CN116205886A - Point cloud quality assessment method based on relative entropy - Google Patents

Point cloud quality assessment method based on relative entropy Download PDF

Info

Publication number
CN116205886A
CN116205886A CN202310209556.0A CN202310209556A CN116205886A CN 116205886 A CN116205886 A CN 116205886A CN 202310209556 A CN202310209556 A CN 202310209556A CN 116205886 A CN116205886 A CN 116205886A
Authority
CN
China
Prior art keywords
point cloud
point
sub
geometric
relative entropy
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
CN202310209556.0A
Other languages
Chinese (zh)
Inventor
罗雷
武文欣
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.)
Chongqing University of Post and Telecommunications
Original Assignee
Chongqing University of Post and Telecommunications
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 Chongqing University of Post and Telecommunications filed Critical Chongqing University of Post and Telecommunications
Priority to CN202310209556.0A priority Critical patent/CN116205886A/en
Publication of CN116205886A publication Critical patent/CN116205886A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G06V10/765—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects using rules for classification or partitioning the feature space
    • 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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)

Abstract

The invention relates to a point cloud quality assessment method based on relative entropy, which belongs to the field of multimedia information processing, and comprises the steps of downsampling a reference point cloud to obtain a geometric framework of the point cloud; constructing sub-point clouds taking a geometric framework as a center in the reference point clouds and the distorted point clouds, calculating a rapid point characteristic histogram and relative entropy of the sub-point clouds, and taking the rapid point characteristic histogram and the relative entropy as a geometric quality result of the sub-point clouds; converting the color attribute of the sub-point cloud into an LMN color space, and calculating the zeroth moment, the first moment and the second moment of the color space to obtain a color quality result of the sub-point cloud; and weighting and aggregating the geometric quality result and the color quality result of the sub-point cloud to obtain a final point cloud quality evaluation model based on relative entropy.

Description

Point cloud quality assessment method based on relative entropy
Technical Field
The invention belongs to the field of multimedia information processing, and relates to a point cloud quality assessment method based on relative entropy.
Background
The point cloud is an emerging multimedia type, and is widely focused in the fields of intelligent transportation, unmanned driving, virtual reality and the like. A typical point cloud consists of millions of unstructured, non-uniformly distributed 3D points, each containing at least location information and color information.
Similar to the traditional image processing method, the processing of the point cloud is subjected to the steps of sampling, compression, transmission, reconstruction, rendering, analysis and the like. Due to technical limitations, the quality of the point cloud must be compromised during the above processing, for example, the lossy compression may cause distortion, the sampling process may be polluted by noise, etc., so it is important to accurately evaluate the quality of the lossy point cloud.
The quality evaluation of the point cloud can be divided into subjective evaluation and objective evaluation, wherein the subjective evaluation evaluates the point cloud according to subjective impression of a viewer, and a common subjective quality evaluation method gives an original point cloud (reference point cloud) and a distorted point cloud (point cloud to be evaluated), so that the viewer scores an image to be evaluated, sums all subjective scores and averages the subjective scores to obtain an average subjective score; objective assessment is the use of a specific mathematical model to give a quantitative value of the difference between the reference image and the image to be assessed. Subjective evaluation is influenced by factors such as subjective preference of viewers, viewing equipment sites and the like, and has high cost and low efficiency; the score of objective evaluation is not influenced by subjective factors, and is widely applied to the field of point cloud quality evaluation. The applicant finds that the existing point cloud objective quality assessment algorithm is not accurate enough, so that the efficiency of the subsequent processing steps is low.
Disclosure of Invention
In view of the above, the present invention aims to provide a point cloud quality evaluation method based on relative entropy. The method considers the geometric characteristics and the color characteristics of the point cloud, firstly, extracts geometric key points by resampling the reference point cloud, and forms a geometric skeleton. Then, respectively constructing sub-point clouds in the reference point cloud and the distortion point cloud by taking the key points as centers, and evaluating the sub-point clouds of the reference point cloud and the distortion point cloud from two angles of geometric distortion and color distortion: firstly, calculating fast point characteristic histograms of sub-point clouds of a reference point cloud and a distorted point cloud, and then calculating relative entropy between the point characteristic histograms to measure geometric distortion degree so as to obtain geometric quality of the sub-point clouds; and calculating the statistical moment of the color gradient of the sub-point cloud of the reference point cloud and the distorted point cloud to measure the degree of color distortion, and obtaining the color quality of the sub-point cloud. And finally, carrying out linear weighting on the geometric quality and the color quality of all the sub-point clouds to obtain a tensor point cloud quality evaluation model based on the relative entropy.
In order to achieve the above purpose, the present invention provides the following technical solutions:
a point cloud quality assessment method based on relative entropy comprises the following steps:
s1: and downsampling the reference point cloud, and acquiring the geometric framework of the downsampled reference point cloud through a high-pass diagram filter.
S2: constructing sub-point clouds taking geometric architecture as center in the reference point cloud and the distorted point cloud by using a k nearest neighbor classification algorithm, and respectively recording as S r and Sd ；
S3: calculation S r and Sd Fast point feature histogram F of (1) r and Fd Then calculate F r and Fd The relative entropy is taken as a geometric quality result of the sub-point cloud;
s4: will S r and Sd The color attributes of the sub-point cloud are converted from an RGB color space to an LMN color space, the zeroth moment, the first moment and the second moment of the color space are calculated, and the color quality result of the sub-point cloud is obtained through weighted calculation;
s5: weighting and aggregating the geometric quality result and the color quality result of the sub-point cloud to obtain a final tensor point cloud quality evaluation model based on relative entropy;
further, the step S1 includes the steps of:
s11: the reference point cloud consists of N points, each point having K attributes, noted as:
Figure BDA0004112126090000021
wherein ,
Figure BDA0004112126090000022
represents the 1 st to N th points, +.>
Figure BDA0004112126090000023
Representing a dimension of N×Space of K;
s12: using geometric distances
Figure BDA0004112126090000024
To calculate the sub-point cloud midpoint +.>
Figure BDA0004112126090000025
and />
Figure BDA0004112126090000026
The connection weight between the two is defined as follows:
Figure BDA0004112126090000027
wherein ,σ2 Representing the variance between points, τ represents the euclidean distance threshold for clustering neighboring points into the same sub-point cloud.
S13: measuring edge density of each vertex using diagonal matrix D, points
Figure BDA0004112126090000028
Is->
Figure BDA0004112126090000029
The edge density D is defined as follows:
Figure BDA00041121260900000210
s14: the high pass filter through which the reference point cloud passes is expressed as:
Figure BDA00041121260900000211
where L is the filter length, h l Is the first filter coefficient;
Figure BDA00041121260900000212
is a sub-point cloud shift operator defined as：
A＝D -1 W
S15: on the basis of h (A), selecting a Haar-like local area filter h HH (A) The high pass filtering is realized as follows:
Figure BDA0004112126090000031
wherein λ1 ～λ N Is the characteristic value of A, U is the characteristic vector of A;
s16: each point passing through the filter
Figure BDA0004112126090000032
The frequency response of (2) is:
Figure BDA0004112126090000033
wherein
Figure BDA0004112126090000034
Representation dot->
Figure BDA0004112126090000035
Is a group of adjacent points->
Figure BDA0004112126090000036
Representation dot->
Figure BDA0004112126090000037
and />
Figure BDA0004112126090000038
Is a feature covariance matrix of (1); />
Let the sampling frequency of the filter be f s The total point number of the sampled point cloud
Figure BDA0004112126090000039
Resampled reference point cloud +.>
Figure BDA00041121260900000310
Is a geometric architecture.
Further, in step S2, a Euclidean distance threshold gamma is given to the reference point cloud and the distorted point cloud
Figure BDA00041121260900000311
Geometric key of->
Figure BDA00041121260900000312
The sub-point cloud is built for the center using a k-nearest neighbor search algorithm, defined as follows:
Figure BDA00041121260900000313
Figure BDA00041121260900000314
wherein Sr Representing a geometrically centered sub-point cloud built in a reference point cloud,
Figure BDA00041121260900000315
representing a point in a reference point cloud, +.>
Figure BDA00041121260900000316
Representing a resampled reference point cloud, S d Representing a geometrical architecture centered sub-point cloud built in a distorted point cloud, ++>
Figure BDA00041121260900000317
Representing points in a distorted point cloud, +.>
Figure BDA00041121260900000318
Representing the resampled distorted point cloud;
the points in the sub-point cloud are arranged from small to large according to the distance from the geometric key point.
Further, the specific step of regarding the relative entropy as the geometric quality result of the sub-point cloud in step S3 includes:
s31: from S r Selecting distance geometric key points
Figure BDA00041121260900000319
Nearest k r A point of which is k r The maximum distance of a point from a geometrical key point is denoted +.>
Figure BDA00041121260900000320
From S d Selecting k nearest to the geometric key point d A point of which is k d The maximum distance of a point from a geometric key point is recorded as
Figure BDA00041121260900000321
S32: at S r In (3) calculating
Figure BDA00041121260900000322
Every pair within the distance +.>
Figure BDA00041121260900000323
And its adjacent point->
Figure BDA00041121260900000324
Three elements of a point feature histogram of (a)
Figure BDA00041121260900000325
Then statistically, a simple point feature histogram +.>
Figure BDA00041121260900000326
At S d In (3) calculating
Figure BDA00041121260900000327
Every pair within the distance +.>
Figure BDA00041121260900000328
And its adjacent point->
Figure BDA00041121260900000329
Is a three-element point feature histogram>
Figure BDA00041121260900000330
Then statistically, a simple point feature histogram +.>
Figure BDA0004112126090000041
S33: at S r In determining each point
Figure BDA0004112126090000042
At->
Figure BDA00041121260900000418
Adjacent points within the distance and calculating each +.>
Figure BDA0004112126090000043
Is a three-element point feature histogram>
Figure BDA0004112126090000044
Then statistically, a simple point feature histogram +.>
Figure BDA0004112126090000045
At S d In determining each point
Figure BDA0004112126090000046
At->
Figure BDA0004112126090000047
Three elements of a Point characteristic histogram within a distance +.>
Figure BDA0004112126090000048
Then statistically, a simple point feature histogram +.>
Figure BDA0004112126090000049
S34: calculation S r and Sd Fast point feature histogram FPFH of (a) r and FPFHd The definition is as follows:
Figure BDA00041121260900000410
Figure BDA00041121260900000411
wherein di Is the Euclidean distance between points;
s35: calculate FPFH r and FPFHd The relative entropy between the two to obtain a geometric quality evaluation index D i The relative entropy between histograms is defined as:
Figure BDA00041121260900000412
Figure BDA00041121260900000413
further, the calculating in step S4 obtains the color quality result of the sub-point cloud, which specifically includes the following steps:
s41: will S r and Sd Is converted from RGB space to LMN space:
Figure BDA00041121260900000414
s42: defining the zeroth moment, the first moment and the second moment as the total number m g Mean mu g Variance sigma g ；
S43: definition S r and Sd The 3 metric values of the color attribute of (a) are:
Figure BDA00041121260900000415
Figure BDA00041121260900000416
Figure BDA00041121260900000417
N 0 ,N 1 ,N 3 three small non-zero numbers, preventing the denominator from being 0;
s44: multiplying the above 3 metrics to obtain S r and Sd Quality metric results for color attributes in between:
Figure BDA0004112126090000051
further, in step S5, the geometric quality result and the color quality result of the weighted aggregate sub-point cloud are obtained, and the final tensor point cloud quality evaluation model based on the relative entropy is obtained as follows:
Figure BDA0004112126090000052
wherein ,
Figure BDA0004112126090000053
and />
Figure BDA0004112126090000054
The weight factors of the geometric quality assessment index and the color quality assessment index are respectively.
The invention has the beneficial effects that: the model not only evaluates the color quality of the point cloud, but also accurately evaluates the geometric tensor of the point cloud in theory of relative entropy.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objects and other advantages of the invention may be realized and obtained by means of the instrumentalities and combinations particularly pointed out in the specification.
Drawings
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in the following preferred detail with reference to the accompanying drawings, in which:
FIG. 1 is a flow chart of a system employed in the present invention;
FIG. 2 is a schematic diagram of the point cloud resampling in step S1 of the method of the present invention;
FIG. 3 is a schematic diagram of the method of the present invention after clustering according to the geometric architecture in step S2;
fig. 4 is a region diagram of a fast point feature histogram affecting a sub-point cloud in step S3 of the method of the present invention.
Detailed Description
Other advantages and effects of the present invention will become apparent to those skilled in the art from the following disclosure, which describes the embodiments of the present invention with reference to specific examples. The invention may be practiced or carried out in other embodiments that depart from the specific details, and the details of the present description may be modified or varied from the spirit and scope of the present invention. It should be noted that the illustrations provided in the following embodiments merely illustrate the basic idea of the present invention by way of illustration, and the following embodiments and features in the embodiments may be combined with each other without conflict.
Wherein the drawings are for illustrative purposes only and are shown in schematic, non-physical, and not intended to limit the invention; for the purpose of better illustrating embodiments of the invention, certain elements of the drawings may be omitted, enlarged or reduced and do not represent the size of the actual product; it will be appreciated by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The same or similar reference numbers in the drawings of embodiments of the invention correspond to the same or similar components; in the description of the present invention, it should be understood that, if there are terms such as "upper", "lower", "left", "right", "front", "rear", etc., that indicate an azimuth or a positional relationship based on the azimuth or the positional relationship shown in the drawings, it is only for convenience of describing the present invention and simplifying the description, but not for indicating or suggesting that the referred device or element must have a specific azimuth, be constructed and operated in a specific azimuth, so that the terms describing the positional relationship in the drawings are merely for exemplary illustration and should not be construed as limiting the present invention, and that the specific meaning of the above terms may be understood by those of ordinary skill in the art according to the specific circumstances.
As shown in fig. 1, the invention provides a point cloud quality evaluation method based on relative entropy, which comprises the following steps:
s1: as shown in fig. 2, the point cloud is resampled by a high pass filter to obtain a geometric architecture.
The point cloud consists of N points, each with K attributes. Such point clouds are noted as
Figure BDA0004112126090000061
Using geometric distances
Figure BDA0004112126090000062
To calculate the sub-point cloud midpoint +.>
Figure BDA0004112126090000063
and />
Figure BDA0004112126090000064
The connection weight between the two is defined as follows:
Figure BDA0004112126090000065
the edge density of each vertex is measured using a diagonal matrix D. Point(s)
Figure BDA0004112126090000066
Is->
Figure BDA0004112126090000067
The edge density D is defined as follows:
Figure BDA0004112126090000068
the high pass filter through which the reference point cloud passes is denoted as
Figure BDA0004112126090000069
Where L is the filter length, h l Is the first filter coefficient. />
Figure BDA00041121260900000610
Is a sub-point cloud shift operator defined as:
A＝D -1 W
on the basis of h (A), selecting a Haar-like local area filter h HH (A) The high pass filtering is realized as follows:
Figure BDA00041121260900000611
wherein λi Is the eigenvalue of a, and U is the eigenvector of a.
Each point passing through the filter
Figure BDA0004112126090000071
The frequency response of (2) is:
Figure BDA0004112126090000072
let the sampling frequency of the filter be f s The total point number of the sampled point cloud
Figure BDA0004112126090000073
Resampled point cloud->
Figure BDA0004112126090000074
Is a geometric architecture.
Further, in the step S2, a euclidean distance threshold γ is given to the reference point cloud and the distorted point cloud
Figure BDA0004112126090000075
Geometric key of->
Figure BDA0004112126090000076
Constructing a sub-point cloud shown in fig. 3 by using a k-nearest neighbor search algorithm for the center, wherein the definition is as follows;
Figure BDA0004112126090000077
Figure BDA0004112126090000078
the points in the sub-point cloud are arranged from small to large according to the distance from the geometric key point.
S3, the specific step of regarding the relative entropy as the geometric quality result of the sub-point cloud comprises the following steps:
s31: from S r Selecting distance geometric key points
Figure BDA00041121260900000730
Nearest k r A point of which is k r The maximum distance of a point from a geometrical key point is denoted +.>
Figure BDA0004112126090000079
From S d Selecting k nearest to the geometric key point d A point of which is k d The maximum distance of a point from a geometric key point is recorded as
Figure BDA00041121260900000710
S32: at S r In (3) calculating
Figure BDA00041121260900000711
Every pair within the distance +.>
Figure BDA00041121260900000712
And its adjacent point->
Figure BDA00041121260900000713
Three elements of a point feature histogram of (a)
Figure BDA00041121260900000714
Then statistically, a simple point feature histogram +.>
Figure BDA00041121260900000715
At S d In (3) calculating
Figure BDA00041121260900000716
Every pair within the distance +.>
Figure BDA00041121260900000717
And its adjacent point->
Figure BDA00041121260900000718
Is a three-element point feature histogram>
Figure BDA00041121260900000719
Then statistically, a simple point feature histogram +.>
Figure BDA00041121260900000720
S33: at S r In determining each point
Figure BDA00041121260900000721
At->
Figure BDA00041121260900000722
Adjacent points within the distance and calculating each +.>
Figure BDA00041121260900000723
Is a three-element point feature histogram>
Figure BDA00041121260900000724
Then statistically, a simple point feature histogram +.>
Figure BDA00041121260900000725
At S d In determining each point
Figure BDA00041121260900000726
At->
Figure BDA00041121260900000727
Three elements of a Point characteristic histogram within a distance +.>
Figure BDA00041121260900000728
Then statistically, a simple point feature histogram +.>
Figure BDA00041121260900000729
/>
S44: the area of the fast point feature histogram affecting the first principal component is shown in fig. 4. Calculation S r and Sd Fast point feature histogram FPFH of (a) r and FPFHd The definition is as follows:
Figure BDA0004112126090000081
Figure BDA0004112126090000082
wherein di Is the euclidean distance between points.
S35: calculate FPFH r and FPFHd The relative entropy between the two to obtain a geometric quality evaluation index D i . The relative entropy between histograms is defined as:
Figure BDA0004112126090000083
further, the specific steps for obtaining the color quality result of the sub-point cloud in S4 are as follows:
s41: will S r and Sd Is converted from RGB space to LMN space.
Figure BDA0004112126090000084
S42: defining the zeroth moment, the first moment and the second moment as the total number m g Mean mu g Variance sigma g 。
S43: definition S r and Sd The 3 metric values of the color attribute of (a) are:
Figure BDA0004112126090000085
Figure BDA0004112126090000086
Figure BDA0004112126090000087
N 0 ,N 1 ,N 3 is three small non-zero numbers, preventing the denominator from being 0.
S44: multiplying the above 3 metrics to obtain S r and Sd Quality metrics of color attributes in between.
Figure BDA0004112126090000088
Further, in the step S5, the geometric quality result and the color quality result of the sub-point cloud are weighted and aggregated, and the final tensor point cloud quality evaluation model based on the relative entropy is obtained as follows:
Figure BDA0004112126090000089
wherein ,
Figure BDA00041121260900000810
and />
Figure BDA00041121260900000811
The weight factors of the geometric quality assessment index and the color quality assessment index are respectively.
Finally, it is noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made thereto without departing from the spirit and scope of the present invention, which is intended to be covered by the claims of the present invention.

Claims (6)

1. A point cloud quality assessment method based on relative entropy is characterized by comprising the following steps: the method comprises the following steps:
s1: downsampling the reference point cloud, and acquiring the geometric framework of the downsampled reference point cloud through a high-pass diagram filter;
s2: constructing sub-point clouds taking geometric architecture as center in the reference point cloud and the distorted point cloud by using a k nearest neighbor classification algorithm, and respectively recording as S r and Sd ；
S3: calculation S r and Sd Fast point feature histogram F of (1) r and Fd Then calculate F r and Fd The relative entropy is taken as a geometric quality result of the sub-point cloud;
S4：will S r and Sd The color attributes of the sub-point cloud are converted from an RGB color space to an LMN color space, the zeroth moment, the first moment and the second moment of the color space are calculated, and the color quality result of the sub-point cloud is obtained through weighted calculation;
s5: and weighting and aggregating the geometric quality result and the color quality result of the sub-point cloud to obtain a final point cloud quality evaluation model based on relative entropy.
2. The point cloud quality assessment method based on relative entropy according to claim 1, wherein: the step S1 comprises the following steps:
s11: the reference point cloud consists of N points, each point having K attributes, noted as:
Figure FDA0004112126080000011
wherein ,
Figure FDA0004112126080000012
represents the 1 st to N th points, +.>
Figure FDA0004112126080000013
Representing a space of dimension N x K;
s12: using geometric distances
Figure FDA00041121260800000113
To calculate the sub-point cloud midpoint +.>
Figure FDA0004112126080000015
and />
Figure FDA0004112126080000016
The connection weight between the two is defined as follows:
Figure FDA0004112126080000017
wherein ,σ2 Representing the variance between points, τ representing the euclidean distance threshold for clustering adjacent points into the same sub-point cloud;
s13: measuring edge density of each vertex using diagonal matrix D, points
Figure FDA0004112126080000018
Is->
Figure FDA0004112126080000019
The edge density D is defined as follows:
Figure FDA00041121260800000110
s14: the high pass filter through which the reference point cloud passes is expressed as:
Figure FDA00041121260800000111
where L is the filter length, h l Is the first filter coefficient;
Figure FDA00041121260800000112
is a sub-point cloud shift operator defined as:
A＝D -1 W
s15: on the basis of h (A), selecting a Haar-like local area filter h HH (A) The high pass filtering is realized as follows:
Figure FDA0004112126080000021
wherein λ1 ～λ N Is the characteristic value of A, U is the characteristic vector of A;
s16: each point passing through the filter
Figure FDA0004112126080000022
The frequency response of (2) is:
Figure FDA0004112126080000023
wherein
Figure FDA0004112126080000024
Representation dot->
Figure FDA0004112126080000025
Is a group of adjacent points->
Figure FDA0004112126080000026
Representation dot->
Figure FDA00041121260800000222
and />
Figure FDA00041121260800000223
Is a feature covariance matrix of (1);
let the sampling frequency of the filter be f s The total point number of the sampled point cloud
Figure FDA0004112126080000029
Resampled reference point cloud +.>
Figure FDA00041121260800000210
Is a geometric architecture.
3. The point cloud quality assessment method based on relative entropy according to claim 2, wherein: in step S2, a Euclidean distance threshold gamma is given, and resampled reference point clouds are used in the reference point clouds and the distorted point clouds
Figure FDA00041121260800000211
Geometric key of->
Figure FDA00041121260800000212
The sub-point cloud is built for the center using a k-nearest neighbor search algorithm, defined as follows:
Figure FDA00041121260800000213
Figure FDA00041121260800000214
wherein Sr Representing a geometrically centered sub-point cloud built in a reference point cloud,
Figure FDA00041121260800000215
representing points in the reference point cloud,
Figure FDA00041121260800000216
representing a resampled reference point cloud, S d Representing a geometrical architecture centered sub-point cloud built in a distorted point cloud, ++>
Figure FDA00041121260800000217
Representing points in a distorted point cloud, +.>
Figure FDA00041121260800000218
Representing the resampled distorted point cloud;
the points in the sub-point cloud are arranged from small to large according to the distance from the geometric key point.
4. The point cloud quality assessment method based on relative entropy according to claim 3, wherein: the specific step of using the relative entropy as the geometric quality result of the sub-point cloud in the step S3 includes:
s31: from the slaveS r Selecting distance geometric key points
Figure FDA00041121260800000219
Nearest k r A point of which is k r The maximum distance of a point from a geometrical key point is denoted +.>
Figure FDA00041121260800000220
From S d Selecting k nearest to the geometric key point d A point of which is k d The maximum distance of a point from a geometric key point is recorded as
Figure FDA00041121260800000221
S32: at S r In (3) calculating
Figure FDA0004112126080000031
Every pair within the distance +.>
Figure FDA0004112126080000032
And its adjacent point->
Figure FDA0004112126080000033
Is a three-element point feature histogram>
Figure FDA0004112126080000034
Then statistically, a simple point feature histogram +.>
Figure FDA0004112126080000035
At S d In (3) calculating
Figure FDA0004112126080000036
Every pair within the distance +.>
Figure FDA0004112126080000037
And its adjacent point->
Figure FDA0004112126080000038
Is a three-element point feature histogram>
Figure FDA0004112126080000039
Then statistically, a simple point feature histogram +.>
Figure FDA00041121260800000310
S33: at S r In determining each point
Figure FDA00041121260800000311
At->
Figure FDA00041121260800000312
Adjacent points within the distance and calculating each +.>
Figure FDA00041121260800000313
Is a three-element point feature histogram>
Figure FDA00041121260800000314
Then statistically, a simple point feature histogram +.>
Figure FDA00041121260800000315
At the position of
Figure FDA00041121260800000316
In (c) determining each point +.>
Figure FDA00041121260800000317
At->
Figure FDA00041121260800000318
Three elements of a Point characteristic histogram within a distance +.>
Figure FDA00041121260800000319
Then statistically, a simple point feature histogram +.>
Figure FDA00041121260800000320
S34: calculation S r and Sd Fast point feature histogram FPFH of (a) r and FPFHd The definition is as follows:
Figure FDA00041121260800000321
Figure FDA00041121260800000322
wherein di Is the Euclidean distance between points;
s35: calculate FPFH r and FPFHd The relative entropy between the two to obtain a geometric quality evaluation index D i The relative entropy between histograms is defined as:
Figure FDA00041121260800000323
5. the method for evaluating the quality of the point cloud based on the relative entropy according to claim 4, wherein the method comprises the following steps of: the calculating in the step S4 obtains the color quality result of the sub-point cloud, which specifically includes the following steps:
s41: will S r and Sd Is converted from RGB space to LMN space:
Figure FDA00041121260800000324
s42: definition of the zeroth moment, the first momentThe two moments are respectively the total number m g Mean mu g Variance sigma g ；
S43: definition S r and Sd The 3 metric values of the color attribute of (a) are:
Figure FDA0004112126080000041
Figure FDA0004112126080000042
Figure FDA0004112126080000043
N 0 ,N 1 ,N 3 three small non-zero numbers, preventing the denominator from being 0;
s44: multiplying the above 3 metrics to obtain S r and Sd Quality metric results for color attributes in between:
Figure FDA0004112126080000044
6. the method for evaluating the quality of the point cloud based on the relative entropy according to claim 5, wherein the method comprises the following steps of: in step S5, the geometric quality result and the color quality result of the weighted aggregate sub-point cloud are obtained, and the final point cloud quality evaluation model based on the relative entropy is obtained as follows:
Figure FDA0004112126080000045
wherein ,
Figure FDA0004112126080000046
and />
Figure FDA0004112126080000047
The weight factors of the geometric quality assessment index and the color quality assessment index are respectively. />
CN202310209556.0A 2023-03-07 2023-03-07 Point cloud quality assessment method based on relative entropy Pending CN116205886A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310209556.0A CN116205886A (en) 2023-03-07 2023-03-07 Point cloud quality assessment method based on relative entropy

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310209556.0A CN116205886A (en) 2023-03-07 2023-03-07 Point cloud quality assessment method based on relative entropy

Publications (1)

Publication Number Publication Date
CN116205886A true CN116205886A (en) 2023-06-02

Family

ID=86512634

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310209556.0A Pending CN116205886A (en) 2023-03-07 2023-03-07 Point cloud quality assessment method based on relative entropy

Country Status (1)

Country Link
CN (1) CN116205886A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117011299A (en) * 2023-10-07 2023-11-07 华侨大学 Reference point cloud quality assessment method and system integrating graph resampling and gradient characteristics
CN117789198A (en) * 2024-02-28 2024-03-29 上海几何伙伴智能驾驶有限公司 Method for realizing point cloud degradation detection based on 4D millimeter wave imaging radar

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117011299A (en) * 2023-10-07 2023-11-07 华侨大学 Reference point cloud quality assessment method and system integrating graph resampling and gradient characteristics
CN117011299B (en) * 2023-10-07 2024-02-20 华侨大学 Reference point cloud quality assessment method and system integrating graph resampling and gradient characteristics
CN117789198A (en) * 2024-02-28 2024-03-29 上海几何伙伴智能驾驶有限公司 Method for realizing point cloud degradation detection based on 4D millimeter wave imaging radar
CN117789198B (en) * 2024-02-28 2024-05-14 上海几何伙伴智能驾驶有限公司 Method for realizing point cloud degradation detection based on 4D millimeter wave imaging radar

Similar Documents

Publication Publication Date Title
CN116205886A (en) Point cloud quality assessment method based on relative entropy
CN115082467B (en) Building material welding surface defect detection method based on computer vision
CN109284786B (en) SAR image terrain classification method for generating countermeasure network based on distribution and structure matching
CN101976444B (en) Pixel type based objective assessment method of image quality by utilizing structural similarity
CN109035196B (en) Saliency-based image local blur detection method
CN103955709B (en) Weighted synthetic kernel and triple markov field (TMF) based polarimetric synthetic aperture radar (SAR) image classification method
CN108960142B (en) Pedestrian re-identification method based on global feature loss function
CN105976351B (en) Stereo image quality evaluation method based on central offset
CN109788275A (en) Naturality, structure and binocular asymmetry are without reference stereo image quality evaluation method
CN106651834A (en) Method and device for evaluating quality of substation equipment infrared thermal image with no reference image
CN110059713A (en) Precipitation type identification method based on precipitation particle multi-feature parameters
CN111507426A (en) No-reference image quality grading evaluation method and device based on visual fusion characteristics
CN111783589A (en) Complex scene crowd counting method based on scene classification and multi-scale feature fusion
CN112184672A (en) No-reference image quality evaluation method and system
Mahajan et al. Perceptual quality evaluation of hazy natural images
CN105654452A (en) Background clutter measurement method based on edge structures
CN114331989A (en) Full-reference 3D point cloud quality evaluation method based on point feature histogram geodesic distance
CN108682005B (en) Semi-reference 3D synthetic image quality evaluation method based on covariance matrix characteristics
Lv et al. Blind dehazed image quality assessment: a deep CNN-based approach
CN110910497B (en) Method and system for realizing augmented reality map
Yan et al. No-reference remote sensing image quality assessment based on gradient-weighted natural scene statistics in spatial domain
CN113408342B (en) Target detection method for determining intersection ratio threshold based on features
CN114693743A (en) Data cleaning method for target tracking training data set
CN112801950A (en) Image adaptation quality evaluation method based on geometric distortion measurement
CN109635874A (en) A kind of Image Blind evaluation method and device of characteristics dictionary mapping

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