A typical terrain construction method based on array synthetic aperture radar 3D point cloud

Abstract

The present invention provides a typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud, the method comprising: (1) obtaining a three-dimensional point cloud data set through an array synthetic aperture radar; (2) preprocessing the three-dimensional point cloud data set to obtain a new three-dimensional point cloud data set; (3) extracting ground points through the new three-dimensional point cloud data set; (4) extracting typical terrain ground points; (5) constructing a tree height feature operator based on the extracted typical terrain ground points; (6) segmenting the typical terrain ground points, and reconstructing the ground points to obtain a new typical terrain ground point data set; (7) constructing a terrain model. The present invention proposes a typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud, which is used to remove the tree height in the typical terrain area; extracting and processing the ground points in the typical terrain area, and calculating and obtaining the typical terrain ground points based on the ground point extraction, and reconstructing the ground points. The method solves the problem of insufficient point cloud data integrity and digital elevation model (DEM) generation loopholes due to observation conditions in mountainous gully terrain areas covered by vegetation.

Description

Technical Field

[0001] The invention belongs to the technical field of data processing, and in particular relates to a typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud. Background Art

[0002] Synthetic aperture radar (SAR) technology has unique advantages in terrain mapping applications: it can be used for terrain mapping in any weather conditions, and is not affected by weather factors such as clouds, rain and fog. It can achieve one-time mapping of a large area of ​​the surface, which is suitable for monitoring and analyzing regional terrain changes. For areas with obvious terrain undulations, SAR technology can provide high-precision terrain mapping results. This technology is a non-contact remote sensing mapping method that does not require field data collection, reducing manpower and material costs and measurement risks.

[0003] At the same time, SAR technology has strong visibility on the surface, but for mountainous gully terrain areas covered by vegetation, due to limited observation conditions, it is impossible to obtain accurate terrain information under conditions of undulating terrain and obstructions, and the problem of overlapping cannot be solved.

[0004] At present, there are the following difficulties in constructing point clouds and generating digital terrain models (DEM) in mountainous gully areas: (1) Occlusion and shadow: There are obstructions such as cliffs and trees in mountainous terrain, which make some areas unable to be observed by sensors, forming shadow areas; these obstructions and shadows will affect the integrity and quality of point cloud data. (2) Ground classification: In mountainous gully terrain, the degree of mixing between ground points and non-ground points may be high, especially in canyons and gullies, where there are a large number of non-ground points such as vegetation. Summary of the invention

[0005] The present invention provides a typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud, the improvement of which is that the method includes

[0006] Step (1): Acquire a three-dimensional point cloud data set through array synthetic aperture radar;

[0007] Step (2): preprocessing the three-dimensional point cloud dataset to obtain a new three-dimensional point cloud dataset;

[0008] Step (3): extract ground points using the new 3D point cloud dataset;

[0009] Step (4): extracting typical terrain ground points;

[0010] Step (5): construct a tree height feature operator based on the extracted typical terrain ground points;

[0011] Step (6): Segment typical terrain ground points and reconstruct the ground points to obtain a ground point dataset;

[0012] Step (7): Construct terrain model.

[0013] Furthermore, the step (1) includes acquiring array SAR images by means of an array synthetic aperture radar antenna mounted on an airborne platform, acquiring a three-dimensional point cloud by means of a tomographic SAR technique, and integrating multiple sets of data on the three-dimensional point cloud to form a three-dimensional point cloud data set.

[0014] Furthermore, the preprocessing of the three-dimensional point cloud data set in step (2) includes de-redundancy of overlapping areas of the three-dimensional point cloud data set and removal of low points and flying points.

[0015] Furthermore, the processing of removing redundancy in overlapping areas and removing low points and flying points of the three-dimensional point cloud data set is achieved by the following steps:

[0016] (2.1) For each 3D point cloud point P in the 3D point cloud dataset i , calculate and other points P j The distance between ij ;

[0017] (2.2) Setting the threshold Where D is the overlapping width of adjacent strips;

[0018] (2.3) For each 3D point cloud point P i , count the number of its neighboring points N i , set the threshold N i ≤100;

[0019] (2.4) When d ij ≤S and N i ≤100, then point P j It's point P i 's neighboring points;

[0020] (2.5) When d ij >S or N i >100, then point P j It's point P i Non-neighboring points of

[0021] (2.6) Repeat steps (2.1)-(2.5) until all points in the three-dimensional point cloud data set in the overlapping area are calculated once, and each point is determined to be a neighboring point or a non-neighboring point;

[0022] (2.7) Delete non-neighboring points;

[0023] (2.8) Remove low points and flying points from the data obtained in step (2.7) to obtain a new three-dimensional point cloud data set.

[0024] Furthermore, the step (3) comprises

[0025] The new 3D point cloud dataset is divided into ground points and non-ground points to form a ground point dataset {P g}.

[0026] Furthermore, the step (4) comprises

[0027] (4.1) Extract ground points to generate digital elevation model;

[0028] (4.2) Calculate the ground tilt angle based on the extracted digital elevation model;

[0029] (4.3) Obtaining a synthetic aperture radar intensity map of the same phase and area as the three-dimensional point cloud;

[0030] (4.4) Extract tree range based on synthetic aperture radar intensity map;

[0031] (4.5) According to the ground inclination angle and the tree range, we select the ground points with ground inclination angle greater than 6° and within the tree range to form a typical terrain ground point dataset {P gd}.

[0032] Furthermore, the step (5) comprises

[0033] (5.1) Construct a profile of typical terrain points using vertical contour lines to obtain a two-dimensional point cloud;

[0034] (5.2) Construct the tree height feature operator and obtain the tree height H.

[0035] Furthermore, let point A and point B be the two endpoints with the longest distance from the bottom of the tree crown, point C be the highest point of the tree crown, and point D be the ground point closest to the tree crown. The coordinates of points A, B, C, and D of the sunny slope tree are (x A阳 ,y A阳 ),(x B阳 ,y B阳 ),(x C阳 ,y C阳 ),

[0036] (x D阳 ,y D阳 ); the coordinates of points A, B, C, and D of the shady slope tree are (x A阴 ,y A阴 ),(x B阴 ,y B阴 ),(x C阴 ,y C阴 ),(x D阴 ,y D阴 );

[0037] Assume that the tree crown is a triangular cone parallel to the ground and grows symmetrically, that is, AC = CB; taking a single parabola as an example, the height of a single tree is CC′ + C′D; where point C' is the midpoint of points A and B, point D is the ground point, and C′D is perpendicular to the ground;

[0038]

[0039] H 阳 =CC′ 阳 +C′D 阳 ;

[0040] H 阴 =CC′ 阴 +C′D 阴 ;

[0041]

[0042] Furthermore, the step (6) comprises

[0043] (6.1) In the horizontal direction, the typical terrain ground points are divided into several units by using rectangles of size R×R; each unit contains multiple ground points;

[0044] (6.2) Taking a rectangle of size R×R, the R segmentation formula is:

[0045]

[0046] (6.3) The highest point P in the vertical direction in each unit is divided into several units. h Extract, extract several P h Points constitute point cloud dataset {P h};

[0047] (6.4){P h}Remove the tree height H along the normal of the ground slope to obtain a new typical terrain ground point dataset {P hg}.

[0048] Further, the step (7) includes merging the new typical terrain ground point dataset and the atypical ground point dataset to generate a new ground point dataset, and constructing a three-dimensional terrain model according to the new ground point dataset; wherein the atypical terrain ground point dataset is obtained by removing the typical terrain ground point dataset from the ground point dataset, that is, {P g}-{P gd}.

[0049] Beneficial effects:

[0050] The present invention proposes a typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud, which is used to remove the tree height in the typical terrain area; extract and process the ground points in the typical terrain area, calculate and obtain the typical terrain ground points based on the ground point extraction, and reconstruct the ground points. It solves the problem of insufficient point cloud data integrity and digital elevation model (DEM) generation loopholes in the mountainous gully terrain area covered by vegetation due to observation conditions.

[0051] The present invention optimizes the data preprocessing step. In the step of removing data redundancy in overlapping areas, a threshold is set based on the distance filtering algorithm, which reduces a large amount of repeated information, noise and errors in the point cloud data, improves the quality and accuracy of the entire point cloud data, makes it more in line with actual scenarios, and achieves a better de-redundancy purpose.

[0052] It should be understood that the above general description and the following specific embodiments are only exemplary and illustrative and do not limit the scope of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A flowchart of a typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud according to the present invention;

[0054] It should be understood that the drawings are not necessarily drawn to scale, presenting somewhat simplified representations of various features illustrative of the basic principles of the present disclosure. The specific design features of the present invention as disclosed herein, including, for example, specific dimensions, orientations, locations, and shapes will be determined in part by the particular intended application and use environment.

[0055] In the drawings, reference numbers refer to the same or equivalent parts of the present invention throughout the several figures of the drawing. DETAILED DESCRIPTION

[0056] Reference will now be made in detail to various embodiments of the present invention, examples of which are illustrated in the accompanying drawings and described below. Although the present invention will be described in conjunction with exemplary embodiments of the present invention, it should be understood that this specification is not intended to limit the present invention to those exemplary embodiments. On the other hand, the present invention is intended to cover not only the exemplary embodiments of the present invention, but also various alternatives, modifiers, equivalents and other embodiments, which may be included in the spirit and scope of the present invention as defined by the appended claims.

[0057] Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. The specific structures and functions described in the exemplary embodiments of the present invention are for illustrative purposes only. Embodiments according to the concepts of the present invention can be implemented in various forms, and it should be understood that they should not be interpreted as being limited to the exemplary embodiments described in the exemplary embodiments, but include all modifications, equivalents or substitutes included in the spirit and scope of the present invention.

[0058] Throughout the specification, the professional terms used herein are only for describing various exemplary embodiments and are not intended to be limiting. It will be further understood that the terms "include", "comprising", "having", etc., when used in the exemplary embodiments, specifically refer to the presence of the stated parts, steps, operations or elements, but do not exclude the presence or addition of one or more other parts, steps, operations or elements.

[0059] The present invention optimizes the data preprocessing step. In the step of removing data redundancy in overlapping areas, a threshold is set based on the distance filtering algorithm, which reduces a large amount of repeated information, noise and errors in the point cloud data, improves the quality and accuracy of the entire point cloud data, makes it more in line with actual scenarios, and achieves a better de-redundancy purpose.

[0060] The present invention proposes a typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud, which is used to remove the tree height in the typical terrain area; extract and process the ground points in the typical terrain area, calculate and obtain the typical terrain ground points based on the ground point extraction, and reconstruct the ground points. It solves the problem of insufficient point cloud data integrity and digital elevation model (DEM) generation loopholes in the mountainous gully terrain area covered by vegetation due to observation conditions.

[0061] like Figure 1 As shown, the present invention provides a typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud, the method comprising:

[0062] Step (1): Acquire a three-dimensional point cloud data set through array synthetic aperture radar;

[0063] Array SAR images are acquired using an airborne platform equipped with an array synthetic aperture radar antenna, and a three-dimensional point cloud is obtained through tomographic SAR technology; multiple groups of data on the acquired three-dimensional point cloud are integrated to form a three-dimensional point cloud dataset.

[0064] Step (2): preprocessing the three-dimensional point cloud dataset to obtain a new three-dimensional point cloud dataset;

[0065] Since data is affected by various factors during scanning and acquisition, a large number of noise points will appear in the 3D point cloud data, including jitter caused by airflow during flight, unstable gimbal, operator errors, occlusion by foreign objects such as birds, insufficient scanning, etc. Since 3D point cloud data is discrete, it is difficult to distinguish and process it through general methods; currently, the following two types of noise need to be considered:

[0066] Redundant points: mainly refer to the redundant data generated by repeatedly scanning the same area, or the excessive point density in the overlapping area after the 3D point cloud is stitched. In addition to the above reasons, insufficient scanning of objects at the edge of the scanning area will also lead to the appearance of redundant points;

[0067] Low points and flying points: They are the most typical 3D point cloud noise points, mainly referring to sparse points or small point clouds far away from the main body of the 3D point cloud. Such 3D point clouds have little correlation with the main 3D point cloud and contain less information, but they interfere greatly with the algorithm.

[0068] The present invention first removes redundancy of points in the overlapping area based on the three-dimensional point cloud data set obtained in step (1), and then removes low points and flying points; thereby obtaining a new three-dimensional point cloud data set.

[0069] The data in synthetic aperture radar three-dimensional point cloud data is acquired in units of flight strips. There is a large amount of repeated information in the point cloud data in the overlapping areas of adjacent strips. The redundant data in the overlapping areas may introduce noise and errors. Removing these redundant data can effectively reduce the amount of data, reduce the cost of data processing and storage, and improve the quality and accuracy of the entire point cloud data, making it more in line with actual scenarios.

[0070] The overlapping area de-redundancy and removal of low points and flying points of the 3D point cloud dataset are processed by the following steps:

[0071] (2.1) For each 3D point cloud point P in the 3D point cloud dataset obtained in step (1), i , calculate its and other points P j The distance between ij ;

[0072] (2.2) Setting the threshold Where D is the overlapping width of adjacent strips;

[0073] (2.3) For each 3D point cloud point P i , count the number of its neighboring points N i , set the threshold N i ≤100;

[0074] (2.4) When d ij ≤S and N i ≤100, then point Pj It's point P i 's neighboring points;

[0075] (2.5) When d ij >S or N i >100, then point P j It's point P i Non-adjacent points; non-adjacent points are redundant points;

[0076] (2.6) Repeat steps (2.1)-(2.5) until all points in the three-dimensional point cloud data set in the overlapping area are calculated once, and each point is determined to be a neighboring point or a non-neighboring point;

[0077] (2.7) Delete non-neighboring points;

[0078] Among them, the present invention improves the performance of the filtering algorithm, optimizes the parameters in the filtering algorithm by innovatively defining the distance threshold, makes the filtering algorithm more adaptable to solving problems, improves filtering efficiency, and solves the problem of redundant point clouds in overlapping area strips caused by repeated scanning;

[0079] (2.8) removing low points and flying points from the data obtained in step (2.7) to obtain a new three-dimensional point cloud data set;

[0080] The purpose of removing low points and flying points in the present invention is to clean and purify the original three-dimensional point cloud data so as to make it more suitable for subsequent analysis and application; the purpose of removing low points is to filter out points below the ground from the three-dimensional point cloud data so as to reduce the amount of data and improve the efficiency of subsequent processing; flying points refer to outliers in the three-dimensional point cloud that are inconsistent with the surrounding environment, which may be caused by sensor errors, environmental interference or data acquisition errors; flying points will affect the accuracy and reliability of the three-dimensional point cloud data; the purpose of removing flying points is to identify and filter out these abnormal points, thereby improving the quality of the three-dimensional point cloud data.

[0081] Step (3): extract ground points using the new 3D point cloud dataset;

[0082] Ground point extraction refers to the process of separating the ground plane from the mixed point cloud data. This includes ground points and non-ground points; non-ground points are points outside the ground plane; the pre-processed 3D point cloud dataset in step (2) is divided into two categories: ground points and non-ground points, forming a ground point dataset {P g};

[0083] Step (4): extracting typical terrain ground points;

[0084] (4.1) Set the “ground tilt angle” attribute of the ground point and extract the ground point to generate a digital elevation model;

[0085] Generate a digital elevation model (DEM) based on the ground points extracted in step (3). DEM is a physical ground model that represents ground elevation in the form of a set of ordered numerical arrays (grid cells); wherein each ground point corresponds to a unique grid cell;

[0086] (4.2) Calculate the ground tilt angle based on the extracted digital elevation model;

[0087] Calculate the ground tilt angle based on the DEM extracted in step (4.1); the slope of the surface unit is the angle between the normal direction of its tangent plane and the Z axis; select a 3x3 grid unit for slope calculation; calculate the slope of the grid unit in eight directions and then take the average value; the slope is equal to the tangent value of the tilt angle, and then calculate the tilt angle; obtain the tilt angle of each grid unit;

[0088] The ground inclination angle of the ground point, that is, the ground inclination angle of the grid unit corresponding to the ground point;

[0089] (4.3) Obtain a synthetic aperture radar (SAR) intensity map of the same phase and area as the three-dimensional point cloud; each ground point can correspond to a point on the intensity map on a horizontal two-dimensional plane;

[0090] (4.4) Define the attribute of "whether it is a tree" of the ground point and extract the range of trees based on the synthetic aperture radar intensity map;

[0091] (4.4.1) Extract tree range based on synthetic aperture radar intensity map;

[0092] (4.4.2) Extract features from SAR intensity images to describe different types of targets; including statistical features (such as mean, variance), texture features (such as gray-level co-occurrence matrix) and shape features (such as contour);

[0093] (4.4.3) Divide the 3D point dataset into two categories: training set and test set;

[0094] (4.4.4) Use the training set data to train the classification model and adjust the model parameters to optimize performance;

[0095] (4.4.5) Use the test set data to evaluate the performance of the trained model;

[0096] (4.4.6) Optimize the model based on the evaluation results to improve the classification effect of the model;

[0097] (4.4.7) The optimized classification model is used to classify new regional SAR intensity images;

[0098] (4.4.8) Based on the synthetic aperture radar intensity map, it is divided into two categories: "trees" and "non-trees";

[0099] (4.4.9) Let the attribute of ground points within the “tree” category on the synthetic aperture radar intensity map be “is a tree”.

[0100] (4.5) According to the ground inclination angle and the range of trees, ground points with a ground inclination angle greater than 6° and the attribute of "is a tree" are selected to form a typical terrain ground point dataset {P gd}.

[0101] Step (5): construct a tree height feature operator based on the extracted typical terrain ground points to obtain the tree height H;

[0102] The vertical contour lines of the typical terrain ground points are used to construct the profile to obtain a two-dimensional point cloud; a tree is selected on the sunny slope and the shady slope in the valley gully. Taking a single tree as an example, the two endpoints with the longest distance from the bottom of the tree crown are selected, point A and point B; the highest point of the tree crown is selected, point C; the ground point closest to the tree crown is selected, point D; the coordinates of points A, B, C, and D of the sunny slope tree are (x A阳 ,y A阳 ),(x B阳 ,y B阳 ),(x C阳 ,y C阳 ),(x D阳 ,y D阳 The coordinates of points A, B, C, and D of the shady slope tree are (x A阴 ,y A阴 ),(x B阴 ,y B阴 ),(x C阴 ,y C阴 ),(x D阴 ,y D阴 );

[0103] Since trees exist in three-dimensional form in point cloud data and in irregular form in nature, in the modeling process, the tree crown is assumed to be a triangular cone parallel to the ground and grow symmetrically, that is, the height difference between point A and point B can be ignored, and AC = CB. Taking a single parabola as an example, it can be concluded that the height of a single tree is CC'+C'D. Among them, point C' is the midpoint of points A and B, point D is a ground point, and C'D is perpendicular to the ground.

[0104]

[0105] H 阳 =CC′ 阳 +C′D 阳 ;

[0106] H 阴 =CC′ 阴 +C′D 阴 ;

[0107]

[0108] Step (6): Segment typical terrain ground points and reconstruct the ground points to obtain a ground point dataset;

[0109] (6.1) The typical terrain ground points are divided into blocks in the horizontal direction using rectangles of size R×R; they are divided into a number of units; each unit contains multiple ground points; the mathematical datum used in the “horizontal direction” here is the same as the mathematical datum of the “typical terrain ground point dataset”.

[0110] (6.2) Taking a rectangle of size R×R, the R segmentation formula is:

[0111]

[0112] (6.3) Based on the result obtained in step (6.1), among the units of ground point segmentation, each unit has a highest point P in the vertical direction. h , there are several P h Points, forming a point cloud dataset {P h};

[0113] (6.4){P h}Remove the tree height H along the normal of the ground slope to obtain a new typical terrain ground point dataset {P hg}.

[0114] Step (7): constructing a terrain model;

[0115] Merge the new typical terrain ground point dataset (ground point dataset {P hg}) and atypical terrain ground point dataset (the atypical terrain ground point dataset is obtained by removing the typical terrain ground point dataset from the ground point dataset, that is, {P g}-{P gd}), generate a new ground point dataset, build a three-dimensional terrain model based on the dataset, and generate a DEM (digital elevation model).

[0116] The above-mentioned embodiments of the present application can be implemented in various hardware, software coding or a combination of the two. For example, the embodiments of the present application can also represent the program code for executing the above-mentioned method in a data signal processor. The present application can also relate to the various functions performed by a computer processor, a digital signal processor, a microprocessor or a field programmable gate array. The above-mentioned processor can be configured according to the present application to perform specific tasks, which are completed by executing machine-readable software code or firmware code that defines the specific method disclosed by the present application. The software code or firmware code development can represent different programming languages ​​and different formats or forms. Different target platform compiled software codes can also be represented. However, the different code styles, types and languages ​​of the software code for performing tasks according to the present application and other types of configuration codes do not depart from the spirit and scope of the present application.

[0117] The foregoing description of specific exemplary embodiments of the present invention has been presented for the purpose of illustration and description. It is not intended to exclude or limit the present invention to the precise form disclosed, and it is obvious that many modifications and changes are feasible in view of the above teachings. Exemplary embodiments are selected and described to explain certain principles of the present invention and their practical application, so that other technicians in the field can make or use various exemplary embodiments of the present invention, and various alternatives and modifications thereof. It is intended that the scope of the present invention will be defined by the claims attached hereto and their equivalents.

[0118] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud, characterized in that: The method comprises Step (1): Acquire a three-dimensional point cloud dataset through array synthetic aperture radar; Step (2): preprocessing the three-dimensional point cloud dataset to obtain a new three-dimensional point cloud dataset; Step (3): Extract ground points using the new 3D point cloud dataset; Step (4): Extract typical terrain ground points; Step (5): construct a tree height feature operator based on the extracted typical terrain ground points; (5.1) Construct a profile of typical terrain points by vertical contour lines to obtain a two-dimensional point cloud; (5.2) Construct the tree height feature operator and obtain the tree height H; Step (6): Segment typical terrain ground points and reconstruct the ground points to obtain a ground point dataset; (6.1) In the horizontal direction Rectangles of different sizes are used to divide typical terrain points into several units, each of which contains multiple ground points. (6.2) The R segmentation formula for a rectangle of size is: ； ； ； (6.3) The highest point in each unit of the ground point is Extract, extract several Points constitute point cloud dataset ; (6.4) Remove the tree height H along the normal direction of the ground slope to obtain a new typical terrain ground point dataset ; Step (7): constructing a terrain model; Among them, the point , point B is the two endpoints with the longest distance from the bottom of the tree crown, point C is the highest point of the tree crown, and point D is the ground point closest to the tree crown. The coordinates of points A, B, C, and D of the sunny slope tree are ( ),( ),( ),( ); The coordinates of points A, B, C, and D of the shady slope tree are ( ),( ),( ),( ).

2. The typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud according to claim 1 is characterized in that: The step (1) includes acquiring an array SAR image by using an array synthetic aperture radar antenna mounted on an airborne platform, acquiring a three-dimensional point cloud by using a tomographic SAR technique, and integrating multiple sets of data on the three-dimensional point cloud to form a three-dimensional point cloud data set.

3. The typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud according to claim 1 is characterized in that: The preprocessing of the three-dimensional point cloud data set in step (2) includes removing redundancy in overlapping areas and removing low points and flying points of the three-dimensional point cloud data set.

4. The typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud according to claim 3 is characterized in that: The processing of removing redundancy in overlapping areas and removing low points and flying points of the three-dimensional point cloud data set is achieved by the following steps: (2.1) For each 3D point cloud point in the 3D point cloud dataset , calculate and other points The distance between ; (2.2) Setting the threshold , where D is the overlapping width of adjacent strips; (2.3) For each 3D point cloud point , count the number of its neighboring points , set the threshold ; (2.4) When and , then it is considered that point Yes 's neighboring points; (2.5) When or , then it is considered that point Yes Non-neighboring points of (2.6) Repeat steps (2.1) to (2.5) until all points in the three-dimensional point cloud data set in the overlapping area are calculated once, and each point is determined to be a neighboring point or a non-neighboring point; (2.7) Delete non-neighboring points; (2.8) Remove low points and flying points from the data obtained in step (2.7) to obtain a new three-dimensional point cloud dataset.

5. The typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud according to claim 1 is characterized in that: The step (3) includes dividing the new three-dimensional point cloud data set into ground points and non-ground points to form a ground point data set. .

6. The typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud according to claim 1, characterized in that: The step (4) includes (4.1) Extract ground points to generate digital elevation model; (4.2) Calculate the ground tilt angle based on the extracted digital elevation model; (4.3) Obtain a synthetic aperture radar intensity map of the same phase and area as the 3D point cloud; (4.4) Extract tree range based on synthetic aperture radar intensity map; (4.5) According to the ground inclination angle and tree range, select ground points with ground inclination angle greater than 6° and within the tree range to form a typical terrain ground point dataset .

7. The typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud according to claim 1, characterized in that: Assume that the tree crown is a triangular cone parallel to the ground and grows symmetrically, that is, ; The height of a single tree is + ; Point C' is the midpoint between points A and B, and point D is the ground point. perpendicular to the ground; ； ； ； ； ； ； 。 8. The typical terrain construction method based on array synthetic aperture radar three-dimensional point cloud according to claim 1, characterized in that: The step (7) includes merging a new typical terrain ground point dataset and an atypical ground point dataset to generate a new ground point dataset, and constructing a three-dimensional terrain model based on the new ground point dataset; wherein the atypical terrain ground point dataset is obtained by removing the typical terrain ground point dataset from the ground point dataset. .

Images