Local map generation method and system based on point cloud image fusion

Abstract

The application discloses a local map generation method and system based on point cloud image fusion, relates to the technical field of point cloud image fusion, and comprises the following steps: establishing initial regional information set of a target region, recording collection point coordinates of point cloud data and image data, inputting the collection point coordinates and the initial regional information set into a reset correction point evaluation network, and determining N reset correction points. The point cloud data and the image data are subjected to adaptive data preprocessing of a collection environment, a global coordinate system is established according to a preprocessing result, the preprocessing result is subjected to preliminary data fusion in the global coordinate system, data is subjected to synchronous alignment correction through the N reset correction nodes, and a local map of the target region is generated according to a synchronous alignment correction result. The application solves the technical problems of low data updating speed and low precision in the prior art, and achieves the technical effects of improving the precision and real-time performance of map generation.

Description

Technical Field

[0001] This invention relates to the field of point cloud image fusion technology, and more specifically to a method and system for generating local maps based on point cloud image fusion. Background Technology

[0002] With the acceleration of urbanization and the widespread application of geographic information, the generation and management of city maps have become an important part of urban planning and development. However, traditional map generation methods are often based on two-dimensional images or latitude and longitude coordinates, making it difficult to accurately reflect the three-dimensional structure and geographic information of cities. Furthermore, the production and management of city maps also face problems such as untimely data updates and low accuracy, hindering the development of urban planning and management. Existing local map generation methods suffer from technical problems such as untimely data updates and low accuracy. Summary of the Invention

[0003] This application provides a method and system for generating local maps based on point cloud image fusion, which is used to address the technical problems of untimely data updates and low accuracy in the prior art.

[0004] In view of the above problems, this application provides a method and system for generating local maps based on point cloud image fusion.

[0005] The first aspect of this application provides a method for generating local maps based on point cloud image fusion, the method comprising:

[0006] An initial region information set for the target region is established, wherein the initial region information set is constructed by interacting with pre-stored data of the target region;

[0007] Record the coordinates of the acquisition points for point cloud data and image data, wherein the point cloud data and the image data are data generated by performing target area data acquisition through corresponding sensors;

[0008] The coordinates of the collection points and the initial area information set are input into the reset correction point evaluation network to determine N reset correction points.

[0009] Adaptive data preprocessing for the acquisition environment is performed on the point cloud data and the image data, and a global coordinate system is established based on the preprocessing results;

[0010] The preprocessed results are initially fused using the global coordinate system, and the data is synchronously aligned and corrected using N reset and correction nodes. A local map of the target area is then generated based on the synchronous alignment and correction results.

[0011] A second aspect of this application provides a local map generation system based on point cloud image fusion, the system comprising:

[0012] A target area data integration module is used to establish an initial area information set for the target area, wherein the initial area information set is constructed by interacting with pre-stored data of the target area;

[0013] The acquisition point coordinate capture module is used to record the acquisition point coordinates of point cloud data and image data, wherein the point cloud data and the image data are data generated by performing target area data acquisition through corresponding sensors;

[0014] A reset calibration point determination module is used to input the coordinates of the acquisition points and the initial area information set into a reset calibration point evaluation network to determine N reset calibration points.

[0015] An adaptive data preprocessing module is used to perform adaptive data preprocessing on the point cloud data and the image data according to the acquisition environment, and to establish a global coordinate system based on the preprocessing results.

[0016] The data fusion and local map construction module is used to perform preliminary data fusion of the preprocessed results using the global coordinate system, and to perform synchronous alignment correction of the data through N reset correction nodes, and to generate a local map of the target area based on the synchronous alignment correction results.

[0017] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0018] An initial regional information set for the target area is established, recording the coordinates of the acquisition points for point cloud data and image data. These acquisition point coordinates and the initial regional information set are input into a reset correction point evaluation network to determine N reset correction points. The point cloud data and image data undergo adaptive data preprocessing based on the acquisition environment. A global coordinate system is established based on the preprocessing results, and preliminary data fusion of the preprocessing results is performed using this global coordinate system. Synchronous alignment correction is then performed through the N reset correction nodes, and a local map of the target area is generated based on the synchronization alignment correction results. This achieves the technical effect of improving the accuracy and real-time performance of map generation. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1A schematic diagram of the local map generation method based on point cloud image fusion provided in the embodiments of this application;

[0021] Figure 2 A schematic diagram illustrating the process of making decisions on constructing N reset and correction points in the local map generation method based on point cloud image fusion provided in this application embodiment;

[0022] Figure 3 This is a schematic diagram illustrating the adaptive data preprocessing process in the local map generation method based on point cloud image fusion provided in the embodiments of this application.

[0023] Figure 4 This is a schematic diagram of the structure of a local map generation system based on point cloud image fusion provided in an embodiment of this application.

[0024] Figure labeling: Target area data integration module 11, acquisition point coordinate capture module 12, reset correction point determination module 13, adaptive data preprocessing module 14, data fusion and local map construction module 15. Detailed Implementation

[0025] This application provides a local map generation method based on point cloud image fusion, which addresses the technical problems of untimely data updates and low accuracy in existing local map generation methods.

[0026] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0027] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0028] Example 1

[0029] like Figure 1As shown, this application provides a method for generating local maps based on point cloud image fusion, the method comprising:

[0030] Step S100: Establish an initial region information set for the target region, wherein the initial region information set is constructed by interacting with pre-stored data of the target region;

[0031] In this embodiment, an initial region information set is constructed by using pre-stored data of the target region. This pre-stored data may come from previous measurements, satellite images, images taken by drones, etc.

[0032] Specifically, the first step is to collect pre-existing data about the target area. This data may include various types of geographic information, such as topography, building distribution, and vegetation. The collected pre-existing data may require preprocessing to improve its accuracy and usability. Preprocessing may include data cleaning, format conversion, and coordinate transformation. The preprocessed data is then integrated to construct an initial regional information set, which will contain basic information about the target area, such as topography. Finally, the initial regional information set needs to be verified and revised, either through field investigation or further data analysis.

[0033] Step S200: Record the coordinates of the acquisition points of the point cloud data and the image data, wherein the point cloud data and the image data are data generated by performing target area data acquisition through the corresponding sensor;

[0034] In this embodiment, when recording the coordinates of acquisition points for point cloud data and image data, firstly, appropriate sensors are selected to perform data acquisition based on the characteristics and requirements of the target area. These sensors may include LiDAR, cameras, infrared sensors, etc. Next, the selected sensors are deployed in the target area, and data acquisition operations are performed. For point cloud data, the LiDAR emits a laser beam and receives the reflected laser beam to obtain 3D point cloud data of the target area. For image data, the camera captures an image of the target area, acquiring two-dimensional image data. Simultaneously, during the data acquisition process, the sensors record the coordinate information of each data point. For point cloud data, each data point has corresponding X, Y, and Z coordinates; for image data, each data point has corresponding two-dimensional coordinates (X, Y).

[0035] Step S300: Input the coordinates of the acquisition points and the initial area information set into the reset correction point evaluation network, and determine N reset correction points;

[0036] In this embodiment, the coordinates of the collected points and the initial region information set are input into the reset correction point evaluation network. This data provides basic information and location references about the target region. A convolutional neural network model is established, and the information set containing point cloud data, image data, and other relevant information of the target region is input into the convolutional neural network model for training. During training, the model is optimized to improve its performance and generalization ability. Finally, a validation set or test set is used to evaluate the model's performance, understand its strengths and weaknesses, and further optimize it. The trained model can then be deployed to a real-world application scenario. The input data is analyzed to determine which collected point coordinates need to be used as reset correction points. During the evaluation process, factors such as the accuracy and reliability of the collected point coordinates, as well as terrain, building, and other feature information from the initial region information set, are considered to comprehensively determine N reset correction points. These points may be selected based on data distribution, geometric features, or other factors for further data processing and map generation.

[0037] Step S400: Perform adaptive data preprocessing on the point cloud data and the image data to suit the acquisition environment, and establish a global coordinate system based on the preprocessing results;

[0038] In this embodiment, point cloud data is preprocessed to remove noise, as point cloud data typically contains noise caused by factors such as sensor noise and environmental interference. This noise can be removed using filters. Simultaneously, point cloud data can be very dense, containing a large number of points. To reduce the number of points and highlight important features, data simplification techniques, such as uniform grid sampling, can be used to reduce the number of data points, thus achieving data simplification. When dealing with multi-view or time-series point cloud data, inverse registration and regressive mapping are required, which can be achieved using algorithms such as iterative nearest-point iteration and least squares methods.

[0039] Image data preprocessing requires noise reduction, which can be achieved using median filtering. Contrast enhancement and brightness adjustment can also be performed to improve image quality. When the image is tilted or perspective-dependent, geometric correction is necessary to accurately represent the object's geometry, typically achieved through affine or homography transformations.

[0040] To establish a global coordinate system, first select a global reference, typically a geographic coordinate system or a map projection coordinate system. Next, transform the point cloud data and image data from their original local coordinates to the global coordinate system. This can be achieved through rigid transformation, affine transformation, or homography transformation. Finally, calibrate the global coordinate system using known control points, such as GPS measurement points, to ensure accurate correspondence with the actual geographic environment.

[0041] Step S500: Perform preliminary data fusion of the preprocessed results using the global coordinate system, and perform synchronous alignment correction of the data through N reset correction nodes, and generate a local map of the target area based on the synchronous alignment correction results.

[0042] In this embodiment, a global coordinate system is used as a reference to initially fuse point cloud data and image data. This can be achieved by assigning coordinates in the global coordinate system to each point in the point cloud data and converting the image data into a projected coordinate system aligned with the global coordinate system. During the data fusion process, some additional processing may be required, such as alignment operations and data interpolation, to ensure data continuity and consistency.

[0043] During synchronous alignment correction, N reset correction nodes are used to perform synchronous alignment correction on the initially fused data. By defining these reset correction nodes, coordinate correction points can be distributed according to the actual environmental conditions, allowing for a more accurate combination of the collected information and achieving synchronous coordinate correction. Using reset correction nodes enables precise data alignment and calibration, involving matching the data with the coordinates of the reset correction nodes and performing necessary transformations and adjustments.

[0044] Finally, based on the results of the synchronization alignment and correction, a local map of the target area is generated. This may include rendering and overlaying the corrected point cloud data and image data to generate a visually readable map.

[0045] Furthermore, such as Figure 2 As shown, step S300 in the method provided in this application embodiment further includes:

[0046] Configure the regional interest points of the target area, and initialize the complexity calculation model based on the regional interest points;

[0047] The complexity calculation model is coupled to the reset correction point evaluation network, and the initial region information set is synchronized to the initialized complexity calculation model.

[0048] The interest complexity of the target region is calculated based on the complexity calculation model, and the interest complexity calculation result is transmitted to the reset correction point evaluation network to complete the construction decision of N reset correction points.

[0049] In this embodiment, representative points of interest (POIs) are identified by analyzing point cloud data and image data of the target area. These POIs may include key locations such as landmarks, intersections, and buildings. Each POI is assigned a corresponding label or feature to characterize its importance or function on the map. These labels or features may include terrain information, traffic flow, building type, etc. A suitable complexity calculation model is selected based on the extracted POIs. This model can be a deep learning model, such as a convolutional neural network or a recurrent neural network. The model parameters are initialized based on the selected model architecture and labeled data; these parameters may be weights, biases, or other model parameters, adjusted according to specific circumstances. The model is initially trained using existing training data or partially labeled data, enabling the model to evaluate complexity based on the features of the POIs.

[0050] The initial complexity calculation model and the reset correction point evaluation network are coupled by using the output of the complexity calculation model as the input to the reset correction point evaluation network. Based on the coupled network structure, the reset correction point evaluation network is further trained and optimized. The initialized and coupled complexity calculation model is synchronized with the initial region information set by updating the parameters or states of the initial region information set into the complexity calculation model.

[0051] Using the established complexity calculation model, the interest complexity of the target region is calculated. Based on the calculation results, the interest complexity of the target region is analyzed, including its spatial distribution, density, and other relevant attributes. The calculated interest complexity results are used as input to the reset correction point evaluation network. The network is then further trained and optimized using these results to determine the locations of N reset correction points. Finally, based on the decision results of the reset correction point evaluation network, the locations of the N reset correction points are determined. These locations typically correspond to areas with high interest complexity or key functions within the target region.

[0052] Furthermore, the previous steps can be referenced, for example, coupling the complexity calculation model to the reset correction point evaluation network, and synchronizing the initial region information set to the initialized complexity calculation model. The method further includes:

[0053] The initial regional information set is analyzed to determine the regional spatial data;

[0054] Read the precision constraints generated from the local map, and construct a uniformly distributed grid using the regional spatial data and the precision constraints;

[0055] The decision to construct N reset and correction points is made using the uniformly distributed grid, the interest complexity calculation results, and the coordinates of the collection points.

[0056] In this embodiment, the initial regional information set is parsed to extract spatially relevant data, such as terrain elevation, building distribution, and vegetation information. This data will be used to construct a uniformly distributed grid. The parsed regional spatial data is then processed and formatted to make it suitable for subsequent grid construction and map generation.

[0057] Relevant accuracy constraints are read from the stored or received local map generation system. These constraints may include map resolution requirements, error ranges for point cloud data, etc. Based on these accuracy constraints, the feasibility of the data acquisition equipment and methods used is evaluated to ensure the generated map meets accuracy requirements. Using the parsed regional spatial data and the read accuracy constraints, a uniformly distributed grid is constructed using appropriate algorithms or methods, such as Voronoi diagrams. The grid size and density are adjusted according to the characteristics of the regional targets and accuracy requirements to ensure the map's accuracy and level of detail.

[0058] Finally, using the constructed uniformly distributed grid, combined with the interest complexity calculation results and the coordinate information of the collected points, a decision is made to construct N reset and correction points. Based on the interest complexity of each cell in the grid and the coordinate information of the collected points, suitable locations for reset and correction points are determined. This can be achieved through weight allocation, probabilistic models, or other decision algorithms. The locations and related attributes of the N reset and correction points are recorded and output for subsequent map generation.

[0059] Furthermore, the previous steps can be referenced, for example, reading the accuracy constraints generated by the local map, and constructing a uniformly distributed grid using the regional spatial data and the accuracy constraints. The method further includes:

[0060] Establish basic distribution constraints, wherein the basic distribution constraints are the basic number constraints of reset correction points within the grid;

[0061] The feature evaluation within the uniformly distributed grid is performed based on the interest complexity calculation results and the coordinates of the collection points.

[0062] Based on the feature evaluation results and the basic number constraint of the reset correction points, a decision is made to construct N reset correction points.

[0063] In this embodiment, a basic quantity constraint for reset correction points within the mesh is determined. This constraint can be a fixed value or calculated based on the size, shape, and other factors of the target area. This basic quantity constraint serves as a fundamental reference for subsequent construction decisions.

[0064] Using interest complexity results and coordinate information of collection points, the characteristics and inverse behavior within each grid cell of a uniformly distributed grid are evaluated. Feature evaluation can include interest complexity, terrain features, building distribution, traffic flow, etc. The specific evaluation method depends on the specific characteristics of the target area and application requirements.

[0065] Based on the feature evaluation results of each grid cell, determine the number of reset correction points required within that grid cell. This can be done through weighted calculations based on the feature evaluation results; for example, regions with higher complexity require more reset correction points. Simultaneously, consider the basic quantity constraint to ensure that the overall basic reset correction point requirement within the grid is met. Combining these two factors, make a decision for each grid cell to determine the specific location of the reset correction points. Repeat this step until the locations of the reset correction points within all grid cells are determined.

[0066] Furthermore, the method also includes:

[0067] Configure a breakthrough threshold, which is a combined breakthrough threshold, including a distance breakthrough threshold and an eigenvalue breakthrough threshold, wherein the distance breakthrough threshold is established by the grid distance of the uniformly distributed grid;

[0068] Initial reset correction points are established based on the feature evaluation results and the basic quantity constraints of the reset correction points.

[0069] The initial reset correction point is evaluated for failure based on the failure threshold.

[0070] If there is a new reset correction point that satisfies the threshold between any initial reset correction points, then a new instruction is generated;

[0071] The newly added reset correction point is inserted into the initial reset correction point according to the new instruction, thus completing the construction of N reset correction points.

[0072] In this embodiment, a combined breach threshold is established, comprising a distance breach threshold and an eigenvalue breach threshold. The distance breach threshold is established using the grid distance of a uniformly distributed grid. This can be calculated based on the average distance between grid cells or a threshold set for a specific application. The eigenvalue breach threshold is set based on feature evaluation results; for example, a grid cell may be considered to meet the breach threshold if its eigenvalue exceeds a set threshold. Based on the feature evaluation results of each grid cell, an initial reset correction point is assigned to each grid cell. The locations of these initial reset correction points can be determined based on the feature evaluation results and the basic number constraints of reset correction points. For example, complex regions may require more reset correction points.

[0073] For each initial reset correction point, check whether it meets the distance breach threshold and the eigenvalue breach threshold. If an initial reset correction point meets either breach threshold, it is marked as meeting the breach condition. If a new reset correction point that meets the breach threshold exists between any two initial reset correction points that meet the breach condition, a new instruction is generated. The new instruction may include the location of the new reset correction point and other relevant information.

[0074] Finally, based on the newly added instructions, the new reset correction points are inserted into the initial reset correction points. This can be done by adding new reset correction points around the initial reset correction points, or by adjusting the position of the initial reset correction points as needed. These steps are repeated until all initial reset correction points within all grid cells meet the threshold or the preset maximum number of iterations is reached. The final output includes the positions and related properties of N reset correction points.

[0075] Furthermore, such as Figure 3 As shown, step S400 in the method provided in this application embodiment further includes:

[0076] Environmental data is collected from the target area to establish an environmental dataset, wherein the environmental dataset is a set of environmental data that has a time node mapping with point cloud data and image data;

[0077] Using the environmental dataset as matching features, parameter matching for data correction is performed to generate correction compensation results;

[0078] The point cloud data and image data are preprocessed using the correction and compensation results.

[0079] In this embodiment, when collecting environmental data from the target area, appropriate methods and devices are used, which may include lidar, cameras, ultrasonic sensors, etc., to acquire point cloud data, image data, etc. The collected data is then organized, filtered, and labeled to establish an environmental dataset that maps the point cloud data and image data to specific time points. This dataset should contain sufficient information to identify and describe the environment and features within the target area.

[0080] The established environmental dataset is used as matching features to perform alignment and correction with point cloud and image data. Parametric matching methods, such as feature matching and deep learning models, are used for inverse registration and localization of the point cloud and image data. This can include geometric transformations such as rotation and translation to achieve alignment and correction. Based on the matching results, correction compensation parameters are generated, which can be used for further refinement and optimization of the point cloud and image data.

[0081] Based on the generated correction and compensation results, adaptive data preprocessing is performed on the point cloud data and image data to improve their acquisition environment. Point cloud data undergoes resampling, noise removal, and voxelization by applying correction and compensation parameters to improve its quality. For image data, image enhancement, denoising, and color correction are performed to enhance image quality and make it more suitable for subsequent analysis and processing.

[0082] In summary, the embodiments of this application have at least the following technical effects:

[0083] This application establishes an initial regional information set for the target area, records the coordinates of acquisition points for point cloud data and image data, and inputs the acquisition point coordinates and the initial regional information set into a reset correction point evaluation network to determine N reset correction points. The point cloud data and image data undergo adaptive data preprocessing based on the acquisition environment, and a global coordinate system is established based on the preprocessing results. Preliminary data fusion of the preprocessing results is performed using this global coordinate system, and synchronous alignment correction is performed through the N reset correction nodes. A local map of the target area is generated based on the synchronous alignment correction results. This achieves the technical effect of improving the accuracy and real-time performance of map generation.

[0084] Example 2

[0085] Based on the same inventive concept as the local map generation method based on point cloud image fusion in the foregoing embodiments, such as Figure 4 As shown, this application provides a local map generation system based on point cloud image fusion. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0086] The target area data integration module 11 is used to establish an initial area information set for the target area, wherein the initial area information set is constructed by interacting with pre-stored data of the target area;

[0087] The acquisition point coordinate capture module 12 is used to record the acquisition point coordinates of point cloud data and image data, wherein the point cloud data and the image data are data generated by performing target area data acquisition through corresponding sensors;

[0088] The reset calibration point determination module 13 is used to input the coordinates of the acquisition point and the initial area information set into the reset calibration point evaluation network to determine N reset calibration points.

[0089] An adaptive data preprocessing module 14 is used to perform adaptive data preprocessing on the point cloud data and the image data according to the acquisition environment, and to establish a global coordinate system based on the preprocessing results.

[0090] The data fusion and local map construction module 15 is used to perform preliminary data fusion of the preprocessed results in the global coordinate system, and to perform synchronous alignment correction of the data through N reset correction nodes, and to generate a local map of the target area based on the synchronous alignment correction results.

[0091] Furthermore, the reset calibration point determination module 13 is used to perform the following methods:

[0092] Configure the regional interest points of the target area, and initialize the complexity calculation model based on the regional interest points;

[0093] The complexity calculation model is coupled to the reset correction point evaluation network, and the initial region information set is synchronized to the initialized complexity calculation model.

[0094] The interest complexity of the target region is calculated based on the complexity calculation model, and the interest complexity calculation result is transmitted to the reset correction point evaluation network to complete the construction decision of N reset correction points.

[0095] Furthermore, the system also includes:

[0096] The initial regional information set is analyzed to determine the regional spatial data;

[0097] Read the precision constraints generated from the local map, and construct a uniformly distributed grid using the regional spatial data and the precision constraints;

[0098] The decision to construct N reset and correction points is made using the uniformly distributed grid, the interest complexity calculation results, and the coordinates of the collection points.

[0099] Furthermore, the system also includes:

[0100] Establish basic distribution constraints, wherein the basic distribution constraints are the basic number constraints of reset correction points within the grid;

[0101] The feature evaluation within the uniformly distributed grid is performed based on the interest complexity calculation results and the coordinates of the collection points.

[0102] Based on the feature evaluation results and the basic number constraint of the reset correction points, a decision is made to construct N reset correction points.

[0103] Furthermore, the system also includes:

[0104] Configure a breakthrough threshold, which is a combined breakthrough threshold, including a distance breakthrough threshold and an eigenvalue breakthrough threshold, wherein the distance breakthrough threshold is established by the grid distance of the uniformly distributed grid;

[0105] Initial reset correction points are established based on the feature evaluation results and the basic quantity constraints of the reset correction points.

[0106] The initial reset correction point is evaluated for failure based on the failure threshold.

[0107] If there is a new reset correction point that satisfies the threshold between any initial reset correction points, then a new instruction is generated;

[0108] The newly added reset correction point is inserted into the initial reset correction point according to the new instruction, thus completing the construction of N reset correction points.

[0109] Furthermore, the adaptive data preprocessing module 14 is used to perform the following methods:

[0110] Environmental data is collected from the target area to establish an environmental dataset, wherein the environmental dataset is a set of environmental data that has a time node mapping with point cloud data and image data;

[0111] Using the environmental dataset as matching features, parameter matching for data correction is performed to generate correction compensation results;

[0112] The correction and compensation results are used to perform adaptive data preprocessing for the acquisition environment of the point cloud data and the image data.

[0113] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, specific embodiments have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims can be performed in a different order than that shown in the embodiments and still achieve the desired result. Additionally, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0114] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0115] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for generating local maps based on point cloud image fusion, characterized in that, The method includes: An initial region information set for the target region is established, wherein the initial region information set is constructed by interacting with pre-stored data of the target region; Record the coordinates of the acquisition points for point cloud data and image data, wherein the point cloud data and the image data are data generated by performing target area data acquisition through corresponding sensors; The coordinates of the collection points and the initial area information set are input into the reset correction point evaluation network to determine N reset correction points. Adaptive data preprocessing for the acquisition environment is performed on the point cloud data and the image data, and a global coordinate system is established based on the preprocessing results; The preprocessed results are initially fused using the global coordinate system, and the data is synchronously aligned and corrected using N reset and correction nodes. A local map of the target area is generated based on the synchronous alignment and correction results. The method further includes: Configure the regional interest points of the target area, and initialize the complexity calculation model based on the regional interest points; The complexity calculation model is coupled to the reset correction point evaluation network, and the initial region information set is synchronized to the initialized complexity calculation model. The interest complexity of the target region is calculated based on the complexity calculation model, and the interest complexity calculation result is transmitted to the reset correction point evaluation network to complete the construction decision of N reset correction points.

2. The method as described in claim 1, characterized in that, The method further includes: The initial regional information set is analyzed to determine the regional spatial data; Read the precision constraints generated from the local map, and construct a uniformly distributed grid using the regional spatial data and the precision constraints; The decision to construct N reset and correction points is made using the uniformly distributed grid, the interest complexity calculation results, and the coordinates of the collection points.

3. The method as described in claim 2, characterized in that, The method further includes: Establish basic distribution constraints, wherein the basic distribution constraints are the basic number constraints of reset correction points within the grid; The feature evaluation within the uniformly distributed grid is performed based on the interest complexity calculation results and the coordinates of the collection points. Based on the feature evaluation results and the basic number constraint of the reset correction points, a decision is made to construct N reset correction points.

4. The method as described in claim 3, characterized in that, The method further includes: Configure a breakthrough threshold, which is a combined breakthrough threshold, including a distance breakthrough threshold and an eigenvalue breakthrough threshold, wherein the distance breakthrough threshold is established by the grid distance of the uniformly distributed grid; Initial reset correction points are established based on the feature evaluation results and the basic quantity constraints of the reset correction points. The initial reset correction point is evaluated for failure based on the failure threshold. If there is a new reset correction point that satisfies the threshold between any initial reset correction points, then a new instruction is generated; The newly added reset correction point is inserted into the initial reset correction point according to the new instruction, thus completing the construction of N reset correction points.

5. The method as described in claim 1, characterized in that, The method further includes: Environmental data is collected from the target area to establish an environmental dataset, wherein the environmental dataset is a set of environmental data that has a time node mapping with point cloud data and image data; Using the environmental dataset as matching features, parameter matching for data correction is performed to generate correction compensation results; The correction and compensation results are used to perform adaptive data preprocessing for the acquisition environment of the point cloud data and the image data.

6. A local map generation system based on point cloud image fusion, characterized in that, The system includes: A target area data integration module is used to establish an initial area information set for the target area, wherein the initial area information set is constructed by interacting with pre-stored data of the target area; The acquisition point coordinate capture module is used to record the acquisition point coordinates of point cloud data and image data, wherein the point cloud data and the image data are data generated by performing target area data acquisition through corresponding sensors; A reset calibration point determination module is used to input the coordinates of the acquisition points and the initial area information set into a reset calibration point evaluation network to determine N reset calibration points. An adaptive data preprocessing module is used to perform adaptive data preprocessing on the point cloud data and the image data according to the acquisition environment, and to establish a global coordinate system based on the preprocessing results. The data fusion and local map construction module is used to perform preliminary data fusion of the preprocessed results using the global coordinate system, and to perform synchronous alignment correction of the data through N reset correction nodes, and to generate a local map of the target area based on the synchronous alignment correction results. The adaptive data preprocessing module is also used to configure the regional interest points of the target area and initialize the complexity calculation model based on the regional interest points; The complexity calculation model is coupled to the reset correction point evaluation network, and the initial region information set is synchronized to the initialized complexity calculation model. The interest complexity of the target region is calculated based on the complexity calculation model, and the interest complexity calculation result is transmitted to the reset correction point evaluation network to complete the construction decision of N reset correction points.

Images