A terrain real scene modeling system and method based on multi-source surveying and mapping data fusion

Abstract

The present invention discloses a terrain real scene modeling system and method based on multi-source surveying and mapping data fusion, and relates to the field of three-dimensional modeling. The present invention generates a terrain real scene model by adopting multi-source surveying and mapping data of three dimensions, namely, selected control points, aerial image data, and ground-shot image data. After the model is generated, the length and height of objects in the model and the coordinate positions of multiple points are compared with corresponding data in an actual environment, and the coordinates of the control points and the shooting parameters when collecting the aerial image data and the ground-shot image data are adjusted according to the comparison results, so that the accuracy of the generated model is higher, wherein a certain number of control points are arranged and the geodetic coordinates of the control points are measured, and then the ground measurement results and the aerial measurement data are fused in the subsequent post-processing of the aerial photographs, thereby improving the geographical accuracy of the terrain real scene modeling.

Description

Technical Field

[0001] The present invention belongs to the field of three-dimensional modeling, and in particular, relates to a terrain real scene modeling system and method based on multi-source surveying and mapping data fusion. Background Art

[0002] With the development of surveying and mapping technology, multi-source data integration has become possible, but how to effectively integrate data from different sources and with different resolutions to construct a high-precision, high-detail three-dimensional terrain model has become a research hotspot.

[0003] The Chinese patent with announcement number CN116778105B discloses a modeling method based on multi-precision three-dimensional surveying and mapping data fusion. The method obtains a surveying image at a surveying point, screens the surveying image, and screens or reorganizes a first candidate image to determine a target image; establishes a terrain measurement data set, and processes the plane real-life map into three dimensions, fuses the two generated three-dimensional terrain models to generate a candidate model, and if there is a fusion error, screens out the error area; collects terrain data at the surveying point to correct the candidate model.

[0004] Existing models are often built using a single data source, which often cannot provide comprehensive terrain information. For example, satellite images alone may not be able to capture the fine texture and height changes of the terrain. In addition, the accuracy of a single data source may be affected by a variety of factors, resulting in low accuracy of the generated model, which cannot accurately reflect the scenes in the real environment. Summary of the invention

[0005] In view of the problems in the related art, the present invention proposes a terrain real scene modeling system and method based on multi-source surveying and mapping data fusion to overcome the above technical problems existing in the existing related art.

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

[0007] The present invention is a terrain real scene modeling method based on multi-source surveying and mapping data fusion, comprising the following steps:

[0008] S1. Preliminarily select several control points on the ground within the modeling coverage area and measure the coordinates to obtain an initial control point coordinate set; set parameter data corresponding to several shooting parameter types when collecting aerial image data and ground image data to obtain an aerial initial shooting parameter data matrix and a ground initial shooting parameter data matrix;

[0009] S2, collecting and preprocessing the aerial image data and the ground image data in cooperation with the aerial initial shooting parameter data matrix and the ground initial shooting parameter data matrix, to obtain an initial aerial image data classification matrix and a screened initial ground image data set;

[0010] S3, generating an initial terrain real scene model of the coverage area according to the initial layout control point coordinate set, the initial aerial image data classification matrix and the screened initial ground image data set, and obtaining the height data, length data and coordinate data of the corresponding objects at the corresponding check points, and obtaining the model height data set, the model length data set and the model actual coordinate set;

[0011] S4, adjusting the geometric accuracy of the initial terrain real scene model of the coverage area according to the model height data set and the model length data set to obtain a first adjusted terrain real scene model;

[0012] S5, adjusting the plane accuracy of the first adjusted terrain real scene model according to the actual coordinate set of the model to obtain a second adjusted terrain real scene model;

[0013] By selecting a number of control points, collecting aerial image data and ground image data, a terrain real-scene model of the initial modeling coverage area is constructed; the object length data and height data of the corresponding verification points on the constructed initial terrain real-scene model of the modeling coverage area are compared with the actual object length data and height data. When the data error is greater than or equal to the set threshold, the set control points and the shooting parameters when collecting the aerial image data are iteratively optimized, so that the geometric accuracy of the objects in the terrain real-scene model of the modeling coverage area is optimized, so that the accuracy of the model is improved; in addition, multiple coordinate positions on the ground of the terrain real-scene model of the modeling coverage area are collected, and the coordinate positions are compared with the actual coordinate positions. When the error value between the two is greater than or equal to the set threshold, the shooting parameters when collecting the ground image data are iteratively optimized, so that the plane accuracy of the objects in the terrain real-scene model of the modeling coverage area is optimized, so that the accuracy of the model is further improved.

[0014] Preferably, the S1 comprises the following steps:

[0015] S11, set the modeling coverage area; preliminarily select a number of layout control points on the ground within the modeling coverage area, and measure the three-dimensional coordinates of each initial layout control point to obtain the initial layout control point coordinate set a 1 ;as follows,

[0016]

[0017] in, They represent the longitude, latitude and altitude data of the i-th selected initial layout control point, respectively. 1 ′ represents the total number of initial layout control points selected;

[0018] Select several plane check points, height check points and length check points on the ground within the modeling coverage area and measure the corresponding three-dimensional coordinates to obtain the model check point coordinate matrix a 2 ;as follows,

[0019]

[0020] in, Respectively represent the longitude data, latitude data and altitude data of the selected i-th plane check point, Respectively represent the longitude data, latitude data and altitude data of the selected i-th altitude check point, Respectively represent the longitude data, latitude data and altitude data of the selected i-th length check point, a′ 2 、a 3 ′、a′ 4 Respectively represent the total number of selected plane checkpoints, height checkpoints and length checkpoints;

[0021] Preferably, the technology used in S11 to measure the three-dimensional coordinates of each initial layout control point, plane check point, height check point and length check point includes GPSPIK technology;

[0022] S12, setting several shooting parameter types when using drones to collect aerial image data, and obtaining an aerial image data shooting parameter type set; then obtaining shooting parameter data of several drones for aerial image data collection according to the aerial image data shooting parameter type set, and obtaining an initial drone set and an aerial initial shooting parameter data matrix as follows,

[0023]

[0024] in, represents the j-th type of shooting parameter data when the i-th UAV collects aerial image data. Indicates the total number of aerial image data shooting parameter types that are set. Indicates the total number of drones collecting aerial image data;

[0025] S13, setting several shooting parameter types when collecting ground image data to obtain a ground image data shooting parameter type set; then obtaining shooting parameter data of several devices when collecting ground image data according to the ground image data shooting parameter type set to obtain a ground shooting device set and a ground initial shooting parameter data matrix as follows,

[0026]

[0027] in, It represents the jth type of shooting parameter data when the i-th device collects ground image data. Indicates the total number of shooting parameter types set when collecting ground image data. Indicates the total number of devices collecting ground image data;

[0028] Preferably, the types of devices in the ground shooting device set in S13 include mobile phones, SLR cameras, and pan / tilt camera devices, etc.;

[0029] Since the accuracy of using drones directly for shooting and modeling is poor, in special scenarios (such as battlefield terrain), its accuracy is far from meeting the requirements. Therefore, in this solution, a certain number of control points are first deployed in the modeling coverage area and the geodetic coordinates of the control points are measured. Later, in the post-processing of aerial photos, the ground measurement results and aerial measurement data are integrated to improve the geographic accuracy of terrain real-scene modeling; the aerial image data shooting parameter types include flight routes, flight altitudes, overlaps of adjacent images, aerial lighting intensity, exposure rates of shooting cameras, etc.; the ground image data shooting parameter types include overlaps between consecutive images, device focal lengths, and resolutions of device fixed-focus lenses, etc.

[0030] Preferably, S2 comprises the following steps:

[0031] S21, using each drone in the initial drone set to collect aerial image data to obtain an initial aerial image data set b 1i represents the i-th initial aerial image data collected, b 1 ′ represents the total number of initial aerial image data collected; then the ground shooting equipment is used to collect ground image data from each device to obtain the initial ground image data set b 2i represents the i-th initial ground image data collected, b 2 ′ represents the total number of initial ground image data collected;

[0032] S22, performing a preprocessing operation on the initial aerial image dataset and the initial ground image dataset;

[0033] By preprocessing the initial aerial image dataset and the initial ground image dataset, it is more convenient and efficient to use these data for terrain real scene modeling.

[0034] Preferably, the S22 comprises the following steps:

[0035] S221, using a pre-trained deep convolutional generative adversarial network to perform a clustering operation on the initial aerial image data set to obtain an initial aerial image data classification matrix b 3 ;as follows,

[0036]

[0037] Among them, b 3i1 represents the jth initial aerial image data in the i-th group of aerial image data classification obtained by clustering the initial aerial image data set, represents the total number of initial aerial image data in the i-th group of aerial image data classification obtained by clustering the initial aerial image data set, represents the total number of aerial image data classifications obtained by clustering the initial aerial image data set;

[0038] S222, setting a blur level threshold; obtaining blur level data of each initial ground image data in the initial ground image data set to obtain a blur level data set; removing the initial ground image data corresponding to blur level data greater than or equal to the blur level threshold in the blur level data set to obtain a filtered initial ground image data set;

[0039] Before using smart3D modeling software to process the aerial image data obtained by the drone, it should be screened and classified, and the front, back, left, right and orthophoto photo data should be classified and written into the POS information respectively. Based on this, this scheme performs clustering operations on the initial aerial image data set to screen and classify the aerial image data; by eliminating the initial ground image data corresponding to the blur degree data greater than or equal to the blur degree threshold in the blur degree data set, the subsequently generated terrain real scene model is clearer and more accurate.

[0040] Preferably, S3 comprises the following steps:

[0041] S31, inputting the initial layout control point coordinate set, the initial aerial image data classification matrix and the screened initial ground image data set into the modeling software to generate an initial terrain real scene model of the modeling coverage area, recorded as the initial terrain real scene model of the coverage area;

[0042] S32, selecting the height data of the object corresponding to each height check point in the modeled coverage area and the initial terrain real scene model of the coverage area in conjunction with the model check point coordinate matrix to obtain the actual height data set c 1 ={c 11 ,...,c 1i ,...,c 1a′3} and model height dataset c2 ={c 21 ,...,c 2i ,...,c 2a′3},c 1i 、c 2i Respectively represent the height data of the corresponding objects in the modeling coverage area and the initial terrain real scene model of the coverage area measured at the selected i-th height verification point;

[0043] At each length check point in the modeled coverage area and the initial terrain real scene model of the coverage area, the length data of the corresponding object located at the length check point is selected to obtain an actual length data set

[0044] c 3 ={c 31 ,...,c 3i ,...,c 3a′4} and model length dataset c 4 ={c 41 ,...,c 4i ,...,c 4a′4},c 3i 、c 4i Respectively represent the length data of the corresponding objects in the modeling coverage area and the initial terrain real scene model of the coverage area measured at the selected i-th length verification point;

[0045] Collect the three-dimensional coordinates of each plane check point in the initial terrain real scene model of the coverage area to obtain the actual coordinate set c of the model 5 ;as follows,

[0046]

[0047] in, respectively represent the longitude data, latitude data and altitude data of the i-th plane check point in the initial terrain real scene model of the coverage area;

[0048] Preferably, the modeling software includes smart3D software;

[0049] Collect the object length data, height data and coordinate point data of each verification point in the initial terrain real-scene model of the coverage area, compare these data with the corresponding actual data, so as to determine whether the accuracy of the currently generated initial terrain real-scene model of the coverage area meets the requirements. If it does not meet the requirements, adjust the model.

[0050] Preferably, S4 comprises the following steps:

[0051] S41, calculating the Euclidean distance between the actual height data set and the model height data set and between the actual length data set and the model length data set, to obtain the height error Euclidean distance data c 1 ′ and length error Euclidean distance data c′ 2 The calculation formulas are as follows:

[0052]

[0053] S42, setting a height error threshold and a length error threshold; when the height error Euclidean distance data is greater than or equal to the height error threshold or the length error Euclidean distance data is greater than or equal to the length error threshold, performing geometric accuracy adjustment on the initial terrain real scene model of the coverage area to obtain a first adjusted terrain real scene model; otherwise, there is no need to perform geometric accuracy adjustment on the initial terrain real scene model of the coverage area, and the initial terrain real scene model of the coverage area is used as the first adjusted terrain real scene model;

[0054] Geometric accuracy adjustment refers to the error between the size of each object in the model and the size of the corresponding object in the actual environment; geometric accuracy has a great influence on the model's reflection of the real environment; and geometric accuracy has a great relationship with the position of the initially selected control points and the aerial image data taken; among them, by setting the height error threshold and the length error threshold, a quantitative judgment basis is provided for determining whether the geometric accuracy of the generated model meets the requirements.

[0055] Preferably, the step of adjusting the geometric accuracy of the initial terrain real scene model of the coverage area in S42 to obtain a first adjusted terrain real scene model comprises the following steps:

[0056] S421, construct geometric precision to regulate the pelican population represents the i-th pelican in the geometric precision regulated pelican population, Indicates the size of the pelican population adjusted by the geometric precision; sets the maximum number of iterations of the pelican population adjusted by the geometric precision to d 1 And the current iteration number is d 2 , respectively recorded as the maximum iteration number of geometric precision adjustment and the current iteration number of geometric precision adjustment; the search space dimension of the geometric precision adjustment pelican population is set to

[0057] S422, setting the value interval of each coordinate in the initial layout control point coordinate set and the value interval of the parameter corresponding to each aerial image data shooting parameter type in the aerial image data shooting parameter type set, to obtain the control point coordinate value interval set d 1 ′={[d 1 '11 ,d 1 ' 12 ],[d 1 ' 21 ,d 1 ' 22 ],[d 1 ' 31 ,d 1 ' 32 ]} and the aerial image data shooting parameter value interval set d 1 ' 11 d 1 ' 21 d 1 ' 31 They represent the lower limits of the longitude, latitude and altitude data of the control point, respectively. 1 ' 12 d 1 ' 22 d 1 ' 32 Respectively represent the upper limits of the longitude, latitude and altitude data of the control point; d 2 ' i1 d 2 ' i2 They respectively represent the lower limit and upper limit of the value of the parameter corresponding to the i-th aerial image data shooting parameter type;

[0058] The initial position of each pelican in the pelican population is adjusted according to the control point coordinate value interval set and the aerial image data shooting parameter value interval set to obtain a first initial position matrix as follows,

[0059]

[0060] in, Respectively represent the position component set of the initial position of the kth pelican in the pelican population adjusted by the geometric accuracy in the control point coordinate dimension and the position component matrix in the aerial image data shooting parameter data dimension; express The position component of the j-th type of shooting parameter data dimension when the i-th UAV collects aerial image data; Respectively The position components in the dimensions of longitude, latitude and altitude of the i-th control point; The calculation formulas are as follows:

[0061]

[0062] In the formula, rand k2ij Respectively for Generate a random number between 0 and 1;

[0063] S423, according to the height error Euclidean distance data c 1 ′ and length error Euclidean distance data c′ 2 Constructing a fitness function for regulating the pelican population with geometric precision as follows,

[0064]

[0065] In the formula, β 1 is a positive number, indicating the first protection parameter;

[0066] S424, start iteration, before iteration, set the current iteration number of the geometric precision adjustment to 1; in the first round of iteration, use the geometric precision to adjust the fitness function of the pelican population Calculate the fitness value of the initial position of each pelican in the first initial position matrix to obtain a first fitness value set; use the maximum fitness in the first fitness value set and the initial position of the corresponding pelican as the first global optimal fitness and the first global optimal position respectively; update the initial position of each pelican in the first initial position matrix according to the first global optimal fitness and the first global optimal position; after the update is completed, add 1 to the current iteration number of the geometric accuracy adjustment and enter the next iteration;

[0067] In each other iteration, the geometric accuracy is used to adjust the fitness function of the pelican population. Calculate the fitness value of the position of each pelican in the pelican population adjusted by geometric precision obtained in the previous round of iterations to obtain a second fitness value set; use the maximum fitness and the corresponding pelican position in the second fitness value set as the second global optimal fitness and the second global optimal position respectively; update the position of each pelican in the pelican population adjusted by geometric precision obtained in the previous round of iterations according to the second global optimal fitness and the second global optimal position; after the update is completed, add 1 to the current iteration number of the geometric precision adjustment and enter the next iteration;

[0068] S425, when d 2 ≥d 1 When d , stop the iteration and get the first final global optimal position; otherwise, continue to iterate until d 2 ≥d 1until the time; taking the position component of the first final global optimal position in the control point coordinate dimension and the position component in the aerial image data shooting parameter data dimension as the optimized control point coordinate set and the aerial optimized shooting parameter data matrix respectively; selecting the control points for deployment on the ground in the modeling coverage area according to the optimized control point coordinate set and the aerial optimized shooting parameter data matrix, setting the shooting parameter data when using the drone for aerial image data collection, and combining with the ground initial shooting parameter data matrix to generate the terrain real scene model of the modeling coverage area, to obtain the optimized terrain real scene model of the first coverage area;

[0069] S426, selecting the height data of the object corresponding to each height check point in the optimized terrain real scene model of the first coverage area at each height check point and selecting the length data of the object corresponding to each length check point at each length check point in accordance with the model check point coordinate matrix, to obtain an optimized model height data set and an optimized model length data set; calculating the Euclidean distance between the actual height data set and the optimized model height data set and between the actual length data set and the optimized model length data set, to obtain optimized height error Euclidean distance data and optimized length error Euclidean distance data;

[0070] When the optimized height error Euclidean distance data is less than the height error threshold or the optimized length error Euclidean distance data is less than the length error threshold, the optimized terrain real scene model of the first coverage area is used as the first adjusted terrain real scene model; otherwise, return to S424 and continue to iterate until the optimized height error Euclidean distance data is less than the height error threshold or the optimized length error Euclidean distance data is less than the length error threshold;

[0071] The Pelican Optimization Algorithm can effectively explore different areas of the solution space, because in the stage of moving toward the prey, the pelican identifies the position of the prey and moves toward the prey; it can usually find a better solution within a smaller number of iterations; based on the above advantages, the Pelican Optimization Algorithm is used in this scheme to simultaneously perform multiple iterations on the coordinates of the multiple control points initially selected and the multiple shooting parameters when shooting aerial images, and the error between the length and height of the object at each check point in the generated model and the actual length and height of the object is used as the fitness function; therefore, as the iteration proceeds, the error value between the length and height of the object at each check point in the generated model and the actual length and height of the object becomes smaller and smaller, which indicates that the geometric accuracy of the model is getting higher and higher.

[0072] Preferably, S5 comprises the following steps:

[0073] S51, setting a model coordinate error threshold; calculating the Euclidean distance between the actual model coordinate set and the first row of data in the model check point coordinate matrix, and obtaining the initial model coordinate error Euclidean distance c 3 ′;

[0074] S52: when the Euclidean distance of the initial model coordinate error is greater than or equal to the model coordinate error threshold, performing plane accuracy adjustment on the first adjusted terrain real scene model to obtain a second adjusted terrain real scene model; otherwise, there is no need to perform plane accuracy adjustment on the first adjusted terrain real scene model, and the first adjusted terrain real scene model is used as the second adjusted terrain real scene model;

[0075] By setting the model coordinate error threshold, a quantitative basis is provided for determining whether the plane accuracy of the generated model meets the requirements.

[0076] A terrain real scene modeling method based on multi-source surveying and mapping data fusion, including an initial control point coordinate setting module, a check point setting module, a shooting initial parameter setting module, an image data acquisition module, a preprocessing module, an initial model generation module, a model check point initial data acquisition module, a model geometry accuracy adjustment module and a model plane accuracy adjustment module.

[0077] The present invention has the following beneficial effects:

[0078] 1. In the present invention, a terrain real scene model is generated by using multi-source surveying and mapping data in three dimensions, namely, selected control points, aerial image data, and ground-shot image data; after the model is generated, the length, height, and coordinate positions of multiple points of the objects in the model are compared with the corresponding data in the actual environment, and the coordinates of the control points, the shooting parameters when collecting the aerial image data, and the ground-shot image data are adjusted according to the comparison results, so that the generated model has higher accuracy.

[0079] 2. In the present invention, a certain number of control points are arranged in the modeling coverage area and the geodetic coordinates of the control points are measured. In the subsequent post-processing of the aerial photographs, the ground measurement results and the aerial measurement data are integrated, thereby improving the geographic accuracy of the terrain real-scene modeling.

[0080] 3. In the present invention, by preprocessing the initial aerial image data set and the initial ground image data set, the subsequent use of these data for terrain real scene modeling is more convenient and more efficient.

[0081] 4. The present invention adopts the Pelican optimization algorithm to simultaneously perform multiple iterations on the coordinates of the multiple control points initially selected and the multiple shooting parameters during aerial image shooting, and adopts the error between the length and height of the object at each check point in the generated model and the length and height of the actual object as the fitness function; therefore, as the iteration proceeds, the error value between the length and height of the object at each check point in the generated model and the length and height of the actual object becomes smaller and smaller, which indicates that the geometric accuracy of the model is getting higher and higher.

[0082] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0084] Figure 1 The present invention is a flowchart of a terrain real scene modeling method based on multi-source surveying and mapping data fusion. DETAILED DESCRIPTION

[0085] The following will be combined with the drawings in the embodiments of the invention to clearly and completely describe the technical solutions in the embodiments of the invention. Obviously, the described embodiments are only part of the embodiments of the invention, not all of the embodiments. Based on the embodiments in the invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the invention.

[0086] Embodiment 1

[0087] See also Figure 1 This embodiment is a terrain real scene modeling method based on multi-source surveying and mapping data fusion, comprising the following steps:

[0088] S1. Preliminarily select several control points on the ground within the modeling coverage area and measure the coordinates to obtain an initial control point coordinate set; set parameter data corresponding to several shooting parameter types when collecting aerial image data and ground image data to obtain an aerial initial shooting parameter data matrix and a ground initial shooting parameter data matrix;

[0089] The S1 comprises the following steps:

[0090] S11, set the modeling coverage area; preliminarily select a number of layout control points on the ground within the modeling coverage area, and measure the three-dimensional coordinates of each initial layout control point to obtain the initial layout control point coordinate set a 1 ;as follows,

[0091]

[0092] in, They represent the longitude, latitude and altitude data of the i-th initial control point selected, respectively. 1 ′ represents the total number of initial layout control points selected;

[0093] Select several plane check points, height check points and length check points on the ground within the modeling coverage area and measure the corresponding three-dimensional coordinates to obtain the model check point coordinate matrix a 2 ;as follows,

[0094]

[0095] in, Respectively represent the longitude data, latitude data and altitude data of the selected i-th plane check point, Respectively represent the longitude data, latitude data and altitude data of the selected i-th altitude check point, Respectively represent the longitude data, latitude data and altitude data of the selected i-th length check point, a′ 2 、a 3 ′、a′ 4 Respectively represent the total number of selected plane checkpoints, height checkpoints and length checkpoints;

[0096] S12, setting several shooting parameter types when using drones to collect aerial image data, and obtaining an aerial image data shooting parameter type set; then obtaining shooting parameter data of several drones for aerial image data collection according to the aerial image data shooting parameter type set, and obtaining an initial drone set and an aerial initial shooting parameter data matrix as follows,

[0097]

[0098] in, represents the j-th type of shooting parameter data when the i-th UAV collects aerial image data. Indicates the total number of aerial image data shooting parameter types that are set. Indicates the total number of drones collecting aerial image data;

[0099] S13, setting several shooting parameter types when collecting ground image data to obtain a ground image data shooting parameter type set; then obtaining shooting parameter data of several devices when collecting ground image data according to the ground image data shooting parameter type set to obtain a ground shooting device set and a ground initial shooting parameter data matrix as follows,

[0100]

[0101] in, It represents the jth type of shooting parameter data when the i-th device collects ground image data. Indicates the total number of shooting parameter types set when collecting ground image data. Indicates the total number of devices collecting ground image data;

[0102] S2, collecting and preprocessing the aerial image data and the ground image data in cooperation with the aerial initial shooting parameter data matrix and the ground initial shooting parameter data matrix, to obtain an initial aerial image data classification matrix and a screened initial ground image data set;

[0103] The S2 comprises the following steps:

[0104] S21, using each drone in the initial drone set to collect aerial image data to obtain an initial aerial image data set b 1i represents the i-th initial aerial image data collected, b 1 ′ represents the total number of initial aerial image data collected; then the ground shooting equipment is used to collect ground image data from each device to obtain the initial ground image data set b 2i represents the i-th initial ground image data collected, b 2 ′ represents the total number of initial ground image data collected;

[0105] S22, performing a preprocessing operation on the initial aerial image dataset and the initial ground image dataset;

[0106] The S22 comprises the following steps:

[0107] S221, using a pre-trained deep convolutional generative adversarial network to perform a clustering operation on the initial aerial image data set to obtain an initial aerial image data classification matrix b 3 ;as follows,

[0108]

[0109] Among them, b 3i1represents the jth initial aerial image data in the i-th group of aerial image data classification obtained by clustering the initial aerial image data set, represents the total number of initial aerial image data in the i-th group of aerial image data classification obtained by clustering the initial aerial image data set, represents the total number of aerial image data classifications obtained by clustering the initial aerial image data set;

[0110] S222, setting a blur level threshold; obtaining blur level data of each initial ground image data in the initial ground image data set to obtain a blur level data set; removing the initial ground image data corresponding to blur level data greater than or equal to the blur level threshold in the blur level data set to obtain a filtered initial ground image data set;

[0111] Obtaining blur degree data of each initial ground image data in the initial ground image data set in S222, and obtaining the blur degree data set comprises the following steps:

[0112] S2221, collecting a plurality of historical ground image data and corresponding blur degree data to obtain a historical ground image data set and a historical blur degree data set;

[0113] S2222, constructing an initial ground image classification model and setting a training data ratio; dividing the historical ground image data set and the historical blur degree data set according to the training data ratio to obtain a historical ground image training data set, a historical blur degree training data set, a historical ground image test data set, and a historical blur degree test data set;

[0114] S2223, setting a training error threshold; inputting the historical ground image training data set and the historical blur level training data set as training data and training labels respectively into the initial ground image classification model for training; during the training process, when the training error is less than the training error threshold, obtaining a trained ground image classification model; otherwise, continuing the training until the training error is less than the training error threshold;

[0115] S2224, setting a test accuracy threshold; inputting the historical ground image test data set and the historical blur test data set as test data and test labels respectively into the trained ground image classification model for testing; after the test is completed, obtaining the test accuracy data; when the test accuracy data is greater than or equal to the test accuracy threshold, using the trained ground image classification model as the final ground image classification model; otherwise, returning to S2223 to continue training the trained ground image classification model until the test accuracy data is greater than or equal to the test accuracy threshold;

[0116] S2225, inputting each initial ground image data set in the initial ground image data set into the final ground image classification model for classification operation to obtain a blur degree data set;

[0117] The final ground image classification model is preferably a generative adversarial network model;

[0118] S3, generating an initial terrain real scene model of the coverage area according to the initial layout control point coordinate set, the initial aerial image data classification matrix and the screened initial ground image data set, and obtaining the height data, length data and coordinate data of the corresponding objects at the corresponding check points, and obtaining the model height data set, the model length data set and the model actual coordinate set;

[0119] The S3 comprises the following steps:

[0120] S31, inputting the initial layout control point coordinate set, the initial aerial image data classification matrix and the screened initial ground image data set into the modeling software to generate an initial terrain real scene model of the modeling coverage area, recorded as the initial terrain real scene model of the coverage area;

[0121] S32, in conjunction with the model checkpoint coordinate matrix, at each height checkpoint in the modeled coverage area and the initial terrain real scene model of the coverage area, select the height data of the object corresponding to the height checkpoint at the height checkpoint to obtain an actual height data set And the model height dataset c 1i 、c 2i Respectively represent the height data of the corresponding objects in the modeling coverage area and the initial terrain real scene model of the coverage area measured at the selected i-th height verification point;

[0122] At each length check point in the modeled coverage area and the initial terrain real scene model of the coverage area, the length data of the corresponding object located at the length check point is selected to obtain an actual length data set And the model length dataset c 3i 、c 4i Respectively represent the length data of the corresponding objects in the modeling coverage area and the initial terrain real scene model of the coverage area measured at the selected i-th length verification point;

[0123] Collect the three-dimensional coordinates of each plane check point in the initial terrain real scene model of the coverage area to obtain the actual coordinate set c of the model 5 ;as follows,

[0124]

[0125] in, respectively represent the longitude data, latitude data and altitude data of the i-th plane check point in the initial terrain real scene model of the coverage area;

[0126] S4, adjusting the geometric accuracy of the initial terrain real scene model of the coverage area according to the model height data set and the model length data set to obtain a first adjusted terrain real scene model;

[0127] The S4 comprises the following steps:

[0128] S41, calculating the Euclidean distance between the actual height data set and the model height data set and between the actual length data set and the model length data set, to obtain the height error Euclidean distance data c 1 ′ and length error Euclidean distance data c′ 2 ; The calculation formulas are as follows:

[0129]

[0130] S42, setting a height error threshold and a length error threshold; when the height error Euclidean distance data is greater than or equal to the height error threshold or the length error Euclidean distance data is greater than or equal to the length error threshold, performing geometric accuracy adjustment on the initial terrain real scene model of the coverage area to obtain a first adjusted terrain real scene model; otherwise, there is no need to perform geometric accuracy adjustment on the initial terrain real scene model of the coverage area, and the initial terrain real scene model of the coverage area is used as the first adjusted terrain real scene model;

[0131] The step S42 of adjusting the geometric accuracy of the initial terrain real scene model of the coverage area to obtain a first adjusted terrain real scene model comprises the following steps:

[0132] S421, construct geometric precision to regulate the pelican population represents the i-th pelican in the geometric precision regulated pelican population, Indicates the size of the pelican population adjusted by the geometric precision; sets the maximum number of iterations of the pelican population adjusted by the geometric precision to d 1 And the current iteration number is d 2 , respectively recorded as the maximum iteration number of geometric precision adjustment and the current iteration number of geometric precision adjustment; the search space dimension of the geometric precision adjustment pelican population is set to

[0133] S422, setting the value interval of each coordinate in the initial layout control point coordinate set and the value interval of the parameter corresponding to each aerial image data shooting parameter type in the aerial image data shooting parameter type set, to obtain the control point coordinate value interval set d 1 ′={[d1 ' 11 ,d 1 ' 12 ],[d 1 ' 21 ,d 1 ' 22 ],[d 1 ' 31 ,d 1 ' 32 ]} and the aerial image data shooting parameter value interval set d 1 ' 11 ,d 1 ' 21 ,d 1 ' 31 They represent the lower limits of the longitude, latitude and altitude data of the control point, respectively. 1 ' 12 ,d 1 ' 22 ,d 1 ' 32 Respectively represent the upper limits of the longitude, latitude and altitude data of the control point; d 2 ' i1 ,d 2 ' i2 They respectively represent the lower limit and upper limit of the value of the parameter corresponding to the i-th aerial image data shooting parameter type;

[0134] The initial position of each pelican in the pelican population is adjusted according to the control point coordinate value interval set and the aerial image data shooting parameter value interval set to obtain a first initial position matrix as follows,

[0135]

[0136] in, Respectively represent the position component set of the initial position of the kth pelican in the pelican population adjusted by the geometric accuracy in the control point coordinate dimension and the position component matrix in the aerial image data shooting parameter data dimension; express The position component of the j-th type of shooting parameter data dimension when the i-th UAV collects aerial image data; Respectively The position components in the dimensions of longitude data, latitude data and altitude data of the i-th deployed control point; The calculation formulas are as follows:

[0137]

[0138] In the formula, rand k2ij Respectively for Generate a random number between 0 and 1;

[0139] S423, according to the height error Euclidean distance data c 1 ′ and length error Euclidean distance data c′ 2 Constructing a fitness function for regulating the pelican population with geometric precision as follows,

[0140]

[0141] In the formula, β 1 is a positive number, indicating the first protection parameter;

[0142] S424, start iteration, before iteration, set the current iteration number of the geometric precision adjustment to 1; in the first round of iteration, use the geometric precision to adjust the fitness function of the pelican population Calculate the fitness value of the initial position of each pelican in the first initial position matrix to obtain a first fitness value set; use the maximum fitness in the first fitness value set and the initial position of the corresponding pelican as the first global optimal fitness and the first global optimal position respectively; update the initial position of each pelican in the first initial position matrix according to the first global optimal fitness and the first global optimal position; after the update is completed, add 1 to the current iteration number of the geometric accuracy adjustment and enter the next iteration;

[0143] In each other iteration, the geometric accuracy is used to adjust the fitness function of the pelican population. Calculate the fitness value of the position of each pelican in the pelican population adjusted by geometric precision obtained in the previous round of iterations to obtain a second fitness value set; use the maximum fitness and the corresponding pelican position in the second fitness value set as the second global optimal fitness and the second global optimal position respectively; update the position of each pelican in the pelican population adjusted by geometric precision obtained in the previous round of iterations according to the second global optimal fitness and the second global optimal position; after the update is completed, add 1 to the current iteration number of the geometric precision adjustment and enter the next iteration;

[0144] S425, when d 2 ≥d 1 When d , stop the iteration and get the first final global optimal position; otherwise, continue to iterate until d 2 ≥d 1until the time; taking the position component of the first final global optimal position in the control point coordinate dimension and the position component in the aerial image data shooting parameter data dimension as the optimized control point coordinate set and the aerial optimized shooting parameter data matrix respectively; selecting the control points for deployment on the ground in the modeling coverage area according to the optimized control point coordinate set and the aerial optimized shooting parameter data matrix, setting the shooting parameter data when using the drone for aerial image data collection, and combining with the ground initial shooting parameter data matrix to generate the terrain real scene model of the modeling coverage area, to obtain the optimized terrain real scene model of the first coverage area;

[0145] S426, selecting the height data of the object corresponding to each height check point in the optimized terrain real scene model of the first coverage area at each height check point and selecting the length data of the object corresponding to each length check point at each length check point in accordance with the model check point coordinate matrix, to obtain an optimized model height data set and an optimized model length data set; calculating the Euclidean distance between the actual height data set and the optimized model height data set and between the actual length data set and the optimized model length data set, to obtain optimized height error Euclidean distance data and optimized length error Euclidean distance data;

[0146] When the optimized height error Euclidean distance data is less than the height error threshold or the optimized length error Euclidean distance data is less than the length error threshold, the optimized terrain real scene model of the first coverage area is used as the first adjusted terrain real scene model; otherwise, return to S424 and continue to iterate until the optimized height error Euclidean distance data is less than the height error threshold or the optimized length error Euclidean distance data is less than the length error threshold;

[0147] S5, adjusting the plane accuracy of the first adjusted terrain real scene model according to the actual coordinate set of the model to obtain a second adjusted terrain real scene model;

[0148] The S5 comprises the following steps:

[0149] S51, setting a model coordinate error threshold; calculating the Euclidean distance between the actual model coordinate set and the first row of data in the model check point coordinate matrix, and obtaining the initial model coordinate error Euclidean distance c 3 ′;

[0150] S52: when the Euclidean distance of the initial model coordinate error is greater than or equal to the model coordinate error threshold, performing plane accuracy adjustment on the first adjusted terrain real scene model to obtain a second adjusted terrain real scene model; otherwise, there is no need to perform plane accuracy adjustment on the first adjusted terrain real scene model, and the first adjusted terrain real scene model is used as the second adjusted terrain real scene model;

[0151] In S52, the plane accuracy of the first adjusted terrain real scene model is adjusted to obtain the second adjusted terrain real scene model, which includes the following steps:

[0152] S521, build plane precision to adjust the pelican population represents the i-th pelican in the plane precision adjustment pelican population, Indicates the size of the pelican population adjusted by the plane precision; Sets the maximum number of iterations of the pelican population adjusted by the plane precision to d 3 And the current iteration number is d 4 , respectively recorded as the maximum iteration number of plane precision adjustment and the current iteration number of plane precision adjustment; the search space dimension of the plane precision adjustment pelican population is set to

[0153] S522: Set the value interval of each ground image data shooting parameter type in the ground image data shooting parameter type set to obtain a ground image data shooting parameter value interval set d 3 ';as follows,

[0154] d 3 ′={[d 3 ' 11 ,d 3 ' 12 ],...,[d 3 ' i1 ,d 3 ' i2 ],...,[d 3 ' a～31 ,d 3 ' a～32 ]};

[0155] Among them, d 3 ' i1 ,d 3 ' i2 They respectively represent the lower limit and upper limit of the value of the parameter corresponding to the i-th ground image data shooting parameter type;

[0156] The plane precision is set according to the ground image data shooting parameter value interval set to adjust the initial position of each pelican in the pelican population, and a second initial position matrix set is obtained. Represents the initial position matrix of the kth pelican in the plane precision adjustment pelican population; as follows,

[0157]

[0158] in, express The position component of the j-th type of shooting parameter data dimension when the i-th device collects ground image data;

[0159] S523, according to the Euclidean distance c of the initial model coordinate error 3 ′Construct the fitness function of the pelican population to adjust the plane precision as follows,

[0160]

[0161] In the formula, β 2 is a positive number, indicating the second protection parameter;

[0162] S524, start iteration, before iteration, set the current iteration number of the plane precision adjustment to 1; in the first round of iteration, use the plane precision adjustment fitness function of the pelican population Calculate the fitness value of the initial position matrix of each pelican in the second initial position matrix set to obtain a third fitness value set; use the maximum fitness in the third fitness value set and the initial position matrix of the corresponding pelican as the third global optimal fitness and the third global optimal position respectively; update the initial position matrix of each pelican in the second initial position matrix set according to the third global optimal fitness and the third global optimal position; after the update is completed, add 1 to the current iteration number of the plane accuracy adjustment and enter the next round of iteration;

[0163] In each other iteration process, the plane accuracy is used to adjust the fitness function of the pelican population. Calculate the fitness value of the position matrix of each pelican in the pelican population for plane precision adjustment obtained in the last round of iteration to obtain a fourth fitness value set; use the maximum fitness in the fourth fitness value set and the corresponding pelican position matrix as the fourth global optimal fitness and the fourth global optimal position respectively; update the position matrix of each pelican in the pelican population for plane precision adjustment obtained in the last round of iteration according to the fourth global optimal fitness and the fourth global optimal position; after the update is completed, add 1 to the current iteration number of the plane precision adjustment and enter the next round of iteration;

[0164] S525, when d 4 ≥d 3 When d , the iteration is stopped to obtain the second final global optimal position; otherwise, the iteration is continued until d 2 ≥d 1 The second final global optimal position is used as the ground optimized shooting parameter data matrix, and the shooting parameter data when collecting ground image data is set according to the ground optimized shooting parameter data matrix, and a terrain real scene model of the modeling coverage area is generated to obtain a second coverage area optimized terrain real scene model;

[0165] S526, selecting coordinate data of an object located at each plane check point in the optimized terrain real scene model of the second coverage area in accordance with the model check point coordinate matrix, to obtain an actual coordinate set of the optimized model; calculating the Euclidean distance between the actual coordinate set of the optimized model and the first row of data in the model check point coordinate matrix, to obtain the Euclidean distance of the coordinate error of the optimized model;

[0166] When the Euclidean distance of the optimized model coordinate error is less than the model coordinate error threshold, the optimized terrain real scene model of the second coverage area is used as the second adjusted terrain real scene model; otherwise, return to S524 and continue to iterate until the Euclidean distance of the optimized model coordinate error is less than the model coordinate error threshold;

[0167] The Pelican optimization algorithm is used to iteratively optimize multiple shooting parameters when shooting ground image data, and the error between the coordinates of each check point in the generated model and the coordinates of the corresponding position in the actual environment is used as the fitness function; as the iteration proceeds, the error value between the coordinates of each check point in the generated model and the coordinates of the corresponding position in the actual environment becomes smaller and smaller, which indicates that the planar accuracy of the model is getting higher and higher.

[0168] Embodiment 2

[0169] This embodiment discloses a terrain real scene modeling system based on multi-source surveying and mapping data fusion, the system can implement the method of the above embodiment, including an initial control point coordinate setting module, a check point setting module, a shooting initial parameter setting module, an image data acquisition module, a preprocessing module, an initial model generation module, a model check point initial data acquisition module, a model geometry accuracy adjustment module and a model plane accuracy adjustment module;

[0170] The initial control point coordinate setting module is used to preliminarily select a number of layout control points on the ground within the modeling coverage area and measure the coordinates to obtain an initial layout control point coordinate set;

[0171] The checkpoint setting module is used to select a number of plane checkpoints, height checkpoints and length checkpoints on the ground within the modeling coverage area and measure the corresponding three-dimensional coordinates to obtain a model checkpoint coordinate matrix;

[0172] The shooting initial parameter setting module is used to set parameter data corresponding to several shooting parameter types when collecting aerial image data and ground image data, and obtain an aerial initial shooting parameter data matrix and a ground initial shooting parameter data matrix;

[0173] The image data acquisition module is used to acquire the aerial image data and the ground image data in cooperation with the aerial initial shooting parameter data matrix and the ground initial shooting parameter data matrix to obtain the initial aerial image data set and the initial ground image data set;

[0174] The preprocessing module is used to perform preprocessing operations on the initial aerial image data set and the initial ground image data set to obtain the initial aerial image data classification matrix and the screened initial ground image data set;

[0175] The initial model generation module is used to generate an initial terrain real scene model of the coverage area according to the initial layout control point coordinate set, the initial aerial image data classification matrix and the screened initial ground image data set;

[0176] The model checkpoint initial data acquisition module is used to cooperate with the model checkpoint coordinate matrix to collect the height data, length data and coordinate data of the corresponding objects at the corresponding checkpoints of the modeled coverage area and the initial terrain real scene model of the coverage area, and obtain the actual height data set, model height data set, actual length data set, model length data set and model actual coordinate set;

[0177] The model geometric accuracy adjustment module is used to adjust the geometric accuracy of the initial terrain real scene model of the coverage area according to the actual height data set, the model height data set, the actual length data set and the model length data set to obtain a first adjusted terrain real scene model;

[0178] The model plane accuracy adjustment module is used to adjust the plane accuracy of the first adjusted terrain real scene model according to the model actual coordinate set to obtain the second adjusted terrain real scene model.

[0179] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0180] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. A terrain real scene modeling method based on multi-source surveying and mapping data fusion, characterized in that: The following steps are involved: S1. Preliminarily select several control points on the ground within the modeling coverage area and measure the coordinates to obtain an initial control point coordinate set; set parameter data corresponding to several shooting parameter types when collecting aerial image data and ground image data to obtain an aerial initial shooting parameter data matrix and a ground initial shooting parameter data matrix; S2, collecting and preprocessing the aerial image data and the ground image data in cooperation with the aerial initial shooting parameter data matrix and the ground initial shooting parameter data matrix, to obtain an initial aerial image data classification matrix and a screened initial ground image data set; S3, generating an initial terrain real scene model of the coverage area according to the initial layout control point coordinate set, the initial aerial image data classification matrix and the screened initial ground image data set, and obtaining the height data, length data and coordinate data of the corresponding objects at the corresponding check points, and obtaining the model height data set, the model length data set and the model actual coordinate set; S4, adjusting the geometric accuracy of the initial terrain real scene model of the coverage area according to the model height data set and the model length data set to obtain a first adjusted terrain real scene model; S5. Adjust the plane accuracy of the first adjusted terrain real scene model according to the actual coordinate set of the model to obtain a second adjusted terrain real scene model.

2. The method for terrain real scene modeling based on multi-source surveying and mapping data fusion according to claim 1 is characterized in that: The S1 comprises the following steps: S11, setting a modeling coverage area; preliminarily selecting a number of layout control points on the ground within the modeling coverage area, and measuring the three-dimensional coordinates of each initial layout control point to obtain an initial layout control point coordinate set; S12, acquiring shooting parameter data of several drones for aerial image data collection, and obtaining an initial drone set and an initial aerial shooting parameter data matrix; S13. Set several shooting parameter types when collecting ground image data to obtain a ground image data shooting parameter type set; then obtain shooting parameter data of several devices when collecting ground image data according to the ground image data shooting parameter type set to obtain a ground shooting device set and a ground initial shooting parameter data matrix.

3. The method for terrain real scene modeling based on multi-source surveying and mapping data fusion according to claim 2 is characterized in that: The S2 comprises the following steps: S21, using each drone in the initial drone set to collect aerial image data to obtain an initial aerial image data set; and then using each device in the ground shooting device set to collect ground image data to obtain an initial ground image data set; S22: performing preprocessing operations on the initial aerial image dataset and the initial ground image dataset.

4. The method for terrain real scene modeling based on multi-source surveying and mapping data fusion according to claim 3 is characterized in that: The S22 comprises the following steps: S221, performing a clustering operation on the initial aerial image data set to obtain an initial aerial image data classification matrix; S222, setting a blur level threshold; obtaining blur level data of each initial ground image data in the initial ground image data set and filtering and eliminating the data therein to obtain a filtered initial ground image data set.

5. The method for terrain real scene modeling based on multi-source surveying and mapping data fusion according to claim 4 is characterized in that: The S3 comprises the following steps: S31, inputting the initial layout control point coordinate set, the initial aerial image data classification matrix and the screened initial ground image data set into the modeling software to generate an initial terrain real scene model of the modeling coverage area, recorded as the initial terrain real scene model of the coverage area; S32. At each height check point and length check point in the modeled coverage area and the initial terrain real-scene model of the coverage area, select the height data and length data of the object located at the height check point and the length check point to obtain an actual height data set, a model height data set, an actual length data set and a model length data set; collect the three-dimensional coordinates of each plane check point in the initial terrain real-scene model of the coverage area to obtain a model actual coordinate set.

6. The method for terrain real scene modeling based on multi-source surveying and mapping data fusion according to claim 5 is characterized in that: The S4 comprises the following steps: S41, calculating the Euclidean distance between the actual height dataset and the model height dataset and between the actual length dataset and the model length dataset to obtain height error Euclidean distance data and length error Euclidean distance data; S42, setting a height error threshold and a length error threshold; when the height error Euclidean distance data is greater than or equal to the height error threshold or the length error Euclidean distance data is greater than or equal to the length error threshold, performing geometric accuracy adjustment on the initial terrain real scene model of the coverage area to obtain a first adjusted terrain real scene model; otherwise, there is no need to perform geometric accuracy adjustment on the initial terrain real scene model of the coverage area, and the initial terrain real scene model of the coverage area is used as the first adjusted terrain real scene model.

7. The method for terrain real scene modeling based on multi-source surveying and mapping data fusion according to claim 6 is characterized in that: The step S42 of adjusting the geometric accuracy of the initial terrain real scene model of the coverage area to obtain a first adjusted terrain real scene model comprises the following steps: S421, constructing a pelican population for geometric precision adjustment; setting the maximum number of iterations of the pelican population for geometric precision adjustment to d1 and the current number of iterations to d2, which are recorded as the maximum number of iterations for geometric precision adjustment and the current number of iterations for geometric precision adjustment respectively; S422, setting the value range of each coordinate in the initial layout control point coordinate set and the value range of the parameter corresponding to each aerial image data shooting parameter type and setting the geometric accuracy to adjust the initial position of each pelican in the pelican population to obtain a first initial position matrix; S423, constructing a fitness function for adjusting the pelican population with geometric accuracy according to the height error Euclidean distance data and the length error Euclidean distance data; S424, start iteration; in each round of iteration, use the fitness function of the pelican population adjusted by geometric precision to calculate the fitness value of the position of each pelican in the pelican population adjusted by geometric precision obtained in the previous round of iteration, and update the position of each pelican in the pelican population adjusted by geometric precision obtained in the previous round of iteration; S425. When d2≥d1, stop iteration and obtain the first final global optimal position; otherwise, continue iteration until d2≥d1; use the position component of the first final global optimal position in the control point coordinate dimension and the position component in the aerial image data shooting parameter data dimension as the optimized control point coordinate set and the aerial optimized shooting parameter data matrix respectively; select the control points for deployment on the ground in the modeling coverage area according to the optimized control point coordinate set and the aerial optimized shooting parameter data matrix, set the shooting parameter data when using a drone for aerial image data collection, and generate a terrain real scene model of the modeling coverage area in combination with the ground initial shooting parameter data matrix, so as to obtain the optimized terrain real scene model of the first coverage area; S426, selecting the height data of the object corresponding to each height check point in the optimized terrain real scene model of the first coverage area at the height check point and selecting the length data of the object corresponding to each length check point at the length check point to obtain an optimized model height data set and an optimized model length data set; calculating the Euclidean distance between the actual height data set and the optimized model height data set and between the actual length data set and the optimized model length data set to obtain optimized height error Euclidean distance data and optimized length error Euclidean distance data; When the optimized height error Euclidean distance data is less than the height error threshold or the optimized length error Euclidean distance data is less than the length error threshold, the optimized terrain real scene model of the first coverage area is used as the first adjusted terrain real scene model; otherwise, return to S424 and continue to iterate until the optimized height error Euclidean distance data is less than the height error threshold or the optimized length error Euclidean distance data is less than the length error threshold.

8. The method for terrain real scene modeling based on multi-source surveying and mapping data fusion according to claim 7 is characterized in that: The S5 comprises the following steps: S51, setting a model coordinate error threshold; calculating the Euclidean distance between the actual coordinate set of the model and the corresponding coordinates in the actual environment to obtain the Euclidean distance of the initial model coordinate error; S52: When the Euclidean distance of the initial model coordinate error is greater than or equal to the model coordinate error threshold, the plane accuracy of the first adjusted terrain real scene model is adjusted to obtain a second adjusted terrain real scene model; otherwise, the plane accuracy of the first adjusted terrain real scene model does not need to be adjusted, and the first adjusted terrain real scene model is used as the second adjusted terrain real scene model.

9. The method for terrain real scene modeling based on multi-source surveying and mapping data fusion according to claim 8 is characterized in that: In S52, the plane accuracy of the first adjusted terrain real scene model is adjusted using the Pelican optimization algorithm.

10. A system for implementing a terrain real scene modeling method based on multi-source surveying and mapping data fusion as described in any one of claims 1 to 9.

Images