring euclidean distance of each point in the second point cloud and the average neighboring euclidean distance.
Alternatively, the detailed function and the extended function of the program may be as described above.
The point cloud processing apparatus provided in the embodiments of the present application is described below, and the point cloud processing apparatus described below and the point cloud processing method described above may be referred to in correspondence with each other.
Referring to fig. 12, a point cloud processing apparatus provided in an embodiment of the present application may include:
a first removing unit 41, configured to remove an abnormal value of the point cloud to be processed by using the method for removing an abnormal value of a point cloud provided in each of the above embodiments, so as to obtain a roughly processed point cloud;
a second removing unit 42, configured to remove an abnormal value of the roughly processed point cloud by using the method for removing an abnormal value of a point cloud provided in each of the above embodiments, so as to obtain an incomplete point cloud;
a first obtaining unit 43, configured to calculate a difference between the roughly processed point cloud and the incomplete point cloud by using a frame difference method, so as to obtain a plurality of filling points;
a second obtaining unit 44, configured to obtain a second neighboring point set of each filling point in the incomplete point cloud, where the number of second neighboring points included in the second neighboring point set of each filling point is the same, and an euclidean distance between the second neighboring point and the filling point is smaller than an euclidean distance between any point outside the second neighboring point set and the filling point;
a third obtaining unit 45, configured to obtain a mean value and a covariance of each second neighboring point set, where the mean value of the second neighboring point set is a coordinate mean value of each second neighboring point in the second neighboring point set, and the covariance of the second neighboring point set is a covariance of each second neighboring point in the second neighboring point set;
and a filling unit 46, configured to determine whether to incorporate the filling points into the incomplete point cloud based on a mean and a covariance of the second neighboring point set of each filling point, to obtain a processed incomplete point cloud, and output the processed incomplete point cloud.
In some embodiments of the present application, the process of the filling unit 46 determining whether to incorporate the filling point into the incomplete point cloud based on the mean and covariance of the second set of neighboring points of each filling point may include:
the fill point f is calculated using the following equation i To curved surface Q i Mahalanobis distance M (i):
wherein, c i Is a filling point f i The average value of the second neighboring point set is calculated as follows:
cov(Q i ) Is a filling point f i The covariance of the second neighboring point set is calculated as follows:
wherein, the curved surface Q i Is a filling point f i Of the second set of neighboring points, k 2 Is the number of second neighbors in the second set of neighbors, (x) i,j ,y i,j ,z i,j ) Is a curved surface Q i Point j of (c) i,j Three-dimensional coordinates of (a);
if M (i) is less than the preset threshold, filling the point f i Incorporated into the incomplete point cloud.
The point cloud processing device provided by the embodiment of the application can be applied to point cloud processing equipment such as a computer. Alternatively, fig. 13 shows a block diagram of a hardware structure of the point cloud processing apparatus, and referring to fig. 13, the hardware structure of the point cloud processing apparatus may include: at least one processor 31, at least one communication interface 32, at least one memory 33 and at least one communication bus 34.
In the embodiment of the present application, the number of the processor 31, the communication interface 32, the memory 33 and the communication bus 34 is at least one, and the processor 31, the communication interface 32 and the memory 33 complete the communication with each other through the communication bus 34;
the processor 31 may be a central processing unit CPU, or an Application Specific Integrated Circuit ASIC (Application Specific Integrated Circuit), or one or more Integrated circuits configured to implement the embodiments of the present Application, etc.;
the memory 32 may comprise a high-speed RAM memory, and may further comprise a non-volatile memory (non-volatile memory) or the like, such as at least one disk memory;
wherein the memory 33 stores a program and the processor 31 may invoke the program stored in the memory 33, the program being for:
removing the abnormal value of the point cloud to be processed by using the method for removing the abnormal value of the point cloud provided by each embodiment to obtain a roughly processed point cloud;
removing abnormal values of the roughly processed point cloud by using the method for removing the abnormal values of the point cloud provided by each embodiment to obtain incomplete point cloud;
calculating the difference value between the roughly processed point cloud and the incomplete point cloud by using a frame difference method to obtain a plurality of filling points;
acquiring a second adjacent point set of each filling point in the incomplete point cloud, wherein the number of second adjacent points contained in the second adjacent point set of each filling point is the same, and the Euclidean distance from the second adjacent points to the filling point is smaller than that from any point outside the second adjacent point set to the filling point;
acquiring a mean value and a covariance of each second neighbor point set, wherein the mean value of each second neighbor point set is a coordinate mean value of each second neighbor point in the second neighbor point set, and the covariance of each second neighbor point set is a covariance of each second neighbor point in the second neighbor point set;
and determining whether the filling points are merged into the incomplete point cloud or not based on the mean and covariance of the second neighboring point set of each filling point to obtain a processed incomplete point cloud, and outputting the processed incomplete point cloud.
Alternatively, the detailed function and the extended function of the program may be as described above.
Embodiments of the present application further provide a storage medium, where a program suitable for execution by a processor may be stored, where the program is configured to:
removing abnormal values of the point cloud to be processed by using the method for removing the abnormal values of the point cloud provided by each embodiment to obtain a roughly processed point cloud;
removing abnormal values of the roughly processed point cloud by using the method for removing the abnormal values of the point cloud provided by each embodiment to obtain incomplete point cloud;
calculating the difference value between the roughly processed point cloud and the incomplete point cloud by using a frame difference method to obtain a plurality of filling points;
acquiring a second adjacent point set of each filling point in the incomplete point cloud, wherein the number of second adjacent points contained in the second adjacent point set of each filling point is the same, and the Euclidean distance from the second adjacent points to the filling point is smaller than that from any point outside the second adjacent point set to the filling point;
acquiring a mean value and a covariance of each second neighbor point set, wherein the mean value of each second neighbor point set is a coordinate mean value of each second neighbor point in the second neighbor point set, and the covariance of each second neighbor point set is a covariance of each second neighbor point in the second neighbor point set;
and determining whether the filling points are merged into the incomplete point cloud or not based on the mean and covariance of the second neighboring point set of each filling point to obtain a processed incomplete point cloud, and outputting the processed incomplete point cloud.
Alternatively, the detailed function and the extended function of the program may be as described above.
In summary, the following steps:
the method comprises the steps of firstly, uniformly dividing a point cloud to be processed into a plurality of layers along a reference coordinate axis. Then, for each point in each layer of the point cloud, an enclosing frame with the point as a center is arranged, and if the enclosing frame only contains the point, the point is removed, and a second point cloud is obtained. It is understood that each point in the object point cloud should exist continuously within a certain range, and when the surrounding frame is set to be large enough, only discrete points exist in the surrounding frame, so that the discrete points can be confirmed to be abnormal points, and the removing operation should be performed on the discrete points. Then, for each point in the second point cloud, a first set of neighboring points for the point is determined. The number of the first neighbor points contained in the first neighbor point set of each point is the same, and the Euclidean distance from the first neighbor point to the point is smaller than the Euclidean distance from any point outside the first neighbor point set to the point. Next, neighboring euclidean distances of each point in the second point cloud are obtained, and an average neighboring euclidean distance of the second point cloud is obtained. Wherein the neighbor euclidean distance is an average of the euclidean distances between the point and each of the first neighbor points of the point, which reflects the average distance between a certain point in the point cloud and its first neighbor point; the average neighbor euclidean distance is the average of the neighbor euclidean distances of each point in the second point cloud, reflecting the overall average distance of each point in the point cloud to its first neighbor. Finally, for each point in the second point cloud, determining whether to remove the point based on a difference of the neighboring euclidean distance of the point and the average neighboring euclidean distance, and a standard deviation of the neighboring euclidean distance of each point in the second point cloud and the average neighboring euclidean distance. Through the steps, the removal of the abnormal value in the point cloud can be achieved visually and conveniently on the premise that the detail characteristics are reserved.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising one of 8230; \8230;" 8230; "does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises the element.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, the embodiments may be combined as needed, and the same and similar parts may be referred to each other.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A method for removing a point cloud abnormal value is characterized by comprising the following steps:
uniformly dividing the point cloud to be processed into a plurality of layers along a reference coordinate axis;
setting an enclosing frame taking the point as a center for each point in each layer of the point cloud, and removing the point if the enclosing frame only contains the point to obtain a second point cloud;
determining a first adjacent point set of the points for each point in the second point cloud, wherein the number of first adjacent points contained in the first adjacent point set of each point is the same, and the Euclidean distance from the first adjacent points to the points is smaller than that from any point outside the first adjacent point set to the points;
obtaining neighboring euclidean distances of each point in the second point cloud, and obtaining an average neighboring euclidean distance of the second point cloud, wherein the neighboring euclidean distances are an average of the euclidean distances of each first neighboring point to the point, and the average neighboring euclidean distance is an average of the neighboring euclidean distances of each point in the second point cloud;
for each point in the second point cloud, determining whether to remove the point based on a difference of the neighbor euclidean distance of the point and the average neighbor euclidean distance, and a standard deviation of the neighbor euclidean distance of each point in the second point cloud and the average neighbor euclidean distance.
2. The method of claim 1, wherein the process of uniformly dividing the point cloud to be processed into a plurality of layers along a reference coordinate axis comprises:
the thickness value δ is calculated using the following equation:
wherein n is 1 As the total number of points of the point cloud, S xoy 、S yoz And S xoz Respectively the projection areas of the point cloud on the xoy plane, the yoz plane and the xoz plane;
based on the thickness value δ, the point cloud to be processed is uniformly divided into l layers along the Z-axis:
l＝|z max -z min |/δ
wherein z is max Is the maximum value of the point cloud in the Z axis, Z min And the minimum value of the point cloud in the Z axis is obtained.
3. The method of claim 2, wherein the step of setting the bounding box centered on the point comprises:
setting a square by taking the point as a center, determining the square as the surrounding frame, and calculating the side length xi of the square by the following equation:
wherein alpha is a preset bounding box threshold value, n 1 As the total number of points of the point cloud, S xoy 、S yoz And S xoz The projection areas of the point cloud on the xoy plane, the yoz plane and the xoz plane are respectively.
4. The method of claim 3, wherein determining, for each point in the second point cloud, a first set of neighboring points for the point comprises:
for each point in the second point cloud, calculating the Euclidean distance between the point and each other point in the second point cloud to obtain (n) 2 -1) distance values;
from the (n) 2 -1) the smaller of the determined values of distance k 1 A distance value of k 1 The distance values constitute a first set of neighboring points of said point;
wherein n is 2 Is the total number of points, k, of the second point cloud 1 Is a preset natural number.
5. The method of claim 4, wherein determining, for each point in the second point cloud, whether to remove the point based on a difference of the neighbor Euclidean distance of the point and the average neighbor Euclidean distance and a standard deviation of the neighbor Euclidean distance of each point in the second point cloud and the average neighbor Euclidean distance comprises:
judging the neighbor Euclidean distance D (i) of any point i in the second point cloud and the average neighbor Euclidean distance
Whether the difference satisfies the following equation:
if not, removing the point i, wherein sigma is a preset standard deviation threshold value, and n 2 The total point number of the second point cloud is obtained.
6. A point cloud processing method, comprising:
removing abnormal values of the point cloud to be processed by using the method for removing abnormal values of the point cloud according to any one of claims 1 to 5 to obtain a roughly processed point cloud;
removing the abnormal value of the roughly processed point cloud by using the method for removing the abnormal value of the point cloud according to any one of claims 1 to 5 to obtain an incomplete point cloud;
calculating the difference value between the roughly processed point cloud and the incomplete point cloud by using a frame difference method to obtain a plurality of filling points;
acquiring a second adjacent point set of each filling point in the incomplete point cloud, wherein the number of second adjacent points contained in the second adjacent point set of each filling point is the same, and the Euclidean distance from the second adjacent points to the filling point is smaller than that from any point outside the second adjacent point set to the filling point;
acquiring the mean value and covariance of each second neighbor point set, wherein the mean value of each second neighbor point set is the coordinate mean value of each second neighbor point in the second neighbor point set, and the covariance of each second neighbor point set is the covariance of each second neighbor point in the second neighbor point set;
and determining whether the filling points are merged into the incomplete point cloud or not based on the mean and covariance of the second neighboring point set of each filling point to obtain a processed incomplete point cloud, and outputting the processed incomplete point cloud.
7. The method of claim 6, wherein the determining whether to incorporate the filled point into the incomplete point cloud based on a mean and a covariance of the second set of neighbor points for each filled point comprises:
the fill point f is calculated using the following equation i To curved surface Q i Mahalanobis distance M (i):
wherein, c i Is a filling point f i The average value of the second neighboring point set is calculated as follows:
cov(Q i ) Is a filling point f i The covariance of the second neighboring point set is calculated as follows:
wherein, the curved surface Q i Is a filling point f i Of the second set of neighboring points, k 2 Is the number of second neighbors in the second set of neighbors, (x) i,j ,y i,j ,z i,j ) Is a curved surface Q i Point j of (c) i,j Three-dimensional coordinates of (a);
if M (i) is less than the preset threshold, filling the point f i Incorporated into the incomplete point cloud.
8. A device for removing a point cloud abnormal value is characterized by comprising:
the point cloud layering unit is used for uniformly dividing the point cloud to be processed into a plurality of layers along a reference coordinate axis;
a first removing unit, configured to set, for each point in each layer of the point cloud, an enclosure frame centered on the point, and if only the point is included in the enclosure frame, remove the point to obtain a second point cloud;
a neighbor point determining unit, configured to determine, for each point in the second point cloud, a first neighbor point set of the points, where the number of first neighbor points included in the first neighbor point set of each point is the same, and an euclidean distance between the first neighbor point and the point is smaller than an euclidean distance between any point outside the first neighbor point set and the point;
a distance determining unit, configured to obtain neighboring euclidean distances of each point in the second point cloud, and obtain an average neighboring euclidean distance of the second point cloud, where the neighboring euclidean distance is an average of the euclidean distances of each first neighboring point to the point, and the average neighboring euclidean distance is an average of the neighboring euclidean distances of each point in the second point cloud;
a second removing unit configured to determine, for each point in the second point cloud, whether to remove the point based on a difference value of the neighboring euclidean distance of the point and the average neighboring euclidean distance, and a standard deviation of the neighboring euclidean distance of each point in the second point cloud and the average neighboring euclidean distance.
9. A point cloud processing apparatus, comprising:
a first removing unit, configured to remove the abnormal value of the point cloud to be processed by using the method for removing the abnormal value of the point cloud according to any one of claims 1 to 5, so as to obtain a roughly processed point cloud;
a second removing unit, configured to remove the abnormal value of the roughly processed point cloud by using the method for removing the abnormal value of the point cloud according to any one of claims 1 to 5, so as to obtain an incomplete point cloud;
the first acquisition unit is used for calculating the difference value between the roughly processed point cloud and the incomplete point cloud by using a frame difference method to obtain a plurality of filling points;
a second obtaining unit, configured to obtain a second neighboring point set of each filling point in the incomplete point cloud, where the number of second neighboring points included in the second neighboring point set of each filling point is the same, and an euclidean distance between the second neighboring point and the filling point is smaller than an euclidean distance between any point outside the second neighboring point set and the filling point;
a third obtaining unit, configured to obtain a mean value and a covariance of each second neighboring point set, where the mean value of the second neighboring point set is a coordinate mean value of each second neighboring point in the second neighboring point set, and the covariance of the second neighboring point set is a covariance of each second neighboring point in the second neighboring point set;
and the filling unit is used for determining whether the filling points are merged into the incomplete point cloud or not based on the mean value and the covariance of the second adjacent point set of each filling point to obtain the processed incomplete point cloud and outputting the processed incomplete point cloud.
10. A point cloud processing apparatus, comprising: a memory and a processor;
the memory is used for storing programs;
the processor is used for executing the program and realizing the method for removing the point cloud abnormal value according to any one of claims 1 to 5 or the steps of the point cloud processing method according to claim 6 or 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210983097.7A CN115661421A (en) | 2022-08-16 | 2022-08-16 | Method for removing abnormal point cloud values, method and device for processing point cloud and related equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210983097.7A CN115661421A (en) | 2022-08-16 | 2022-08-16 | Method for removing abnormal point cloud values, method and device for processing point cloud and related equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115661421A true CN115661421A (en) | 2023-01-31 |
Family
ID=85024157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210983097.7A Pending CN115661421A (en) | 2022-08-16 | 2022-08-16 | Method for removing abnormal point cloud values, method and device for processing point cloud and related equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115661421A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116882035A (en) * | 2023-09-07 | 2023-10-13 | 湖南省国土资源规划院 | Space object recognition and modeling method based on artificial intelligence and related equipment |
-
2022
- 2022-08-16 CN CN202210983097.7A patent/CN115661421A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116882035A (en) * | 2023-09-07 | 2023-10-13 | 湖南省国土资源规划院 | Space object recognition and modeling method based on artificial intelligence and related equipment |
| CN116882035B (en) * | 2023-09-07 | 2023-11-21 | 湖南省国土资源规划院 | Space object recognition and modeling method based on artificial intelligence and related equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Nelson | Finding line segments by stick growing | |
| US8477147B2 (en) | Methods and systems of comparing face models for recognition | |
| CN107851332B (en) | Consistent subdivision via topology aware surface tracking | |
| CN112927353A (en) | Three-dimensional scene reconstruction method based on two-dimensional target detection and model alignment, storage medium and terminal | |
| Song et al. | Volumetric stereo and silhouette fusion for image-based modeling | |
| CN109767391A (en) | Point cloud denoising method, image processing equipment and the device with store function | |
| CN115661421A (en) | Method for removing abnormal point cloud values, method and device for processing point cloud and related equipment | |
| CN112258474A (en) | Wall surface anomaly detection method and device | |
| CN112419460A (en) | Method, apparatus, computer device and storage medium for baking model charting | |
| CN110147460B (en) | Three-dimensional model retrieval method and device based on convolutional neural network and multi-view map | |
| CN114842213A (en) | Obstacle contour detection method and device, terminal equipment and storage medium | |
| CN113436223B (en) | Point cloud data segmentation method and device, computer equipment and storage medium | |
| US11816857B2 (en) | Methods and apparatus for generating point cloud histograms | |
| Liu | Robust segmentation of raw point clouds into consistent surfaces | |
| Dimiccoli et al. | Exploiting t-junctions for depth segregation in single images | |
| CN113379826A (en) | Method and device for measuring volume of logistics piece | |
| Kordelas et al. | Viewpoint independent object recognition in cluttered scenes exploiting ray-triangle intersection and SIFT algorithms | |
| CN114219958B (en) | Multi-view remote sensing image classification method, device, equipment and storage medium | |
| CN113610971B (en) | Fine-grained three-dimensional model construction method and device and electronic equipment | |
| CN113744416B (en) | Global point cloud filtering method, equipment and storage medium based on mask | |
| Dimiccoli et al. | Hierarchical region-based representation for segmentation and filtering with depth in single images | |
| CN115131384A (en) | Bionic robot 3D printing method, device and medium based on edge preservation | |
| CN113111741A (en) | Assembly state identification method based on three-dimensional feature points | |
| Lin et al. | Mesh segmentation by local depth | |
| Avatavului et al. | A Hierarchical Cluster Tree Approach Leveraging Delaunay Triangulation |
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 |