Point cloud noise reduction processing methodTechnical Field
The invention belongs to the technical field of digital measurement, and particularly relates to a point cloud noise reduction processing method.
Background
Along with the continuous development of the aviation manufacturing industry to the intelligent and digital directions, the quality detection requirements on aviation parts are also continuously improved. The degree of fit between the manufactured form factor and the design form factor of the aircraft component directly reflects the quality of the aircraft product. Therefore, high-precision detection of the external dimensions of the aircraft parts will play a vital role in improving the quality of the aircraft products. The traditional aircraft part detection means still rely on conventional tools such as vernier calipers and feelers to measure the external dimensions, so that the external dimensions of the parts cannot be accurately detected, and meanwhile, large-scale measurement data are difficult to form for guiding and optimizing the processing process of the parts.
With the development of three-dimensional measurement technology, digital detection technologies such as photogrammetry, flight time measurement and structured light measurement are gradually applied to the detection links of aircraft parts. The structured light measurement technology mainly obtains point cloud data of the surface of the part through three-dimensional scanning equipment, and carries out data processing on the point cloud to improve measurement accuracy. In the structured light measurement technology, the noise reduction of the point cloud is the most critical link of the data processing of the point cloud, and the processing precision of the noise reduction of the point cloud directly influences the precision of the digital measurement of the structured light, thereby influencing the product quality of an airplane.
Patent document CN113177897a discloses a fast lossless filtering method of disordered 3D point cloud, which performs gridding segmentation on an external cube of the point cloud, and converts the dense disordered point cloud into regular binary three-dimensional point cloud, so as to reduce the calculation amount. In addition, patent document CN103853840a discloses a method for filtering non-uniform scattered point cloud data, which is implemented by organizing index of scattered point cloud data, forming regular square grid models from the scattered point clouds, and obtaining scattered points contained in each grid. However, in the existing point cloud noise reduction method, the influence caused by the step edge on the point cloud noise reduction precision is not considered, the edge protection performance of the point cloud noise reduction is affected, and the noise reduction precision is difficult to ensure.
Disclosure of Invention
The invention aims to provide a point cloud noise reduction processing method, which can solve the problem that the step edge affects the point cloud noise reduction, improve the edge protection performance of the point cloud noise reduction at the step edge and improve the point cloud noise reduction precision.
The invention is realized by the following technical scheme:
the point cloud noise reduction processing method comprises the following steps:
dividing the neighborhood point cloud into a plurality of sub-neighborhood point clouds by adopting a predefined frame cube and taking the point cloud to be filtered as the center;
Carrying out bilateral filtering treatment on each sub-neighborhood point cloud data to obtain a new point cloud;
and carrying out radius filtering treatment on the point cloud to be filtered and the new point cloud by taking the point cloud to be filtered as a center, and finishing the noise reduction treatment of the point cloud.
In some embodiments, the method for acquiring the neighborhood point cloud data of the point cloud to be filtered includes: and acquiring original point cloud data, and acquiring neighborhood point cloud data of the point cloud to be filtered by adopting a K neighbor algorithm.
In some embodiments, a structured light system is adopted to obtain point cloud data of an object to be processed under different view angles, and the point cloud data under different view angles are spliced by a point cloud splicing algorithm to obtain original point cloud data.
In some embodiments, the method for dividing the neighborhood point cloud into a plurality of sub-neighborhood point clouds comprises the following steps: eight bounding box cubes are defined, and are placed together, and the common intersection point of the eight bounding box cubes is the point cloud to be filtered.
In some embodiments, the side length of each bounding box is greater than the radius of the neighborhood point cloud.
In some embodiments, the method for filtering each sub-neighborhood point cloud data to obtain a new point cloud by adopting bilateral filtering includes:
Setting the point cloud to be filtered as p, setting a neighborhood point cloud set of p as N (p), and setting each sub-neighborhood point cloud as Ni (p);
the normal vector of bilateral filtering at point cloud p is:
wherein,The unit normal vector of the point cloud p, the position is denoted as lp,Is the unit normal vector of the neighborhood point cloud of the point cloud p; wC is a standard gaussian filter,WS is a standard gaussian filter,
Is a unit normal vectorAnd (3) withThe intensity difference between them, expressed as:
After bilateral filtering is performed on each sub-neighborhood point cloud set, the obtained new point cloud data can be expressed as:
wherein,Representing the sub-neighborhood point cloud location.
In some embodiments, when the radius filtering process is performed on the point cloud to be filtered and the new point cloud, when the point cloud data in the circle is equal to 1, the point cloud is rejected.
In some embodiments, when the radius filtering process is performed on the point cloud to be filtered and the new point cloud, when the number of the point clouds in the circle is greater than 1, the bilateral filtering process is performed on the point cloud data in the circle again.
Compared with the prior art, the invention has the following advantages:
According to the invention, the neighborhood points of the point cloud to be filtered are divided into different sub-neighborhood points and are respectively subjected to filtering treatment, so that the edge protection performance of filtering and the accuracy of the filtering treatment can be improved to a great extent, the method has good practicality and practical significance in the dimension measurement application of a large number of products in aircraft manufacturing, the treatment speed in the detection process can be greatly improved, the detection accuracy is improved, the detection efficiency is improved, and the lead time of parts is ensured.
According to the method, the neighborhood of each point cloud is directly divided into different areas to select the low-frequency area and the high-frequency area at the step edge, and bilateral filtering is respectively carried out on the low-frequency area and the high-frequency area, so that the complexity of a processing method can be reduced, the robustness of the processing method can be improved, and edge-preserving filtering can be well achieved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following description will briefly describe the drawings in the embodiments, it being understood that the following drawings only illustrate some embodiments of the present invention and should not be considered as limiting the scope, and that other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of a point cloud noise reduction processing method of the present invention.
Fig. 2 is a schematic diagram of processing a neighborhood point cloud of a point cloud to be filtered using a bounding box cube according to the present invention.
Fig. 3 a) is a schematic diagram of a case where the present invention performs radius filtering processing on a point cloud to be filtered and a new point cloud.
Fig. 3 b) is a schematic diagram of another case of the present invention where the radius filtering is performed on the point cloud to be filtered and the new point cloud.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the following detailed description of specific embodiments of the present application is given with reference to the accompanying drawings. It is to be understood that the specific embodiments described herein are merely illustrative of the application and are not limiting thereof. It should be further noted that, for convenience of description, only some, but not all of the matters related to the present application are shown in the accompanying drawings. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although a flowchart depicts operations (or steps) as a sequential process, many of the operations can be performed in parallel, concurrently, or at the same time. Furthermore, the order of the operations may be rearranged. The process may be terminated when its operations are completed, but may have additional steps not included in the figures. The processes may correspond to methods, functions, procedures, subroutines, and the like.
In the structured light measurement technology, when the point cloud data is acquired, due to the influence of equipment precision, operator experience environmental factors and the diffraction property of electromagnetic waves, the surface property of the measured object changes and the influence of the data stitching registration operation process, some noise will inevitably appear in the point cloud data, and the sampling resolution is different, so that the original point cloud data acquired through the structured light system often contains a large number of hash points and isolated points, and therefore, the point cloud noise reduction processing is required.
According to the point cloud noise reduction processing method, the neighborhood point cloud data of the point cloud to be processed is divided into different sub-neighborhood point cloud data based on the predefined frame cube, and bilateral filtering is carried out on the sub-neighborhood point cloud data, so that point cloud noise reduction with outstanding edge protection performance and high precision is achieved, and complexity of the method is reduced.
Referring to fig. 1, the point cloud noise reduction processing method of the present invention is described with reference to a specific embodiment, and includes the following steps:
S1, acquiring point cloud data of an object to be processed under different view angles by adopting a structured light system, and performing splicing processing on the point cloud data under the different view angles by using a point cloud splicing algorithm to obtain original point cloud data;
s2, acquiring point cloud data, acquiring neighborhood point cloud data of the point cloud to be filtered by adopting a K neighbor algorithm, and defining a neighborhood point cloud data set, wherein the neighborhood point cloud data set is expressed as N (p);
S3, defining eight frame cubes, and referring to FIG. 2, each frame cube is respectively defined as 1,2,3,4,5,6,7 and 8;
placing eight frame cubes together, and setting a common intersection point as a point cloud to be filtered; dividing the point cloud to be filtered into eight sub-neighborhood point clouds by utilizing eight frame cubes, wherein the eight sub-neighborhood point clouds are respectively represented as Ni (p), namely each frame cube comprises part of the neighborhood point clouds, and the side length of the frame cube is larger than the radius of the neighborhood point clouds;
S4, carrying out bilateral filtering processing on the point cloud data contained in the eight frame cubes to obtain each new point cloud data;
Is provided withFor the unit normal vector of the point cloud p, the position is denoted as lp, and the normal vector of the bilateral filtering at the point cloud p to be processed can be expressed as:
wherein,Is the unit normal vector of the neighborhood point cloud of the point cloud p; wC is a standard Gaussian filter, in whichWS is also a standard Gaussian filter, where
Is a unit normal vectorAnd (3) withThe intensity difference between them, expressed as:
After bilateral filtering is performed on each sub-neighborhood point cloud set, the obtained new point cloud data can be expressed as:
wherein,Representing the sub-neighborhood point cloud location.
S5, filtering the point cloud to be filtered and eight new point cloud data by utilizing radius filtering;
Setting a radius filter center as a point cloud to be filtered, wherein the radius is R;
If the circle set by radius filtering only contains the point cloud to be filtered, the number of the point clouds in the circle is equal to 1, and the point clouds are removed, as shown in fig. 3 a);
if the number of point clouds contained in the circle set by the radius filtering is greater than 1, performing bilateral filtering on the point clouds contained in the circle again to obtain a final filtering result, and referring to fig. 3 b);
At this time, the noise reduction processing of the point cloud is completed.
The foregoing description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and any simple modification, equivalent variation, etc. of the above embodiment according to the technical matter of the present invention fall within the scope of the present invention.