Intelligent surveying and mapping system for urban surveying and mapping based on Internet of Things

Abstract

The invention relates to the technical field of intelligent surveying and mapping, and discloses an intelligent surveying and mapping system for urban surveying and mapping based on Internet of Things, which comprises a data acquisition module, a surveying and mapping data processing and integration module, a surveying and mapping data analysis module, an intelligent identification module and a result output module, through the arrangement of the surveying and mapping data analysis module, point cloud data and image data in surveying and mapping data are processed and integrated to improve the accuracy of the surveying and mapping data, the surveying and mapping data of different data sources and data types are analyzed in corresponding modes, and then a point cloud processing model and an image processing model are constructed to improve the accuracy of the surveying and mapping data. According to the method, the point cloud data and the image data are analyzed respectively, and then the point cloud data and the image data are combined to obtain the final earth surface object category, so that the accuracy of urban earth surface object recognition by the surveying and mapping data is improved, and meanwhile, a foundation is laid for subsequent intelligent recognition of suspicious targets.

Description

Technical Field

[0001] The present invention relates to the field of intelligent surveying and mapping technology, and more particularly to an intelligent surveying and mapping system for urban surveying and mapping based on the Internet of Things. Background Art

[0002] With the acceleration of urbanization, urban surveying and mapping has become increasingly important; however, traditional surveying and mapping technologies often require a lot of manpower and material resources, are inefficient, and have difficulty in monitoring urban changes in real time.

[0003] The public document with publication number CN115979229B discloses an intelligent surveying and mapping system for urban surveying based on the Internet of Things, which is used to solve the problem that the current three-dimensional data of buildings is collected too long ago and has no reference value during urban surveying work. The system includes a structural analysis module, a comprehensive actual acquisition module, a comprehensive surveying and mapping module and a three-dimensional matching module. The structural analysis module is used to determine the structural complexity level of the target surveying and mapping object. The structural analysis module is also used to set the number of surveying and mapping samples of the target surveying and mapping object. The comprehensive surveying and mapping module is used to perform comprehensive surveying and mapping on the real-time three-dimensional virtual map of the target surveying and mapping object. The three-dimensional matching module is used to compare and match the three-dimensional virtual map of the target surveying and mapping object. The three-dimensional matching module also compares the real-time three-dimensional virtual map with the verification three-dimensional virtual map. The present invention realizes the accurate update of the three-dimensional data of the target surveying and mapping object based on the historical characteristic factors of the building.

[0004] However, existing urban surveying and mapping generates a large amount of data, and the data sources and data types are different. Therefore, different methods need to be used for processing during analysis to obtain corresponding analysis results. Cities are changing every day and the urban environment is becoming more and more complex. Therefore, how to make more effective use of high-precision data is extremely important. The existing surveying and mapping system only performs surveying and mapping functions, and does not make effective and in-depth use of the surveying and mapping data after obtaining it.

[0005] In view of this, the present invention proposes an intelligent surveying and mapping system for urban surveying based on the Internet of Things, which makes full use of surveying and mapping data and adopts intelligent data processing technology to effectively analyze the surveying and mapping data to improve the accuracy and efficiency of urban surveying and mapping. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides an intelligent surveying and mapping system for urban surveying based on the Internet of Things to solve the problems existing in the above-mentioned background technology.

[0007] The present invention provides the following technical solutions: an intelligent surveying and mapping system for urban surveying and mapping based on the Internet of Things, comprising a data acquisition module, a surveying and mapping data processing and integration module, a surveying and mapping data analysis module, an intelligent recognition module, and a result output module;

[0008] The data acquisition module is used to collect real-time surveying and mapping data from multiple data sources;

[0009] The surveying and mapping data processing and integration module is used to receive and pre-process the data from the data acquisition module, record the pre-processed real-time surveying and mapping data as target data, and perform data integration on the target data;

[0010] The surveying and mapping data analysis module is used to receive data from the surveying and mapping data processing and integration module, build a deep learning model to analyze the real-time surveying and mapping data, obtain the surface objects corresponding to the real-time surveying and mapping data, and perform suspicious target identification;

[0011] The intelligent recognition module is used to intelligently identify suspicious targets and distinguish them into normal targets and warning targets;

[0012] The result output module is used to provide early warning prompts to early warning targets and to update the surveying and mapping data in the geographic information system in real time.

[0013] Preferably, the surveying and mapping data include three-dimensional point cloud data and surface image data of urban topography, landforms, urban buildings, urban roads, and vegetation; the surface image data include satellite remote sensing data and urban image data, and high-resolution satellite images are used to obtain the topography, landforms, and land use of the city, namely, satellite remote sensing data, and drones are used to take pictures to obtain urban images, namely, urban image data; three-dimensional point cloud data are collected through three-dimensional laser scanning, and the three-dimensional laser scanning uses lidar technology to quickly obtain three-dimensional structural information of urban buildings, roads, and vegetation.

[0014] Preferably, the surface objects include surface buildings, roads and vegetation; the suspicious targets are targets that are different from existing surveying and mapping data; the existing surveying and mapping data are surveying and mapping data in a geographic information system; the normal targets are surface objects with normal changes, and the warning targets are surface objects with violations.

[0015] Preferably, the surveying and mapping data processing and integration module obtains more accurate real-time surveying and mapping data by performing point cloud data processing on three-dimensional point cloud data and image feature extraction on surface image data;

[0016] The point cloud data processing includes:

[0017] Get the coordinates of the 3D point cloud;

[0018] Filter the noise of 3D point cloud coordinate data to remove abnormal outliers in the point cloud data;

[0019] Perform point cloud registration and use the iterative closest point algorithm to optimize the rigid body transformation matrix;

[0020] The image feature extraction includes:

[0021] Preprocess the image and use the Canny edge detection algorithm to extract the contours of surface objects;

[0022] Then, the road layout is identified based on the constructed U-Net semantic segmentation network.

[0023] Preferably, extracting the contours of surface objects using the Canny edge detection algorithm includes:

[0024] Step S01: using Gaussian filtering to perform Gaussian smoothing on the input image;

[0025] Step S02: Calculate the gradient direction and gradient magnitude using the Sobel operator; the gradient magnitude formula is expressed as: Among them, G is the gradient amplitude, G x Represents the horizontal gradient, G y The gradient in the vertical direction; the formula for the gradient direction is expressed as: Among them, θ represents the gradient direction;

[0026] Step S03: Perform non-maximum suppression, comparing the gradient amplitude of the current pixel with the pixels on both sides along the gradient direction θ. If the gradient value of the current pixel is not the maximum value in this direction, it is set to zero.

[0027] Step S04: distinguishing strong edges, weak edges and non-edge areas;

[0028] Step S05: extract long straight line segments and fit the contour of the surface object: ρ = xcosθ + ysinθ; where ρ represents the distance from the straight line to the origin, θ represents the angle of the straight line, and x and y represent pixel points.

[0029] Preferably, the target data is integrated in the following manner:

[0030] The target data is fused to form a unified data set, merging data from different sources and formats; the formula is: Among them, S ij is the i-th category data of the j-th data source after merging, k j is the weight of the j-th data source, s ij is the i-th type of data of the j-th data source before merging, β is the time synchronization compensation coefficient, is the clock deviation between data sources, and J is the total number of data sources.

[0031] Preferably, the specific manner in which the surveying and mapping data analysis module obtains the surface objects corresponding to the real-time surveying and mapping data is:

[0032] The target data after data integration is divided into two categories: point cloud data and image data. The point cloud data are marked as D1, D2, D3, ..., D n ; Image data is marked as X1, X2, X3, ..., X n ;Point cloud data corresponds to image data one by one;

[0033] Input the point cloud data into the point cloud processing model, input the image data into the image processing model, annotate the target data, annotate the point cloud data by assigning category labels, train the point cloud processing model and the image processing model respectively, and perform weighted fusion of the outputs of the point cloud processing model and the image processing model through the attention mechanism to obtain the final output. The formula is expressed as:

[0034] P r_final =ω·P r_point +(1-ω)·P r_image ;

[0035] Among them, P r_fianl is the final output recognition and classification result of the rth data, P r_point is the recognition and classification result output by the point cloud processing model for the rth data, P r_image is the recognition and classification result of the rth data output by the image processing model, ω represents the learnable parameter; the P r_point Specifically, it is the probability that the r-th point cloud data belongs to the label corresponding to each type of surface object. r_image Specifically, it is the probability that the rth image data belongs to each type of surface object; the P r_final Specifically, it is the probability that the rth data finally belongs to each type of surface object; take P r_final The surface object category corresponding to the maximum value of is the final surface object.

[0036] Preferably, the surveying and mapping data analysis module determines suspicious targets in the following manner:

[0037] Obtain existing surveying and mapping data, compare the target data with the existing surveying and mapping data, mark the point cloud data and surface objects that are different; and treat the marked surface objects and the surface objects corresponding to the marked point cloud data as suspicious targets.

[0038] Preferably, the specific process of the intelligent identification module distinguishing suspicious targets into normal targets and warning targets is as follows:

[0039] Obtaining point cloud data and warning data of suspicious targets, the warning data mainly includes indicator data of building red lines, land use nature, building height and land use scope;

[0040] Select the comparison data in the point cloud data of the suspicious target that is consistent with the warning data. If the comparison data does not meet the indicator range specified in the indicator data of the warning data, the suspicious target corresponding to the comparison data is distinguished as a warning target. If the comparison data meets the indicator range specified in the indicator data of the warning data, the suspicious target corresponding to the comparison data is distinguished as a normal target.

[0041] Preferably, the content of the warning prompt provided by the result output module includes the warning data type of the warning target, displays the warning target and data category, verifies and measures the warning target, removes the surveying and mapping data of the warning target in the real-time surveying and mapping data, and enters other real-time surveying and mapping data into the geographic information system for real-time updating.

[0042] Technical effects and advantages of the present invention:

[0043] (1) The present invention is provided with a surveying and mapping data analysis module, which is conducive to improving the accuracy of surveying and mapping data by processing and integrating the point cloud data and image data in the surveying and mapping data separately, analyzing the surveying and mapping data of different data sources and data types in a corresponding manner, and then analyzing the point cloud data and image data separately by constructing a point cloud processing model and an image processing model, and then combining the point cloud data and the image data to obtain the final surface object category, thereby improving the accuracy of surveying and mapping data in identifying urban surface objects, and at the same time laying the foundation for the subsequent intelligent identification of suspicious targets.

[0044] (2) The present invention is provided with an intelligent recognition module, which is conducive to intelligently identifying suspicious targets, distinguishing suspicious targets into normal targets and warning targets, and making more efficient use of high-precision surveying and mapping data. While surveying and mapping data is used for surveying, the surveying and mapping data is further analyzed in depth to obtain surface buildings or surface objects with violations, so as to conduct timely inspections and measurements to prevent the occurrence of violations, and to assist staff in conducting violation detection with scientific data to ensure the standardization and compliance of land use. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a structural diagram of the intelligent surveying and mapping system for urban surveying based on the Internet of Things of the present invention. DETAILED DESCRIPTION

[0046] The technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the present invention. In addition, the forms of the various structures described in the following embodiments are merely examples. The intelligent surveying and mapping system for urban surveying based on the Internet of Things involved in the present invention is not limited to the various structures described in the following embodiments. All other embodiments obtained by ordinary technicians in this field without making creative work fall within the scope of protection of the present invention.

[0047] like Figure 1 As shown, the present invention provides an intelligent surveying and mapping system for urban surveying and mapping based on the Internet of Things, including a data acquisition module, a surveying and mapping data processing and integration module, a surveying and mapping data analysis module, an intelligent recognition module, and a result output module;

[0048] The data acquisition module is used to collect real-time surveying and mapping data from multiple data sources; the surveying and mapping data includes, but is not limited to, three-dimensional point cloud data and surface image data of urban topography, landforms, urban buildings, urban roads, vegetation, etc.; the surface image data includes satellite remote sensing data and urban image data, using high-resolution satellite images to obtain information data such as urban topography, landforms, and land use, namely satellite remote sensing data, and using drones to obtain urban images, namely urban image data; 3D point cloud data is collected through 3D laser scanning, which can use lidar technology to quickly obtain 3D structural information of urban buildings, roads, vegetation, etc.

[0049] The surveying and mapping data processing and integration module is used to receive data from the data acquisition module and perform preprocessing to obtain more accurate real-time surveying and mapping data, record the preprocessed real-time surveying and mapping data as target data, and perform data integration on the target data;

[0050] The surveying and mapping data analysis module is used to receive data from the surveying and mapping data processing and integration module, build a deep learning model to analyze the real-time surveying and mapping data, obtain surface objects corresponding to the real-time surveying and mapping data, and determine suspicious targets; the surface objects include but are not limited to surface buildings, roads, and vegetation; the suspicious targets are targets that differ from existing surveying and mapping data, such as buildings and land use that are suspected of violating regulations; the existing surveying and mapping data is the surveying and mapping data in the geographic information system;

[0051] The intelligent recognition module is used to intelligently identify suspicious targets and distinguish them into normal targets and warning targets; the normal targets are surface objects with normal changes, and the warning targets are surface objects with irregularities;

[0052] The result output module is used to provide early warning prompts to early warning targets and to update the surveying and mapping data in the geographic information system in real time.

[0053] In this embodiment, it should be specifically explained that the mapping data processing and integration module obtains more accurate real-time mapping data by performing point cloud data processing on three-dimensional point cloud data and image feature extraction on surface image data;

[0054] The point cloud data is generated using a multi-beam (e.g., 128-beam) rotating laser radar with a scanning frequency of ≥20 Hz, a wavelength of 905 nm, and a maximum range of 300 m. The laser pulse transmitter emits laser light at the target, and the receiver records the time difference of the reflected signal.

[0055] The point cloud data processing includes: obtaining three-dimensional point cloud coordinates:

[0056] Among them, x d is the x-axis coordinate of the three-dimensional point, y d is the y-axis coordinate of the three-dimensional point cloud, z d is the z-axis coordinate of the three-dimensional point cloud, R represents the laser ranging value, θ represents the vertical angle, i.e. the elevation angle, and φ represents the horizontal angle, i.e. the azimuth angle. Where c is the speed of light and Δt is the time difference;

[0057] Noise filtering is performed on the three-dimensional point cloud coordinate data to remove abnormal outliers in the point cloud data. If there is a point p that satisfies the removal algorithm formula, then point p is removed. The removal algorithm formula is expressed as:

[0058] Among them, p a represents the ath neighboring point of point p in the neighborhood, A represents the total number of neighboring points, μ l represents the mean distance between points in the domain, σ l represents the standard deviation of the distance between points in the domain, and α represents the threshold coefficient, which is usually 1 to 2;

[0059] Perform point cloud registration and use the iterative closest point algorithm to optimize the rigid body transformation matrix Z. The formula is expressed as:

[0060] in, represents the value of Z when the objective function f(Z) on the right is minimized, f(Z) is the objective function, p′ b is the b-th source point cloud, is the nearest neighbor point of the bth target point cloud, B represents the total number of point clouds, b=1, 2, 3, ..., B; the source point cloud is the point cloud data in the existing surveying and mapping data, and the target point cloud is the point cloud data in the real-time surveying and mapping data.

[0061] In this embodiment, it should be specifically noted that the image feature extraction uses a computer vision algorithm to extract key features; the image is preprocessed and the Canny edge detection algorithm is used to extract the contours of surface objects, and then the road layout is recognized based on the constructed U-Net semantic segmentation network;

[0062] The image preprocessing includes fusing multispectral images. Since urban mapping requires vegetation information, multispectral images can provide color information of surface objects. After fusing multispectral images, a simulated true color image can be obtained. The fusion formula is expressed as:

[0063] I fused =λ1I RGB +λ2I NIR ; Among them, I fused Represents the fused multispectral image band, I RGB Represents the visible light band, I NIR represents the near-infrared band, λ1 and λ2 are corresponding weight coefficients respectively; satisfying 0<λ1<1; 0<λ2<1; the specific values of λ1 and λ2 can be set in detail by those skilled in the art according to the vegetation index of different cities, and this embodiment does not specifically limit these specific values;

[0064] The extraction of the contour of the surface object by using the Canny edge detection algorithm includes:

[0065] Step S01: Gaussian smoothing is performed on the input image using Gaussian filtering to suppress noise interference and prevent high-frequency noise from being misjudged as edges;

[0066] Step S02: Use the Sobel operator to calculate the gradient direction and gradient magnitude to detect the intensity and direction of grayscale changes in the image; the formula for the gradient magnitude is expressed as: Among them, G is the gradient amplitude, G x Represents the horizontal gradient, G y The gradient in the vertical direction; the formula for the gradient direction is expressed as: Among them, θ represents the gradient direction;

[0067] Step S03: Perform non-maximum suppression to refine the edges and retain only the pixels with the local maximum in the gradient direction; compare the gradient amplitude of the current pixel with the pixels on both sides along the gradient direction θ. If the gradient value of the current pixel is not the maximum in this direction, set it to zero; for example, if the current pixel is (x, y), if θ = 0°, compare the gradient values of the left pixel (x-1, y) with the right pixel (x+1, y); if θ = 45°, compare the gradient values of the upper left pixel (x-1, y+1) with the lower right pixel (x+1, y-1);

[0068] Step S04: Distinguish strong edges, weak edges and non-edge areas; set a high threshold Y high With low threshold Y low , high threshold is to retain obvious edges, low threshold is to retain potential edges, Y low =0.5×Yhigh ; Y high The value of can be set by those skilled in the art. high Take 30% of the maximum gradient amplitude; when G(x,y)≥Y high , then the pixel point (x, y) is a strong edge and is directly retained; when Y low ≤G(x,y) <Y high When G(x,y) is connected to a strong edge, the pixel (x,y) is retained. In this case, the pixel (x,y) is a weak edge. <Y low When , the pixel point (x, y) is not an edge point, and the pixel point is removed; the isolated weak edge point is connected to the strong edge;

[0069] Step S05: extract long straight line segments and fit the contour of the surface object: ρ = xcosθ + ysinθ; where ρ represents the distance from the straight line to the origin, and θ represents the angle of the straight line;

[0070] The U-Net semantic segmentation network mainly includes an input layer, an encoder, a decoder, a loss function, and an output layer; the input layer is used to input data; the encoder is used to extract multi-scale features and gradually compress the spatial dimension. The encoder includes 4 to 5 downsampling blocks, each of which contains a two-dimensional convolution layer and a 2×2 maximum pooling layer; the decoder is used to restore spatial resolution and accurately locate road boundaries. The decoder includes transposed convolution and jump connections. The jump connection splices the encoder feature map with the decoder feature map at the same layer to retain detailed information; the output layer is used to output the probability of each pixel belonging to the road, P(x,y)∈[0,1]; where P(x,y) represents the probability that the pixel (x,y) belongs to the road;

[0071] The loss function is expressed as:

[0072] LOSS=λ3LOSS Dice +(1-λ3)LOSS CE ;

[0073] Among them, LOSS represents the comprehensive loss, λ3 represents the weight factor, and the value range is 0.5 to 0.7. Dice Represents the Dice LOSS loss function, LOSS CE represents the cross entropy loss;

[0074] LOSS CE =-∑Y i logP i +(1-Y i )log(1-P i );

[0075] Among them, Pi represents the predicted probability, Y i represents the true label, and ∈ represents the smoothing term.

[0076] In this embodiment, it should be specifically explained that the method of integrating the target data is:

[0077] The target data is cleaned and normalized, and the target data is formatted and coordinates are unified to ensure that data with different characteristics are scaled to the same dimension range, laying the foundation for subsequent data integration; the normalization process can be any of data standardization, normalization, and nonlinear normalization; the data standardization converts the characteristic attributes of the data into a standard normal distribution with a mean of 0 and a variance of 1 by subtracting the mean and dividing by the variance. The formula is expressed as: is the normalized data, s is the original data, μ is the mean of the original data, and σ is the variance of the original data; the normalization linearly scales the data to between 0 and 1, and the formula is expressed as: Among them, s′ represents the normalized data, s is the original data, and s min Indicates the minimum value of the original data, s max Represents the maximum value of the original data; the nonlinear normalization maps the original data value through, for example, a logarithmic function, an inverse tangent function, etc.;

[0078] The target data is fused to form a unified data set, merging data from different sources and formats; the formula is: Among them, S ij is the i-th category data of the j-th data source after merging, k j is the weight of the j-th data source, s ij is the i-th type of data of the j-th data source before merging, β is the time synchronization compensation coefficient, is the clock deviation between data sources, J is the total number of data sources;

[0079] The integrated target data is verified to ensure its accuracy and consistency, and then the merged target data is integrated into the geographic information system (GIS) to achieve dynamic updates.

[0080] In this embodiment, it should be specifically explained that the specific manner in which the surveying and mapping data analysis module obtains the surface objects corresponding to the real-time surveying and mapping data is:

[0081] The target data after data integration is divided into two categories: point cloud data and image data. The point cloud data are marked as D1, D2, D3, ..., D n ; Image data is marked as X1, X2, X3, ..., X n; Point cloud data and image data correspond one to one, that is, image data X of the same position or the same object r The corresponding point cloud data is D r ;

[0082] Input the point cloud data into the point cloud processing model, input the image data into the image processing model, and perform data annotation on the target data. The point cloud data is annotated by assigning category labels, for example, building = 1, road = 2, vegetation = 3; the image data is annotated. The point cloud processing model may be a PointNet model, and the image processing model may be a DeepLabv3 model.

[0083] The point cloud processing model and image processing model are trained separately, and the outputs of the point cloud processing model and the image processing model are weightedly fused through the attention mechanism to obtain the final output. The formula is expressed as:

[0084] P r_final =ω·P r_point +(1-ω)·P r_image ;

[0085] Among them, P r_final is the final output recognition and classification result of the rth data, P r_point is the recognition and classification result output by the point cloud processing model for the rth data, P r_image is the recognition and classification result of the rth data output by the image processing model, ω represents the learnable parameter; the P r_point Specifically, it is the probability that the r-th point cloud data belongs to the label corresponding to each type of surface object. r_image Specifically, it is the probability that the rth image data belongs to each type of surface object; the P r_final Specifically, it is the probability that the rth data finally belongs to each type of surface object; take P r_final The surface object category corresponding to the maximum value of is the final surface object.

[0086] In this embodiment, it should be specifically explained that the surveying and mapping data analysis module determines suspicious targets in the following manner:

[0087] Obtain existing surveying and mapping data, compare the target data with the existing surveying and mapping data, and mark the point cloud data and surface objects with differences; treat the marked surface objects and the surface objects corresponding to the marked point cloud data as suspicious targets; for example, if a surface object at a certain location in the existing surveying and mapping data is vegetation, and the surface object at that location in the target data is a building, then mark the surface object at that location; if the building area of the point cloud data of a building in the existing surveying and mapping data is 500, and the building area of the point cloud data of the building in the target data is 600, then mark the building corresponding to the point cloud data.

[0088] In this embodiment, it should be specifically explained that the specific process of the intelligent recognition module distinguishing suspicious targets into normal targets and warning targets is as follows:

[0089] Obtain point cloud data and warning data for suspicious targets. The warning data includes indicators such as building red lines, land use nature, building height, and land use scope. The warning data is collected from multi-source heterogeneous data using big data and natural language processing technologies, including news, social media, academic papers, and government work reports. A synonym dictionary is constructed to clean the text and identify the context of the suspicious targets. Pre-trained models such as the BERT word vector model are used to annotate entities related to the suspicious targets, identify the modification relationship between indicators and values related to the suspicious targets, and accurately extract metadata such as the name, value, unit, and time range of the indicators related to the suspicious targets. A unit mapping table is established to convert the extracted indicators related to the suspicious targets into a format usable by databases or analysis tools.

[0090] Select the comparison data that is consistent with the warning data in the point cloud data of the suspicious target. If the comparison data does not meet the index range specified in the index data of the warning data, the suspicious target corresponding to the comparison data is classified as a warning target. If the comparison data meets the index range specified in the index data of the warning data, the suspicious target corresponding to the comparison data is classified as a normal target.

[0091] For example, if the suspicious target is building K, the comparison data in K's point cloud data that is consistent with the warning data is the height, the indicator data of the warning data is that the height must not be greater than 50m, and the comparison data is a height equal to 48m, then the comparison data meets the indicator data of the warning data, that is, the building K corresponding to the comparison data is a normal target; if the comparison data of building K is a height equal to 55m, then the comparison data does not meet the indicator data of the warning data, that is, the building K corresponding to the comparison data is a warning target;

[0092] If the suspicious target is a surface object K, and the comparison data in the point cloud data of K that is consistent with the warning data is the land use nature, the land use nature data can be marked, and different land use natures can be assigned different labels, such as building land is 1, green land is 2, road land is 3, etc.; the land use nature of the indicator data of the warning data is building land, which cannot be changed, that is, label 1, and cannot be other labels; the land use nature of the comparison data is green land, that is, label 2, then the comparison data does not meet the indicator data of the warning data, that is, the land use nature is not building land, that is, label is not 1, that is, the surface object K corresponding to the comparison data is the warning target; if the land use nature of the indicator data of the warning data is building land, which can be changed to green land, that is, label 1 or 2, and the land use nature of the comparison data is green land, that is, label 2, then the comparison data meets the indicator data of the warning data, that is, the land use nature is changeable green land, that is, label can be changed from 1 to 2, then the surface object K corresponding to the comparison data is a normal target.

[0093] In this embodiment, it should be specifically explained that the content of the warning prompt provided by the result output module includes the warning data type of the warning target, and displays the warning target and data category, for example, the land use nature of building K, the land use range of road K, etc.; verifies and measures the warning target to determine whether there is any violation; removes the surveying and mapping data of the warning target in the real-time surveying and mapping data, and enters other real-time surveying and mapping data into the geographic information system for real-time updating.

[0094] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0095] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. An intelligent surveying and mapping system for urban surveying based on the Internet of Things, characterized by: It includes data acquisition module, surveying and mapping data processing and integration module, surveying and mapping data analysis module, intelligent recognition module and result output module; The data acquisition module is used to collect real-time surveying and mapping data from multiple data sources; The surveying and mapping data processing and integration module is used to receive and pre-process the data from the data acquisition module, record the pre-processed real-time surveying and mapping data as target data, and perform data integration on the target data; The surveying and mapping data analysis module is used to receive data from the surveying and mapping data processing and integration module, build a deep learning model to analyze the real-time surveying and mapping data, obtain the surface objects corresponding to the real-time surveying and mapping data, and perform suspicious target identification; The intelligent recognition module is used to intelligently identify suspicious targets and distinguish them into normal targets and warning targets; The result output module is used to provide early warning prompts to early warning targets and to update the surveying and mapping data in the geographic information system in real time.

2. The intelligent urban surveying and mapping system based on the Internet of Things according to claim 1, characterized in that: The surveying and mapping data includes three-dimensional point cloud data and surface image data of urban terrain, landforms, urban buildings, urban roads, and vegetation; the surface image data includes satellite remote sensing data and urban image data. High-resolution satellite images are used to obtain the city's terrain, landforms, and land use, namely satellite remote sensing data, and drones are used to capture urban images, namely urban image data; three-dimensional point cloud data is collected through three-dimensional laser scanning, and the three-dimensional laser scanning uses lidar technology to quickly obtain three-dimensional structural information of urban buildings, roads, and vegetation.

3. The intelligent surveying and mapping system for urban surveying and mapping based on the Internet of Things according to claim 2, characterized in that: The surface objects include surface buildings, roads and vegetation; the suspicious targets are targets that are different from the existing surveying and mapping data; the existing surveying and mapping data are the surveying and mapping data in the geographic information system; the normal targets are surface objects with normal changes, and the warning targets are surface objects with violations.

4. The intelligent urban surveying and mapping system based on the Internet of Things according to claim 3 is characterized by: The surveying and mapping data processing and integration module obtains more accurate real-time surveying and mapping data by performing point cloud data processing on three-dimensional point cloud data and image feature extraction on surface image data; The point cloud data processing includes: Get the coordinates of the 3D point cloud; Filter the noise of 3D point cloud coordinate data to remove abnormal outliers in the point cloud data; Perform point cloud registration and use the iterative closest point algorithm to optimize the rigid body transformation matrix; The image feature extraction includes: Preprocess the image and use the Canny edge detection algorithm to extract the contours of surface objects; Then, the road layout is identified based on the constructed U-Net semantic segmentation network.

5. The intelligent surveying and mapping system for urban surveying and mapping based on the Internet of Things according to claim 4, characterized in that: The extraction of the contour of the surface object by using the Canny edge detection algorithm includes: Step S01: using Gaussian filtering to perform Gaussian smoothing on the input image; Step S02: Calculate the gradient direction and gradient magnitude using the Sobel operator; the gradient magnitude formula is expressed as: Among them, G is the gradient amplitude, G x Represents the horizontal gradient, G y The gradient in the vertical direction; the formula for the gradient direction is expressed as: Among them, θ represents the gradient direction; Step S03: Perform non-maximum suppression, comparing the gradient amplitude of the current pixel with the pixels on both sides along the gradient direction θ. If the gradient value of the current pixel is not the maximum value in this direction, it is set to zero. Step S04: distinguishing strong edges, weak edges and non-edge areas; Step S05: extract long straight line segments and fit the contour of the surface object: ρ = xcosθ + ysinθ; where ρ represents the distance from the straight line to the origin, θ represents the angle of the straight line, and x and y represent pixel points.

6. The intelligent urban surveying and mapping system based on the Internet of Things according to claim 5, characterized in that: The method of performing data integration on the target data is: The target data is fused to form a unified data set, merging data from different sources and formats; the formula is: Among them, S ij is the i-th category data of the j-th data source after merging, k j is the weight of the j-th data source, s ij is the i-th type of data of the j-th data source before merging, β is the time synchronization compensation coefficient, is the clock deviation between data sources, and J is the total number of data sources.

7. The intelligent urban surveying and mapping system based on the Internet of Things according to claim 6, characterized in that: The specific method for the surveying and mapping data analysis module to obtain the surface objects corresponding to the real-time surveying and mapping data is: The target data after data integration is divided into two categories: point cloud data and image data. The point cloud data are marked as D1, D2, D3, ..., D n ; Image data is marked as X1, X2, X3, ..., X n ;Point cloud data corresponds to image data one by one; Input the point cloud data into the point cloud processing model, input the image data into the image processing model, annotate the target data, annotate the point cloud data by assigning category labels, train the point cloud processing model and the image processing model respectively, and perform weighted fusion of the outputs of the point cloud processing model and the image processing model through the attention mechanism to obtain the final output. The formula is expressed as: P r_final =ω·P r_point +(1-ω)·P r_image ; Among them, P r_final is the final output recognition and classification result of the rth data, P r_point is the recognition and classification result output by the point cloud processing model for the rth data, P r_image is the recognition and classification result of the rth data output by the image processing model, ω represents the learnable parameter; the P r_point Specifically, it is the probability that the r-th point cloud data belongs to the label corresponding to each type of surface object. r_image Specifically, it is the probability that the rth image data belongs to each type of surface object; the P r_final Specifically, it is the probability that the rth data finally belongs to each type of surface object; take P r_final The surface object category corresponding to the maximum value of is the final surface object.

8. The intelligent urban surveying and mapping system based on the Internet of Things according to claim 7, characterized in that: The surveying and mapping data analysis module determines suspicious targets in the following way: Obtain existing surveying and mapping data, compare the target data with the existing surveying and mapping data, mark the point cloud data and surface objects that are different; and treat the marked surface objects and the surface objects corresponding to the marked point cloud data as suspicious targets.

9. The intelligent urban surveying and mapping system based on the Internet of Things according to claim 8, characterized in that: The specific process of the intelligent recognition module distinguishing suspicious targets into normal targets and warning targets is as follows: Obtaining point cloud data and warning data of suspicious targets, the warning data mainly includes indicator data of building red lines, land use nature, building height and land use scope; Select the comparison data in the point cloud data of the suspicious target that is consistent with the warning data. If the comparison data does not meet the indicator range specified in the indicator data of the warning data, the suspicious target corresponding to the comparison data is distinguished as a warning target. If the comparison data meets the indicator range specified in the indicator data of the warning data, the suspicious target corresponding to the comparison data is distinguished as a normal target.

10. The intelligent surveying and mapping system for urban surveying and mapping based on the Internet of Things according to claim 9, characterized in that: The content of the warning prompt provided by the result output module includes the warning data type of the warning target, displays the warning target and data category, verifies and measures the warning target, removes the surveying and mapping data of the warning target in the real-time surveying and mapping data, and enters other real-time surveying and mapping data into the geographic information system for real-time updating.

Images