CN117392569B - Airborne Lidar point cloud ground point extraction method integrating DOM image and three-dimensional live-action model - Google Patents

Abstract

The invention discloses an airborne Lidar point cloud ground point extraction method, an airborne Lidar point cloud ground point extraction device, an electronic product and a computer readable storage medium which are used for fusing DOM images and three-dimensional live-action models. According to the method and the device, the slope ridge region is judged according to the texture and the gradient information of the target region, and accurate partition processing is carried out on the project region, so that the data processing time can be saved, the data processing efficiency is improved, and the time waste caused by conventional region data fusion is avoided. According to the invention, the slope surface of the three-dimensional live-action model is used as an initial reference surface, the reference point is determined according to the distance, and the characteristic reference point at the slope turning position is increased, so that the point cloud classification reference surface which is maximally attached to the ground surface is constructed, the completeness of the characteristic points at the top and the bottom of the slope can be ensured, and the accuracy of point cloud classification is improved.

Description

Method for extracting ground points from airborne Lidar point cloud by integrating DOM images and 3D real scene models

Technical Field

The present invention belongs to the field of laser radar point cloud data processing, and specifically relates to an airborne lidar point cloud ground point extraction method, device, electronic product and computer-readable storage medium that integrates DOM images and three-dimensional real scene models.

Background Art

As important information in terrain surveying and mapping, the accuracy and point density of ground elevation points will have a significant impact on subsequent design and construction. Mountainous areas have large elevation differences, steep terrain, difficult access for personnel, and poor safety of operations. The efficiency of actual measurement is low and the cost is high. Aerial photogrammetry technology cannot obtain ground elevation values in vegetation-covered areas. However, airborne Lidar scanning technology has a certain degree of penetration into vegetation and is less restricted by terrain and traffic conditions. It can quickly obtain complete three-dimensional point cloud data of the target area. Therefore, in terms of current technical means, airborne Lidar scanning technology is the first choice for quickly obtaining ground points. It has obvious advantages and is irreplaceable.

The original Lidar point cloud data obtained by airborne Lidar scanning technology contains not only ground points, but also ground points such as buildings and vegetation, which need to be classified before the ground point results can be obtained. At present, most of the Lidar point cloud classification methods are based on mathematical filtering algorithms, which perform iterative operations based on distance and angle.

However, these classification methods generally have limitations in slopes. First, the accuracy of the results of these methods is greatly affected by the distance and angle thresholds. Different thresholds need to be set for different slopes and different surface conditions, which has the problem of poor versatility. Secondly, it is difficult to finely divide the area and set a suitable threshold for complex natural landforms. On the one hand, setting a threshold that is too small will cause ground points to be misclassified as ground object points, resulting in missing results data and missing key terrain information; on the other hand, setting a threshold that is too large will cause ground object points (such as building walls, tree trunks, etc.) to be misclassified as ground points, causing abnormal terrain undulations and distorting the real terrain information. Due to improper threshold selection, there is a common problem of misclassification. Third, since the classification reference surface in these classification methods is greatly affected by parameter settings, the grid size, boundaries, etc. are different and have great randomness, it is difficult to reflect the basic real undulation of the terrain. In areas where the surface slope changes suddenly, such as slopes, there is a common problem of missing data of unidentified ground points. Fourth, the current airborne Lidar point cloud classification results have a certain degree of elevation deviation. In summary, the current Lidar point cloud classification algorithm is difficult to meet the needs of ground point extraction.

Summary of the invention

In view of this, the purpose of the present invention is to realize a method for extracting ground points from an airborne Lidar point cloud, so as to solve the problems in the above-mentioned background technology, such as the difficulty in selecting thresholds at slopes, misclassification caused by improper threshold selection, data omissions caused by the randomness of the classification reference surface, and elevation deviation in the classification results.

To achieve the above objectives, in a first aspect, the present invention provides a method for extracting ground points from an airborne Lidar point cloud by integrating DOM images and a three-dimensional real scene model, comprising:

Acquire DOM images, three-dimensional real scene models and Lidar point cloud data of the target area, wherein the DOM images, three-dimensional real scene models and Lidar point cloud data have the same coordinate system and elevation system and have the same range;

Using the DOM image and 3D real scene model of the target area, search and determine the slope area in the target area, extract and save the closed range line of the slope area;

Cut and save the Lidar point cloud data of the slope area according to the closed range line of the slope area;

For each slope area, a search starting reference surface is constructed, and the initial ground points of the Lidar point cloud data are determined based on the search starting reference surface. The Lidar point cloud data classification reference surface is constructed according to the determined initial ground points. Iterative classification is performed based on the Lidar point cloud data classification reference surface to obtain all the Lidar point cloud ground points in the slope area.

In some embodiments of the present invention, DOM images are obtained by processing the original photo data, POS data and image control point data obtained by the airborne camera to obtain DOM image data with real geographic coordinates; three-dimensional real scene models are obtained by processing the original photo data, POS data and image control point data obtained by the drone to obtain a high-precision three-dimensional real scene model with real geographic coordinates; Lidar point cloud data are obtained by obtaining the obtained airborne Lidar point cloud measured data through trajectory solution, flight strip stitching and coordinate conversion.

In some embodiments of the present invention, a search is performed based on the texture information of the DOM image of the target area and the surface undulation of the three-dimensional real scene model, and the closed range line of the slope area is extracted and saved according to a preset buffer radius. The buffer radius is a redundancy of the closed range line set to compensate for the omission of point cloud classification results caused by errors in determining the closed range line.

In some embodiments of the present invention, the construction of the search starting reference plane specifically involves first removing the objects on the surface of the three-dimensional real scene model through a filtering algorithm, and then constructing the search starting reference plane for the slope area based on the slope data and position data of the slope area in the three-dimensional real scene model. The search starting reference plane is a mathematical surface with the same slope as the three-dimensional real scene model, and the spatial position is determined based on the concentrated area of the elevation values of the Lidar point cloud data.

In some embodiments of the present invention, determining the initial ground point of the Lidar point cloud data based on the search starting reference plane includes: if the distance between the Lidar point cloud data and the search reference plane in the slope area is within a preset flatness threshold range, then the point with the lowest elevation value of the Lidar point cloud data is the initial ground point; otherwise, searching is performed in the slope area based on the position of the Lidar point cloud data and the constructed search starting reference plane: if the Lidar point cloud data is completely located on one side of the search starting reference plane, then the initial ground point is the point with the shortest vertical distance to the reference plane; if the Lidar point cloud data is completely located on both sides of the search starting reference plane, first distinguish the object side and the ground side of the search starting reference plane, and take the point on the ground side that is farthest from the reference plane as the initial ground point; in addition, at the place where the slope of the search starting reference plane changes, the plumb line passing through the turning point is used as the search reference line, and the lowest Lidar point cloud data close to the search reference line is found as the initial ground point.

In some embodiments of the present invention, a Lidar point cloud data classification reference surface is constructed based on the determined initial ground points, and a TIN grid is formed by all the initial ground points. The TIN grid serves as the Lidar point cloud data classification reference surface.

In some embodiments of the present invention, iterative classification based on the Lidar point cloud data classification reference plane includes: for each Lidar point cloud data, determining whether its vertical distance and angle from the Lidar point cloud data classification reference plane meet the judgment conditions; if the conditions are met, classifying it as a ground point, and adding the ground point to the Lidar point cloud data classification reference plane, that is, updating the Lidar point cloud data classification reference plane, traversing all Lidar point cloud data to perform the above processing, and iteratively updating the Lidar point cloud data classification reference plane while classifying, until all point clouds are traversed and classification is completed; the judgment condition is within the range defined by the preset distance threshold and angle threshold.

In a second aspect, the present invention provides an airborne Lidar point cloud ground point extraction device that integrates DOM images and three-dimensional real scene models, comprising:

A data acquisition unit, which is used to acquire DOM images, three-dimensional real scene models and Lidar point cloud data of the target area, wherein the DOM images, three-dimensional real scene models and Lidar point cloud data have the same coordinate system and elevation system and have the same range;

A slope area determination unit, which is used to use the DOM image and the three-dimensional real scene model of the target area to search and determine the slope area in the target area, and extract and save the closed range line of the slope area;

A data cutting unit, which is used to cut and save the Lidar point cloud data of the slope area according to the closed range line of the slope area;

The ground point extraction unit is used to construct a search starting reference plane for each slope area, determine the initial ground points of the Lidar point cloud data based on the search starting reference plane, construct the Lidar point cloud data classification reference plane based on the determined initial ground points, and perform iterative classification based on the Lidar point cloud data classification reference plane to obtain all Lidar point cloud ground points in the slope area.

In a third aspect, the present invention provides an electronic product, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method for extracting ground points from an airborne Lidar point cloud that integrates DOM images and three-dimensional real scene models.

In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the method for extracting ground points from an airborne Lidar point cloud that integrates DOM images and three-dimensional real scene models is implemented.

Beneficial Effects

Compared with the traditional Lidar point cloud classification method, the present invention combines high-precision orthophotos and three-dimensional real scene models to classify Lidar point clouds, which has the following beneficial effects:

(1) According to the texture and slope information of the target area, the slope area is judged and the project area is accurately zoned. This can save data processing time, improve data processing efficiency, and avoid the waste of time in conventional regional data fusion.

(2) The slope surface of the 3D real-life model is used as the starting reference surface. The reference points are determined based on the distance, and characteristic reference points are added at the slope turning points. This constructs a point cloud classification reference surface that fits the surface as closely as possible. This ensures the integrity of the characteristic points at the top and bottom of the slope and improves the accuracy of point cloud classification.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of a method for extracting ground points from an airborne Lidar point cloud in an embodiment of the present invention.

FIG. 2 is a schematic diagram of a method for selecting initial ground points in a relatively flat area according to an embodiment of the present invention.

FIG3 is a schematic diagram of a method for selecting an initial ground point in a single-sided situation according to an embodiment of the present invention.

FIG. 4 is a schematic diagram of a method for selecting initial ground points in a double-sided situation in an embodiment of the present invention.

FIG. 5 is a schematic diagram of the point cloud classification principle in an embodiment of the present invention.

FIG6 is a schematic diagram showing the composition of an airborne Lidar point cloud ground point extraction device in an embodiment of the present invention.

FIG. 7 is a schematic diagram of the composition of an electronic product in an embodiment of the present invention.

Reference numerals:

1. Search area for Lidar point cloud classification, 2. Lidar point cloud data, 3. Initial ground point search reference line, 4. Search starting datum surface, 5. Lidar point cloud data classification datum surface

DETAILED DESCRIPTION

The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

In order to solve the problems existing in the existing Lidar point cloud classification, as shown in FIG. 1 to FIG. 5 , the present invention in one embodiment of the present invention integrates DOM images and three-dimensional real scene models to extract ground points of airborne Lidar point clouds, including:

S100, acquiring a DOM image, a three-dimensional real scene model, and Lidar point cloud data of a target area, wherein the DOM image, the three-dimensional real scene model, and Lidar point cloud data have the same coordinate system and elevation system and have the same range;

S200, using the DOM image and the three-dimensional real scene model of the target area, searching and determining the slope area in the target area, extracting and saving the closed range line of the slope area;

S300, cutting and saving Lidar point cloud data of the slope area according to the closed range line of the slope area;

S400, for each slope area, construct a search starting reference surface, determine the initial ground points of the Lidar point cloud data based on the search starting reference surface, construct a Lidar point cloud data classification reference surface based on the determined initial ground points, perform iterative classification based on the Lidar point cloud data classification reference surface, and obtain all Lidar point cloud ground points in the slope area.

Referring to Figures 2-5, Figure 2 is a schematic diagram of the method for selecting initial ground points in a relatively flat area in an embodiment of the present invention; Figure 3 is a schematic diagram of the method for selecting initial ground points in a single-sided situation in an embodiment of the present invention; and Figure 4 is a schematic diagram of the method for selecting initial ground points in a double-sided situation in an embodiment of the present invention.

FIG5 is a schematic diagram of the point cloud classification principle in an embodiment of the present invention. The radar point cloud data is stereo data in three-dimensional space, and FIG2-FIG4 is illustrated by a cross-section slice in the stereo data. Among them, 1 represents the search area (three-dimensional area, shown as a box in the cross-section slice diagram) for point cloud classification, which is the basic unit for searching the initial ground point of the point cloud, 2 represents the Lidar point cloud data, and 3 represents the reference line for searching the initial ground point. Among them, FIG1 takes the lowest point as the starting ground point, and the reference line is in the horizontal direction. The lowest point can be determined as the initial ground point by comparing the elevation value of the point cloud. In FIG2 and FIG3, the vertical distance to the starting reference plane is used as the basis for determining the initial ground point, so the reference line is a line segment perpendicular to the starting reference plane; 4 represents the search starting reference plane; 5 represents the Lidar point cloud data classification reference plane constructed according to the initial ground point.

Specifically, in S100, the acquisition of DOM images, in some embodiments of the present invention, is to process the original photo data, POS data (data recording the flight position and posture information of the drone, each photo corresponds to a record in the POS data) and image control point data acquired by the airborne camera through aerial triangulation calculation, image stitching and other processing to produce a high-precision DOM image data result with real geographic coordinates. The acquisition of three-dimensional real-scene models, in some embodiments of the present invention, is to process the original photos, POS data and image control point data acquired by the drone through connection point extraction, triangulation network construction, texture mapping and other processing to produce a high-precision three-dimensional real-scene model with real geographic coordinates. When producing high-precision DOM and three-dimensional real-scene models, the same plane and elevation references are used, and the existing results can also be converted into coordinates to achieve spatial position alignment. The acquisition of Lidar point cloud data, in some embodiments of the present invention, is to obtain the acquired airborne Lidar point cloud measured data through trajectory solution, flight strip splicing, and coordinate conversion. The coordinate conversion is to convert the coordinate system used for airborne Lidar measurement to the actual required coordinate system. The plane generally adopts the CGCS2000 coordinate system, and the elevation generally adopts the 1985 National Yellow Sea Elevation System. The specific situation shall prevail. High-precision DOM images, three-dimensional real-scene models, and Lidar point cloud data have the same coordinate system and elevation system, otherwise coordinate conversion preprocessing is performed, and the three ranges are consistent.

Specifically, in S200, a search is performed based on the texture information of the DOM image of the target area and the surface undulation of the three-dimensional real scene model. For example, if a sudden change in the surface slope is found in a certain area of the target area, the area is determined to be a slope area, and the closed range line of the slope area is extracted and saved according to a preset buffer radius. The buffer radius is a closed range line redundancy set to compensate for the omission of point cloud classification results caused by the error in determining the closed range line.

Specifically, in S300, the Lidar point cloud data of the slope area is cut and saved according to the closed range line of the slope area, and the Lidar point cloud data of the entire target area is divided into two parts: the Lidar point cloud data of the slope area and the Lidar point cloud data of the general area (non-slope area). The Lidar point cloud data of the general area is processed according to the conventional point cloud classification method.

Specifically, in S400, vegetation, buildings and other objects on the surface of the three-dimensional real scene model are first removed through a filtering algorithm, and then a search starting reference plane for the slope area is constructed according to the slope data and position data of the slope area in the three-dimensional real scene model. The search starting reference plane is a mathematical surface with the same slope as the three-dimensional real scene model. The spatial position is determined according to the area where the elevation values of the Lidar point cloud data are concentrated. Generally, the spatial position is set at a position that intersects with the lowest area in the elevation value concentration of the Lidar point cloud data or is close to the lower part of the lowest area in the elevation value concentration of the Lidar point cloud data, as shown by the dotted lines in Figures 2 to 4. It should be noted that the search starting reference plane 4 is a three-dimensional surface patch without texture information, and the accuracy of the surface slope information at the vegetation coverage is not high.

The present invention uses three-dimensional real-scene model data to construct a search start reference plane instead of the reference plane constructed by the lowest elevation point in the existing conventional method, which can ensure the integrity of the characteristic points of the top and bottom of the slope and improve the accuracy of point cloud classification.

Specifically, in S400, determining the initial ground point of the Lidar point cloud data based on the search starting reference plane includes:

If the Lidar point cloud data is relatively flat relative to the search starting reference plane in the slope area, for example, the distance between the Lidar point cloud data and the search reference plane is within the preset flatness threshold range, which can be obtained based on experience or experiments, then the point with the lowest elevation value of the Lidar point cloud data is the initial ground point, as shown in Figure 2.

Otherwise, search is performed in the slope area based on the position of the Lidar point cloud data and the constructed search starting datum plane: if the Lidar point cloud data is completely located on one side of the search starting datum plane, the point with the shortest vertical distance to the datum plane is the initial ground point, as shown in Figure 3; if the Lidar point cloud data is located on both sides of the search starting datum plane, first distinguish the object side and the ground side of the search starting datum plane, and take the point on the ground side farthest from the datum plane as the initial ground point; in addition, at the slope change of the search starting datum plane, take the plumb line passing through the turning point as the search reference line 3, and find the lowest Lidar point cloud data close to the search reference line 3 as the initial ground point. If there are N slope changes in the same basic unit, select N+1 initial ground points. Add initial ground points according to the above rules, as shown in Figure 4. There are two ways to distinguish the object side and the ground side of the search starting reference plane. One is to define the side of the Lidar point cloud data that deviates farther from the search starting reference plane and has a larger number as the object side, and the other side as the ground side. The other is to distinguish based on the direction of the search reference plane, that is, the side facing the sky is the object side, and the side facing away from the sky is the ground side.

Specifically, in S400, a Lidar point cloud data classification reference plane is constructed according to the determined initial ground points, and a tin grid is formed by all the initial ground points. The tin grid is used as a Lidar point cloud data classification reference plane, and its accuracy is much higher than the search starting reference plane.

Specifically, in S400, iterative classification based on the Lidar point cloud data classification reference plane includes:

As shown in Figure 5, for each Lidar point cloud data, determine whether its vertical distance and angle from the Lidar point cloud data classification reference plane meet the judgment conditions. As shown in Figure 5, if the conditions are met, it is classified as a ground point, and the ground point is added to the Lidar point cloud data classification reference plane, that is, the Lidar point cloud data classification reference plane is updated, and all Lidar point cloud data are traversed to perform the above processing. While classifying, the Lidar point cloud data classification reference plane is iteratively updated until all point clouds are traversed and the classification is completed. The data processor can set the distance threshold and angle threshold as the judgment condition according to the actual situation of the study area.

After the above processing, all the Lidar point cloud ground points in the slope area are obtained. In addition, the Lidar point cloud data in the general area is obtained according to the conventional point cloud classification method. The Lidar point cloud ground points in the target area are extracted and made into various data products, which can be set in various formats such as *.csv, *.txt, *.dat, etc. according to actual needs.

As shown in FIG6 , an airborne Lidar point cloud ground point extraction device for fusing DOM images and three-dimensional real scene models in one embodiment of the present invention includes:

A data acquisition unit, which is used to acquire DOM images, three-dimensional real scene models and Lidar point cloud data of the target area, wherein the DOM images, three-dimensional real scene models and Lidar point cloud data have the same coordinate system and elevation system and have the same range;

A slope area determination unit, which is used to use the DOM image and the three-dimensional real scene model of the target area to search and determine the slope area in the target area, and extract and save the closed range line of the slope area;

A data cutting unit, which is used to cut and save the Lidar point cloud data of the slope area according to the closed range line of the slope area;

The ground point extraction unit is used to construct a search starting reference plane for each slope area, determine the initial ground points of the Lidar point cloud data based on the search starting reference plane, construct the Lidar point cloud data classification reference plane based on the determined initial ground points, and perform iterative classification based on the Lidar point cloud data classification reference plane to obtain all Lidar point cloud ground points in the slope area.

The present invention provides an electronic product in one embodiment, as shown in Figure 7, the electronic device includes at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned airborne Lidar point cloud ground point extraction method that integrates DOM images and three-dimensional real scene models.

Among them, the memory and the processor are connected in a bus manner, and the bus may include any number of interconnected buses and bridges, and the bus connects various circuits of one or more processors and memories together. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits through interfaces, which are all well known in the art. The interface provides an interface between the bus and the transceiver, such as a communication interface and a user interface. The transceiver can be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium. The data processed by the processor is transmitted on a wireless medium through an antenna, and further, the antenna also receives data and transmits the data to the processor.

The processor is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory can be used to store data used by the processor when performing operations.

In one embodiment, the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the method for extracting ground points from an airborne Lidar point cloud that integrates DOM images and three-dimensional real scene models is implemented.

Those skilled in the art can understand from the above description that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including several instructions for making a device (which can be a single-chip microcomputer, chip, etc.) or a processor (processor) execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes but is not limited to various media that can store program codes, such as USB flash drives, mobile hard disks, magnetic storage devices, and optical storage devices.

In the several embodiments provided in the present application, it should be understood that the disclosed system, device or method can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of modules/units is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules or units, which can be electrical, mechanical or other forms.

The modules/units described as separate components may or may not be physically separated, and the components displayed as modules/units may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules/units may be selected according to actual needs to achieve the purpose of the embodiments of the present application. For example, the functional modules/units in the various embodiments of the present application may be integrated into one processing module, or each module/unit may exist physically separately, or two or more modules/units may be integrated into one module/unit.

Those of ordinary skill in the art should further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

The descriptions of the processes or structures corresponding to the above-mentioned figures have different emphases. For parts that are not described in detail in a certain process or structure, please refer to the relevant descriptions of other processes or structures.

The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims (8)

1. A method for extracting ground points from airborne Lidar point clouds by integrating DOM images and three-dimensional real scene models, comprising: Acquire DOM images, three-dimensional real scene models and Lidar point cloud data of the target area, wherein the DOM images, three-dimensional real scene models and Lidar point cloud data have the same coordinate system and elevation system and have the same range; Using the DOM image and 3D real scene model of the target area, search and determine the slope area in the target area, extract and save the closed range line of the slope area; According to the closed range line of the slope area, the Lidar point cloud data of the slope area is cut and saved, and the Lidar point cloud data of the entire target area is divided into two parts: the Lidar point cloud data of the slope area and the Lidar point cloud data of the non-slope area; For each slope area, a search starting reference plane is constructed, and the initial ground point of the Lidar point cloud data of the slope area is determined based on the search starting reference plane. A classification reference plane of the Lidar point cloud data of the slope area is constructed according to the determined initial ground point. Iterative classification is performed based on the classification reference plane of the Lidar point cloud data of the slope area to obtain all the Lidar point cloud ground points of the slope area, wherein the search starting reference plane is a mathematical plane with the same slope as the three-dimensional real scene model; The constructing of the search starting reference plane comprises: First, the ground objects on the surface of the three-dimensional real scene model are removed by a filtering algorithm, and then a search starting reference plane of the slope area is constructed according to the slope data and position data of the slope area in the three-dimensional real scene model. The spatial position of the search starting reference plane is determined according to the concentrated area of the Lidar point cloud data elevation value of the slope area, and the spatial position is set at a position intersecting with the lowest area in the Lidar point cloud data elevation value concentration or close to the lower part of the lowest area in the Lidar point cloud data elevation value concentration; Determining the initial ground point of the Lidar point cloud data in the slope area based on the search starting reference plane includes: If the distance between the Lidar point cloud data in the slope area and the search start reference plane is within the preset flatness threshold range, the point with the lowest elevation value of the Lidar point cloud data in the slope area is the initial ground point; Otherwise, a search is performed in the slope area based on the position of the slope area Lidar point cloud data and the constructed search starting datum plane: if the slope area Lidar point cloud data is completely located on one side of the search starting datum plane, the initial ground point is the point with the shortest vertical distance to the search starting datum plane; if the slope area Lidar point cloud data is located on both sides of the search starting datum plane, the object side and the ground side of the search starting datum plane are first distinguished, and the point on the ground side that is farthest from the search starting datum plane is taken as the initial ground point; in addition, at the place where the slope of the search starting datum plane changes, the plumb line passing through the turning point is used as the search reference line, and the lowest slope area Lidar point cloud data close to the search reference line is found as the initial ground point; the side with the larger number of Lidar point cloud data deviating from the search starting datum plane is taken as the object side, and the other side is taken as the ground side; or according to the direction of the search starting datum plane, the side facing the sky is taken as the object side, and the side facing away from the sky is taken as the ground side. 2. The method for extracting ground points from airborne Lidar point clouds by integrating DOM images and three-dimensional real scene models as claimed in claim 1, wherein: The acquisition of DOM images is to obtain DOM image data with real geographic coordinates by processing the original photo data, POS data and image control point data acquired by the onboard camera; The acquisition of the 3D real scene model is to obtain a high-precision 3D real scene model with real geographic coordinates by processing the original photo data, POS data and image control point data acquired by the drone; The acquisition of Lidar point cloud data is obtained by obtaining the airborne Lidar point cloud measured data through trajectory solution, flight track stitching and coordinate conversion. 3. The method for extracting ground points from airborne Lidar point clouds by integrating DOM images and three-dimensional real scene models as claimed in claim 1, wherein: A search is performed based on the texture information of the DOM image of the target area and the surface undulation of the three-dimensional real-scene model. The closed range line of the slope area is extracted and saved according to the preset buffer radius. The buffer radius is the redundancy of the closed range line set to compensate for the omission of point cloud classification results caused by the error in determining the closed range line. 4. The method for extracting ground points from airborne Lidar point clouds by fusing DOM images and three-dimensional real scene models as claimed in claim 1, wherein a Lidar point cloud data classification reference surface is constructed based on the determined initial ground points, a tin grid is formed by all the initial ground points, and the tin grid is used as a Lidar point cloud data classification reference surface in a slope area. 5. The method for extracting ground points from airborne Lidar point clouds by integrating DOM images and three-dimensional real scene models as claimed in claim 1, wherein: Iterative classification based on the Lidar point cloud data classification benchmark in the slope area includes: For each Lidar point cloud data in the slope area, determine whether its vertical distance and angle from the classification reference plane of the Lidar point cloud data in the slope area meet the judgment conditions. If the conditions are met, classify it as a ground point, and add the ground point to the classification reference plane of the Lidar point cloud data in the slope area, that is, update the classification reference plane of the Lidar point cloud data in the slope area, traverse all the Lidar point cloud data in the slope area to perform the above processing, and iteratively update the classification reference plane of the Lidar point cloud data in the slope area while classifying until all point clouds are traversed and classification is completed; the judgment condition is within the range defined by the preset distance threshold and angle threshold. 6. An airborne Lidar point cloud ground point extraction device integrating DOM images and three-dimensional real scene models, comprising: A data acquisition unit, which is used to acquire DOM images, three-dimensional real scene models and Lidar point cloud data of the target area, wherein the DOM images, three-dimensional real scene models and Lidar point cloud data have the same coordinate system and elevation system and have the same range; A slope area determination unit, which is used to use the DOM image and the three-dimensional real scene model of the target area to search and determine the slope area in the target area, and extract and save the closed range line of the slope area; A data cutting unit, which is used to cut and save the Lidar point cloud data of the slope area according to the closed range line of the slope area, and divide the Lidar point cloud data of the entire target area into two parts: the Lidar point cloud data of the slope area and the Lidar point cloud data of the non-slope area; A ground point extraction unit is used to construct a search starting reference plane for each slope area, determine the initial ground points of the Lidar point cloud data of the slope area based on the search starting reference plane, construct a classification reference plane of the Lidar point cloud data of the slope area according to the determined initial ground points, perform iterative classification based on the classification reference plane of the Lidar point cloud data of the slope area, and obtain all the Lidar point cloud ground points of the slope area, wherein the search starting reference plane is a mathematical surface with the same slope as the three-dimensional real scene model; The constructing of the search starting reference plane comprises: First, the ground objects on the surface of the three-dimensional real scene model are removed by a filtering algorithm, and then a search starting reference plane of the slope area is constructed according to the slope data and position data of the slope area in the three-dimensional real scene model. The spatial position of the search starting reference plane is determined according to the concentrated area of the Lidar point cloud data elevation value of the slope area, and the spatial position is set at a position intersecting with the lowest area in the Lidar point cloud data elevation value concentration or close to the lower part of the lowest area in the Lidar point cloud data elevation value concentration; Determining the initial ground point of the Lidar point cloud data in the slope area based on the search starting reference plane includes: If the distance between the Lidar point cloud data in the slope area and the search reference surface is within the preset flatness threshold range, the point with the lowest elevation value of the Lidar point cloud data in the slope area is the initial ground point; Otherwise, a search is performed in the slope area based on the position of the slope area Lidar point cloud data and the constructed search starting datum plane: if the slope area Lidar point cloud data is completely located on one side of the search starting datum plane, the initial ground point is the point with the shortest vertical distance to the datum plane; if the slope area Lidar point cloud data is located on both sides of the search starting datum plane, the object side and the ground side of the search starting datum plane are first distinguished, and the point on the ground side that is farthest from the datum plane is used as the initial ground point; in addition, at the place where the slope of the search starting datum plane changes, the plumb line passing through the turning point is used as the search reference line, and the lowest slope area Lidar point cloud data close to the search reference line is found as the initial ground point; the side with the larger number of Lidar point cloud data deviating from the search starting datum plane is used as the object side, and the other side is used as the ground side; or according to the direction of the search datum plane, the side facing the sky is used as the object side, and the side facing away from the sky is used as the ground side. 7. An electronic product comprising: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the airborne Lidar point cloud ground point extraction method that integrates DOM images and three-dimensional real scene models as described in any one of claims 1 to 5. 8. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the method for extracting ground points from an airborne Lidar point cloud by fusing DOM images and a three-dimensional real scene model according to any one of claims 1 to 5 is implemented.