An interactive point cloud data editing method on a mobile terminal

Abstract

The application discloses an interactive point cloud data editing method on a mobile terminal and relates to the technical field of point cloud data processing, which comprises the following steps: collecting point cloud data in a transformer substation by using a handheld laser radar, performing preliminary processing on the point cloud data by using a mobile terminal, and segmenting the point cloud data by using a fragmentation algorithm; transmitting the fragmented point cloud data to an edge computing device through a high-speed wireless network, filtering noise of the point cloud data on the edge computing device by using an adaptive learning algorithm, and dynamically adjusting filtering parameters; extracting geometric features of the point cloud data filtered by noise on the edge computing device, extracting geometric feature points and performing compression processing; receiving the compressed geometric feature point data returned from the edge computing device by the mobile terminal, decompressing and reconstructing the point cloud data by using an inverse sparse matrix decomposition algorithm, and enabling a user to perform interactive editing operation by using the mobile terminal; and adjusting a display perspective of a point cloud model by the user through a gesture operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of point cloud data processing, and particularly relates to an interactive point cloud data editing method on a mobile terminal. BACKGROUND

[0002] With the rapid development of three-dimensional scanning technology and sensor technology, point cloud data collection and processing technology has been widely applied in many fields. Point cloud data is an important three-dimensional space information expression method and is widely used in building information modeling (BIM), automatic driving, virtual reality, geographic information system (GIS) and industrial detection field. In particular, the progress of handheld laser radar (LiDAR) technology makes it possible to quickly obtain high-precision three-dimensional point cloud data in a complex environment. In the maintenance and management of power systems, the demand for three-dimensional modeling of some key infrastructures in substations is increasingly prominent. Point cloud data can clearly express the spatial position, size and mutual relationship of equipment. However, due to the large amount of point cloud data, many noise points and serious data redundancy, how to efficiently process and edit point cloud data and realize real-time interactive operation on a mobile terminal has become an important issue in current technical development.

[0003] Existing point cloud data processing technology mainly focuses on desktop computing platforms, and massive point cloud data is processed through high-performance computing resources. However, with the improvement of mobile device performance and the popularization of 5G communication network, more application scenarios begin to shift to mobile terminals. However, the existing technology has many limitations when extending point cloud data processing to mobile terminals, mainly in the following aspects: first, the computing resources of mobile terminals are limited and cannot efficiently process large-scale point cloud data; second, the noise filtering and compression algorithm of point cloud data is relatively complex, and the existing method is difficult to realize real-time requirements on a mobile terminal; third, the existing technology fails to fully utilize the advantages of edge computing in the data transmission process, resulting in low efficiency of data interaction between the mobile terminal and the server and failing to meet the demand of real-time operation. In addition, in the editing process of point cloud data, how to realize the rapid adjustment and accurate labeling of three-dimensional models through intuitive interactive methods (such as gesture operation) is also a difficult problem to be solved in current technology. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides an interactive point cloud data editing method on a mobile terminal to solve the problem that the noise filtering and compression algorithm of point cloud data is relatively complex and the existing method is difficult to realize real-time requirements on a mobile terminal.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides an interactive point cloud data editing method on a mobile terminal, comprising,

[0008] Point cloud data is collected in a substation using a handheld laser radar, and the point cloud data is preliminarily processed by a mobile terminal, and the point cloud data is segmented using a fragmentation algorithm;

[0009] The fragmented point cloud data is transmitted to an edge computing device through a high-speed wireless network, and an adaptive learning algorithm is used to filter noise from the point cloud data on the edge computing device, and the filtering parameters are dynamically adjusted;

[0010] On the edge computing device, the point cloud data filtered by noise is subjected to geometric feature extraction, and the geometric feature points are extracted and compressed;

[0011] The mobile terminal receives the compressed geometric feature point data returned from the edge computing device, decompresses and reconstructs the point cloud data through an inverse sparse matrix decomposition algorithm, and obtains complete three-dimensional point cloud data;

[0012] The decompressed and reconstructed point cloud data is displayed in real time on the mobile terminal, and the user can perform interactive editing operations through the mobile terminal;

[0013] The user adjusts the display perspective of the point cloud model through gesture operation, and performs distance measurement and region annotation operation through the editing tool.

[0014] As a preferred scheme of the interactive point cloud data editing method on the mobile terminal of the present application, wherein: the point cloud data is collected in the substation using a handheld laser radar, and the point cloud data is preliminarily processed by a mobile terminal, and the point cloud data is segmented using a fragmentation algorithm, and the specific steps are,

[0015] A handheld laser radar LiDAR device is used to perform three-dimensional space modeling in a substation, and the laser radar measures the three-dimensional coordinates of objects by emitting laser beams and receiving reflected signals to generate high-density point cloud data;

[0016] An octree fragmentation algorithm based on spatial position is used to partition the collected point cloud data according to spatial coordinates, generate multiple fragments, and obtain the volume of each fragment after segmentation, and the expression is:

[0017]

[0018] Wherein, V p represents the volume of the pth fragment after n times of segmentation, p is an index variable, n represents the number of divisions, and V0 represents the initial volume;

[0019] The fragmented point cloud data is transmitted to an edge computing device through a high-speed wireless network 5G.

[0020] As a preferred scheme of the interactive point cloud data editing method on the mobile terminal of the application, wherein: the fragmented point cloud data is transmitted to the edge computing device through the high-speed wireless network, the adaptive learning algorithm is used to filter the noise of the point cloud data on the edge computing device, and the filtering parameter is dynamically adjusted, specifically,

[0021] Adaptive filtering algorithm is used to filter out noise points in the collection process in real time, and the expression is:

[0022]

[0023] Wherein, P(t) represents the effective data probability at time t, P0 represents the validity probability of the initial point, e represents the natural constant, β represents the filtering parameter, d(t) represents the distance between the current point and its neighborhood points, and d0 represents the distance threshold;

[0024] On the edge device, the adaptive learning algorithm optimizes the noise filtering parameter through multiple iterations, and after each iteration, the adaptive learning algorithm analyzes the geometric distribution of each point and its neighborhood points in the point cloud data, and gradually updates the filtering parameter β;

[0025] Based on the initial conditions of the collection environment, set the initial filtering parameter β and d0, and perform the first filtering;

[0026] After the initial filtering, analyze the neighborhood distribution of the remaining point cloud data to detect whether there are still abnormal points;

[0027] According to the distribution of noise points, gradually increase or decrease the filtering parameter β, and the expression is:

[0028]

[0029] Wherein, β n+1 represents the filtering parameter of the n+1th iteration, β n represents the filtering parameter of the nth iteration, γ represents the adjustment step, N represents the total number of points in the point cloud data, and T represents the threshold of valid points.

[0030] As a preferred scheme of the interactive point cloud data editing method on the mobile terminal of the application, wherein: the geometric feature extraction is performed on the point cloud data filtered by noise on the edge computing device, and the geometric feature points are extracted and compressed, and the specific steps are,

[0031] The point cloud data filtered by noise is extracted by using the geometric feature descriptor, and the geometric feature points are extracted, and the expression is:

[0032]

[0033] wherein G i represents the geometric feature value of the i-th point cloud data, j is an index variable, o represents the number of neighborhood points, d ij represents the distance between point i and its neighborhood point j, f i and f j represent the normal vector of point i and its neighborhood point j.

[0034] As a preferred scheme of the interactive point cloud data editing method on the mobile terminal according to the present application, wherein: the mobile terminal receives the compressed geometric feature point data returned from the edge computing device, decompresses and reconstructs the point cloud data through an inverse sparse matrix decomposition algorithm to obtain complete three-dimensional point cloud data, including,

[0035] After the feature extraction is completed, the edge computing device performs sparse matrix decomposition and data compression according to the geometric distribution of the feature points;

[0036] On the edge device, the point cloud data is organized into a sparse matrix M, and the expression of the sparse matrix M is:

[0037]

[0038] wherein x N represents the horizontal coordinate of the N-th point, y N represents the vertical coordinate of the N-th point, z N represents the height coordinate of the M-th point, and N is the total number of points in the point cloud data;

[0039] The point cloud data is compressed using sparse singular value decomposition (SVD), and the expression is:

[0040] M = UDV T ;

[0041] wherein U is an N x k matrix containing left singular vectors, D is a k x k diagonal matrix containing singular values of the matrix, and V T is a k x 3 matrix, and k represents the compressed dimension;

[0042] The compressed dimension k is calculated according to the formula:

[0043]

[0044] wherein C represents a scene complexity factor, log represents a logarithmic function, q i represents the number of neighborhood points of the i-th point within a radius r, and V r represents the volume of the spherical neighborhood corresponding to the radius r.

[0045] As a preferred scheme of the interactive point cloud data editing method on the mobile terminal, the decompressed and reconstructed point cloud data is displayed on the mobile terminal in real time, and the user can perform interactive editing operation through the mobile terminal, and the specific steps are as follows,

[0046] The point cloud data compressed through sparse matrix decomposition is transmitted to the mobile terminal through a wireless network;

[0047] After the mobile terminal receives the compressed data, the point cloud data is decompressed and reconstructed through an inverse sparse matrix decomposition algorithm to restore complete three-dimensional point cloud data, and the compression rate is adjusted according to C, and the expression is as follows:

[0048] M' = UDV T ;

[0049] Wherein, M' represents the decompressed point cloud data matrix, and represents the approximate reconstruction result of the original point cloud data.

[0050] As a preferred scheme of the interactive point cloud data editing method on the mobile terminal, the user adjusts the display perspective of the point cloud model through a gesture operation, and performs distance measurement and region annotation operation through an editing tool, and the specific steps are as follows,

[0051] After reconstruction, the mobile terminal uses WebGL for real-time rendering, and the user performs various interactive operations through the touch screen of the mobile terminal, and the user can perform various editing operations on the point cloud data through the editing tool on the mobile terminal;

[0052] The user performs various interactive operations through the touch screen of the mobile terminal, and the specific operations include:

[0053] Rotation operation: the user can rotate the point cloud model through single-finger sliding to adjust the perspective to view three-dimensional data at different angles;

[0054] Zoom operation: the user can adjust the size of the point cloud model through a two-finger zoom gesture to perform local zoom-in or zoom-out operation;

[0055] Translation operation: the user can translate the point cloud model through two-finger sliding operation to view different model regions on the display screen of the mobile terminal;

[0056] Perspective reset: the system can automatically reset the perspective of the point cloud model to the initial state through double-clicking the screen;

[0057] The user can perform various editing operations on the point cloud data through the editing tool on the mobile terminal, and the specific operations include:

[0058] Point cloud data selection: the user can select any point or region in the point cloud model through the touch screen, and the system highlights the selected point cloud data;

[0059] Deletion of point cloud data: after the user selects a specific point or region in the point cloud model, the unwanted data points can be removed through the deletion operation;

[0060] Labeling of point cloud data: the user can place a marker point at any position of the point cloud model to record or mark the information of the specific position, and the marker point can be associated with external data;

[0061] Measurement of point cloud data: the user can select any two points to realize the distance measurement function by calculating the Euclidean distance between the two points.

[0062] In a second aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the computer program, when executed by the processor, implements any step of the interactive point cloud data editing method on the mobile terminal according to the first aspect of the present application.

[0063] In a third aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any step of the interactive point cloud data editing method on the mobile terminal according to the first aspect of the present application.

[0064] The present application has the following advantages: the present application collects point cloud data by a handheld laser radar and performs preliminary processing using an octree slicing algorithm, thereby reducing the computing burden of the mobile terminal. Secondly, the adaptive learning algorithm is used to dynamically adjust the noise filtering parameters, effectively eliminate noise, retain key points, improve data quality, and then use the sparse matrix decomposition technology for dynamic compression, adjust the compression rate in combination with the scene complexity, significantly reduce the data volume and ensure the information integrity. After transmission through the wireless network, the mobile terminal can efficiently decompress the point cloud data and reconstruct the three-dimensional model. The user can interact in real time through WebGL and ensure smooth display by using the dynamically adjusted rendering resolution. In addition, the system enhances the practicality through distance measurement and other functions. The overall technical solution improves the processing efficiency, transmission speed and user interaction experience of large-scale point cloud data, and is particularly suitable for three-dimensional space modeling and data management in complex environments. BRIEF DESCRIPTION OF DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0066] Figure 1 The flowchart of the interactive point cloud data editing method on the mobile terminal in embodiment 1.

[0067] Figure 2 The flowchart of decompression and reconstruction of the mobile terminal in Example 1. DETAILED DESCRIPTION

[0068] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0069] In the following description, a large number of specific details are set forth in order to facilitate a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the spirit of the present application, therefore the present application is not limited to the specific embodiments disclosed below.

[0070] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an embodiment that is separate or alternative to other embodiments.

[0071] Example 1, refer to Figure 1 and Figure 2 , the first embodiment of the present application, the embodiment provides an interactive point cloud data editing method on a mobile terminal, comprising the following steps:

[0072] S1, using a handheld laser radar to collect point cloud data in a substation, and performing preliminary processing on the point cloud data through a mobile terminal, and using a slicing algorithm to segment the point cloud data;

[0073] Using a handheld laser radar LiDAR device to perform three-dimensional space modeling in a substation, the laser radar measures the three-dimensional coordinates of objects by emitting laser beams and receiving reflected signals, and generates high-density point cloud data;

[0074] During the collection process, the shielding, reflecting surface and light conditions of power equipment can cause noise and abnormal points, therefore, the collection process must be optimized to ensure the accuracy of the data;

[0075] Due to the limited computing and storage capacity of the mobile terminal, the overall size of the point cloud data may exceed the processing capacity of the mobile terminal, therefore, the data must be sliced to reduce the computing burden of the mobile terminal;

[0076] An octree slicing algorithm based on spatial position is used to partition the collected point cloud data according to spatial coordinates, generate multiple slices, and obtain the volume of each slice after segmentation, the expression is:

[0077]

[0078] wherein V p represents the volume of the p-th segment after n times of segmentation, p is an index variable, n represents the number of times of segmentation, and V0represents the initial volume;

[0079] The segmented point cloud data after preliminary processing is transmitted to the edge computing device through a high-speed wireless network 5G;

[0080] During transmission, each segment should contain complete three-dimensional point cloud coordinate information and the validity mark of each point, to ensure that the edge device can further process the point cloud data.

[0081] S2, the segmented point cloud data is transmitted to the edge computing device through a high-speed wireless network, and an adaptive learning algorithm is used to filter noise on the point cloud data on the edge computing device, and the filtering parameter is dynamically adjusted; on the edge computing device, the point cloud data after noise filtering is subjected to geometric feature extraction, and the geometric feature points are extracted and compressed;

[0082] An adaptive filtering algorithm is used to filter out obvious noise points, i.e. reflection interference, in the collection process in real time, and the expression is:

[0083]

[0084] wherein P(t) represents the effective data probability at time t, P0represents the validity probability of the initial point, which is obtained by collecting in an open area without interference, e represents a natural constant, β represents a filtering parameter, which can be set according to requirements, the more complex the environment, the larger the value of α, to accelerate the attenuation of noise points, and vice versa, d(t) represents the distance between the current point and its neighborhood points, and d0represents a distance threshold, points exceeding the threshold are considered to be invalid points;

[0085] The value range of the effective data probability P(t) is [0, 1], the larger the value, the more effective the point, which can be retained in the subsequent processing link; the smaller the value, the more noise the point is, which should be removed, the formula dynamically adjusts the validity probability of the point cloud data to ensure the noise filtering effect in a complex environment;

[0086] On the edge device, the adaptive learning algorithm optimizes the noise filtering parameter through multiple iterations, and after each iteration, the adaptive learning algorithm analyzes the geometric distribution of each point and its neighborhood points in the point cloud data, and gradually updates the filtering parameter β to improve the filtering efficiency;

[0087] Based on the initial conditions of the collection environment (such as noise expectation, device shielding, and light reflection), the initial parameters β and d0are set, and the first filtering is performed;

[0088] After the initial filtering, the neighborhood distribution of the remaining point cloud data is analyzed to detect whether there are still abnormal points (such as isolated points or points with excessive distance);

[0089] According to the distribution of noise points, the filtering parameter β is gradually increased or decreased to enhance or weaken the suppression effect of noise points, and the expression is:

[0090]

[0091] where β n+1 represents the filtering parameter of the n+1th iteration, β n represents the filtering parameter of the nth iteration, γ represents the adjustment step, N represents the total number of points in the point cloud data, and T represents the threshold of valid points;

[0092] The point cloud data filtered by noise is described by a geometric feature descriptor, and geometric feature points are extracted. These feature points will be used for subsequent three-dimensional reconstruction and data compression, and the expression is:

[0093]

[0094] where G i represents the geometric feature value of the ith point cloud data, the geometric feature value includes the distance between the neighborhood points and the similarity between the normal vectors, j is the index variable, o represents the number of neighborhood points, d ij represents the distance between point i and its neighborhood point j, f i and f j represent the normal vector of point i and its neighborhood point j;

[0095] where G i is a positive number, and the larger the value, the higher the importance of the point;

[0096] To improve transmission efficiency, data compression technology is used to compress the fragmented data. The compression algorithm dynamically adjusts according to the sparsity of the point cloud data, ensuring the efficiency and integrity of data transmission.

[0097] S3, the mobile terminal receives the compressed geometric feature point data returned from the edge computing device, decompresses and reconstructs the point cloud data through the inverse sparse matrix decomposition algorithm, and obtains the complete three-dimensional point cloud data;

[0098] After feature extraction is completed, the edge computing device performs sparse matrix decomposition and data compression according to the geometric distribution of the feature points, takes the geometric feature value as the input of the compression algorithm, and ensures that only the key points with the most information are retained to reduce the burden of data transmission and storage;

[0099] On the edge device, the point cloud data is organized as a sparse matrix M, the point cloud data is usually a set of three-dimensional space coordinates, each point is represented by (x, y, z) coordinates, the distance between points can be represented by the neighborhood relationship matrix, therefore, the point cloud data can be combined into a sparse matrix M, the sparsity of the sparse matrix M comes from the sparse distribution characteristics of the point cloud, the expression of the sparse matrix M is:

[0100]

[0101] Wherein, x N represents the horizontal coordinate of the Nth point, y N represents the vertical coordinate of the Nth point, z N represents the height coordinate of the Nth point, N is the total number of points in the point cloud data;

[0102] The point cloud data is compressed using sparse singular value decomposition SVD, and the expression is:

[0103] M=UDV T ;

[0104] Wherein, U is an N×k matrix containing left singular vectors, D is a k×k diagonal matrix containing singular values of the matrix, V T is a k×3 matrix, k represents the compressed dimension;

[0105] The sparse matrix decomposition technique and dynamic compression according to scene complexity adjustment are the core creative points of the present application. First, the sparse matrix decomposition SVD has outstanding advantages in point cloud data compression, SVD decomposition can express the point cloud data as UDV T , wherein U and V T are singular vector matrices respectively, and D is a singular value matrix, this decomposition process can effectively extract the principal components of the point cloud data, retain geometric feature points and reduce redundant data, which is particularly useful in the case of limited computing power of mobile terminals, because it can significantly reduce the burden of data processing and storage;

[0106] The compressed dimension k is calculated by the formula:

[0107]

[0108] Wherein, C represents the scene complexity factor, the value of C can be calculated according to the number of objects in the scene, the geometric structure complexity, the occlusion rate, the larger the value, the more complex the scene, the lower the compression rate, on the contrary, the lower the compression rate, log represents the logarithmic function, q i represents the number of neighborhood points of the ith point within the radius r, V r represents the volume of the spherical neighborhood corresponding to the radius r;

[0109] The application dynamically adjusts the compression rate through a scene complexity factor C, has a significant technical breakthrough, and the prior art generally adopts a static compression rate and cannot be adjusted according to the scene complexity, so that key information may be lost in a complex scene, and bandwidth is wasted in a simple scene, and the application can adjust the compression dimension kkk according to the number of objects, the geometric structure complexity and the occlusion rate in the scene through the introduction of C, so that adaptive compression is realized.

[0110] The point cloud data compressed through the sparse matrix decomposition is transmitted to the mobile terminal through a wireless network;

[0111] After the mobile terminal receives the compressed data, the point cloud data is decompressed and reconstructed through an inverse sparse matrix decomposition algorithm, and complete three-dimensional point cloud data is restored, and the compression rate is dynamically adjusted according to C, and the expression is as follows:

[0112] M' = UDV T ;

[0113] Wherein, M' represents the decompressed point cloud data matrix, represents the approximate reconstruction result of the original point cloud data, M≠M', since the dimension k of the matrix is reduced in the compression process, there is an error between M' and the sparse matrix M, and the compression accuracy depends on the scene complexity factor C and the compression dimension k;

[0114] The application solves the problems of inflexible compression rate, low data transmission efficiency and waste of computing resources in the prior art through the sparse matrix decomposition and dynamic adjustment of the scene complexity, and the method is particularly suitable for processing large-scale point cloud data on a mobile terminal, and significantly improves the processing efficiency and interactive experience.

[0115] S4, the decompressed and reconstructed point cloud data is displayed on the mobile terminal in real time, and the user can perform interactive editing operation through the mobile terminal; the user adjusts the display perspective of the point cloud model through a gesture operation, and performs distance measurement and region annotation operation through an editing tool.

[0116] After reconstruction, the mobile terminal uses WebGL for real-time rendering, and the user performs real-time interaction with the point cloud data through rotation operation, scaling operation, translation operation and distance measurement;

[0117] The user performs various interactive operations through the touch screen of the mobile terminal, and the operations specifically include:

[0118] Rotation operation: the user can rotate the point cloud model through single-finger sliding to adjust the perspective to view three-dimensional data at different angles;

[0119] Scaling operation: the user can adjust the size of the point cloud model through a two-finger scaling gesture to perform local zoom-in or zoom-out operation;

[0120] Translation operation: the user can translate the point cloud model by a two-finger sliding operation to view different model regions on the display screen of the mobile terminal;

[0121] View angle resetting: by double-clicking the screen, the system can automatically reset the view angle of the point cloud model to the initial state;

[0122] The user can perform various editing operations on the point cloud data through the editing tools on the mobile terminal, including:

[0123] Selection of point cloud data: the user can select any point or region in the point cloud model by touching the screen, and the system highlights the selected point cloud data;

[0124] Deletion of point cloud data: after the user selects a specific point or region in the point cloud model, the user can remove the unwanted data points through the deletion operation;

[0125] Labeling of point cloud data: the user can place a label point at any position of the point cloud model to record or mark the information of the specific position, and the label point can be associated with external data;

[0126] Measurement of point cloud data: the user can select any two points to realize the distance measurement function by calculating the Euclidean distance between the two points, and the calculation formula is:

[0127]

[0128] where d represents the distance between the two points, (x1, y1, z1) and (x2, y2, z2) represent the three-dimensional coordinates of the two selected points;

[0129] Region annotation: the user can frame a region of the point cloud model through a gesture and add annotation information to the region using a text box tool;

[0130] The standard formula for calculating the Euclidean distance is widely used to measure the straight-line distance between two points in three-dimensional space. In the present application, the user can select any two points and calculate their distance to realize accurate measurement of certain objects or regions in the point cloud. Specifically, (x1, y1, z1) and (x2, y2, z2) represent the three-dimensional coordinates of the two points selected by the user, and these coordinates are directly derived from the point cloud data. In the application scenarios of point cloud data, such as three-dimensional modeling of substations or measurement of complex structures, it is very meaningful to measure the distance in real time and accurately. For example, the user can use this function to measure the distance between devices to assist in decision-making for maintenance and repair. Therefore, the application of the Euclidean distance formula here is not only reasonable, but also provides a key user interaction function;

[0131] To ensure that in different scene complexity conditions, the resolution of rendering can be adaptively adjusted, a dynamic rendering optimization formula is constructed according to the scene complexity factor C, and the expression is:

[0132]

[0133] Wherein, R represents the current rendering resolution, and R0 represents the maximum resolution.

[0134] The embodiment also provides a computer device suitable for the case of the interactive point cloud data editing method on a mobile terminal, including a memory and a processor; the memory is used for storing computer executable instructions, and the processor is used for executing the computer executable instructions to realize the interactive point cloud data editing method on the mobile terminal as proposed in the above embodiment;

[0135] The key innovation points of the application are as follows: (1) dynamically adjusting the rendering resolution to adapt to the scene complexity, and (2) providing multiple user interaction modes to improve the user experience; compared with the static rendering resolution setting in the prior art, the application realizes dynamic adjustment by introducing a scene complexity factor, and specifically, if the scene is relatively simple, C is low, and the rendering resolution is close to the maximum value, at this time, the user can see higher precision point cloud data; if the scene is complex, the rendering resolution will be automatically reduced to ensure the smoothness of operation; through this adaptive adjustment mechanism, the application effectively solves the problems of limited computing power of the mobile terminal and large changes in scene complexity;

[0136] In terms of user interaction, rotation, scaling, translation and distance measurement operations are all realized through intuitive gestures, and combined with the real-time rendering function of WebGL, the user can freely explore three-dimensional point cloud data and obtain an immersive experience; these operations are also common in other three-dimensional data visualization applications, but in the application, combined with the scene complexity and dynamic rendering technology, these interactive operations are still smooth in complex scenes, ensuring the consistency of the user experience;

[0137] In summary, the dynamic rendering optimization mechanism and multiple user interaction modes in the application are not only reasonable but also have technical innovation, especially the design of adjusting the rendering resolution combined with the scene complexity, which can effectively deal with the data processing and display problems in complex scenes and improve the performance and user experience of the system.

[0138] The computer device can be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected by a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved by WIFI, a carrier network, NFC (Near Field Communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, a trackball or a touchpad arranged on the shell of the computer device, or an external keyboard, a touchpad or a mouse, etc.

[0139] The embodiment also provides a storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the interactive point cloud data editing method on a mobile terminal as described in the above embodiment. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage, a flash memory, a magnetic disk or an optical disk.

[0140] To sum up, the application collects point cloud data by a handheld laser radar and preliminarily processes the point cloud data by using an octree slicing algorithm, thereby reducing the computing burden of the mobile terminal, secondly, the adaptive learning algorithm is used to dynamically adjust the noise filtering parameters, effectively eliminate the noise, retain the key points, improve the data quality, then, the sparse matrix decomposition technology is used for dynamic compression, the compression rate is adjusted in combination with the scene complexity, thereby significantly reducing the data amount and ensuring the information integrity, after the wireless network transmission, the mobile terminal can efficiently decompress the point cloud data and reconstruct the three-dimensional model, the user can interact in real time through WebGL, and the dynamically adjusted rendering resolution is used to ensure the smooth display, in addition, the system enhances the practicability through the distance measurement and other functions, the overall technical scheme improves the processing efficiency, transmission speed and user interaction experience of the large-scale point cloud data, and is particularly suitable for three-dimensional space modeling and data management in a complex environment.

[0141] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application but not limit the present application, although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, and all should be covered in the scope of the claims of the present application.

Claims

1. An interactive point cloud data editing method on a mobile terminal, characterized in that: include, Handheld lidar was used to collect point cloud data in the substation. The point cloud data was initially processed by a mobile terminal and segmented using a segmentation algorithm. The segmented point cloud data is transmitted to the edge computing device via a high-speed wireless network. An adaptive learning algorithm is used to filter noise in the point cloud data on the edge computing device and dynamically adjust the filtering parameters. On edge computing devices, geometric features are extracted from noise-filtered point cloud data, and geometric feature points are extracted and compressed. The mobile terminal receives compressed geometric feature point data returned from the edge computing device, decompresses and reconstructs the point cloud data using an inverse sparse matrix factorization algorithm, and obtains complete 3D point cloud data, including... After feature extraction is completed, the edge computing device performs sparse matrix decomposition and data compression based on the geometric distribution of feature points; On edge devices, point cloud data is organized into sparse matrices. sparse matrix The expression is: ； in, Indicates the first The x-coordinates of the points Indicates the first The ordinates of the points Indicates the first The height coordinates of each point The total number of points in the point cloud data; Sparse Singular Value Decomposition (SVD) is used to compress point cloud data; the expression is: ； in, for A matrix containing left singular vectors. for A diagonal matrix containing the singular values ​​of the matrix. for The matrix, Indicates the dimension of compression; Compressed dimensions The calculation formula is: ； ； in, This represents the scenario complexity factor. Represents the logarithmic function. Indicates the first The number of neighborhood points of a given point within radius r. This represents the volume of the spherical neighborhood corresponding to radius r; The decompressed and reconstructed point cloud data is displayed in real time on the mobile terminal, and users can perform interactive editing operations through the mobile terminal; Users can adjust the display view of the point cloud model using gestures, and perform distance measurement and region labeling operations using editing tools.

2. The interactive point cloud data editing method on a mobile terminal as described in claim 1, characterized in that: The process involves using a handheld lidar to collect point cloud data in the substation, performing preliminary processing on the point cloud data via a mobile terminal, and segmenting the point cloud data using a data segmentation algorithm. The specific steps are as follows: Using a handheld LiDAR device to perform 3D spatial modeling in a substation, the LiDAR measures the 3D coordinates of an object by emitting a laser beam and receiving the reflected signal, generating high-density point cloud data. An octree-based partitioning algorithm based on spatial location is used to partition the collected point cloud data according to spatial coordinates, generating multiple partitions. The volume of each partition is then obtained, expressed as: ； in, Indicates the process After the first segmentation The size of each slice For index variables, Indicates the number of splits. Indicates the initial volume; The pre-processed point cloud data is segmented and transmitted to edge computing devices via a high-speed 5G wireless network.

3. The interactive point cloud data editing method on a mobile terminal as described in claim 2, characterized in that: The process involves transmitting the segmented point cloud data to an edge computing device via a high-speed wireless network, and then using an adaptive learning algorithm to filter noise from the point cloud data on the edge computing device, dynamically adjusting the filtering parameters. Specifically... An adaptive filtering algorithm is used to filter out noise points during the acquisition process in real time. The expression is: ； in, Indicates time The probability of valid data for a point. This represents the probability of the initial point being valid. Represents the natural constant. Indicates the filtering parameters. This represents the distance between the current point and its neighboring points. Indicates the distance threshold; On edge devices, the adaptive learning algorithm optimizes the noise filtering parameters through multiple iterations. After each iteration, the adaptive learning algorithm analyzes the geometric distribution of each point and its neighboring points in the point cloud data and gradually updates the filtering parameters. ; Based on the initial conditions of the data acquisition environment, set the initial filtering parameters. and And perform initial filtering; After the initial filtering, the neighborhood distribution of the remaining point cloud data is analyzed to detect whether there are still outliers. Based on the distribution of noise points, gradually increase or decrease the filtering parameters. The expression is: ； in, Indicates the first The filtering parameters for the next iteration. Indicates the first The filtering parameters for the next iteration. Indicates adjusting the step size. This represents the total number of points in the point cloud data. The threshold representing the valid points.

4. The interactive point cloud data editing method on a mobile terminal as described in claim 3, characterized in that: On the edge computing device, geometric features are extracted from the noise-filtered point cloud data. The extracted geometric feature points are then compressed. The specific steps are as follows: The point cloud data after noise filtering is processed using a geometric feature descriptor to extract geometric feature points, expressed as follows: ； in, Indicates the first Geometric feature values ​​of a point cloud dataset For index variables, Indicates the number of neighboring points. Point and its neighboring points The distance between them and Point and its neighboring points The normal vector.

5. The interactive point cloud data editing method on a mobile terminal as described in claim 4, characterized in that: The decompressed and reconstructed point cloud data is displayed in real time on a mobile terminal, allowing users to perform interactive editing operations. The specific steps are as follows: Point cloud data compressed by sparse matrix decomposition is transmitted to the mobile terminal via wireless network. After receiving the compressed data, the mobile terminal decompresses and reconstructs the point cloud data using an inverse sparse matrix factorization algorithm to restore the complete 3D point cloud data. The compression ratio is dynamically adjusted using the following expression: ； in, This represents the decompressed point cloud data matrix, while this represents the approximate reconstruction result of the original point cloud data.

6. The interactive point cloud data editing method on a mobile terminal as described in claim 5, characterized in that: The user adjusts the display view of the point cloud model through gestures and performs distance measurement and region labeling operations using editing tools. Specifically, After reconstruction, the mobile terminal uses WebGL for real-time rendering, and users can perform various interactive operations through the touch screen of the mobile terminal. Users can also perform various editing operations on the point cloud data through the editing tools on the mobile terminal. The user performs various interactive operations through the touchscreen of the mobile terminal, specifically including: Rotation operation: Users can rotate the point cloud model by swiping with one finger to adjust the viewing angle and view the 3D data from different angles; Zooming operation: Users can adjust the size of the point cloud model by using a two-finger zooming gesture to zoom in or out locally; Panning operation: Users can pan the point cloud model by swiping with two fingers to view different model areas on the mobile terminal's display screen; View reset: By double-tapping the screen, the system can automatically reset the view of the point cloud model to its initial state; The user can perform various editing operations on the point cloud data through the editing tools on the mobile terminal, specifically including: Point cloud data selection: Users can select any point or region in the point cloud model by touching the screen, and the system will highlight the selected point cloud data. Deleting point cloud data: After selecting a specific point or region in the point cloud model, the user can remove the unwanted data points through the delete operation; Point cloud data labeling: Users can place marker points at any location in the point cloud model to record or label information at a specific location. These marker points can be associated with external data. Point cloud data measurement: Users can select any two points and calculate the Euclidean distance between them to achieve distance measurement.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the interactive point cloud data editing method on the mobile terminal according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the interactive point cloud data editing method on the mobile terminal according to any one of claims 1 to 6.

Images