High-precision digital scanning system based on GIS+BIM technology

Abstract

The application relates to the technical field of data processing, in particular to a high-precision digital scanning system based on GIS+BIM technology, which comprises the following modules: a point cloud data acquisition module, which is used for acquiring a point cloud data set obtained by scanning any city building each time; a point cloud data denoising module, which is used for simplifying point cloud data as points, acquiring the same attribute quantity and a first filtering radius of each point, acquiring the similarity between any two points, acquiring a second filtering radius of each point, obtaining the matching degree between each point cloud data set and each other point cloud data set, acquiring a third filtering radius of each point in each point cloud data set, and performing denoising processing on each point cloud data set; and a digital scanning model generation module, which is used for generating a digital scanning model by adopting GIS+BIM technology. The application aims to improve the denoising effect of point cloud data, and further improve the precision of a digital scanning model generated by adopting GIS+BIM technology subsequently.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of data processing, in particular to a high-precision digital scanning system based on GIS+BIM technology. BACKGROUND

[0002] With the acceleration of urbanization and the continuous development of infrastructure construction, the demand for efficient and accurate management and decision support is growing. Traditional GIS platforms are good at processing macro geographic spatial data such as terrain, landform, and transportation network, but they lack in expressing and managing micro building level information. BIM models are excellent in fine modeling and information integration of individual buildings, but they lack the ability to grasp the macro view of large-scale geographic space. Therefore, it is necessary to integrate GIS and BIM technology to build a high-precision digital scanning model.

[0003] However, the collected point cloud data often introduces noise points due to factors such as laser reflection, sensor error, or environmental interference, affecting subsequent tasks such as terrain modeling and target recognition. Currently, when using the radius filtering algorithm to denoise point cloud data, a fixed filtering radius is often used, which does not fully consider the distribution density changes and intensity information of point cloud data, resulting in poor denoising effect and affecting the accuracy of the digital scanning model. SUMMARY

[0004] In view of the above, it is necessary to provide a high-precision digital scanning system based on GIS+BIM technology. Compared with traditional GIS+BIM-based digital scanning systems, the system improves the denoising effect of point cloud data, and thus improves the accuracy of the digital scanning model generated by subsequent GIS+BIM technology.

[0005] The high-precision digital scanning system based on GIS+BIM technology of the application adopts the following technical solutions:

[0006] One embodiment of the application provides a high-precision digital scanning system based on GIS+BIM technology, which comprises:

[0007] A point cloud data acquisition module for acquiring a point cloud data set obtained by each scan of any city building, wherein the point cloud data has an intensity value;

[0008] The point cloud data denoising module is configured to, for each point cloud data set and each point cloud data therein, obtain a same attribute quantity of each point by intensity values of points in a preset neighborhood of each point, obtain a first filtering radius of each point by distribution of the same attribute quantities of all points and in combination with a total number of points in the preset neighborhood of each point, obtain a similarity between any two points by a number of common points and intensity value difference in a preset neighborhood between the any two points, and obtain a second filtering radius of each point by the similarity between each point and each preset neighbor point of the each point and the first filtering radius of each preset neighbor point.

[0009] The feature points in each point cloud data set are obtained respectively, a matching factor between each feature point in each point cloud data set and each feature point in each other point cloud data set is obtained by spatial distance, intensity value difference and same attribute quantity difference between the each feature point and the each feature point, and a matching degree between each point cloud data set and each other point cloud data set is obtained, a third filtering radius of each point in each point cloud data set is obtained in combination with the first filtering radius of points in a preset neighborhood of each point in each point cloud data set in each other point cloud data set, and each point cloud data set is denoised in combination with the first filtering radius and the second filtering radius of each point in each point cloud data set.

[0010] The digital scanning model generation module is configured to generate a digital scanning model by using GIS+BIM technology according to the denoised point cloud data set.

[0011] In one embodiment, the same attribute quantity obtaining includes:

[0012] The clustering algorithm is used to divide all points into classes according to intensity values of all points for each point cloud data set.

[0013] The same attribute quantity is a number of points in a preset neighborhood of each point and belonging to a same class as each point.

[0014] In one embodiment, the first filtering radius obtaining includes:

[0015] The number of points in a preset neighborhood of each point is counted.

[0016] The anomaly detection algorithm is used to obtain an anomaly score of the same attribute quantity of each point by the same attribute quantities of all points in a class to which each point belongs.

[0017] The sum of the same attribute quantity of each point and the number is calculated, and a product of the sum and the anomaly score is calculated.

[0018] The first filtering radius is directly proportional to a preset filtering radius and inversely proportional to the product.

[0019] In one embodiment, the first filtering radius is calculated by mapping the product to a first positive number, and the first filtering radius is a ratio of the preset filtering radius and the first positive number.

[0020] In one embodiment, the similarity is obtained by:

[0021] calculating a difference between the intensity values of the two points, and mapping the difference to a second positive number;

[0022] the similarity is a normalized value of a ratio of the number of common points and the second positive number.

[0023] In one embodiment, the second filtering radius is obtained by:

[0024] a weighted sum of the first filtering radius of all points in a preset neighborhood of each point is taken as the second filtering radius of each point, wherein the weight of the first filtering radius of each point is a normalized value of the similarity corresponding to each point, and the sum of the weights of the first filtering radius of all points is 1.

[0025] In one embodiment, the matching factor is obtained by:

[0026] calculating a difference between the intensity values of the feature points and each feature point, and taking an inverse of a sum of the difference, the spatial distance and a preset positive number as a first inverse;

[0027] calculating a deviation between the number of same attributes of the feature points and each feature point, and taking an inverse of a sum of the deviation and a preset positive number as a second inverse;

[0028] the matching factor is a product of the first inverse and the second inverse.

[0029] In one embodiment, the matching degree is obtained by:

[0030] all the matching factors obtained between each point cloud data set and each of the remaining point cloud data sets are taken as input of the Hungarian algorithm to obtain all the matched feature point pairs between each point cloud data set and each of the remaining point cloud data sets, and an average of the matching factors between the all the matched feature point pairs is taken as the matching degree between each point cloud data set and each of the remaining point cloud data sets.

[0031] In one embodiment, the third filtering radius is obtained by:

[0032] The average value of the first filtering radius of each point in each point cloud data set in the preset neighborhood of all points in each of the remaining point cloud data sets is calculated, and a product value of the matching degree and the average value is calculated, and the cumulative value of the product value between each point in each point cloud data set and all the remaining point cloud data sets is taken as the third filtering radius of each point in each point cloud data set.

[0033] In one of the embodiments, the denoising processing of the point cloud data sets comprises:

[0034] The average value of the first filtering radius, the second filtering radius and the third filtering radius of each point in each point cloud data set is taken as the filtering radius of the radius filtering algorithm for denoising processing of each point in each point cloud data set.

[0035] The present application has at least the following beneficial effects:

[0036] The present application considers the distribution density of the point cloud data and the similarity of the intensity values of the point cloud data and the neighboring point cloud data by calculating the first filtering radius, and can dynamically calculate the first filtering radius according to the actual situation of the neighboring area of each point cloud data, which is beneficial to improve the calculation accuracy of the filtering radius in the subsequent radius filtering algorithm.

[0037] Further, considering that the distribution density of the point cloud data has the characteristic of continuous change, the common neighborhood point number and the intensity value difference of the neighboring point cloud data are introduced to provide a reliable weight value for the first filtering radius of the point cloud data, and to ensure that the filtering radius can change continuously in space.

[0038] Further, for the point cloud data sets obtained by multiple scans, the first filtering radius calculated is adjusted through mutual verification of the multiple point cloud data sets, and then the first filtering radius, the second filtering radius and the third filtering radius are comprehensively considered to obtain the final filtering radius of the radius filtering algorithm, so that the denoising processing can adapt to various complex scenes, improve the denoising effect of the point cloud data, and further improve the accuracy of the digital scanning model generated by using the GIS+BIM technology. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed in the following embodiment or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0040] Figure 1 The block diagram of the high-precision digital scanning system based on GIS+BIM technology is provided.

[0041] Figure 2 A flowchart for obtaining a filtering radius of a radius filtering algorithm;

[0042] Figure 3 A flowchart of a high-precision digital scanning system based on GIS+BIM technology. DETAILED DESCRIPTION

[0043] In the description of the embodiments of the present application, the words "exemplary", "or", "for example" are used to mean as an example, instance, or illustration. Any embodiment or design presented as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or advantageous than other embodiments or design solutions. In fact, the use of "exemplary", "or", "for example" is intended to present concepts in a specific way.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. It is to be understood that the use of "or" in the present application is meant to be an "or" inclusive of the possibilities of "and" unless otherwise stated.

[0045] It should also be noted that the terms "first", "second" and "third" in the present application are used to distinguish similar objects, and are not intended to describe a specific order or sequence.

[0046] The specific scheme of the high-precision digital scanning system based on GIS+BIM technology provided by the present application will be described in detail below in combination with the drawings.

[0047] Please refer to Figure 1 , which shows a block diagram of a high-precision digital scanning system based on GIS+BIM technology provided by an embodiment of the present application, which includes a point cloud data acquisition module 101, a point cloud data denoising module 102, and a digital scanning model generation module 103.

[0048] The point cloud data acquisition module 101 is configured to acquire a point cloud data set obtained by each scanning of any urban building, wherein the point cloud data set includes intensity values of point cloud data.

[0049] The point cloud data set obtained by each scanning of any urban building is acquired from a GIS platform, wherein the point cloud data set is obtained by laser radar scanning and stored on the GIS platform, and the format of the point cloud data in the point cloud data set is (x, y, z, H), x, y, and z represent three-dimensional coordinates of the point cloud data, and H represents an intensity value of the point cloud data, specifically an intensity value of a return signal received by the laser radar.

[0050] The point cloud data denoising module 102 is configured to denoise each point cloud data set by analyzing the distribution density and intensity value of the point cloud data in each point cloud data set.

[0051] The existing radius filtering algorithm usually adopts a fixed filtering radius when denoising the point cloud data, and thus, the phenomenon of false deletion or omission is likely to occur due to the difference in the density or intensity value distribution, specifically, if the fixed filtering radius is too small, the dense point cloud is likely to be misjudged as noise point cloud, and if the fixed filtering radius is too large, the real noise is likely to be not identified; for the point cloud data with similar intensity values but far spatial distance from the remaining point cloud data, the point cloud data is likely to be isolated object data, and is easily misdeleted during denoising, and the noise point cloud with large intensity value difference but dense spatial distribution is likely to be retained. Thus, the completeness and accuracy of the point cloud data are affected, and therefore, when filtering the point cloud data, the point cloud data needs to be denoised according to the distribution density of the point cloud data and the intensity information of the point cloud data.

[0052] (1) For each point cloud data set and each point cloud data therein, the point cloud data is simply referred to as a point, the number of points with the same attribute of each point is obtained through the intensity values of the points in the preset neighborhood of each point, and the first filtering radius of each point is obtained through the distribution of the number of points with the same attribute of all points and the total number of points in the preset neighborhood of each point.

[0053] First, taking the i th point cloud data set and the j th point cloud data therein as an example, the number of point cloud data in the preset neighborhood of the j th point cloud data in the i th point cloud data set is taken as the distribution density factor of the j th point cloud data in the i th point cloud data set.

[0054] In this embodiment, the preset neighborhood refers to a spherical neighborhood with the j th point cloud data as the center and a radius of 15, wherein 15 is only one embodiment of the present application, and the implementer can set the specific value according to the actual situation.

[0055] At the same time, in order to better analyze the distribution of the point cloud data in the i th point cloud data set, the intensity value of the point cloud data needs to be analyzed.

[0056] Based on the above analysis, the intensity values of all point cloud data in the i-th point cloud data set are clustered, all point cloud data in the i-th point cloud data set is divided into classes, and the number of point cloud data in the same class as the j-th point cloud data in the i-th point cloud data set in the preset neighborhood of the j-th point cloud data is taken as the number of the same attribute of the j-th point cloud data in the i-th point cloud data set. The number of the same attribute can reflect the similarity between the j-th point cloud data and its neighboring point cloud data in terms of intensity value. The greater the number of the same attribute, the greater the similarity between the j-th point cloud data and its neighboring point cloud data in terms of intensity value. Conversely, the smaller the number of the same attribute, the smaller the similarity between the j-th point cloud data and its neighboring point cloud data in terms of intensity value.

[0057] In this embodiment, the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm is used to divide all point cloud data in the i-th point cloud data set into classes, wherein the neighborhood radius in the DBSCAN clustering algorithm is 3, the number of neighborhood samples is 20, and the DBSCAN clustering algorithm is a known technology, which will not be described herein. As other embodiments, as long as the all point cloud data in the i-th point cloud data set can be divided into classes, the implementer can use other existing feasible clustering algorithms, and the present application does not make special limitations.

[0058] Further, the number of the same attribute of all point cloud data in the class in which the j-th point cloud data is located is taken as the input of the anomaly detection algorithm, and the anomaly score of the number of the same attribute of the j-th point cloud data is output. The greater the anomaly score, the more likely the j-th point cloud data is a noise point cloud. Therefore, when denoising the j-th point cloud data, a smaller filtering radius should be used to remove the noise point cloud.

[0059] Based on the above analysis, the first filtering radius of each point cloud data in each point cloud data set is obtained through the distribution density factor, the number of the same attribute, and the anomaly score of the number of the same attribute of each point cloud data in each point cloud data set, and the expression is:

[0060] ; in the formula, represents the first filtering radius of the j-th point cloud data in the i-th point cloud data set; r represents a preset filtering radius; represents the anomaly score of the number of the same attribute of the j-th point cloud data in the i-th point cloud data set; represents the number of the same attribute of the j-th point cloud data in the i-th point cloud data set; represents the distribution density factor of the j-th point cloud data in the i-th point cloud data set; represents a preset positive number, which is used to avoid the denominator being 0, The value of the preset filtering radius is set by a person, and the implementer can set it by himself / herself. In this embodiment, the value of the preset filtering radius is 15, and the value of the preset filtering radius is set by a person. The implementer can set it according to the actual situation, and the application does not make special limitations. The value of the first filtering radius is 0.01. In order to avoid the denoising effect of the point cloud data being poor due to the first filtering radius being too large or too small, when the value of the first filtering radius is greater than 100, the value of the first filtering radius is set to 100, and when the value of the first filtering radius is less than 3, the value of the first filtering radius is set to 3. The value of the second filtering radius is recorded as a first positive number. The value of the second filtering radius is recorded as a first positive number.

[0061] In this embodiment, the value of the preset filtering radius is 15, and the value of the preset filtering radius is set by a person. The implementer can set it according to the actual situation, and the application does not make special limitations.

[0062] It should be noted that: when the distribution density of the point cloud data in the neighborhood range of the jth point cloud data is greater, and the intensity values of the point cloud data are closer to each other, a smaller filtering radius is used in the denoising process to retain the real point cloud data and remove noise data, and the first filtering radius is smaller; when the anomaly score of the same attribute number of the jth point cloud data is greater, the jth data is more likely to be noise data, and a smaller filtering radius needs to be used to remove the jth point cloud data, and the first filtering radius is smaller.

[0063] (2) Obtain the similarity between the two points by the number of common points in the preset neighborhood and the intensity value difference between the two points. Obtain the second filtering radius of each point by the similarity between each point and each preset neighbor point and the first filtering radius of each preset neighbor point.

[0064] Generally, the distribution density of the point cloud data is continuously changed, so the filtering radius should also be continuously changed when using the radius filtering algorithm for denoising.

[0065] Based on the above analysis, the number of common point cloud data in the preset neighborhood between any two point cloud data in the ith point cloud data set is counted, and the difference between the intensity values of the two point cloud data is calculated. The difference is mapped to a second positive number, and the ratio of the number of common point cloud data to the second positive number is taken as the similarity between the two point cloud data.

[0066] In this embodiment, the difference between the intensity values is the absolute value of the difference, and other embodiments can use other calculation methods, such as the square of the difference, to measure the difference between the intensity values, and the application does not make special limitations.

[0067] Further, the second filtering radius of each point cloud data in each point cloud data set is obtained by the similarity between each point cloud data and each preset neighbor point cloud data in each point cloud data set and the first filtering radius of each preset neighbor point cloud data. Specifically,

[0068] The weighted sum of the first filtering radius of all point cloud data in the preset neighborhood of the jth point cloud data is taken as the second filtering radius of the jth point cloud data, wherein the weight of the first filtering radius of each point cloud data is the normalized value of the similarity corresponding to each point cloud data, and the sum of the weights of the first filtering radius of all point cloud data in the preset neighborhood of the jth point cloud data is 1.

[0069] (3) The feature points in each point cloud data set are obtained respectively, and the matching factor between each feature point in each point cloud data set and each feature point in the remaining point cloud data set is obtained according to the spatial distance, intensity value difference and same attribute number difference between each feature point in each point cloud data set and each feature point in the remaining point cloud data set, and then the matching degree between each point cloud data set and the remaining point cloud data set is obtained. The third filtering radius of each point in each point cloud data set is obtained in combination with the first filtering radius of the points in the preset neighborhood of each point in each point cloud data set in the remaining point cloud data set.

[0070] Since the scanning of the any urban building is a multi-time and multi-directional scanning process, a plurality of point cloud data sets are obtained from the same direction for the same position of the any urban building, and therefore the filtering radius of the point cloud data can be comprehensively analyzed through the point cloud data sets obtained by different scanning times.

[0071] Firstly, the similarity between the point cloud data sets obtained by different scanning times needs to be analyzed, specifically as follows: the feature points in each point cloud data set are obtained respectively, and the matching factor between each feature point in each point cloud data set and each feature point in the remaining point cloud data set is obtained according to the spatial distance, intensity value difference and same attribute number difference between each feature point in each point cloud data set and each feature point in the remaining point cloud data set, and then the matching degree between each point cloud data set and the remaining point cloud data set is obtained. The specific process of obtaining the matching degree is as follows:

[0072] The feature points in each point cloud data set are obtained respectively by using a feature point detection algorithm. Taking the wth feature point in the ith point cloud data set and the th feature point in the Tth point cloud data set as examples, the difference value of the intensity values between the wth feature point and the th feature point is calculated, and the spatial distance between the wth feature point and the th feature point is calculated. The reciprocal of the sum of the difference value, the spatial distance and a preset first value greater than 0 is denoted as a first reciprocal. At the same time, the deviation value of the same attribute number between the wth feature point and the th feature point is calculated. The reciprocal of the sum of the deviation value and a preset second value greater than 0 is denoted as a second reciprocal. Further, the product of the first reciprocal and the second reciprocal is taken as the matching factor between the wth feature point and the th feature point.

[0073] The feature matching factor matrix between the ith point cloud data set and the Tth point cloud data set is constructed by all the matching factors obtained between the ith point cloud data set and the Tth point cloud data set, wherein the number of rows of the matrix is equal to the number of feature points in the ith point cloud data set, the number of columns is equal to the number of feature points in the Tth point cloud data set, and the element in the ath row and the bth column of the matrix represents the matching factor between the ath feature point in the ith point cloud data set and the bth feature point in the Tth point cloud data set; according to the feature matching factor matrix, the Hungarian algorithm is used to obtain all the matched feature point pairs between the ith point cloud data set and the Tth point cloud data set, and the average value of the matching factors between the matched feature point pairs is taken as the matching degree between the ith point cloud data set and the Tth point cloud data set. The Hungarian algorithm is a known technology, and will not be described here.

[0074] In this embodiment, the feature point detection algorithm is a SIFT (Scale-Invariant Feature Transform) key point detection algorithm, and the SIFT key point detection algorithm is a known technology, which will not be described here. As other embodiments, on the basis of being able to realize the acquisition of feature points in the point cloud data set, the implementer can use other existing technologies, such as ISS (Intrinsic Shape Signatures) key point detection algorithm, and the present application does not make special limitation.

[0075] In this embodiment, the difference value between the intensity values is the absolute value of the difference, and as other embodiments, on the basis of being able to measure the difference between the intensity values, the implementer can use other calculation methods, such as the square of the difference, and the present application does not make special limitation.

[0076] In this embodiment, the deviation value between the same attribute numbers is the absolute value of the difference, and as other embodiments, on the basis of being able to measure the difference between the same attribute numbers, the implementer can use other calculation methods, such as the square of the difference, and the present application does not make special limitation.

[0077] In this embodiment, the values of the first number greater than 0 and the second number greater than 0 are both 0.01, and the purpose is to avoid the denominator being 0. The values of the first number greater than 0 and the second number greater than 0 are both preset by human, and the implementer can set them by himself, and the present application does not make special limitation.

[0078] Further, the third filtering radius of each point cloud data in each point cloud data set is obtained by the matching degree between each point cloud data set and each point cloud data set, and the first filtering radius of the point cloud data in the preset neighborhood of each point cloud data in each point cloud data set in each point cloud data set, which is specifically:

[0079] The jth point cloud data in the ith point cloud dataset is substituted into the Tth point cloud dataset, and each point cloud data in the preset neighborhood of the jth point cloud data in the Tth point cloud dataset is recorded as a respective reference point cloud data of the jth point cloud data;

[0080] The expression of the third filtering radius of the jth point cloud data in the ith point cloud dataset is:

[0081] In the formula, The third filtering radius of the jth point cloud data in the ith point cloud dataset is represented; M represents the number of point cloud datasets; The matching degree between the ith point cloud dataset and the Tth point cloud dataset is represented. The average value of the first filtering radius of all reference point cloud data of the jth point cloud data in the ith point cloud dataset in the Tth point cloud dataset is represented.

[0082] Further, the point cloud dataset is denoised by using a radius filtering algorithm, wherein the average value of the first filtering radius, the second filtering radius and the third filtering radius of the jth point cloud data in the ith point cloud dataset is used as the filtering radius when the jth point cloud data in the ith point cloud dataset is denoised by using the radius filtering algorithm. The radius filtering algorithm is a known technology, and will not be described herein. The acquisition process of the filtering radius of the radius filtering algorithm is shown in Figure 2 .

[0083] The digital scanning model generation module 103 is configured to generate a digital scanning model by using GIS+BIM technology according to the point cloud dataset after denoising.

[0084] According to the point cloud dataset after denoising, a digital scanning model is generated by using GIS+BIM technology, and the specific process is as follows:

[0085] The coordinate system of the BIM model of any city building is aligned with the coordinate system of the GIS platform by using a three-dimensional similarity transformation method;

[0086] The coordinates of the BIM model are converted into the coordinate system of the GIS platform by using a seven-parameter conversion model;

[0087] The point cloud dataset after denoising is fused with the BIM model to construct a complete three-dimensional scene; a feature point-based matching algorithm is used to ensure the accurate matching of the BIM model and the GIS data in the spatial position;

[0088] The point cloud dataset after denoising, the BIM model and other related geographic spatial data, such as terrain data, are integrated into a unified three-dimensional scene to form a high-precision digital scanning model;

[0089] The generated high-precision digital scanning model is visualized and analyzed using professional GIS and BIM software.

[0090] Among them, the three-dimensional similarity transformation method, the seven-parameter conversion model and the matching algorithm based on feature points are all known technologies, and the present application will not be repeated. The flowchart of the high-precision digital scanning system based on GIS+BIM technology is shown in Figure 3

[0091] In summary, the present application considers the distribution density of point cloud data, the similarity degree of the intensity values of point cloud data and neighboring point cloud data by calculating the first filtering radius, and can dynamically calculate the first filtering radius according to the actual situation of the neighborhood region of each point cloud data, which is beneficial to improve the calculation accuracy of the filtering radius in the subsequent radius filtering algorithm.

[0092] Further, considering that the distribution density of point cloud data has the characteristics of continuous change, by introducing the common neighborhood point number and intensity value difference of adjacent point cloud data, a reliable weight value can be provided for the first filtering radius of the point cloud data, and the filtering radius can be continuously changed in space.

[0093] Further, for the point cloud data sets obtained by multiple scans, the first filtering radius calculated is adjusted by mutual verification of multiple point cloud data sets, and then the first filtering radius, the second filtering radius and the third filtering radius are integrated to obtain the final filtering radius of the radius filtering algorithm, so that the denoising processing can adapt to various complex scenes, improve the denoising effect of the point cloud data, and further improve the precision of the digital scanning model generated by using GIS+BIM technology subsequently.

[0094] ​The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flow diagrams and / or block diagrams.

[0095] It is apparent that a person skilled in the art can make a variety of modifications to the application described above without departing from the spirit and scope of the application. Therefore, the above-described embodiments of the application are intended to be illustrative only and not restrictive.

Claims

1. A high-precision digital scanning system based on GIS+BIM technology, characterized in that, The system comprises: a point cloud data acquisition module, configured to acquire a point cloud data set obtained by each scan of any city building, wherein there is an intensity value of the point cloud data; a point cloud data denoising module, configured to, for any point cloud data set and each point cloud data in the point cloud data set, acquire a same attribute quantity of each point by using intensity values of points in a preset neighborhood of each point, acquire a first filtering radius of each point by using a distribution of the same attribute quantities of all points and combining a total number of points in the preset neighborhood of each point, and acquire a similarity between any two points by using a number of common points in a preset neighborhood between the two points and an intensity value difference; and acquire a second filtering radius of each point by using the similarity between each point and each preset neighbor point of the point and the first filtering radius of each preset neighbor point; acquire feature points in each point cloud data set respectively, acquire a matching factor between each feature point in each point cloud data set and each feature point in each other point cloud data set by using a spatial distance, an intensity value difference and a same attribute quantity difference between the feature points, and further acquire a matching degree between each point cloud data set and each other point cloud data set, acquire a third filtering radius of each point in each point cloud data set by using the first filtering radius of each point in the preset neighborhood of each point in each other point cloud data set, and perform denoising processing on each point cloud data set by combining the first filtering radius and the second filtering radius of each point in each point cloud data set; a digital scanning model generation module, configured to generate a digital scanning model by using GIS+BIM technology according to the point cloud data set after the denoising processing; the acquisition of the same attribute quantity comprises: for the any point cloud data set, all points are divided into classes by using a clustering algorithm according to intensity values of the points; the same attribute quantity is a number of points in a preset neighborhood of each point and belonging to a same class as the point; the acquisition of the first filtering radius comprises: counting a number of points in a preset neighborhood of each point; acquiring an abnormal score of the same attribute quantity of each point by using an abnormality detection algorithm according to the same attribute quantity of all points in a class to which each point belongs; calculating a sum value of the same attribute quantity of each point and the number; and calculating a product of the sum value and the abnormal score; the first filtering radius is directly proportional to a preset filtering radius and inversely proportional to the product, and the preset filtering radius is 15; the acquisition of the second filtering radius comprises: using a weighted sum of the first filtering radii of all points in a preset neighborhood of each point as the second filtering radius of the point, wherein a weight value of the first filtering radius of each point is a normalized value of a corresponding similarity of the point, and a sum of the weight values of the first filtering radii of all points is 1; the acquisition of the third filtering radius comprises: calculating an average value of the first filtering radii of all points in a preset neighborhood of each point in each point cloud data set in each other point cloud data set; calculating a product value of the matching degree and the average value; and using an accumulated value of the product value between each point in each point cloud data set and all other point cloud data sets as the third filtering radius of each point in each point cloud data set.

2. The high precision digital scanning system based on GIS+BIM technology as claimed in claim 1, wherein, The first filtering radius is calculated by mapping the product to a first positive number, and the first filtering radius is a ratio of the preset filtering radius and the first positive number.

3. The high precision digital scanning system based on GIS+BIM technology as claimed in claim 1, wherein, The similarity is obtained by: calculating a difference between intensity values of the two points, and mapping the difference to a second positive number; the similarity is a normalized value of a ratio of the number of common points and the second positive number.

4. The high precision digital scanning system based on GIS+BIM technology as claimed in claim 1, wherein, The matching factor is obtained by: calculating a difference between intensity values of the feature points and each of the feature points, and mapping an inverse of a sum of the difference, the spatial distance and a preset positive number to a first inverse; calculating a deviation between the number of same attributes of the feature points and each of the feature points, and mapping an inverse of a sum of the deviation and a preset positive number to a second inverse; the matching factor is a product of the first inverse and the second inverse.

5. The high precision digital scanning system based on GIS+BIM technology as claimed in claim 1, wherein, The matching degree is obtained by: using the Hungarian algorithm on all the matching factors between each point cloud data set and each of the remaining point cloud data sets to obtain all matching feature point pairs between each point cloud data set and each of the remaining point cloud data sets, and taking an average of the matching factors between the all matching feature point pairs as the matching degree between each point cloud data set and each of the remaining point cloud data sets.

6. The high precision digital scanning system based on GIS+BIM technology as claimed in claim 1, wherein, The denoising processing of the point cloud data sets comprises: using a radius filtering algorithm to perform denoising processing on the point cloud data sets, wherein an average of the first filtering radius, the second filtering radius and the third filtering radius of each point in each point cloud data set is taken as a filtering radius of the radius filtering algorithm for performing denoising processing on each point in each point cloud data set.

Images