Highlight building point cloud data management method based on multispectral image

Abstract

The invention discloses a highlight building point cloud data management method based on a multispectral image, and relates to the technical field of building point cloud data management. The highlight building point cloud data management method based on the multispectral image comprises the following steps: carrying out predefined spectral image acquisition on a highlight building through an unmanned aerial vehicle; performing first contrastive analysis on the highlight building multispectral point cloud segmentation data; performing first evaluation analysis on the multispectral point cloud segmentation data of the highlight building; performing second evaluation analysis on the multispectral point cloud segmentation data of the highlight building; and performing third point cloud data management processing according to the third comparative analysis result. According to the invention, through the first evaluation analysis, the second evaluation analysis and hierarchical comparison analysis regulation and control, the night acquisition accuracy of the highlight building point cloud data of the multispectral image is improved, and the problem that the night acquisition accuracy of the highlight building point cloud data based on the multispectral image is not high in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of building point cloud data management, and in particular to a method for managing highlight building point cloud data based on multispectral images. Background Art

[0002] With the rapid development of remote sensing technology, high-resolution remote sensing images have become easy to obtain. These images have rich texture and spectral information, which provide favorable features for the identification and extraction of buildings. They have been widely used in the field of building extraction and recognition. At the same time, LiDAR technology, as a real-time means of acquiring three-dimensional spatial information, is developing at a rapid pace. LiDAR technology effectively broadens the scope of data sources, changes the data acquisition mode, and can quickly acquire point cloud data to build high-resolution digital surface models.

[0003] The existing point cloud data management method for highlighted buildings based on multispectral images is implemented through the following technologies, including: building detection, using machine learning or deep learning algorithms to detect and identify buildings from fused data; highlighted building recognition, identifying highlighted buildings according to specific standards (such as height, reflectivity, etc.); point cloud optimization, optimizing point cloud data through algorithms to improve its accuracy and quality, and using point cloud data to reconstruct the three-dimensional surface model of the building; designing effective data structures to store and manage large amounts of point cloud and multispectral image data; using database management systems to store, retrieve and analyze data; visualization and analysis, using three-dimensional visualization tools to display the point cloud model of highlighted buildings.

[0004] For example, the invention patent application with publication number: CN118428737A discloses a building risk assessment method based on InSAR point cloud and related equipment, which includes: acquiring multi-orbit SAR data of a study area, extracting multi-orbit InSAR point clouds of multiple buildings in the study area; constructing a point cloud registration loss function, registering the multi-orbit InSAR point clouds, and obtaining the registered multi-orbit InSAR point clouds of each building in the multiple buildings; constructing a three-dimensional deformation decomposition model, and calculating the multi-dimensional deformation rate and multi-dimensional deformation time series of each building based on the three-dimensional deformation decomposition model and the registered multi-orbit InSAR point clouds; calculating the deformation feature parameters corresponding to each building based on the multi-dimensional deformation rate and the multi-dimensional deformation time series; and evaluating the risk level of each building based on the deformation feature parameters.

[0005] For example, the invention patent application with publication number: CN118570023A discloses a method for calculating the photovoltaic potential of large-scale buildings based on airborne point cloud data, including: obtaining three-dimensional point cloud data of urban areas through airborne laser radar; performing unsupervised domain adaptation training using public annotated three-dimensional point cloud data of urban areas and three-dimensional point cloud data that needs to be classified; using the trained deep learning model to perform forward reasoning on the three-dimensional point cloud data that needs to be classified to obtain the classification results of the point cloud data; extracting the building category point cloud according to the semantic segmentation results, obtaining the point cloud data of a single building through conditional Euclidean clustering, and performing three-dimensional reconstruction on the point cloud data of a single building to obtain a three-dimensional model of the building; performing radiation simulation based on the three-dimensional model of the building in the area and the nearby weather data, and obtaining the photovoltaic potential of each building according to the solar radiation value.

[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0007] In the existing technology, for the collection of point cloud data of highlight buildings in night environments, some existing technologies, such as structured light scanning and other methods will fail, while other existing technologies, such as airborne lidar measurement and other methods, will be much less effective. In addition, when the light is insufficient at night, the light of the highlight buildings themselves is relatively strong, and there is a problem that the accuracy of the point cloud data of highlight buildings based on multispectral imaging is not high at night. Summary of the invention

[0008] The embodiment of the present application solves the problem in the prior art that the accuracy of highlight building point cloud data collected based on multispectral images at night is low by providing a highlight building point cloud data management method based on multispectral images, thereby achieving the effect of improving the accuracy of highlight building point cloud data collected based on multispectral images at night.

[0009] The embodiment of the present application provides a method for managing point cloud data of highlighted buildings based on multispectral images, comprising the following steps: collecting predefined spectral images of highlighted buildings by using an unmanned aerial vehicle, obtaining original multispectral image data of highlighted buildings and original laser radar data of highlighted buildings and performing preprocessing to obtain multispectral point cloud fusion data of highlighted buildings; segmenting the multispectral point cloud fusion data of highlighted buildings according to the predefined fusion data size to obtain multispectral point cloud segmentation data of highlighted buildings, performing a first comparative analysis on the multispectral point cloud segmentation data of highlighted buildings, and performing a first point cloud data management process according to the first comparative analysis result; A first evaluation and analysis is performed on the multispectral point cloud segmentation data of the highlighted building to obtain a basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building, a second comparative analysis is performed based on the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building, and a second point cloud data management and processing is performed based on the result of the second comparative analysis; a second evaluation and analysis is performed on the multispectral point cloud segmentation data of the highlighted building to obtain a noise correction value of the quality of the multispectral point cloud segmentation of the highlighted building, a comprehensive analysis is performed to obtain a comprehensive value of the quality of the multispectral point cloud segmentation of the highlighted building, a third comparative analysis is performed based on the comprehensive value of the quality of the multispectral point cloud segmentation of the highlighted building, and a third point cloud data management and processing is performed based on the result of the third comparative analysis.

[0010] Furthermore, the multispectral point cloud fusion data of the highlighted building is obtained, and the specific process is: importing the original multispectral image data of the highlighted building collected by the drone into the predefined spectral data processing software, and performing radiation correction, geometric correction and atmospheric correction on the original multispectral image data of the highlighted building through the predefined data processing software to obtain the multispectral image correction data of the highlighted building; importing the original laser radar data of the highlighted building collected by the drone into the predefined point cloud data processing software, and converting the original laser radar data of the highlighted building into point cloud data of the highlighted building and removing abnormal points through the predefined point cloud data processing software to obtain the point cloud correction data of the highlighted building; according to the built-in position positioning system of the drone, the corresponding multispectral image correction data of the highlighted building and the point cloud correction data of the highlighted building are annotated with position data; according to the same position data annotation information, the corresponding multispectral image correction data of the highlighted building and the corresponding point cloud correction data of the highlighted building are fused with data to obtain the multispectral point cloud fusion data of the highlighted building.

[0011] Furthermore, the multispectral point cloud segmentation data of the highlighted buildings is subjected to a first comparative analysis, specifically comprising: processing different multispectral point cloud segmentation data of the highlighted buildings through a predefined contrast software to obtain different degrees of discreteness of the multispectral point cloud segmentation data of the highlighted buildings; processing different multispectral point cloud segmentation data of the highlighted buildings through a predefined contrast software to obtain different contrasts of the multispectral point cloud segmentation data of the highlighted buildings.

[0012] Furthermore, the first point cloud data management processing is performed according to the first comparative analysis result, and the specific process is: if the discrete degree of the multispectral point cloud segmentation data of the highlighted building is less than the discrete degree threshold of the multispectral point cloud segmentation data of the highlighted building and the contrast of the multispectral point cloud segmentation data of the highlighted building is less than the contrast threshold of the multispectral point cloud segmentation data of the highlighted building, then the corresponding segmentation area is recorded as the first point cloud data collection area; if the discrete degree of the multispectral point cloud segmentation data of the highlighted building is greater than or equal to the discrete degree threshold of the multispectral point cloud segmentation data of the highlighted building, then the corresponding segmentation area is recorded as the second point cloud data collection area; if the contrast of the multispectral point cloud segmentation data of the highlighted building is greater than or equal to the multispectral point cloud segmentation data of the highlighted building The contrast threshold of the spectral point cloud segmentation data is set, and the corresponding segmentation area is recorded as the second point cloud data collection area; the position point corresponding to the second point cloud data collection area is obtained by the built-in positioning tool of the drone, and the drone flies to the corresponding position point to collect data again and make a loop judgment until the discrete degree of the multi-spectral point cloud segmentation data of the highlighted building in the loop judgment is less than the discrete degree threshold of the multi-spectral point cloud segmentation data of the highlighted building and the contrast of the multi-spectral point cloud segmentation data of the highlighted building is less than the contrast threshold of the multi-spectral point cloud segmentation data of the highlighted building and the number of loop judgments is less than the loop judgment threshold; if the number of loop judgments is greater than or equal to the loop judgment threshold, the corresponding second point cloud data collection area is notified to the relevant personnel.

[0013] Furthermore, the basic value of the quality fluctuation of the multi-spectral point cloud segmentation of highlighted buildings is obtained, specifically including: classifying and extracting the multi-spectral point cloud segmentation data of highlighted buildings in the first point cloud data collection area, and obtaining the reflectivity of highlighted buildings, the contrast of highlighted buildings and the radiant brightness of highlighted buildings in the first point cloud data collection area; the reflectivity of highlighted buildings includes the maximum reflectivity of highlighted buildings and the minimum reflectivity of highlighted buildings; the contrast of highlighted buildings includes the maximum contrast of highlighted buildings and the minimum contrast of highlighted buildings; directly obtaining the standard value of the reflectivity of highlighted buildings, the standard value of the contrast of highlighted buildings and the standard value of the radiant brightness of highlighted buildings from the multi-spectral image database of highlighted buildings; and analyzing and obtaining the basic value of the quality fluctuation of the multi-spectral point cloud segmentation of highlighted buildings in the first point cloud data collection area.

[0014] Furthermore, the second point cloud data management processing is performed according to the second comparative analysis result, and the specific process is: if the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building is greater than or equal to the multi-spectral point cloud segmentation quality fluctuation threshold of the highlighted building, the UAV reaches the first point cloud data collection area through the built-in positioning system, opens the telescopic sunshade and extends the exposure time to perform predefined spectral image collection, and performs a loop judgment until the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building in the loop judgment is less than the multi-spectral point cloud segmentation quality fluctuation threshold of the highlighted building and the number of loop judgments is less than the loop judgment threshold; if the number of loop judgments is greater than or equal to the loop judgment threshold, the corresponding first point cloud data collection area is notified to relevant personnel; if the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building is less than the multi-spectral point cloud segmentation quality fluctuation threshold of the highlighted building, the first point cloud data collection area is recorded as the third point cloud data collection area.

[0015] Furthermore, the obtained quality noise correction value of the multi-spectral point cloud segmentation of the highlighted buildings specifically includes: classifying and extracting the multi-spectral point cloud segmentation data of the highlighted buildings in the third point cloud data collection area to obtain the surface material reflectance of the highlighted buildings in the third point cloud data collection area; the surface material reflectance of the highlighted buildings includes the maximum value and the minimum value of the surface material reflectance of the highlighted buildings; obtaining the maximum value and the minimum value of the night ambient lighting in the third point cloud data collection area through the lux meter; and analyzing and obtaining the quality noise correction value of the multi-spectral point cloud segmentation of the highlighted buildings in the third point cloud data collection area.

[0016] Furthermore, the specific process of the comprehensive analysis to obtain the comprehensive value of the multispectral point cloud segmentation quality of highlighted buildings is: performing a first evaluation analysis on the multispectral point cloud segmentation data of highlighted buildings in the third point cloud data collection area to obtain a basic value of the multispectral point cloud segmentation quality fluctuation of highlighted buildings in the third point cloud data collection area; obtaining a noise correction value of the multispectral point cloud segmentation quality of highlighted buildings in the third point cloud data collection area; and comprehensively analyzing the basic value of the multispectral point cloud segmentation quality fluctuation and the multispectral point cloud segmentation quality noise correction value of highlighted buildings in the third point cloud data collection area to obtain a comprehensive value of the multispectral point cloud segmentation quality of highlighted buildings.

[0017] Furthermore, the third point cloud data management processing is performed according to the third comparative analysis result, specifically including: if the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building is less than the comprehensive threshold value of the multispectral point cloud segmentation quality of the highlighted building, no adjustment is made; if the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building is greater than or equal to the comprehensive threshold value of the multispectral point cloud segmentation quality of the highlighted building, then the difference between the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building and the comprehensive threshold value of the multispectral point cloud segmentation quality of the highlighted building is recorded as the multispectral point cloud segmentation quality difference of the highlighted building; if the multispectral point cloud segmentation quality difference of the highlighted building is less than the multispectral point cloud segmentation quality difference threshold of the highlighted building, density filtering processing is performed on the multispectral point cloud segmentation data of the highlighted building in the corresponding third point cloud data collection area; if the multispectral point cloud segmentation quality difference of the highlighted building is greater than or equal to the multispectral point cloud segmentation quality difference threshold of the highlighted building, bilateral filtering processing is performed on the multispectral point cloud segmentation data of the highlighted building in the corresponding third point cloud data collection area.

[0018] Furthermore, the third point cloud data management and processing according to the third comparative analysis result also includes: collecting data again according to the third point cloud data collection area through the built-in positioning system of the drone and making a loop judgment until the comprehensive value of the multi-spectral point cloud segmentation quality of the highlighted building in the loop judgment is less than the comprehensive threshold of the multi-spectral point cloud segmentation quality of the highlighted building and the number of loop judgments is less than the loop judgment threshold; if the number of loop judgments is greater than or equal to the loop judgment threshold, the corresponding third point cloud data collection area is notified to relevant personnel.

[0019] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0020] 1. Use drones to collect predefined spectral images of highlighted buildings; perform a first comparative analysis on the multispectral point cloud segmentation data of highlighted buildings; perform a first evaluation analysis on the multispectral point cloud segmentation data of highlighted buildings; perform a second evaluation analysis on the multispectral point cloud segmentation data of highlighted buildings; and perform third point cloud data management and processing based on the third comparative analysis results. Through the first and second evaluation analyses and hierarchical comparative analysis regulation, the accuracy of the point cloud data of highlighted buildings in multispectral images collected at night is improved, which solves the problem of low accuracy of the point cloud data of highlighted buildings in the prior art based on multispectral images collected at night.

[0021] 2. Perform a second point cloud data management process based on the second comparative analysis result. The specific process is as follows: by real-time monitoring of the fluctuation of the multispectral point cloud segmentation quality and taking corresponding measures when the quality decreases, the quality of the collected point cloud data is always maintained at a high level. The built-in positioning system and predefined collection process of the drone are used to automate the point cloud data collection, thereby greatly improving the efficiency of the point cloud data management method for highlighted buildings based on multispectral images.

[0022] 3. By performing the third point cloud data management processing according to the third comparative analysis results, and by comparing the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building with the threshold, the data area that needs further processing can be identified, and then different filtering methods can be selected according to the comparison result of the comprehensive value of the segmentation quality with the threshold, thereby achieving the robustness of the highlighted building point cloud data management method based on multispectral images. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A flow chart of a method for managing point cloud data of highlighted buildings based on multispectral images provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] The embodiment of the present application provides a method for managing point cloud data of highlighted buildings based on multispectral images, thereby solving the problem in the prior art that the accuracy of point cloud data of highlighted buildings based on multispectral images at night is low. Through the first and second evaluation analyses and hierarchical comparative analysis and regulation, the effect of improving the accuracy of point cloud data of highlighted buildings based on multispectral images at night is achieved.

[0025] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0026] like Figure 1As shown, it is a flow chart of a method for managing point cloud data of highlighted buildings based on multispectral imaging provided by an embodiment of the present application, and the method comprises the following steps: collecting predefined spectral images of highlighted buildings by using a drone to obtain original multispectral image data of highlighted buildings and original laser radar data of highlighted buildings and preprocessing them to obtain multispectral point cloud fusion data of highlighted buildings; segmenting the multispectral point cloud fusion data of highlighted buildings according to the predefined fusion data size to obtain multispectral point cloud segmentation data of highlighted buildings, performing a first comparative analysis on the multispectral point cloud segmentation data of highlighted buildings, and performing a first point cloud data management process according to the first comparative analysis result; A first evaluation analysis is performed on the multispectral point cloud segmentation data of the highlighted buildings to obtain a basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted buildings, a second comparative analysis is performed based on the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted buildings, and a second point cloud data management and processing is performed based on the result of the second comparative analysis; a second evaluation analysis is performed on the multispectral point cloud segmentation data of the highlighted buildings to obtain a noise correction value of the quality of the multispectral point cloud segmentation of the highlighted buildings, a comprehensive analysis is performed to obtain a comprehensive value of the quality of the multispectral point cloud segmentation of the highlighted buildings, a third comparative analysis is performed based on the comprehensive value of the quality of the multispectral point cloud segmentation of the highlighted buildings, and a third point cloud data management and processing is performed based on the result of the third comparative analysis.

[0027] Furthermore, multispectral point cloud fusion data of highlighted buildings are obtained, and the specific process is as follows: the original multispectral image data of highlighted buildings collected by the UAV is imported into the predefined spectral data processing software, and the radiation correction, geometric correction and atmospheric correction are performed on the original multispectral image data of highlighted buildings through the predefined data processing software to obtain the multispectral image correction data of highlighted buildings; the original lidar data of highlighted buildings collected by the UAV is imported into the predefined point cloud data processing software, and the original lidar data of highlighted buildings is converted into point cloud data of highlighted buildings and abnormal points are removed through the predefined point cloud data processing software to obtain the point cloud correction data of highlighted buildings; the corresponding multispectral image correction data of highlighted buildings and the point cloud correction data of highlighted buildings are annotated with position data according to the built-in positioning system of the UAV; the corresponding multispectral image correction data of highlighted buildings and the corresponding point cloud correction data of highlighted buildings are fused with data according to the same position data annotation information to obtain the multispectral point cloud fusion data of highlighted buildings.

[0028] In this embodiment, the specific collection angle and height of the drone are predefined settings and are changed according to the specific settings.

[0029] The collected raw data of multispectral images of highlighted buildings are marked with GPS locations.

[0030] Perform radiometric correction on the image to eliminate the impact of the sensor and atmospheric conditions on the image and ensure the consistency of the image's radiometric measurement. Perform geometric correction on the image to eliminate image distortion caused by changes in the UAV's flight attitude. Perform atmospheric correction on the image to reduce the impact of atmospheric scattering and absorption on the image and restore the true reflectivity of the ground object.

[0031] The predefined spectral data processing software is ENVI or ERDAS.

[0032] Use the radiometric calibration parameters provided by the sensor: First, use the radiometric calibration coefficients (including gain and offset) provided by the sensor manufacturer to convert the digital value of the image into radiometric brightness values. This process is called radiometric calibration. Use the conversion formula to convert the raw data of the multispectral image of the highlight building into radiometric brightness values.

[0033] Further calibration using data from a radiation standard: If available, the image can be further calibrated using data from a radiation standard (e.g., a laboratory-calibrated radiation source or a field-set radiation reference plate). The radiation standard provides a known, stable radiance or reflectance value that can be used as a reference to obtain radiometric correction data for high-brightness building multispectral images.

[0034] Geometric correction: Select or import ground control points, use geometric correction algorithms (such as polynomial transformation, affine transformation, etc.) to correct the radiation correction data of the highlighted building multispectral image according to GCPs, and obtain the geometric correction data of the highlighted building multispectral image.

[0035] Atmospheric correction requires the input of parameters required for atmospheric correction, such as imaging time, geographical location, altitude, and atmospheric conditions. The 6S model is selected as the atmospheric correction model.

[0036] Run the atmospheric correction algorithm to generate the corrected image and obtain the multispectral image correction data of the highlighted building.

[0037] Data collection is performed using a lidar system carried by a drone. The lidar system emits laser pulses and receives light signals reflected from the surface of buildings; the lidar calculates the distance to the object by emitting laser pulses and measuring the return time, thereby generating three-dimensional point cloud data.

[0038] The lidar data is synchronized with the GPS and inertial measurement unit data on the drone to facilitate subsequent geolocation and attitude correction. The collected lidar raw data of the highlighted buildings are annotated with GPS locations.

[0039] The collected LiDAR data is preprocessed, including removing invalid points, correcting LiDAR scanning angle and distance deviation, etc. This is achieved through the built-in software of the LiDAR.

[0040] Use predefined point cloud data processing software, such as PDAL (Point Data Abstraction Library), to convert the preprocessed lidar data into point cloud data.

[0041] Export the processed point cloud data to a standard format, such as LAS, which is an open standard format used by the United States Geographic Information System for the exchange of LiDAR data. The LAS file format was developed by ASPRS (American Society for Photogrammetry and Remote Sensing) and has become the de facto standard for LiDAR data exchange.

[0042] According to the same location data annotation information, the corresponding highlighted building multispectral image correction data and the corresponding highlighted building point cloud correction data are fused, specifically including:

[0043] Project the multispectral image data into the space of the point cloud data. This is done by mapping the image pixels to the corresponding points in the point cloud.

[0044] For each point in the point cloud, find its corresponding pixel in the multispectral image and assign the color information of the pixel to the point in the point cloud.

[0045] If some points in the point cloud have no corresponding pixels in the multispectral image, interpolation methods (such as nearest neighbor interpolation, bilinear interpolation, etc.) are used to estimate the colors of these points.

[0046] The point cloud data and multispectral image data are fused together to form a point cloud dataset containing 3D geometric information and multispectral color information.

[0047] Furthermore, the multispectral point cloud segmentation data of the highlighted buildings are subjected to a first comparative analysis, specifically including: processing different multispectral point cloud segmentation data of the highlighted buildings through a predefined contrast software to obtain different degrees of discreteness of the multispectral point cloud segmentation data of the highlighted buildings; processing different multispectral point cloud segmentation data of the highlighted buildings through a predefined contrast software to obtain different contrasts of the multispectral point cloud segmentation data of the highlighted buildings.

[0048] In this embodiment, the standard deviation of the midpoint coordinates of the multispectral point cloud segmentation data of the highlight building is calculated. The standard deviation reflects the discrete degree of the point coordinate distribution, and the points in the point cloud data known to be uniform or low texture areas are selected for calculation. These areas usually contain less structural information, so the impact of noise is more obvious. In the environment of highlight buildings at night, the light is selected by default as the midpoint coordinates of the multispectral point cloud segmentation data of the highlight building.

[0049] The point cloud data is processed by predefined contrast software, for example, using Python's Pandas library and NumPy library to calculate the degree of dispersion.

[0050] import pandas as pd

[0051] import numpy as np

[0052] #Read point cloud data

[0053] point_cloud_data = pd.read_csv('point_cloud_data.csv')

[0054] #Select one or more bands for noise analysis

[0055] #For example, select X, Y, Z coordinates

[0056] x = point_cloud_data['X']

[0057] y = point_cloud_data['Y']

[0058] z = point_cloud_data['Z']

[0059] #Select one or more regions to calculate the noise level

[0060] #For example, select a small area in the point cloud data center region_of_interest=point_cloud_data[(point_cloud_data['X']>100)&(point_cloud_data['X']<200)&point_cloud_data['Y']>100)&(point_cloud_data['Y']<200)&(point_cloud_data['Z']>100)&(point_cloud_data['Z']<200)]

[0061] # Calculate the degree of dispersion

[0062] noise_level_x=np.std(region_of_interest['X'])

[0063] noise_level_y=np.std(region_of_interest['Y'])

[0064] noise_level_z=np.std(region_of_interest['Z'])

[0065] The contrast of the point cloud data is calculated using predefined contrast software, for example, Python's Pandas and NumPy libraries.

[0066] import pandas as pd

[0067] Import numpy as np

[0068] #Read point cloud data

[0069] point_cloud_data=pd.read_csv('point_cloud_data.csv')

[0070] #Select one or more bands for contrast analysis

[0071] #For example, select X, Y, Z coordinates

[0072] x = point_cloud_data1['X']

[0073] y = point_cloud_data2['Y']

[0074] z = point_cloud_data3['Z']

[0075] #Calculate the brightness difference between adjacent points in the point cloud, for example, calculate the contrast of the X coordinate

[0076] contrast_x = np.abs(np.diff(x))

[0077] By calling Python software, the degree of dispersion and contrast can be quickly analyzed.

[0078] Furthermore, a first point cloud data management process is performed according to the first comparative analysis result, and the specific process is: if the discrete degree of the multispectral point cloud segmentation data of the highlighted building is less than the discrete degree threshold of the multispectral point cloud segmentation data of the highlighted building and the contrast of the multispectral point cloud segmentation data of the highlighted building is less than the contrast threshold of the multispectral point cloud segmentation data of the highlighted building, then the corresponding segmentation area is recorded as the first point cloud data collection area; if the discrete degree of the multispectral point cloud segmentation data of the highlighted building is greater than or equal to the discrete degree threshold of the multispectral point cloud segmentation data of the highlighted building, then the corresponding segmentation area is recorded as the second point cloud data collection area; if the contrast of the multispectral point cloud segmentation data of the highlighted building is greater than or equal to the multispectral point cloud segmentation data of the highlighted building The contrast threshold of the spectral point cloud segmentation data is set, and the corresponding segmentation area is recorded as the second point cloud data collection area; the position point corresponding to the second point cloud data collection area is obtained by the built-in positioning tool of the drone, and the drone flies to the corresponding position point to collect data again and make a loop judgment until the discrete degree of the multi-spectral point cloud segmentation data of the highlighted building in the loop judgment is less than the discrete degree threshold of the multi-spectral point cloud segmentation data of the highlighted building and the contrast of the multi-spectral point cloud segmentation data of the highlighted building is less than the contrast threshold of the multi-spectral point cloud segmentation data of the highlighted building and the number of loop judgments is less than the loop judgment threshold; if the number of loop judgments is greater than or equal to the loop judgment threshold, the corresponding second point cloud data collection area is notified to the relevant personnel.

[0079] In this embodiment, if the number of loop judgments is greater than or equal to the loop judgment threshold, it means that in actual conditions, strong light from a building happens to shine into the second point cloud data collection area, or most of the rooms in the building are dark at night, with only a few lights, which makes the contrast of the spectrum in the area obvious, so that the multi-spectral point cloud segmentation data of the highlighted building collected by the drone is always at a low quality level.

[0080] Furthermore, a basic value of quality fluctuation of multi-spectral point cloud segmentation of highlighted buildings is obtained, specifically including: classifying and extracting the multi-spectral point cloud segmentation data of highlighted buildings in the first point cloud data collection area, and obtaining the reflectivity of highlighted buildings, the contrast of highlighted buildings and the radiant brightness of highlighted buildings in the first point cloud data collection area; the reflectivity of highlighted buildings includes the maximum reflectivity of highlighted buildings and the minimum reflectivity of highlighted buildings; the contrast of highlighted buildings includes the maximum contrast of highlighted buildings and the minimum contrast of highlighted buildings; the standard value of reflectivity of highlighted buildings, the standard value of contrast of highlighted buildings and the standard value of radiant brightness of highlighted buildings are directly obtained from the multi-spectral image database of highlighted buildings; and the basic value of quality fluctuation of multi-spectral point cloud segmentation of highlighted buildings in the first point cloud data collection area is obtained by analysis.

[0081] In this embodiment, the multispectral point cloud segmentation data of the highlighted building in the first point cloud data collection area is classified and extracted. Specifically, the reflectivity is calculated for each point of the multispectral point cloud segmentation data of the highlighted building, usually by comparing the intensity of the incident light and the reflected light. The contrast of the highlighted building area is calculated, usually by comparing the reflectivity difference of adjacent points of the multispectral point cloud segmentation data of the highlighted building. The radiant brightness of each point of the multispectral point cloud segmentation data of the highlighted building is obtained, which is usually obtained by sensor calibration and atmospheric correction to analyze the distribution of radiant brightness to determine the radiation characteristics of the highlighted building.

[0082] The basic value of the quality fluctuation of the multispectral point cloud segmentation of highlighted buildings is used to describe the negative degree to which the highlighted buildings are affected by the quality of multispectral point cloud acquisition.

[0083] Number the first point cloud data collection area, Indicates the number of the first point cloud data collection area, Indicates the total number of points in the first point cloud data collection area.

[0084] The point cloud data detection time is divided into different point cloud data detection time periods according to the size of the predefined point cloud data detection time window, and the point cloud data detection time periods are numbered. Indicates the number of the time period for point cloud data detection. Indicates the total number of points in the cloud data detection time period.

[0085] Indicates The basic value of multispectral point cloud segmentation quality fluctuation of the highlighted buildings in the first point cloud data collection area.

[0086] ;

[0087] represents a natural constant;

[0088] Indicates The first point cloud data collection area The maximum reflectivity of the highlighted buildings in the point cloud data detection time period.

[0089] Indicates The first point cloud data collection area The minimum reflectivity of the highlighted buildings in the point cloud data detection time period.

[0090] Indicates The first point cloud data collection area The standard reflectance value of the highlighted building is directly obtained from the highlighted building multispectral image database.

[0091] Indicates The first point cloud data collection area The maximum contrast value of the highlighted buildings in the point cloud data detection time period.

[0092] Indicates The first point cloud data collection area The minimum contrast value of the highlighted buildings in the point cloud data detection time period.

[0093] Indicates The first point cloud data collection area The contrast standard value of the highlighted building is directly obtained from the highlighted building multispectral image database.

[0094] Indicates The first point cloud data collection area The radiant brightness of highlighted buildings during the detection period of point cloud data.

[0095] Indicates The standard value of the radiant brightness of the highlighted buildings in the first point cloud data collection area is directly obtained from the multispectral image database of highlighted buildings.

[0096] Bright buildings may have unusually high radiance, which may be caused by excessive reflectivity or sensor saturation. Certain building materials (such as glass, reflective metal panels, etc.) have high reflectivity, which will cause a large amount of solar radiation to be reflected at certain angles and lighting conditions.

[0097] It represents the weight factor of the difference between the maximum reflectance of the highlighted building and the minimum reflectance of the highlighted building on the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building, and represents the influence of the maximum reflectance of the highlighted building on the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building.

[0098] It represents the weight factor of the difference between the maximum contrast value of the highlighted building and the minimum contrast value of the highlighted building on the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building, and represents the influence of the maximum contrast value of the highlighted building on the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building.

[0099] The weight factor of the maximum reflectance of the highlighted building for the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building and the weight factor of the maximum contrast of the highlighted building for the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building are directly obtained from the multispectral image database of the highlighted building through a preset mapping relationship.

[0100] Use a light meter to directly measure the light intensity at a specific location. The light meter can measure the luminous flux per unit area. For example, a mapping set of ambient light intensity and its corresponding maximum reflectance of a highlighted building for the weight factor of the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building and the maximum contrast of the highlighted building for the weight factor of the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building is constructed. The real-time ambient light intensity is input into the mapping set to obtain the weight factor of the maximum reflectance of the highlighted building for the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building and the weight factor of the maximum contrast of the highlighted building for the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building, wherein the mapping relationship is a one-to-one correspondence or a many-to-one relationship.

[0101] Represents the sensor resolution error factor. The spatial resolution of the sensor affects the detail level of the point cloud. It is obtained from the highlighted building multispectral image database through a pre-set mapping relationship.

[0102] Atmospheric transmittance describes the energy loss of light when it passes through the atmosphere. The lower the transmittance, the stronger the absorption and scattering of light by the atmosphere.

[0103] Indicates The first point cloud data collection area Atmospheric transmittance during the detection period of point cloud data.

[0104] Indicates The first point cloud data collection area The standard value of atmospheric transmittance during the detection period of point cloud data.

[0105] The atmospheric radiation transfer model is used to simulate the propagation of light waves in the atmosphere in order to calculate the atmospheric transmittance.

[0106] Example steps to obtain transmittance using an atmospheric model:

[0107] Choose an appropriate radiation transfer model: Select the model based on the wavelength range and accuracy requirements to be simulated.

[0108] Input atmospheric parameters: including but not limited to aerosol type, aerosol optical depth (AOD), water vapor content, ozone content, ground pressure, temperature, etc.

[0109] Set the observation geometry: including the zenith angle, azimuth angle, etc. of the sun and the observer.

[0110] Run the model: The model will calculate the propagation path of light waves in the atmosphere based on the input parameters and give the transmittance.

[0111] Output results: The model will output the transmittance value of a specific wavelength or band.

[0112] Furthermore, a second point cloud data management process is performed according to the second comparative analysis result, and the specific process is: if the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building is greater than or equal to the multi-spectral point cloud segmentation quality fluctuation threshold of the highlighted building, the UAV reaches the first point cloud data collection area through the built-in positioning system, opens the telescopic sunshade and extends the exposure time to perform predefined spectral image collection, and performs a loop judgment until the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building in the loop judgment is less than the multi-spectral point cloud segmentation quality fluctuation threshold of the highlighted building and the number of loop judgments is less than the loop judgment threshold; if the number of loop judgments is greater than or equal to the loop judgment threshold, the corresponding first point cloud data collection area is notified to relevant personnel; if the basic value of the multi-spectral point cloud segmentation quality fluctuation of the highlighted building is less than the multi-spectral point cloud segmentation quality fluctuation threshold of the highlighted building, the first point cloud data collection area is recorded as the third point cloud data collection area.

[0113] In this embodiment, under night conditions, reflectivity is still obtained in spectral analysis, but some special factors need to be considered. The lighting conditions at night are different from those during the day, and mainly rely on moonlight, artificial light sources (such as street lights, building lighting, etc.) and scattered light in the environment. The intensity and spectral characteristics of these light sources are significantly different from sunlight, including: light sources at night may have different spectral distributions, which will affect the measurement of reflectivity. It is necessary to understand the spectral characteristics of the light source for accurate measurement and correction. The light intensity at night is generally low, which may result in a reduced signal-to-noise ratio for reflectivity measurement. Therefore, the exposure time may be extended to obtain sufficient light signal. The ambient light at night may include scattered light from surrounding objects, which may interfere with the measurement of reflectivity. Measures need to be taken to reduce these interferences, such as using a sunshade.

[0114] Furthermore, a quality noise correction value of the multispectral point cloud segmentation of highlighted buildings is obtained, specifically including: classifying and extracting the multispectral point cloud segmentation data of highlighted buildings in the third point cloud data collection area to obtain the surface material reflectance of highlighted buildings in the third point cloud data collection area; the surface material reflectance of highlighted buildings includes the maximum value and the minimum value of the surface material reflectance of highlighted buildings; the maximum value and the minimum value of the night ambient lighting in the third point cloud data collection area are obtained by collecting through a lux meter; and the quality noise correction value of the multispectral point cloud segmentation of highlighted buildings in the third point cloud data collection area is obtained by analysis.

[0115] In this embodiment, spectral library matching: matching the collected spectral data with a spectral library of known materials to obtain the maximum reflectance of the highlighted building surface material and the minimum reflectance of the highlighted building surface material.

[0116] The noise correction value of the multispectral point cloud segmentation quality of highlighted buildings is used to describe the degree of negative correction of highlighted buildings affected by the multispectral point cloud acquisition quality.

[0117] Number the third point cloud data collection area, Indicates the number of the third point cloud data collection area, Indicates the total number of numbers for the third point cloud data collection area.

[0118] Indicates The noise correction value of the multispectral point cloud segmentation quality of the highlighted buildings in the third point cloud data collection area.

[0119] ;

[0120] Extract predefined non-natural spectral component data from the multispectral image database of highlighted buildings;

[0121] Indicates The third point cloud data collection area The maximum reflectivity of the highlighted building surface material corresponding to the point cloud data detection time period.

[0122] Indicates The third point cloud data collection area The minimum reflectivity of the surface material of the highlighted building in the point cloud data detection time period.

[0123] Indicates The third point cloud data collection area The standard value of the reflectance of the surface material of the highlighted building is directly obtained from the multispectral image database of the highlighted building.

[0124] Surface material of the object being measured: Different surface materials will reflect or absorb light of different wavelengths, thus affecting the quality of the point cloud data. For example, if the reflectivity of different building surface materials varies greatly, the previous contrast evaluation result will be too large.

[0125] Indicates The third point cloud data collection area The maximum night ambient light intensity corresponding to the point cloud data detection time period.

[0126] Indicates The third point cloud data collection area The minimum night ambient light value corresponding to the point cloud data detection time period.

[0127] Indicates The third point cloud data collection area The night ambient lighting standard value corresponding to the detection time period of each point cloud data is directly obtained from the multispectral image database of highlighted buildings.

[0128] Measure the light intensity of artificial light sources, usually using a lux meter. The higher the light intensity, the greater the impact on the point cloud data. Artificial light sources around buildings (such as street lights, billboards, etc.) may cause interference (increase background noise) Spectral interference: Additional lights may introduce unnatural spectral components, interfering with the multispectral camera's capture of the natural spectral characteristics of the building surface.

[0129] It indicates the noise error factor corresponding to the laser radar used for acquisition. The value range is (0, 1). It indicates the influence of the corresponding noise error of the laser radar itself on the noise correction value of the multispectral point cloud segmentation quality of the highlighted building. It is directly obtained from the multispectral image database of highlighted buildings through a pre-set mapping relationship.

[0130] Indicates The first point cloud data collection area The wavelength of the laser during the detection period of each point cloud data. The wavelength affects the laser's penetration ability and reflectivity to different materials. Lasers with longer wavelengths have stronger penetration, but may have lower resolution and are easily affected by atmospheric scattering. Lasers with shorter wavelengths have higher resolution, but weaker penetration and are easily absorbed and scattered by particles such as water vapor and dust.

[0131] Indicates The first point cloud data collection area The laser wavelength standard value for each point cloud data detection period is directly obtained from the high-brightness building multispectral image database.

[0132] Indicates The first point cloud data collection area The laser divergence angle during the detection period of the point cloud data. The laser divergence angle determines the focusing degree of the laser beam, affecting the field of view of the lidar and the accuracy of the point cloud. A larger laser divergence angle will result in reduced accuracy of the point cloud because the reflected signal comes from a larger area. Although a laser divergence angle that is too small can improve accuracy, it will limit the detection range and may require a more complex scanning mechanism to cover a larger area.

[0133] Indicates The first point cloud data collection area The standard value of laser divergence angle in the detection period of point cloud data is directly obtained from the multispectral image database of highlighted buildings.

[0134] Furthermore, the specific process of comprehensively analyzing and obtaining the comprehensive value of the multispectral point cloud segmentation quality of highlighted buildings is: performing a first evaluation analysis on the multispectral point cloud segmentation data of highlighted buildings in the third point cloud data collection area to obtain the basic value of the multispectral point cloud segmentation quality fluctuation of highlighted buildings in the third point cloud data collection area; obtaining the noise correction value of the multispectral point cloud segmentation quality of highlighted buildings in the third point cloud data collection area; comprehensively analyzing the basic value of the multispectral point cloud segmentation quality fluctuation and the noise correction value of the multispectral point cloud segmentation quality of highlighted buildings in the third point cloud data collection area to obtain the comprehensive value of the multispectral point cloud segmentation quality of highlighted buildings.

[0135] In this embodiment, the comprehensive value of the segmentation quality of the multispectral point cloud of the highlighted building is used to describe the comprehensive degree of the negative impact of the multispectral point cloud acquisition quality on the highlighted building.

[0136] Number the third point cloud data collection area, Indicates the number of the third point cloud data collection area, Indicates the total number of numbers for the third point cloud data collection area.

[0137] Indicates The comprehensive value of the multispectral point cloud segmentation quality of the highlighted buildings in the third point cloud data collection area.

[0138] ;

[0139] Indicates The basic value of multispectral point cloud segmentation quality fluctuation of the highlighted buildings in the third point cloud data collection area.

[0140] Indicates The noise correction value of the multispectral point cloud segmentation quality of the highlighted buildings in the third point cloud data collection area.

[0141] It represents the correction factor of the lidar dark current for the comprehensive value of the segmentation quality of the multispectral point cloud of the highlight building, and the value range is (0, 1).

[0142] Take a series of images in a dark environment (for example, by placing the lidar in a completely light-proof box or using a camera cover to cover the lens). These images will contain only the dark current noise generated by the lidar itself.

[0143] To improve the accuracy of the correction, multiple dark current images were taken.

[0144] All dark current point cloud images are added together and then divided by the number of images to obtain the average dark current value.

[0145] The correction factor of the lidar dark current for the comprehensive value of the segmentation quality of the multispectral point cloud of the highlighted building is directly obtained from the multispectral image database of the highlighted building through a pre-set mapping relationship.

[0146] For example, a mapping set of lidar dark current and its corresponding correction factor for the comprehensive value of multi-spectral point cloud segmentation quality of highlighted buildings is constructed, and the average value of lidar dark current is input into the mapping set to obtain the weight factor of the correction factor of lidar dark current for the comprehensive value of multi-spectral point cloud segmentation quality of highlighted buildings, wherein the mapping relationship is a one-to-one correspondence or a many-to-one relationship.

[0147] Furthermore, a third point cloud data management process is performed according to the third comparative analysis result, specifically including: if the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building is less than the comprehensive threshold value of the multispectral point cloud segmentation quality of the highlighted building, no adjustment is made; if the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building is greater than or equal to the comprehensive threshold value of the multispectral point cloud segmentation quality of the highlighted building, then the difference between the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building and the comprehensive threshold value of the multispectral point cloud segmentation quality of the highlighted building is recorded as the multispectral point cloud segmentation quality difference of the highlighted building; if the multispectral point cloud segmentation quality difference of the highlighted building is less than the multispectral point cloud segmentation quality difference threshold of the highlighted building, density filtering is performed on the multispectral point cloud segmentation data of the highlighted building in the corresponding third point cloud data collection area; if the multispectral point cloud segmentation quality difference of the highlighted building is greater than or equal to the multispectral point cloud segmentation quality difference threshold of the highlighted building, bilateral filtering is performed on the multispectral point cloud segmentation data of the highlighted building in the corresponding third point cloud data collection area.

[0148] In this embodiment, density-based filtering: In the shadow area, the point cloud density may be low, and a density-based filtering method is used to identify and remove isolated points or abnormal points. For example, bilateral filtering combines spatial distance and intensity similarity to remove noise points while maintaining edge information.

[0149] Bilateral filtering combines spatial distance and brightness similarity to remove outliers while preserving edge information.

[0150] Furthermore, performing third point cloud data management and processing according to the third comparative analysis result also includes: collecting data again according to the third point cloud data collection area through the built-in positioning system of the drone and making a loop judgment until the comprehensive value of the multi-spectral point cloud segmentation quality of the highlighted building in the loop judgment is less than the comprehensive threshold of the multi-spectral point cloud segmentation quality of the highlighted building and the number of loop judgments is less than the loop judgment threshold; if the number of loop judgments is greater than or equal to the loop judgment threshold, notifying relevant personnel of the corresponding third point cloud data collection area.

[0151] In this embodiment, density-based filtering: In the shadow area, the point cloud density may be low, and a density-based filtering method is used to identify and remove isolated points or abnormal points, for example, bilateral filtering, which combines spatial distance and intensity similarity to remove noise points while maintaining edge information. Bilateral filtering combines spatial distance and brightness similarity to remove abnormal values ​​while maintaining edge information.

[0152] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0153] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0154] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0155] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0156] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0157] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for managing point cloud data of highlighted buildings based on multispectral images, characterized in that: The following steps are involved: The predefined spectral images of the highlighted buildings are collected by drones to obtain the original multispectral image data of the highlighted buildings and the original lidar data of the highlighted buildings, and the multispectral point cloud fusion data of the highlighted buildings are obtained after preprocessing. The multispectral point cloud fusion data of the highlighted building is segmented according to a predefined fusion data size to obtain segmented multispectral point cloud data of the highlighted building, the multispectral point cloud segmented data of the highlighted building is subjected to a first comparative analysis, and a first point cloud data management process is performed according to the first comparative analysis result; Perform a first evaluation analysis on the multispectral point cloud segmentation data of the highlighted building to obtain a basic value of the multispectral point cloud segmentation quality fluctuation of the highlighted building, perform a second comparative analysis based on the basic value of the multispectral point cloud segmentation quality fluctuation of the highlighted building, and perform a second point cloud data management process based on the second comparative analysis result; A second evaluation analysis is performed on the multispectral point cloud segmentation data of highlighted buildings to obtain the noise correction value of the multispectral point cloud segmentation quality of highlighted buildings. A comprehensive analysis is performed to obtain the comprehensive value of the multispectral point cloud segmentation quality of highlighted buildings. A third comparative analysis is performed based on the comprehensive value of the multispectral point cloud segmentation quality of highlighted buildings. The third point cloud data management and processing is performed based on the results of the third comparative analysis.

2. The method for managing point cloud data of highlighted buildings based on multispectral images as claimed in claim 1, characterized in that: The specific process of obtaining the multi-spectral point cloud fusion data of the highlighted building is as follows: The original data of the multispectral image of the highlighted building collected by the UAV is imported into the predefined spectral data processing software, and the radiation correction, geometric correction and atmospheric correction are performed on the original data of the multispectral image of the highlighted building by the predefined data processing software to obtain the corrected data of the multispectral image of the highlighted building; Import the raw data of the laser radar of the highlighted building collected by the drone into the predefined point cloud data processing software, convert the raw data of the laser radar of the highlighted building into the point cloud data of the highlighted building through the predefined point cloud data processing software and remove abnormal points to obtain the point cloud correction data of the highlighted building; According to the built-in positioning system of the drone, the corresponding highlighted building multispectral image correction data and highlighted building point cloud correction data are annotated with position data; According to the same location data annotation information, the corresponding highlighted building multispectral image correction data and the corresponding highlighted building point cloud correction data are fused to obtain the highlighted building multispectral point cloud fusion data.

3. The method for managing point cloud data of highlighted buildings based on multispectral images as claimed in claim 1, characterized in that: The multispectral point cloud segmentation data of the highlight building is subjected to a first comparative analysis, specifically comprising: The multispectral point cloud segmentation data of different highlight buildings are processed by predefined contrast software to obtain the discrete degree of the multispectral point cloud segmentation data of different highlight buildings; The multispectral point cloud segmentation data of different highlighted buildings are processed by predefined contrast software to obtain the contrast of the multispectral point cloud segmentation data of different highlighted buildings.

4. The method for managing point cloud data of highlighted buildings based on multispectral images as claimed in claim 1, characterized in that: The first point cloud data management process is performed according to the first comparative analysis result, and the specific process is as follows: If the discrete degree of the highlighted building multispectral point cloud segmentation data is less than the discrete degree threshold of the highlighted building multispectral point cloud segmentation data and the contrast of the highlighted building multispectral point cloud segmentation data is less than the contrast threshold of the highlighted building multispectral point cloud segmentation data, the corresponding segmentation area is recorded as the first point cloud data collection area; If the discrete degree of the highlighted building multispectral point cloud segmentation data is greater than or equal to the discrete degree threshold of the highlighted building multispectral point cloud segmentation data, the corresponding segmentation area is recorded as the second point cloud data collection area; If the contrast of the highlighted building multispectral point cloud segmentation data is greater than or equal to the contrast threshold of the highlighted building multispectral point cloud segmentation data, the corresponding segmentation area is recorded as the second point cloud data collection area; The position point corresponding to the second point cloud data collection area is obtained by using the built-in positioning tool of the drone, and the drone flies to the corresponding position point to collect data again and perform cyclic judgment until the discrete degree of the multispectral point cloud segmentation data of the highlighted building in the cyclic judgment is less than the discrete degree threshold of the multispectral point cloud segmentation data of the highlighted building and the contrast of the multispectral point cloud segmentation data of the highlighted building is less than the contrast threshold of the multispectral point cloud segmentation data of the highlighted building and the number of cyclic judgments is less than the cyclic judgment threshold; If the number of loop judgment times is greater than or equal to the loop judgment threshold, the corresponding second point cloud data collection area is notified to relevant personnel.

5. The method for managing point cloud data of highlighted buildings based on multispectral images as claimed in claim 1, characterized in that: The obtained quality fluctuation basic value of the multispectral point cloud segmentation of the highlighted building specifically includes: Classify and extract the multispectral point cloud segmentation data of the highlight buildings in the first point cloud data collection area to obtain the reflectivity, contrast and radiance of the highlight buildings in the first point cloud data collection area; The reflectivity of the highlighted building includes the maximum reflectivity of the highlighted building and the minimum reflectivity of the highlighted building; The highlight building contrast includes the maximum highlight building contrast and the minimum highlight building contrast; Directly obtain the standard values ​​of reflectance, contrast and radiance of highlighted buildings from the highlighted building multispectral image database; The basic value of the multispectral point cloud segmentation quality fluctuation of the highlighted buildings in the first point cloud data collection area is obtained through analysis.

6. The method for managing point cloud data of highlighted buildings based on multispectral images as claimed in claim 1, characterized in that: The second point cloud data management process is performed according to the second comparative analysis result, and the specific process is as follows: If the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building is greater than or equal to the quality fluctuation threshold of the multispectral point cloud segmentation of the highlighted building, the UAV reaches the first point cloud data collection area through the built-in position positioning system, opens the telescopic sunshade and extends the exposure time to collect predefined spectral images, and performs a cyclic judgment until the basic value of the quality fluctuation of the multispectral point cloud segmentation of the highlighted building in the cyclic judgment is less than the quality fluctuation threshold of the multispectral point cloud segmentation of the highlighted building and the number of cyclic judgments is less than the cyclic judgment threshold; If the number of loop judgments is greater than or equal to the loop judgment threshold, the corresponding first point cloud data collection area is notified to relevant personnel; If the highlighted building multispectral point cloud segmentation quality fluctuation base value is less than the highlighted building multispectral point cloud segmentation quality fluctuation threshold value, the first point cloud data collection area is recorded as the third point cloud data collection area.

7. The method for managing point cloud data of highlighted buildings based on multispectral images as claimed in claim 1, characterized in that: The obtained high-brightness building multispectral point cloud segmentation quality noise correction value specifically includes: Classify and extract the multispectral point cloud segmentation data of the highlighted buildings in the third point cloud data collection area to obtain the surface material reflectance of the highlighted buildings in the third point cloud data collection area; The reflectivity of the surface material of the highlighted building includes the maximum reflectivity of the surface material of the highlighted building and the minimum reflectivity of the surface material of the highlighted building; The maximum value and the minimum value of the night ambient light in the third point cloud data collection area are obtained by using a lux meter; The noise correction value of the multispectral point cloud segmentation quality of the highlighted building in the third point cloud data collection area is obtained by analysis.

8. The method for managing point cloud data of highlighted buildings based on multispectral images as claimed in claim 1, characterized in that: The specific process of obtaining the comprehensive value of the segmentation quality of the multispectral point cloud of the highlighted building through comprehensive analysis is as follows: Perform a first evaluation analysis on the multispectral point cloud segmentation data of the highlight buildings in the third point cloud data collection area to obtain a basic value of the multispectral point cloud segmentation quality fluctuation of the highlight buildings in the third point cloud data collection area; Obtain the noise correction value of the multispectral point cloud segmentation quality of the highlighted buildings in the third point cloud data collection area; The basic value of the multispectral point cloud segmentation quality fluctuation of the highlighted buildings in the third point cloud data collection area and the noise correction value of the multispectral point cloud segmentation quality of the highlighted buildings are comprehensively analyzed to obtain the comprehensive value of the multispectral point cloud segmentation quality of the highlighted buildings.

9. The method for managing point cloud data of highlighted buildings based on multispectral images as claimed in claim 1, characterized in that: The performing of the third point cloud data management process according to the third comparative analysis result specifically includes: If the comprehensive value of the segmentation quality of the multispectral point cloud of the highlighted building is less than the comprehensive threshold of the segmentation quality of the multispectral point cloud of the highlighted building, no adjustment will be made; If the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building is greater than or equal to the comprehensive threshold value of the multispectral point cloud segmentation quality of the highlighted building, the difference between the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building and the comprehensive threshold value of the multispectral point cloud segmentation quality of the highlighted building is recorded as the multispectral point cloud segmentation quality difference value of the highlighted building; If the multispectral point cloud segmentation quality difference of the highlighted building is less than the multispectral point cloud segmentation quality difference threshold of the highlighted building, density filtering is performed on the multispectral point cloud segmentation data of the highlighted building in the corresponding third point cloud data collection area; If the multispectral point cloud segmentation quality difference of the highlighted building is greater than or equal to the multispectral point cloud segmentation quality difference threshold of the highlighted building, bilateral filtering is performed on the multispectral point cloud segmentation data of the highlighted building in the corresponding third point cloud data collection area.

10. The method for managing point cloud data of highlighted buildings based on multispectral images according to claim 1, characterized in that: The performing third point cloud data management processing according to the third comparative analysis result also includes: The built-in positioning system of the drone collects data again according to the third point cloud data collection area and performs a cyclic judgment until the comprehensive value of the multispectral point cloud segmentation quality of the highlighted building in the cyclic judgment is less than the comprehensive threshold value of the multispectral point cloud segmentation quality of the highlighted building and the number of cyclic judgments is less than the cyclic judgment threshold value; If the number of loop judgments is greater than or equal to the loop judgment threshold, the corresponding third point cloud data collection area is notified to relevant personnel.

Images