CN112651889A - Fusion filtering method suitable for SLAM point cloud denoising, electronic device and storage medium - Google Patents

Fusion filtering method suitable for SLAM point cloud denoising, electronic device and storage medium Download PDF

Info

Publication number
CN112651889A
CN112651889A CN202011495165.2A CN202011495165A CN112651889A CN 112651889 A CN112651889 A CN 112651889A CN 202011495165 A CN202011495165 A CN 202011495165A CN 112651889 A CN112651889 A CN 112651889A
Authority
CN
China
Prior art keywords
point cloud
point
grid
value
points
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
Application number
CN202011495165.2A
Other languages
Chinese (zh)
Inventor
王坚
谢海成
章小明
陈润华
徐昀鹏
宁振伟
范铀
刘小芬
段涛
于淑君
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
South Digital Technology Co ltd
Original Assignee
South Digital Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by South Digital Technology Co ltd filed Critical South Digital Technology Co ltd
Priority to CN202011495165.2A priority Critical patent/CN112651889A/en
Publication of CN112651889A publication Critical patent/CN112651889A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10028—Range image; Depth image; 3D point clouds

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)

Abstract

The invention provides a fusion filtering method suitable for SLAM point cloud denoising, which comprises the following steps: counting the frequency of the point cloud in each elevation interval, and filtering extreme value noise points outside the point cloud main body; determining the number of grids divided in space and establishing a topological relation of point clouds; calculating the number of point clouds in each grid, and if the number of point clouds in each grid is smaller than a threshold value, determining that the point clouds in the grid are discrete noise, and deleting; establishing a K neighborhood of the point cloud through a space grid, and simultaneously removing isolated cluster noise points; and removing the tiny noise attached to the point cloud main body through a bilateral filter to obtain a final denoising result. The invention relates to an electronic device and a storage medium, which are used for executing a fusion filtering method suitable for SLAM point cloud denoising. The method has good denoising effect on the SLAM point cloud with wide noise space distribution and different scales, has higher calculation efficiency, and greatly improves the efficiency and effect of preprocessing the SLAM point cloud data.

Description

Fusion filtering method suitable for SLAM point cloud denoising, electronic device and storage medium
Technical Field
The invention relates to the technical field of modern surveying and mapping data processing, in particular to a fusion filtering method, electronic equipment and a storage medium suitable for SLAM point cloud denoising.
Background
A mobile laser scanning system based on an SLAM (synchronous positioning and mapping) technology is an important means for acquiring 3D point cloud data at present, but due to the influences of equipment, environment, human factors and matching errors, various noise points inevitably exist in the SLAM point cloud data, the visual impression of the three-dimensional point cloud is influenced by the existence of the noise points, and the accuracy of subsequent point cloud surface reconstruction is also directly influenced, so that noise removal is an essential link in SLAM point cloud data processing.
SLAM point cloud has the characteristics of large data volume, wide noise space distribution and inconsistent scale. When removing SLAM point cloud data, the existing denoising means such as K neighborhood statistical filtering, radius filtering, mean filtering, median filtering, Gaussian filtering and the like often have the problems of large calculated amount, high time consumption, low efficiency, lost characteristics and poor denoising effect, and at the moment, a fusion filtering method needs to be used for removing the noise of the SLAM point cloud data according to the characteristics of the SLAM point cloud data.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a fusion filtering method suitable for SLAM point cloud denoising, and solves the problems of large calculated amount, high time consumption, low efficiency, lost characteristics and poor denoising effect of the conventional point cloud denoising method.
The invention provides a fusion filtering method suitable for SLAM point cloud denoising, which comprises the following steps:
counting the frequency of the point cloud, counting the frequency of the point cloud in each elevation interval, and filtering extreme value noise points outside the point cloud main body;
establishing a topological relation, determining the number of grids divided in space, and establishing a topological relation of point cloud;
removing discrete noise, calculating the number of point clouds in each grid, and if the number of point clouds in each grid is smaller than a threshold value, determining the point clouds in the grid as the discrete noise and deleting the point clouds;
establishing a K neighborhood, establishing the K neighborhood of the point cloud through a space grid, and simultaneously removing isolated cluster noise points;
smoothing the fine noise, and removing the fine noise attached to the point cloud main body through a bilateral filter to obtain a final denoising result.
Further, the step of counting the point cloud frequency comprises the following steps:
acquiring information, reading point cloud, inputting elevation interval and judging a threshold value;
counting the quantity, namely counting the quantity of the point clouds in each elevation interval according to the Z coordinate value of the point clouds;
and judging the quantity, namely judging whether the quantity of the point clouds in each elevation interval is smaller than the judgment threshold value, if so, deleting the point cloud data in the elevation interval as noise.
Further, in the step of establishing the topological relation, the topological relation of the point cloud is established by adopting a three-dimensional grid method.
Further, the step of establishing the topological relation includes:
establishing a bounding box, establishing a minimum external cuboid bounding box according to the maximum value and the minimum value of the point cloud coordinate in the direction of three coordinate axes of X, Y, Z, wherein the length, the width and the height of the bounding box are parallel to the three coordinate axes of the three-dimensional point cloud, and the length, the width and the height of the bounding box are expressed as follows:
Xbox＝Xmax-Xmin
Ybox＝Ymax-Ymin
Zbox＝Zmax-Zmin
wherein, Xmax,Ymax,ZmaxMaximum values of X, Y, Z coordinate values, Xmin，Ymin,ZminX, Y, Z minimum values of coordinate values;
selecting an expected point cloud number, determining a value range of the expected point cloud number, traversing each expected point cloud numerical value in the value range, recording the number of small grids containing points corresponding to each expected point cloud numerical value, calculating the difference value between the product of the expected point cloud numerical value and the corresponding number of small grids containing points and the total number of the point cloud, and selecting the expected point cloud numerical value corresponding to the minimum difference value;
dividing a bounding box, equally dividing the bounding box into a plurality of cubic grids, and expressing as follows:
M＝N/C
Figure BDA0002841937430000031
wherein M is the number of the cube grids, L is the side length of the cube grids, N is the total number of the point clouds, and C is the expected number of the point clouds in each small grid;
storing the point cloud, storing the three-dimensional point cloud into corresponding grids, wherein each point has a unique grid corresponding to the point cloud, and the corresponding grids are numbered as follows:
Mi＝(Xi-Xmin)/L
Ni＝(Yi-Ymin)/L
Li＝(Zi-Zmin)/L
wherein Xi is the X coordinate value of point Pi, Yi is the Y coordinate value of point Pi, Zi is the Z coordinate value of point Pi, and Mi, Ni, Li are grid index numbers corresponding to point Pi in the X, Y, Z axis directions.
Further, a bounding box extension is further included between the bounding box establishing step and the expected point cloud number selecting step, a constant is added to each edge of the bounding box, and the length, width and height of the extended bounding box are represented as:
Xbox＝Xmax-Xmin+d
Ybox＝Ymax-Ymin+d
Zbox＝Zmax-Zmin+d
wherein d is a constant.
Further, in the step of selecting the expected point cloud number, the value range of the expected point cloud number is 20-50, and the step length is 5.
Further, in the step of establishing the K neighborhood, searching the K neighborhood inside the small grid where the point is located, and if the number of points in the small grid is greater than K and the point is located in the central area of the grid, searching K nearest points inside the small grid; if the number of points in the small grid is larger than K and the points are located at the edge of the small grid, directly searching in the grid with the total number of the peripheral grids being a first preset grid threshold; if the number of the points in the small grid is less than K, continuously searching in the small grids which are adjacent to each other and the total number of the points is a second preset grid threshold value; if the search range is not enough K points, the points are regarded as outlier cluster noise points and are directly deleted.
Further, the smoothing fine noise step includes:
fitting a tangent plane, and fitting the tangent plane on the K neighborhood of any point by a least square method;
calculating a unit normal vector, and calculating the unit normal vector of the plane according to the fitted tangent plane;
calculating a final coordinate, and calculating the final coordinate by using the unit normal vector and the bilateral filtering factor, wherein the formula is as follows:
P'＝P+n×α
wherein, P is the coordinate of the point, P' is the final coordinate of the point, n is the corresponding unit normal vector, and alpha is the bilateral filtering factor;
Figure BDA0002841937430000041
Figure BDA0002841937430000042
Figure BDA0002841937430000043
wherein, Wc(x)、Ws(x) Respectively for the fairing filter weight function and the feature preserving weight function, sigmacTaking the value as the radius of the point neighborhood, σsTaking the value as the standard deviation of the distance from the sampling point to the neighborhood point, NKThe number of grids is small with dots.
An electronic device, comprising: a processor;
a memory; and a program, wherein the program is stored in the memory and configured to be executed by the processor, the program comprising instructions for performing a fusion filtering method suitable for SLAM point cloud denoising.
A computer-readable storage medium having stored thereon a computer program for executing by a processor a fusion filtering method suitable for SLAM point cloud denoising.
Compared with the prior art, the invention has the beneficial effects that:
the method solves the problems of large calculated amount, high time consumption, low efficiency, characteristic loss and poor denoising effect of the conventional point cloud noise removing method, has good denoising effect on SLAM point clouds with wide noise space distribution and different scales, has high calculation efficiency, and greatly improves the efficiency and effect of preprocessing the SLAM point cloud data.
The foregoing description is only an overview of the technical solutions of the present invention, and in order to make the technical solutions of the present invention more clearly understood and to implement them in accordance with the contents of the description, the following detailed description is given with reference to the preferred embodiments of the present invention and the accompanying drawings. The detailed description of the present invention is given in detail by the following examples and the accompanying drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:
FIG. 1 is a flow chart of a fusion filtering method suitable for SLAM point cloud denoising according to the present invention;
FIG. 2 is an example of raw SLAM point cloud data to be processed in accordance with an embodiment of the present invention;
FIG. 3 is a diagram illustrating the effect of filtering extreme value noise by using a height statistics method according to an embodiment of the present invention;
FIG. 4 is a distribution diagram of discrete noise in the point cloud before the step of removing discrete noise and the step of establishing K neighborhood according to the embodiment of the present invention;
FIG. 5 is a diagram illustrating the denoising effect of discrete noise in point cloud in the step of removing discrete noise and the step of establishing K neighborhood according to the embodiment of the present invention;
FIG. 6 is a diagram illustrating the filtering effect of the bilateral filter on the point cloud noise in the embodiment of the present invention.
Detailed Description
The present invention will be further described with reference to the accompanying drawings and the detailed description, and it should be noted that any combination of the embodiments or technical features described below can be used to form a new embodiment without conflict.
A fusion filtering method suitable for SLAM point cloud denoising, as shown in fig. 1, includes the following steps:
the original point cloud obtained by SLAM means is shown in fig. 2, and noise points floating above and below the main body of the point cloud are extreme value noise of the point cloud.
And counting the frequency of the point cloud, counting the frequency of the point cloud in each elevation interval, and filtering extreme value noise points outside the point cloud main body. Specifically, the method comprises the following steps:
acquiring information, reading point cloud, inputting elevation interval and a judgment threshold, and if the elevation interval is set to be 0.2m, judging that the threshold is 20;
counting the quantity, namely counting the quantity of the point clouds in each elevation interval according to the Z coordinate value of the point clouds;
and judging the quantity, namely judging whether the quantity of the point clouds in each elevation interval is smaller than a judgment threshold value, if so, deleting the point cloud data in the elevation interval as noise. As shown in fig. 3, after the extreme value noise is removed, the spatial distribution range of the point cloud is greatly reduced, and the workload of subsequent calculation processing is reduced.
And establishing a topological relation, determining the number of grids divided in space, and establishing the topological relation of the point cloud. In this embodiment, the establishing of the topological relation of the point cloud based on the adaptive three-dimensional grid method includes:
establishing a bounding box, establishing a minimum external cuboid bounding box according to the maximum value and the minimum value of the point cloud coordinate in the direction of three coordinate axes of X, Y, Z, wherein the length, the width and the height of the bounding box are parallel to the three coordinate axes of the three-dimensional point cloud, and the length, the width and the height of the bounding box are expressed as follows:
Xbox＝Xmax-Xmin
Ybox＝Ymax-Ymin
Zbox＝Zmax-Zmin
wherein, Xmax,Ymax,ZmaxMaximum values of X, Y, Z coordinate values, Xmin，Ymin,ZminX, Y, Z minimum values of coordinate values;
since there may be point clouds at the edge or vertex of the bounding box, the bounding box needs to be expanded, and a constant d is added to each edge of the bounding box, where d is 1 in this embodiment. The expanded bounding box length, width, height are expressed as:
Xbox＝Xmax-Xmin+d
Ybox＝Ymax-Ymin+d
Zbox＝Zmax-Zmin+d
the method selects an expected point cloud number C, and provides a self-adaptive grid number determination method in order to avoid too dense grid division and excessive empty grids without point clouds. Firstly, determining the value range of the point cloud number C in the expected grid, in this embodiment, taking C to be not less than 20 and not more than 50, and the step length to be 5I.e. the candidate C is {20,25,30,35,40,45,50}, traversing each C value, and recording the number N of small grids containing points corresponding to each C valueKCalculating C and NKSelecting a C value corresponding to the minimum difference value from the difference value of the point cloud total number;
dividing the bounding box, equally dividing the bounding box into M cubic grids with the side length of L, and expressing as follows:
M＝N/C
Figure BDA0002841937430000071
wherein M is the number of the cube grids, L is the side length of the cube grids, N is the total number of the point clouds, and C is the expected number of the point clouds in each small grid;
storing the point cloud, storing the three-dimensional point cloud into the corresponding grids, wherein each point Pi has one and only one grid corresponding to the point Pi, and the grid number corresponding to the point Pi is as follows:
Mi＝(Xi-Xmin)/L
Ni＝(Yi-Ymin)/L
Li＝(Zi-Zmin)/L
wherein Xi is the X coordinate value of point Pi, Yi is the Y coordinate value of point Pi, Zi is the Z coordinate value of point Pi, and Mi, Ni, Li are grid index numbers corresponding to point Pi in the X, Y, Z axis directions.
Removing discrete noise, calculating the number of point clouds in each grid, and if the number of point clouds in each grid is smaller than a threshold value, namely 10, determining that the point clouds in the grid are discrete noise, and deleting;
and establishing a K neighborhood, establishing the K neighborhood of the point cloud through a space grid, and taking K as 50. K neighborhood K (P) of point P is the set of K points nearest to point P. The K neighborhood is established based on the space grid, so that all points in the search point cloud can be prevented from being traversed, and the method has the characteristic of high efficiency. And simultaneously, removing isolated cluster noise. In particular, the method comprises the following steps of,
searching K neighborhoods inside the small grids where the point P is located, and searching K nearest points inside the small grids if the number of points in the small grids is larger than K and the point P is located in a grid central area; if the number of points in the small grid is greater than K and the point P is located at the edge of the small grid, searching directly in the grid itself and the grid whose total number of peripheral grids is the first preset grid threshold, in this embodiment, the first preset grid threshold is 27; if the number of the points in the small grid is less than K, continuously searching in the small grids which are adjacent to each other and the total number of which is the second preset grid threshold value, wherein in the embodiment, the second preset grid threshold value is 27-1 to 26; if the search range is not enough K points, the points are regarded as outlier cluster noise points and are directly deleted.
The point cloud before the step of removing the discrete noise and the step of establishing the K neighborhood are shown in FIG. 4, and the point cloud after the step of removing the discrete noise and the step of establishing the K neighborhood are shown in FIG. 5.
Smoothing the fine noise, and removing the fine noise attached to the point cloud main body through a bilateral filter to obtain a final denoising result. Specifically, the method comprises the following steps:
fitting a tangent plane, wherein the K neighborhood of any point P is K (P), and fitting the tangent plane on the K neighborhood K (P) of any point by a least square method, and recording the fitting result as S (P);
and calculating a unit normal vector, and calculating a unit normal vector n of the plane according to the plane S (p) fitted by the K neighborhood points, wherein the unit normal vector n of the fitted plane S (p) is the unit normal vector to be estimated.
Calculating a final coordinate, and calculating the final coordinate by using a unit method vector and a bilateral filtering factor, wherein the formula is as follows:
P'＝P+n×α
wherein, P is the coordinate of the point, P' is the final coordinate of the point, n is the corresponding unit normal vector, and alpha is the bilateral filtering factor;
Figure BDA0002841937430000091
Figure BDA0002841937430000092
Figure BDA0002841937430000093
wherein, Wc(x)、Ws(x) Respectively for the fairing filter weight function and the feature preserving weight function, sigmacTaking the value as the radius of the point neighborhood, σsThe value is the standard deviation of the distance from the sampling point to the neighborhood point.
The point cloud processed by the step of smoothing the fine noise is shown in fig. 6, the fine noise on the edge feature of the point cloud is effectively smoothed, and a good visual effect is obtained.
An electronic device, comprising: a processor;
a memory; and a program, wherein the program is stored in the memory and configured to be executed by the processor, the program comprising instructions for performing a fusion filtering method suitable for SLAM point cloud denoising.
A computer-readable storage medium having stored thereon a computer program for executing by a processor a fusion filtering method suitable for SLAM point cloud denoising.
The method solves the problems of large calculated amount, high time consumption, low efficiency, characteristic loss and poor denoising effect of the conventional point cloud noise removing method, has good denoising effect on SLAM point clouds with wide noise space distribution and different scales, has high calculation efficiency, and greatly improves the efficiency and effect of preprocessing the SLAM point cloud data.
The foregoing is merely a preferred embodiment of the invention and is not intended to limit the invention in any manner; those skilled in the art can readily practice the invention as shown and described in the drawings and detailed description herein; however, those skilled in the art should appreciate that they can readily use the disclosed conception and specific embodiments as a basis for designing or modifying other structures for carrying out the same purposes of the present invention without departing from the scope of the invention as defined by the appended claims; meanwhile, any changes, modifications, and evolutions of the equivalent changes of the above embodiments according to the actual techniques of the present invention are still within the protection scope of the technical solution of the present invention.

Claims (10)

1. A fusion filtering method suitable for SLAM point cloud denoising is characterized by comprising the following steps:
counting the frequency of the point cloud, counting the frequency of the point cloud in each elevation interval, and filtering extreme value noise points outside the point cloud main body;
establishing a topological relation, determining the number of grids divided in space, and establishing a topological relation of point cloud;
removing discrete noise, calculating the number of point clouds in each grid, and if the number of point clouds in each grid is smaller than a threshold value, determining the point clouds in the grid as the discrete noise and deleting the point clouds;
establishing a K neighborhood, establishing the K neighborhood of the point cloud through a space grid, and simultaneously removing isolated cluster noise points;
smoothing the fine noise, and removing the fine noise attached to the point cloud main body through a bilateral filter to obtain a final denoising result.
2. The method of claim 1, wherein the filtering method comprises the following steps: the step of counting the point cloud frequency comprises the following steps:
acquiring information, reading point cloud, inputting elevation interval and judging a threshold value;
counting the quantity, namely counting the quantity of the point clouds in each elevation interval according to the Z coordinate value of the point clouds;
and judging the quantity, namely judging whether the quantity of the point clouds in each elevation interval is smaller than the judgment threshold value, if so, deleting the point cloud data in the elevation interval as noise.
3. The method of claim 1, wherein the filtering method comprises the following steps: in the step of establishing the topological relation, the topological relation of the point cloud is established by adopting a three-dimensional grid method.
4. The method of claim 3, wherein the filtering method comprises the following steps: the step of establishing the topological relation comprises the following steps:
establishing a bounding box, establishing a minimum external cuboid bounding box according to the maximum value and the minimum value of the point cloud coordinate in the direction of three coordinate axes of X, Y, Z, wherein the length, the width and the height of the bounding box are parallel to the three coordinate axes of the three-dimensional point cloud, and the length, the width and the height of the bounding box are expressed as follows:
Xbox＝Xmax-Xmin
Ybox＝Ymax-Ymin
Zbox＝Zmax-Zmin
wherein, Xmax,Ymax,ZmaxMaximum values of X, Y, Z coordinate values, Xmin，Ymin,ZminX, Y, Z minimum values of coordinate values;
selecting an expected point cloud number, determining a value range of the expected point cloud number, traversing each expected point cloud numerical value in the value range, recording the number of small grids containing points corresponding to each expected point cloud numerical value, calculating the difference value between the product of the expected point cloud numerical value and the corresponding number of small grids containing points and the total number of the point cloud, and selecting the expected point cloud numerical value corresponding to the minimum difference value;
dividing a bounding box, equally dividing the bounding box into a plurality of cubic grids, and expressing as follows:
M＝N/C
Figure FDA0002841937420000021
wherein M is the number of the cube grids, L is the side length of the cube grids, N is the total number of the point clouds, and C is the expected number of the point clouds in each small grid;
storing the point cloud, storing the three-dimensional point cloud into corresponding grids, wherein each point has a unique grid corresponding to the point cloud, and the corresponding grids are numbered as follows:
Mi＝(Xi-Xmin)/L
Ni＝(Yi-Ymin)/L
Li＝(Zi-Zmin)/L
wherein Xi is the X coordinate value of point Pi, Yi is the Y coordinate value of point Pi, Zi is the Z coordinate value of point Pi, and Mi, Ni, Li are grid index numbers corresponding to point Pi in the X, Y, Z axis directions.
5. The method of claim 4, wherein the filtering method comprises the following steps: and further comprising bounding box expansion between the step of establishing the bounding box and the step of selecting the expected point cloud number, wherein each edge of the bounding box is added with a constant, and the length, the width and the height of the expanded bounding box are represented as follows:
Xbox＝Xmax-Xmin+d
Ybox＝Ymax-Ymin+d
Zbox＝Zmax-Zmin+d
wherein d is a constant.
6. The method of claim 4, wherein the filtering method comprises the following steps: in the step of selecting the expected point cloud number, the value range of the expected point cloud number is 20-50, and the step length is 5.
7. The method of claim 1, wherein the filtering method comprises the following steps: in the step of establishing the K neighborhood, searching the K neighborhood inside a small grid in which the point is located, and if the number of points in the small grid is greater than K and the point is located in a central area of the grid, searching K nearest points inside the small grid; if the number of points in the small grid is larger than K and the points are located at the edge of the small grid, directly searching in the grid with the total number of the peripheral grids being a first preset grid threshold; if the number of the points in the small grid is less than K, continuously searching in the small grids which are adjacent to each other and the total number of the points is a second preset grid threshold value; if the search range is not enough K points, the points are regarded as outlier cluster noise points and are directly deleted.
8. The method of claim 1, wherein the filtering method comprises the following steps: the smoothing fine noise step includes:
fitting a tangent plane, and fitting the tangent plane on the K neighborhood of any point by a least square method;
calculating a unit normal vector, and calculating the unit normal vector of the plane according to the fitted tangent plane;
calculating a final coordinate, and calculating the final coordinate by using the unit normal vector and the bilateral filtering factor, wherein the formula is as follows:
P'＝P+n×α
wherein, P is the coordinate of the point, P' is the final coordinate of the point, n is the corresponding unit normal vector, and alpha is the bilateral filtering factor;
Figure FDA0002841937420000041
Figure FDA0002841937420000042
Figure FDA0002841937420000043
wherein, Wc(x)、Ws(x) Respectively for the fairing filter weight function and the feature preserving weight function, sigmacTaking the value as the radius of the point neighborhood, σsTaking the value as the standard deviation of the distance from the sampling point to the neighborhood point, NKThe number of grids is small with dots.
9. An electronic device, characterized by comprising: a processor;
a memory; and a program, wherein the program is stored in the memory and configured to be executed by the processor, the program comprising instructions for carrying out the method according to any one of claims 1-8.
10. A computer-readable storage medium having stored thereon a computer program, characterized in that: the computer program is executed by a processor for performing the method according to any of claims 1-8.
CN202011495165.2A 2020-12-17 2020-12-17 Fusion filtering method suitable for SLAM point cloud denoising, electronic device and storage medium Pending CN112651889A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011495165.2A CN112651889A (en) 2020-12-17 2020-12-17 Fusion filtering method suitable for SLAM point cloud denoising, electronic device and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011495165.2A CN112651889A (en) 2020-12-17 2020-12-17 Fusion filtering method suitable for SLAM point cloud denoising, electronic device and storage medium

Publications (1)

Publication Number Publication Date
CN112651889A true CN112651889A (en) 2021-04-13

Family

ID=75355390

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011495165.2A Pending CN112651889A (en) 2020-12-17 2020-12-17 Fusion filtering method suitable for SLAM point cloud denoising, electronic device and storage medium

Country Status (1)

Country Link
CN (1) CN112651889A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113483661A (en) * 2021-07-06 2021-10-08 广东南方数码科技股份有限公司 Point cloud data acquisition method, device, equipment and storage medium
CN113763280A (en) * 2021-09-15 2021-12-07 湖南科技大学 Region growing algorithm based on spatial hierarchical topological relation for point cloud denoising
CN114612627A (en) * 2022-03-11 2022-06-10 广东汇天航空航天科技有限公司 Processing method and device of terrain elevation map, vehicle and medium
CN117132478A (en) * 2023-04-25 2023-11-28 兰州交通大学 Orbit point cloud denoising method based on normal vector two-norm characteristic parameter
CN117392000A (en) * 2023-12-08 2024-01-12 吉咖智能机器人有限公司 Noise removing method and device, electronic equipment and storage medium

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113483661A (en) * 2021-07-06 2021-10-08 广东南方数码科技股份有限公司 Point cloud data acquisition method, device, equipment and storage medium
CN113763280A (en) * 2021-09-15 2021-12-07 湖南科技大学 Region growing algorithm based on spatial hierarchical topological relation for point cloud denoising
CN113763280B (en) * 2021-09-15 2023-11-21 湖南科技大学 Region growing method based on space hierarchy topological relation for point cloud denoising
CN114612627A (en) * 2022-03-11 2022-06-10 广东汇天航空航天科技有限公司 Processing method and device of terrain elevation map, vehicle and medium
CN114612627B (en) * 2022-03-11 2023-03-03 广东汇天航空航天科技有限公司 Processing method and device of terrain elevation map, vehicle and medium
CN117132478A (en) * 2023-04-25 2023-11-28 兰州交通大学 Orbit point cloud denoising method based on normal vector two-norm characteristic parameter
CN117132478B (en) * 2023-04-25 2024-05-17 兰州交通大学 Orbit point cloud denoising method based on normal vector two-norm characteristic parameter
CN117392000A (en) * 2023-12-08 2024-01-12 吉咖智能机器人有限公司 Noise removing method and device, electronic equipment and storage medium
CN117392000B (en) * 2023-12-08 2024-03-08 吉咖智能机器人有限公司 Noise removing method and device, electronic equipment and storage medium

Similar Documents

Publication Publication Date Title
CN112651889A (en) Fusion filtering method suitable for SLAM point cloud denoising, electronic device and storage medium
CN111932688B (en) Indoor plane element extraction method, system and equipment based on three-dimensional point cloud
CN106599129B (en) A kind of multi-beam point cloud data denoising method for taking lineament into account
Shi et al. Adaptive simplification of point cloud using k-means clustering
CN108830931B (en) Laser point cloud simplification method based on dynamic grid k neighborhood search
CN110599506B (en) Point cloud segmentation method for three-dimensional measurement of complex special-shaped curved surface robot
CN115372989A (en) Laser radar-based long-distance real-time positioning system and method for cross-country automatic trolley
CN104616349A (en) Local curved surface change factor based scattered point cloud data compaction processing method
CN115222625A (en) Laser radar point cloud denoising method based on multi-scale noise
CN113538264B (en) Denoising method and device for point cloud data and storage medium
CN112861983A (en) Image matching method, image matching device, electronic equipment and storage medium
CN111581776A (en) Geometric reconstruction model-based isogeometric analysis method
CN112825199B (en) Collision detection method, device, equipment and storage medium
CN113610983A (en) Automatic subdivision method for discrete point space curved surface triangular mesh
CN116152306A (en) Method, device, apparatus and medium for determining masonry quality
CN110675323B (en) Three-dimensional map semantic processing method, system, equipment and computer medium
CN115546440A (en) Curved surface reconstruction method, device, equipment and storage medium
CN114299242A (en) Method, device and equipment for processing images in high-precision map and storage medium
CN116721228B (en) Building elevation extraction method and system based on low-density point cloud
CN105869210A (en) Interpolation data processing method in three-dimensional geological surface model
CN117152341A (en) Three-dimensional surface extraction method, apparatus and storage medium
CN112598799B (en) Probability theory-based two-dimensional point cloud outsourcing contour processing method, device and medium
CN110796735A (en) NURBS curved surface finite element plate shell mesh division method and computer implementation system
CN113469907B (en) Data simplification method and system based on blade profile characteristics
CN113763280A (en) Region growing algorithm based on spatial hierarchical topological relation for point cloud denoising

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