Data seamless connection method for three-dimensional panorama and three-dimensional point cloud analysis based on ICP (Inductively Coupled Plasma)

Abstract

The invention discloses a data seamless connection method for three-dimensional panorama and three-dimensional point cloud analysis based on ICP (Inductively Coupled Plasma), which relates to the technical field of data processing, and comprises the following steps: acquiring and preprocessing point cloud data, and constructing a geometric error minimization model and a feature error minimization model; performing registration processing on the point cloud data by using the geometric error minimization model to obtain first point cloud data; performing registration processing on the point cloud data by using the feature error minimization model to obtain second point cloud data; combining the first point cloud data and the second point cloud data to obtain registration point cloud data; performing data conversion on the registration point cloud data, and constructing a three-dimensional scene model; and carrying out visualization and real-time optimization on the three-dimensional scene model to realize seamless connection of data. According to the method, high-precision registration based on geometric information is realized, color information and reflection intensity information are combined, visual and physical characteristics of point cloud are introduced outside the geometric information, and the adaptability of the system is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method for seamlessly connecting data of a three-dimensional panorama and a three-dimensional point cloud analysis based on ICP. Background Art

[0002] With the rapid development of computer vision and graphics, 3D point cloud technology has been widely used in many fields, including but not limited to autonomous driving, robot navigation, urban modeling, cultural heritage protection, industrial inspection and virtual reality. Point cloud data, as an important form of 3D spatial data representation, is a data set composed of a large number of discrete 3D coordinate points. Each point usually contains spatial coordinates and possible other attribute information. This data form can accurately reflect the geometric shape and texture characteristics of the object surface, providing rich spatial information for various applications. The main advantages of 3D point cloud technology are high precision, high resolution and good expression of complex 3D structures. However, since point cloud data is usually large in size and complex to process, it faces many challenges and limitations in practical applications. These shortcomings seriously restrict the application of point cloud technology in wider and more complex scenarios. In recent years, the application of deep learning technology has brought new opportunities for point cloud processing, such as the PointNet neural network architecture specifically designed for direct processing of point cloud data. In addition, advances in hardware technology, such as higher-performance GPUs and dedicated point cloud processing chips, are also continuously improving the efficiency of point cloud processing.

[0003] However, the common solutions currently available still have many shortcomings. First, traditional registration algorithms, such as the iterative closest point (ICP) algorithm, often require a lot of computing resources and time when processing large-scale point cloud data, which not only limits its use in real-time applications, but also greatly reduces processing efficiency. Second, due to environmental factors, sensor errors, and incomplete sampling, the collected point cloud data usually contains a large number of noise points and outliers. These interference information not only reduces the accuracy of the data, but may also cause errors or deviations in subsequent processing. In addition, the existing technology often uses simple filtering or threshold processing methods to deal with these problems, which makes it difficult to effectively distinguish and retain useful information. Finally, the robustness and stability of the point cloud registration and fusion process is also a difficult problem that needs to be solved urgently. The existing registration algorithm is very sensitive to the selection of initial parameters. When faced with complex scenes, a large number of occlusions, or unclear local features, it is easy to fall into the local optimal solution, resulting in registration failure or inaccurate results. This instability seriously affects the application reliability of 3D point cloud technology in complex environments. Summary of the invention

[0004] In view of the fact that the existing technology requires a lot of computing resources and time to process large-scale point cloud data when seamlessly connecting the three-dimensional panoramic model with the three-dimensional point cloud data, the collected point cloud data usually contains a large number of noise points and outliers, which reduces the accuracy of the data. It is easy to fall into the local optimal solution when facing complex scenes, a large number of occlusions or unclear local features, resulting in problems such as registration failure or inaccurate results. The present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to provide a method for realizing three-dimensional data integration and interaction between different platforms or systems by utilizing the ICP algorithm, realizing data docking and integration by aligning and fusing point cloud data generated by different data sources or platforms, and obtaining more complete, accurate and consistent three-dimensional scene information, so as to provide support for applications in the fields of virtual reality, augmented reality, map making and environmental monitoring.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, an embodiment of the present invention provides a method for seamlessly connecting three-dimensional panoramic data with three-dimensional point cloud analysis based on ICP, which includes acquiring point cloud data and performing preprocessing to construct a geometric error minimization model and a feature error minimization model; performing registration processing on the point cloud data using the geometric error minimization model to obtain first point cloud data; performing registration processing on the point cloud data using the feature error minimization model to obtain second point cloud data; combining the first point cloud data and the second point cloud data to obtain registered point cloud data; performing data conversion on the registered point cloud data to construct a three-dimensional scene model; visualizing the three-dimensional scene model and performing real-time optimization to achieve seamless data connection.

[0008] As a preferred solution of the method for seamlessly connecting three-dimensional panoramic data with three-dimensional point cloud analysis based on ICP of the present invention, wherein: the point cloud data includes source point cloud data and target point cloud data; the source point cloud data includes three-dimensional coordinates, color information and reflection intensity information; the target point cloud data includes three-dimensional coordinates, color information and reflection intensity information; the obtaining of the first point cloud data includes the following steps: based on the ICP algorithm, the rotation matrix R and the translation vector t are used to represent the rigid transformation in the three-dimensional space, the relative distance and angle relationship of the point cloud data are retained, and the change matrix T is obtained; based on the change matrix T, the point p of the source point cloud data is converted into i Convert to the coordinate system of the target point cloud to get the point p of the source point cloud data i Point q with target point cloud data i Geometric error; based on the source point cloud data point p i Point q with target point cloud data iThe geometric error d1 is minimized, and the sum of the geometric errors in the point cloud registration is minimized, that is, the first point cloud data.

[0009] As a preferred solution of the method for seamlessly connecting data of 3D panorama and 3D point cloud analysis based on ICP of the present invention, the calculation formula of the change matrix T is as follows:

[0010] T(p i )=R·p i +t

[0011] Among them, T is the change matrix; R is the rotation matrix; p i is the point of the source point cloud data; t is the translation vector; the point p of the source point cloud data i Point q with target point cloud data i The geometric error is calculated as follows:

[0012] d1=||T(p i )-q i ||

[0013] Among them, d1 is the point p of the source point cloud data i Point q with target point cloud data i The geometric error; T is the change matrix; p i is the point of the source point cloud data; q i is the point of the target point cloud data; the calculation formula of the geometric error minimization model is as follows:

[0014]

[0015] Among them, E geo To minimize the sum of geometric errors in point cloud registration; N is the number of point cloud data; R is the rotation matrix; p i is the point of the source point cloud data; t is the translation vector; q i is the point of the target point cloud data.

[0016] As a preferred solution of the method for seamlessly connecting the data of the three-dimensional panoramic view and the three-dimensional point cloud analysis based on ICP of the present invention, wherein: the obtaining of the second point cloud data comprises the following steps: performing calculation based on the color information and the reflection intensity information to obtain the color error and the reflection intensity error; using the weight coefficient α of the color error i and the weight coefficient β of the reflection intensity error i The color information and the reflection intensity information are combined, and the feature error minimization model is constructed with the goal of minimizing the feature error to obtain the minimized feature error in the point cloud registration, that is, the second point cloud data.

[0017] As a preferred solution of the method for seamlessly connecting the data of the three-dimensional panorama and the three-dimensional point cloud analysis based on ICP of the present invention, the calculation formula of the color error and the reflection intensity error is as follows:

[0018] d2=||c i -c i '||

[0019] d3=||I i -I i '||

[0020] Among them, d2 is the color information of the source point cloud data c i The color information c of the target point cloud data i 'Color error; c i is the color information of the source point cloud data; c i ' is the color information of the target point cloud data; d3 is the reflection intensity information of the source point cloud data i Reflection intensity information I of target point cloud data i 'Reflection intensity error; I i is the reflection intensity information of the source point cloud data; I i ' is the reflection intensity information of the target point cloud data; the calculation formula of the feature error minimization model is as follows:

[0021]

[0022] Among them, E f To minimize the feature error in point cloud registration; N is the number of point cloud data; c i is the color information of the source point cloud data; c i ' is the color information of the target point cloud data; I i is the reflection intensity information of the source point cloud data; I i ' is the reflection intensity information of the target point cloud data.

[0023] As a preferred solution of the method for seamlessly connecting the data of the three-dimensional panorama and the three-dimensional point cloud analysis based on ICP of the present invention, wherein: by combining the first point cloud data and the second point cloud data, and introducing a global weight coefficient for adjustment, the registration point cloud data is obtained, and the calculation formula of the registration point cloud data is as follows:

[0024] E total =λ1E geo +λ2E f

[0025] Among them, E total is the total registration error of point cloud data; λ1 is the weight coefficient of geometric error; λ2 is the weight coefficient of feature error; E geoTo minimize the sum of geometric errors in point cloud registration; E f To minimize the feature error in point cloud registration.

[0026] As a preferred solution of the method for seamlessly connecting 3D panorama and 3D point cloud analysis based on ICP of the present invention, if the total registration error E of the point cloud data is total is 0, it is determined that the source point cloud data and the target point cloud data are completely matched in terms of geometry and feature information, and the registration effect of the point cloud data is the best; if the total registration error E of the point cloud data total If the total registration error E of the point cloud data is greater than 0 and less than the first threshold, it is determined that the source point cloud data and the target point cloud data have a certain degree of matching in terms of geometry and feature information, and the existing weight coefficient λ1 of the geometric error and the weight coefficient λ2 of the feature error are kept unchanged; total If the total registration error E of the point cloud data is greater than or equal to the first threshold and less than the second threshold, it is determined that the matching degree between the source point cloud data and the target point cloud data in terms of geometry and feature information is small, and the weight coefficient λ1 of the geometric error and the weight coefficient λ2 of the feature error are adjusted; total If the value is greater than or equal to the second threshold, it is determined that the source point cloud data and the target point cloud data have a high degree of matching in terms of geometry and feature information, and the point cloud data is collected again.

[0027] In the second aspect, in order to further solve the security problems existing in data processing, the present invention provides a data seamless docking system for 3D panorama and 3D point cloud analysis based on ICP, which includes: a model construction module, which is used to collect 3D coordinates, color information and reflection intensity information and construct a geometric error minimization model and a feature error minimization model; a geometric error module, which is used to use the geometric error minimization model to perform registration processing on the 3D coordinates to obtain the first point cloud data; a feature error module, which is used to use the feature error minimization model to perform registration processing on the color information and the reflection intensity information to obtain the second point cloud data; a point cloud registration module, which is used to combine the first point cloud data and the second point cloud data to obtain registered point cloud data; a 3D scene module, which is used to fuse and integrate the registered point cloud data to construct a 3D scene model; a seamless docking module, which is used to visualize and optimize the 3D scene model in real time to achieve seamless data docking.

[0028] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the method for seamlessly connecting three-dimensional panoramic data with three-dimensional point cloud analysis based on ICP as described in the first aspect of the present invention is implemented.

[0029] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the method for seamlessly connecting data of three-dimensional panoramic and three-dimensional point cloud analysis based on ICP as described in the first aspect of the present invention.

[0030] The beneficial effects of the present invention are as follows: the present invention uses a rotation matrix and a translation vector to represent a rigid transformation in three-dimensional space based on an ICP algorithm, thereby retaining the relative distance and angle relationship of point cloud data, achieving high-precision registration based on geometric information, improving the matching degree of point cloud data in spatial position, and laying a geometric foundation for subsequent data fusion; by combining color information and reflection intensity information, a feature error minimization model is constructed, and the visual and physical features of the point cloud are introduced in addition to the geometric information, thereby improving the accuracy and robustness of the registration, especially when dealing with scenes with rich textures or material differences, better registration effects can be obtained; by introducing a global weight coefficient, the registration results based on geometric information and feature information are fused, thereby achieving an optimal balance between geometric and feature information, improving the overall accuracy and reliability of the registration, and enabling the system to adapt to different types of point cloud data and scene features. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0032] Figure 1 This is a flow chart for seamlessly connecting data between 3D panorama and 3D point cloud analysis based on ICP in Example 1. DETAILED DESCRIPTION

[0033] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0034] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0035] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0036] Example 1

[0037] Reference Figure 1 , which is the first embodiment of the present invention, and provides a method for seamlessly connecting 3D panorama and 3D point cloud analysis data based on ICP, comprising the following steps:

[0038] S1: Obtain point cloud data and preprocess it to build a geometric error minimization model and a feature error minimization model.

[0039] Specifically, Figure 1 The figure shows a data seamless connection flow chart of 3D panorama and 3D point cloud analysis based on ICP. Point cloud data including 3D coordinates, color information and reflection intensity information are collected, and a geometric error minimization model is constructed using the 3D coordinates to obtain the first point cloud data. A feature error minimization model is constructed using the color information and reflection intensity information to obtain the second point cloud data. The first point cloud data and the second point cloud data are combined to obtain registration point cloud data. The registration point cloud data is converted, a 3D scene model is constructed, and visualization and real-time optimization are performed to achieve seamless data connection.

[0040] Preferably, the point cloud data includes source point cloud data and target point cloud data.

[0041] Furthermore, the source point cloud data includes three-dimensional coordinates, color information, and reflection intensity information.

[0042] Furthermore, the target point cloud data includes three-dimensional coordinates, color information, and reflection intensity information.

[0043] Specifically, the preprocessing includes denoising and filtering to improve data quality and consistency.

[0044] S2: Use the geometric error minimization model to perform registration processing on the point cloud data to obtain the first point cloud data.

[0045] Preferably, obtaining the first point cloud data comprises the following steps: based on the ICP algorithm, using the rotation matrix R and the translation vector t to represent the rigid transformation in the three-dimensional space, retaining the relative distance and angle relationship of the point cloud data, avoiding the geometric distortion introduced by the registration error, and obtaining the change matrix T, the specific formula is as follows:

[0046] T(p i )=R·p i+t

[0047] Among them, T is the change matrix; R is the rotation matrix; p i is the point of the source point cloud data; t is the translation vector.

[0048] Based on the change matrix T, the point p of the source point cloud data is i Convert to the coordinate system of the target point cloud to get the point p of the source point cloud data i Point q with target point cloud data i The geometric error is as follows:

[0049] d1=||T(p i )-q i ||

[0050] Among them, d1 is the point p of the source point cloud data i Point q with target point cloud data i The geometric error; T is the change matrix; p i is the point of the source point cloud data; q i is the point of the target point cloud data.

[0051] Based on the source point cloud data, point p i Point q with target point cloud data i The geometric error d1 is minimized as the goal to minimize the sum of geometric errors in the point cloud registration, that is, the first point cloud data. The specific formula is as follows:

[0052]

[0053] Among them, E geo To minimize the sum of geometric errors in point cloud registration, that is, the point p of all source point cloud data i At the transformed position and point q of the target point cloud data i The sum of the squares of the distances between; d1 is the point p of the source point cloud data i Point q with target point cloud data i The geometric error; N is the number of point cloud data, where the number of source point cloud data is equal to that of target point cloud data.

[0054] Specifically, the calculation formula of the geometric error minimization model is as follows:

[0055]

[0056] Among them, E geo To minimize the sum of geometric errors in point cloud registration, that is, the point p of all source point cloud data i At the transformed position and point q of the target point cloud data iThe sum of the squares of the distances between them; N is the number of point cloud data, where the number of source point cloud data is equal to that of target point cloud data; R is the rotation matrix; p i is the point of the source point cloud data; t is the translation vector; q i is the point of the target point cloud data.

[0057] Preferably, by using the geometric error minimization model to perform preliminary registration of the three-dimensional coordinates of the point cloud data, the relative distance and angle relationship of the point cloud data is retained, and the geometric distortion introduced by the registration error is avoided, thereby ensuring the accurate transmission of geometric information during the registration process, achieving a high degree of geometric matching between the source point cloud data and the target point cloud data, obtaining accurate first point cloud data, and improving the accuracy and stability of the geometric registration.

[0058] S3: Use the feature error minimization model to perform registration processing on the point cloud data to obtain the second point cloud data.

[0059] Preferably, the second point cloud data includes the following steps: performing calculation based on the color information and the reflection intensity information to obtain the color error and the reflection intensity error, and the specific formula is as follows:

[0060] d2=||c i -c i '||

[0061] d3=||I i -I i '||

[0062] Among them, d2 is the color information of the source point cloud data c i The color information c of the target point cloud data i 'Color error; c i is the color information of the source point cloud data; c i ' is the color information of the target point cloud data; d3 is the reflection intensity information of the source point cloud data i Reflection intensity information I of target point cloud data i 'Reflection intensity error; I i is the reflection intensity information of the source point cloud data; I i ' is the reflection intensity information of the target point cloud data.

[0063] Using the weight coefficient α of color error i and the weight coefficient β of the reflection intensity error i The color information and the reflection intensity information are combined, and the feature error minimization model is constructed with the goal of minimizing the feature error to obtain the feature error in the minimized point cloud registration, that is, the second point cloud data. The specific formula is as follows:

[0064]

[0065] Among them, E f To minimize the feature error in point cloud registration, that is, the sum of the color error and reflection intensity error of all point pairs; N is the number of point cloud data, where the number of source point cloud data is equal to that of target point cloud data; α i is the weight coefficient of color error; β i is the weight coefficient of the reflection intensity error; d2 is the color information of the source point cloud data c i The color information c of the target point cloud data i 'Color error; d3 is the reflection intensity information I of the source point cloud data i Reflection intensity information I of target point cloud data i 'Reflection intensity error.

[0066] Furthermore, based on the distance between the point pairs, the weight coefficient α of the color error is obtained. i , the specific formula is as follows:

[0067]

[0068] Among them, α i is the weight coefficient of color error; c i is the color information of the source point cloud data; c i ' is the color information of the target point cloud data.

[0069] Furthermore, the weight coefficient β of the reflection intensity error i The influence of color information and reflection intensity information on the total error is at the same order of magnitude, and is calculated based on the difference in reflection intensity. The specific formula is as follows:

[0070]

[0071] Among them, β i is the weight coefficient of the reflection intensity error; I i is the reflection intensity information of the source point cloud data; I i ' is the reflection intensity information of the target point cloud data.

[0072] Specifically, the calculation formula of the feature error minimization model is as follows:

[0073]

[0074] Among them, E f To minimize the feature error in point cloud registration, that is, the sum of the color error and reflection intensity error of all point pairs; N is the number of point cloud data, where the number of source point cloud data is equal to that of target point cloud data; c i is the color information of the source point cloud data; c i' is the color information of the target point cloud data; I i is the reflection intensity information of the source point cloud data; I i ' is the reflection intensity information of the target point cloud data.

[0075] Preferably, the color information and reflection intensity information are accurately aligned by utilizing the feature error minimization model, and the matching of color and reflection intensity information is further optimized on the basis of geometric alignment, so that the point cloud data is also accurately aligned in terms of visual and reflection characteristics, balancing the color error and reflection intensity error, ensuring that the influence of the feature information matches the geometric information, and achieving more comprehensive and accurate point cloud data alignment.

[0076] S4: Combine the first point cloud data and the second point cloud data to obtain registration point cloud data.

[0077] Preferably, the first point cloud data and the second point cloud data are combined and a global weight coefficient is introduced for adjustment to optimize the importance of geometric error and feature error, thereby obtaining the registered point cloud data. The calculation formula of the registered point cloud data is as follows:

[0078] E total =λ1E geo +λ2E f

[0079] Among them, E total is the total registration error of point cloud data; λ1 is the weight coefficient of geometric error; λ2 is the weight coefficient of feature error; E geo To minimize the sum of geometric errors in point cloud registration, that is, the point p of all source point clouds i At the transformed position and point q of the target point cloud i The sum of the squares of the distances between f To minimize the feature error in point cloud registration, that is, the sum of the color error and reflection intensity error of all point pairs.

[0080] Furthermore, the weight coefficient λ1 of the geometric error is used to adjust the geometric error according to the stability of the geometric error. The specific formula is as follows:

[0081]

[0082] Among them, λ1 is the weight coefficient of geometric error; E geo To minimize the sum of geometric errors in point cloud registration, that is, the point p of all source point cloud data i At the transformed position and point q of the target point cloud data i The sum of the squares of the distances between fTo minimize the feature error in point cloud registration, that is, the sum of the color error and reflection intensity error of all point pairs; N is the number of point cloud data, where the number of source point cloud data is equal to that of target point cloud data.

[0083] Furthermore, the weight coefficient λ2 of the feature error is used to adjust the color error and the reflection intensity error so that the influence of the color error and the reflection intensity error is balanced with the geometric error. The specific formula is as follows:

[0084]

[0085] Among them, λ2 is the weight coefficient of feature error; E f To minimize the feature error in point cloud registration, that is, the sum of the color error and reflection intensity error of all point pairs; N is the number of point cloud data, where the number of source point cloud data is equal to that of target point cloud data.

[0086] Preferably, by combining the first point cloud data and the second point cloud data and introducing a global weight coefficient for adjustment, the importance of geometric error and feature error is optimized, and a balanced optimization of geometric error and feature error is achieved, which not only ensures the geometric registration accuracy of point cloud data, but also improves the matching degree of color information and reflection intensity information, improves the accuracy and consistency of the final registered point cloud data, and provides a solid foundation for the subsequent three-dimensional scene model construction and visualization.

[0087] Furthermore, if the total registration error E of the point cloud data is total If it is 0, it is determined that the source point cloud data and the target point cloud data completely match in terms of geometry and feature information, and the registration effect of the point cloud data is the best.

[0088] If the total registration error of the point cloud data is E total If it is greater than 0 and less than the first threshold, it is determined that the source point cloud data and the target point cloud data have a certain degree of matching in terms of geometry and feature information, and the existing weight coefficient λ1 of the geometric error and the weight coefficient λ2 of the feature error are kept unchanged.

[0089] If the total registration error of the point cloud data is E total If it is greater than or equal to the first threshold and less than the second threshold, it is determined that the degree of matching between the source point cloud data and the target point cloud data in terms of geometry and feature information is small, and the weight coefficient λ1 of the geometric error and the weight coefficient λ2 of the feature error are adjusted.

[0090] If the total registration error of the point cloud data is E total If the first threshold and the second threshold are both greater than 0, and the first threshold is less than the second threshold, it is determined that the source point cloud data and the target point cloud data have a large degree of matching in geometry and feature information, and the point cloud data is re-collected.

[0091] Specifically, with regard to the first threshold and the second threshold for determining the total registration error of point cloud data, an initial threshold is set, and a large amount of experimental data is collected for testing. The obtained registered point cloud data is analyzed, and then the first threshold and the second threshold are adjusted to obtain the best registration point cloud data.

[0092] S5: Perform data conversion on the registered point cloud data and construct a three-dimensional scene model.

[0093] Preferably, constructing a three-dimensional scene model comprises the following steps: using a voxel gridding method to segment the registration point cloud data into voxels, i.e., cube grids, the specific formula is as follows:

[0094]

[0095] Where v is the voxel; x i is the point p of the source point cloud data i The x-axis coordinate of i is the point p of the source point cloud data i The y-axis coordinate of i is the point p of the source point cloud data i The z-axis coordinate of voxel v; d is the size of voxel v.

[0096] Calculate the average value of the points in each voxel, obtain the center point position of the corresponding voxel, and construct a three-dimensional scene model. The specific formula is as follows:

[0097]

[0098] Among them, c v is the center point position of voxel v; N(v) is the set of points contained in voxel v; |N(v) is the number of points in voxel v; p i is the point of the source point cloud data.

[0099] S6: Visualize the 3D scene model and optimize it in real time to achieve seamless data connection.

[0100] Specifically, visualization includes importing the 3D scene model into the Unreal Engine, using the rendering and visualization functions of the Unreal Engine to display the 3D scene, and providing user interfaces and tools to support users in analyzing, measuring and interacting with the 3D scene.

[0101] Specifically, real-time optimization includes optimizing algorithms and data processing flows based on application scenarios and data feedback in the actual operating environment, improving the real-time and interactivity of data processing and rendering, achieving real-time data updates and user feedback, and supporting users to conduct real-time data analysis and decision-making in the Unreal Engine.

[0102] The present embodiment also provides a data seamless connection system for 3D panorama and 3D point cloud analysis based on ICP, including: a model construction module, used to collect 3D coordinates, color information and reflection intensity information and construct a geometric error minimization model and a feature error minimization model; a geometric error module, used to use the geometric error minimization model to perform registration processing on the 3D coordinates to obtain the first point cloud data; a feature error module, used to use the feature error minimization model to perform registration processing on the color information and the reflection intensity information to obtain the second point cloud data; a point cloud registration module, used to combine the first point cloud data and the second point cloud data to obtain registered point cloud data; a 3D scene module, used to fuse and integrate the registered point cloud data to construct a 3D scene model; a seamless connection module, used to visualize and optimize the 3D scene model in real time to achieve seamless data connection.

[0103] This embodiment also provides a computer device, which is suitable for the method of seamlessly connecting data of three-dimensional panorama and three-dimensional point cloud analysis based on ICP, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the method of seamlessly connecting data of three-dimensional panorama and three-dimensional point cloud analysis based on ICP as proposed in the above embodiment.

[0104] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used 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 the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0105] The present embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for seamlessly connecting data between three-dimensional panoramic view and three-dimensional point cloud analysis based on ICP as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, disk or optical disk.

[0106] In summary, the present invention uses the rotation matrix and translation vector to represent the rigid transformation in the three-dimensional space based on the ICP algorithm, retains the relative distance and angle relationship of the point cloud data, realizes high-precision registration based on geometric information, improves the matching degree of the point cloud data in spatial position, and lays a geometric foundation for subsequent data fusion; by combining color information and reflection intensity information, a feature error minimization model is constructed, and the visual and physical features of the point cloud are introduced in addition to the geometric information to improve the accuracy and robustness of the registration, especially when dealing with scenes with rich textures or material differences, better registration effects can be obtained; by introducing a global weight coefficient, the registration results based on geometric information and feature information are fused, achieving an optimal balance between geometric and feature information, improving the overall accuracy and reliability of the registration, and enabling the system to adapt to different types of point cloud data and scene features.

[0107] Example 2

[0108] Referring to Tables 1 to 3, which are the second embodiment of the present invention, this embodiment is different from the first embodiment in that, in order to verify its beneficial effects, experimental comparison data between the present invention and the prior art are provided.

[0109] This embodiment simulates three different environments, namely, an indoor office environment, an outdoor urban street, and an industrial plant, and collects corresponding point cloud data, including three-dimensional coordinates, color information, and reflection intensity information, to verify the actual data of the method of the present invention, the traditional ICP algorithm, the color-based ICP algorithm, and the normal vector-based ICP algorithm in these three scenarios. Table 1 shows the comparative data of the indoor office environment.

[0110] Table 1 Comparison data of indoor office environment

[0111]

[0112] It can be seen from Table 1 that in a relatively well-controlled environment, the method of the present invention shows significant advantages and can effectively capture and process details accurately, thus providing the possibility for real-time interactive applications in an office environment.

[0113] Table 2 shows the comparative data of outdoor urban streets. When faced with complex and changing environments, the performance of each method has declined, but the method of the present invention still maintains its leading position, indicating that the method performs well in processing color information under complex lighting conditions and can provide strong support for urban planning and monitoring applications. When faced with an environment with complex geometric structures and diverse materials, such as industrial plants, the specific data are shown in Table 3.

[0114] Table 2 Comparative data of outdoor city streets

[0115]

[0116] Table 3 Comparative data of industrial plants

[0117]

[0118] It can be seen from the above three tables that the present invention maintains a registration error significantly lower than that of other methods in various scenarios, which reflects the stability and reliability of the present invention in different scenarios. At the same time, the present invention effectively improves the geometric integrity and feature retention rate, and can generate a more visually realistic three-dimensional model, thereby achieving seamless data connection between the three-dimensional panorama and the three-dimensional point cloud.

[0119] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention 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 invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for seamlessly connecting 3D panorama and 3D point cloud analysis data based on ICP, characterized by: include: Obtain point cloud data and preprocess it to build a geometric error minimization model and a feature error minimization model; Performing registration processing on the point cloud data using the geometric error minimization model to obtain first point cloud data; Performing registration processing on the point cloud data using the feature error minimization model to obtain second point cloud data; Combining the first point cloud data and the second point cloud data to obtain registration point cloud data; Performing data conversion on the registered point cloud data to construct a three-dimensional scene model; The three-dimensional scene model is visualized and optimized in real time to achieve seamless data connection.

2. The method for seamlessly connecting 3D panorama and 3D point cloud analysis data based on ICP as claimed in claim 1, characterized in that: The point cloud data includes source point cloud data and target point cloud data; The source point cloud data includes three-dimensional coordinates, color information and reflection intensity information; The target point cloud data includes three-dimensional coordinates, color information and reflection intensity information; The obtaining of the first point cloud data comprises the following steps: Based on the ICP algorithm, the rotation matrix R and translation vector t are used to represent the rigid transformation in three-dimensional space, retaining the relative distance and angle relationship of the point cloud data, and obtaining the change matrix T; Based on the change matrix T, the point p of the source point cloud data is i Convert to the coordinate system of the target point cloud to get the point p of the source point cloud data i Point q with target point cloud data i The geometric error of Based on the source point cloud data, point p i Point q with target point cloud data i The geometric error d1 is minimized, and the sum of the geometric errors in the point cloud registration is minimized, that is, the first point cloud data.

3. The method for seamlessly connecting 3D panorama and 3D point cloud analysis data based on ICP as claimed in claim 2, characterized in that: The calculation formula of the change matrix T is as follows: T(p i )=R·p i +t Among them, T is the change matrix; R is the rotation matrix; p i is the point of the source point cloud data; t is the translation vector; The point p of the source point cloud data i Point q with target point cloud data i The geometric error is calculated as follows: d1=||T(p i )-q i || Among them, d1 is the point p of the source point cloud data i Point q with target point cloud data i The geometric error; T is the change matrix; p i is the point of the source point cloud data; q i is the point of the target point cloud data; The calculation formula of the geometric error minimization model is as follows: Among them, E geo To minimize the sum of geometric errors in point cloud registration; N is the number of point cloud data; R is the rotation matrix; p i is the point of the source point cloud data; t is the translation vector; q i is the point of the target point cloud data.

4. The method for seamlessly connecting 3D panorama and 3D point cloud analysis data based on ICP as claimed in claim 3, characterized in that: The obtaining of the second point cloud data comprises the following steps: Perform calculation based on the color information and the reflection intensity information to obtain a color error and a reflection intensity error; Using the weight coefficient α of color error i and the weight coefficient β of the reflection intensity error i The color information and the reflection intensity information are combined, and the feature error minimization model is constructed with the goal of minimizing the feature error to obtain the minimized feature error in the point cloud registration, that is, the second point cloud data.

5. The method for seamlessly connecting 3D panorama and 3D point cloud analysis data based on ICP as claimed in claim 4, characterized in that: The calculation formulas of the color error and the reflection intensity error are as follows: d2=||c i -c’ i || d3=||I i -IN' i || Among them, d2 is the color information of the source point cloud data c i The color information c' of the target point cloud data i Color error; c i is the color information of the source point cloud data; c' i is the color information of the target point cloud data; d3 is the reflection intensity information I of the source point cloud data i Reflection intensity information I' of target point cloud data i Reflection intensity error; I i I' is the reflection intensity information of the source point cloud data; i It is the reflection intensity information of the target point cloud data; The calculation formula of the feature error minimization model is as follows: Among them, E f To minimize the feature error in point cloud registration; N is the number of point cloud data; c i is the color information of the source point cloud data; c' i is the color information of the target point cloud data; I i I' is the reflection intensity information of the source point cloud data; i It is the reflection intensity information of the target point cloud data.

6. The method for seamlessly connecting 3D panorama and 3D point cloud analysis data based on ICP as claimed in claim 5, characterized in that: The first point cloud data and the second point cloud data are combined and a global weight coefficient is introduced for adjustment to obtain the registered point cloud data. The calculation formula of the registered point cloud data is as follows: AND total =λ1E geo +λ2E f Among them, E total is the total registration error of point cloud data; λ1 is the weight coefficient of geometric error; λ2 is the weight coefficient of feature error; E geo To minimize the sum of geometric errors in point cloud registration; E f To minimize the feature error in point cloud registration.

7. The method for seamlessly connecting 3D panorama and 3D point cloud analysis data based on ICP as claimed in claim 6, characterized in that: If the total registration error of the point cloud data is E total If it is 0, it is determined that the source point cloud data and the target point cloud data are completely matched in terms of geometry and feature information, and the registration effect of the point cloud data is the best; If the total registration error of the point cloud data is E total If it is greater than 0 and less than the first threshold, it is determined that the source point cloud data and the target point cloud data have a certain degree of matching in terms of geometry and feature information, and the existing weight coefficient λ1 of the geometric error and the weight coefficient λ2 of the feature error are kept unchanged; If the total registration error of the point cloud data is E total If it is greater than or equal to the first threshold and less than the second threshold, it is determined that the matching degree between the source point cloud data and the target point cloud data in terms of geometry and feature information is small, and the weight coefficient λ1 of the geometric error and the weight coefficient λ2 of the feature error are adjusted; If the total registration error of the point cloud data is E total If the value is greater than or equal to the second threshold, it is determined that the source point cloud data and the target point cloud data have a high degree of matching in terms of geometry and feature information, and the point cloud data is collected again.

8. A data seamless connection system for 3D panoramic view and 3D point cloud analysis based on ICP, based on the data seamless connection method for 3D panoramic view and 3D point cloud analysis based on ICP according to any one of claims 1 to 7, characterized in that: include, A model building module, used to collect three-dimensional coordinates, color information and reflection intensity information and build a geometric error minimization model and a feature error minimization model; A geometric error module is used to perform registration processing on the three-dimensional coordinates using a geometric error minimization model to obtain first point cloud data; A feature error module is used to perform registration processing on the color information and the reflection intensity information using a feature error minimization model to obtain second point cloud data; A point cloud registration module, used for combining the first point cloud data and the second point cloud data to obtain registered point cloud data; The 3D scene module is used to fuse and integrate the registered point cloud data to build a 3D scene model; The seamless docking module is used to visualize and optimize the three-dimensional scene model in real time to achieve seamless data docking.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for seamlessly connecting data of three-dimensional panoramic view and three-dimensional point cloud analysis based on ICP according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for seamlessly connecting data of three-dimensional panoramic view and three-dimensional point cloud analysis based on ICP according to any one of claims 1 to 7 are implemented.

Images