Photogrammetry three-dimensional modeling method and system

Abstract

The invention discloses a photogrammetry three-dimensional modeling method and a photogrammetry three-dimensional modeling system. The photogrammetry three-dimensional modeling method specifically comprises the following steps: S1, acquiring camera parameters; s2, target positioning; s3, analyzing the size of the target; s4, performing three-dimensional modeling; the invention relates to the technical field of photography three-dimensional modeling. According to the photogrammetry three-dimensional modeling method and the photogrammetry three-dimensional modeling system, the plane coordinates are established according to the image pixel points, the image grid coordinate diagrams are acquired, the two-dimensional plane coordinates of the same target in different image grid coordinate diagrams are acquired, and the three-dimensional point cloud of the target is generated in combination with an established camera imaging model; the three-dimensional model of the corresponding target can be conveniently and rapidly generated, the three-dimensional model is rapidly built, the data processing amount is small, the cost needed for building the three-dimensional model can be effectively reduced, and meanwhile the building efficiency of the three-dimensional model can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photographic three-dimensional modeling, and in particular to a photogrammetric three-dimensional modeling method and system. Background Art

[0002] With the popularization of digital cities, 3D modeling technology based on drone image data has been increasingly widely used. The technology mainly uses high-precision cameras, lidar and other sensors mounted on drones to collect high-precision data of ground objects, and then uses computers to process and analyze these data to generate high-precision 3D models. Although drone 3D modeling technology has broad application prospects, it is also easily affected by natural environment such as weather and light. In the face of cloudy, foggy and other weather conditions, traditional drone image 3D modeling technology cannot meet the application conditions. When faced with some situations that require all-weather data collection, the use of traditional modeling technology to collect daytime image data cannot meet actual needs.

[0003] A three-dimensional modeling method, device, system and storage medium as described in application number 202310768554.5, which obtains multiple images to be modeled from a camera and multiple original point cloud data from a lidar transmitter; screens out multiple processed point cloud data from all the original point cloud data; screens out multiple blurred images from all the images to be modeled; performs repair analysis on each blurred image to obtain a target repair image; and obtains a three-dimensional modeling result by constructing all processed point cloud data and all target repair images, fully considering the characteristics of the photogrammetry process, reducing the image quality degradation caused by night light and motion blur, improving the modeling accuracy, and reducing the three-dimensional model voids, thereby generating a more accurate and reliable night-time three-dimensional model.

[0004] Based on the retrieval of the above information, it can be seen that the existing three-dimensional model construction based on photographic images is greatly affected by environmental factors, and different algorithms need to be built for different environments, which makes the three-dimensional modeling process more complicated and requires high-performance hardware support. The calculation cost is high and the three-dimensional model generation efficiency is slow. For this reason, a photogrammetry three-dimensional modeling method and system are proposed. While not being affected by environmental factors, the three-dimensional model can be built through a simple algorithm. While the processing efficiency is high, the cost of building a three-dimensional model is effectively reduced. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a photogrammetry 3D modeling method and system, which solves the problems raised in the above background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a photogrammetry 3D modeling method, specifically comprising the following steps:

[0007] S1. Camera parameter acquisition: obtaining the camera's posture information when taking photos and the camera's position coordinates when taking videos, and storing the camera's photographed images;

[0008] S2, target positioning: establish plane coordinates with the pixel points of the image as coordinates, obtain the image grid coordinate map, screen the feature sub-characteristics with repeated characteristics, and use the feature descriptors to match the feature points, obtain the corresponding feature points of the same target in different images, and eliminate the mismatched points. According to the matching results, extract several two-dimensional plane coordinates of the same target from the image grid coordinate map;

[0009] S3. Target size analysis: A camera imaging model is established based on the camera's intrinsic and extrinsic parameters. The three-dimensional point cloud of the target is reconstructed using the image grid coordinate diagrams at different viewing angles and the camera's position coordinates. The size of the target is measured in three-dimensional space. For targets with irregular shapes, the size of the bounding box is calculated for approximate representation.

[0010] S4, 3D modeling: After denoising, filtering and smoothing the 3D point cloud, the 3D point cloud is converted into a 3D network model, and the texture information in the camera image is mapped to the 3D mesh model to obtain the 3D model of the corresponding image.

[0011] The present invention also discloses a photogrammetric 3D modeling system, comprising a camera 3D modeling platform, wherein the camera 3D modeling platform comprises a basic analysis unit, a target positioning unit, a dimension analysis unit and a 3D modeling unit, wherein the basic analysis unit is connected to the target positioning unit, the target positioning unit is connected to the dimension analysis unit, and the dimension analysis unit is connected to the 3D modeling unit.

[0012] The present invention is further configured as follows: the basic analysis unit includes a position coordinate recording module and a posture recording module, and the position coordinate recording module is connected to the posture recording module.

[0013] The present invention is further configured as follows: the position coordinate recording module records the position coordinates of the camera when the user takes pictures, and stores the camera photographed images;

[0014] The posture recording module is used to obtain the posture information of the camera when taking pictures, wherein the posture information includes a pitch angle, a yaw angle and a roll angle.

[0015] The present invention is further configured as follows: the target positioning unit includes a coordinate simulation module, a target setting module and a similar marking module, the coordinate simulation module is connected to the target setting module, and the target setting module is connected to the similar marking module.

[0016] The present invention is further configured as follows: the coordinate simulation module is used to establish plane coordinates using pixel points of the image as coordinates to obtain an image grid coordinate diagram;

[0017] The target preparatory module is used to screen feature sub-images with repeated characteristics, and use feature descriptors to match feature points, obtain the corresponding feature points of the same target in different images, and eliminate mismatched points;

[0018] The same type marking module is used to extract a number of two-dimensional plane coordinates of the same target from the image grid coordinate map according to the matching result.

[0019] The present invention is further configured as follows: the size analysis unit includes a camera imaging modeling module, a three-dimensional reconstruction module and a size calculation module, the camera imaging modeling module is connected to the three-dimensional reconstruction module, and the three-dimensional reconstruction module is connected to the size calculation module.

[0020] The present invention is further configured as follows: the camera imaging modeling module is used to establish a camera imaging model according to the intrinsic parameters and extrinsic parameters of the camera;

[0021] The three-dimensional reconstruction module is used to reconstruct the three-dimensional point cloud of the target using the image grid coordinate diagrams under different viewing angles and the position coordinates of the camera;

[0022] The size calculation module is used to measure the size of the target in three-dimensional space, wherein for targets with irregular shapes, the size of the bounding box is calculated for approximate representation.

[0023] The present invention is further configured as follows: the three-dimensional modeling unit includes a point cloud conversion module, an optimization processing module and a model verification module, the point cloud conversion module is connected to the optimization processing module, and the optimization processing module is connected to the model verification module.

[0024] The present invention is further configured as follows: the point cloud conversion module is used to convert the three-dimensional point cloud into a three-dimensional network model after performing denoising, filtering and smoothing on the three-dimensional point cloud;

[0025] The optimization processing module is used to map the texture information in the camera image to the three-dimensional mesh model to obtain the three-dimensional model of the corresponding image;

[0026] The model verification module is used to derive a plane image based on the three-dimensional model, obtain a comparison image, compare it with the camera image, and mark missing and extra parts.

[0027] The present invention provides a photogrammetry 3D modeling method and system. It has the following beneficial effects:

[0028] (1) The present invention establishes plane coordinates according to image pixel points to obtain an image grid coordinate map. By obtaining the two-dimensional plane coordinates of the same target in different image grid coordinate maps, the three-dimensional point cloud of the target is generated in combination with the established camera imaging model, so as to facilitate and quickly generate a three-dimensional model of the corresponding target. While the three-dimensional model is quickly built, the amount of data processing is small, which can effectively reduce the cost required for building the three-dimensional model and improve the efficiency of building the three-dimensional model.

[0029] (2) The present invention exports a plane image based on the constructed three-dimensional model, and by comparing it with the reference of the initial camera image, marks the missing and excess parts of the plane image relative to the initial camera image, thereby realizing the accuracy evaluation of the constructed three-dimensional model. In this way, while adapting to different photographic environments, it provides accurate data support for the optimization of the three-dimensional model. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a system principle block diagram of the present invention;

[0031] Figure 2 It is a system principle block diagram of the basic analysis unit of the present invention;

[0032] Figure 3 It is a system principle block diagram of the target positioning unit of the present invention;

[0033] Figure 4 It is a system principle block diagram of the dimension analysis unit of the present invention;

[0034] Figure 5 This is a system principle block diagram of the three-dimensional modeling unit of the present invention.

[0035] In the figure:

[0036] 1. Camera 3D modeling platform;

[0037] 2. Basic analysis unit; 201. Position coordinate recording module; 202. Posture recording module;

[0038] 3. Target positioning unit; 301. Coordinate simulation module; 302. Target setting module; 303. Similar marking module;

[0039] 4. Dimension analysis unit; 401. Camera imaging modeling module; 402. Three-dimensional reconstruction module; 403. Dimension calculation module;

[0040] 5. Three-dimensional modeling unit; 501. Point cloud conversion module; 502. Optimization processing module; 503. Model verification module. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0042] See also Figure 1-5 , the embodiment of the present invention provides the following technical solutions:

[0043] Embodiment 1

[0044] A photogrammetric 3D modeling system includes a camera 3D modeling platform 1, which includes a basic analysis unit 2, a target positioning unit 3, a size analysis unit 4 and a 3D modeling unit 5.

[0045] As a preferred solution, the position and direction information of the camera, i.e., the photographic coordinates, is recorded when different images are collected to provide data support for subsequent three-dimensional modeling and spatial position analysis. The basic analysis unit 2 includes a position coordinate recording module 201 and a posture recording module 202. The posture recording module 202 is used to obtain the posture information of the camera when photographing, wherein the posture information includes the pitch angle, the yaw angle, and the roll angle;

[0046] The position coordinate recording module 201 is connected to the posture recording module 202. The position coordinate recording module 201 records the position coordinates of the camera when shooting, and stores the camera shooting images.

[0047] As a preferred solution, in order to extract the two-dimensional plane coordinates of the same target from different images, the basic analysis unit 2 is connected to the target positioning unit 3, and the target positioning unit 3 includes a coordinate simulation module 301, a target setting module 302 and a similar marking module 303. The coordinate simulation module 301 is used to establish plane coordinates with the pixel points of the image as coordinates to obtain the image grid coordinate map;

[0048] The coordinate simulation module 301 is connected to the target formulation module 302. The target formulation module 302 is used to screen feature sub-elements with repeated characteristics, and use feature descriptors to match feature points, obtain the corresponding feature points of the same target in different images, and eliminate mismatched points;

[0049] The target formulation module 302 is connected to the same type marking module 303, and the same type marking module 303 is used to extract a number of two-dimensional plane coordinates of the same target from the image grid coordinate map according to the matching result.

[0050] As a preferred solution, in order to realize the calculation of the actual size of the target, the target positioning unit 3 is connected to the size analysis unit 4, and the size analysis unit 4 includes a camera imaging modeling module 401, a three-dimensional reconstruction module 402 and a size calculation module 403. The camera imaging modeling module 401 is used to establish a camera imaging model according to the intrinsic parameters and extrinsic parameters of the camera;

[0051] The camera imaging modeling module 401 is connected to the 3D reconstruction module 402, and the 3D reconstruction module 402 is used to reconstruct the 3D point cloud of the target using the image grid coordinate diagrams under different viewing angles and the position coordinates of the camera;

[0052] The three-dimensional reconstruction module 402 is connected to the size calculation module 403. The size calculation module 403 is used to measure the size of the target in the three-dimensional space. For targets with irregular shapes, the size of the bounding box is calculated for approximate representation.

[0053] As a preferred solution, the dimension analysis unit 4 is connected to the 3D modeling unit 5, and the 3D modeling unit 5 includes a point cloud conversion module 501 and an optimization processing module 502. The point cloud conversion module 501 is used to perform denoising, filtering and smoothing on the 3D point cloud, and then convert the 3D point cloud into a 3D network model using Poisson reconstruction or Delaunay triangulation method;

[0054] The point cloud conversion module 501 is connected to the optimization processing module 502. The optimization processing module 502 is used to map the texture information in the camera image to the three-dimensional mesh model to obtain the three-dimensional model of the corresponding image.

[0055] In this embodiment, plane coordinates are established according to image pixels to obtain an image grid coordinate map. The two-dimensional plane coordinates of the same target in different image grid coordinate maps are obtained, and combined with the constructed camera imaging model, a three-dimensional point cloud of the target is generated, so that a three-dimensional model of the corresponding target is generated quickly and conveniently. The three-dimensional model is constructed quickly with a small amount of data processing, which can effectively reduce the cost required for constructing the three-dimensional model and improve the efficiency of constructing the three-dimensional model.

[0056] Embodiment 2

[0057] As an improvement of the previous embodiment, this embodiment is a photogrammetry 3D modeling system, which also includes a model verification module 503. The optimization processing module 502 is connected to the model verification module 503. The model verification module 503 is used to export a plane image according to the 3D model, obtain a comparison image, compare it with the camera image, and mark the missing and excess parts.

[0058] As a detailed explanation, based on the parameter information of the camera when the camera image is acquired, the perspective when exporting the plane image of the three-dimensional model can be determined, ensuring the accurate matching between the exported plane image and the initial camera image, so as to improve the accuracy of the reference comparison.

[0059] The advantage of the second embodiment over the first embodiment is that the missing and excess parts of the plane image relative to the initial camera image are marked to achieve accuracy evaluation of the constructed three-dimensional model, thereby providing accurate data support for the optimization of the three-dimensional model while adapting to different photographic environments.

[0060] A photogrammetric three-dimensional modeling method specifically comprises the following steps:

[0061] S1, camera parameter acquisition: the posture recording module 202 acquires the posture information of the camera when shooting, wherein the posture information includes the pitch angle, the yaw angle and the roll angle, and the position coordinate recording module 201 is used to acquire the position coordinate of the camera when shooting, and the camera shooting image is stored;

[0062] S2, target positioning: the coordinate simulation module 301 uses the pixel points of the image as coordinates to establish plane coordinates, obtains the image grid coordinate map, selects feature sub-points with repeated characteristics through the target formulation module 302, and uses feature descriptors to match feature points, obtains the corresponding feature points of the same target in different images, and removes mismatched points. According to the matching results, the same type marking module 303 extracts several two-dimensional plane coordinates of the same target from the image grid coordinate map;

[0063] S3, target size analysis: the camera imaging modeling module 401 establishes a camera imaging model according to the intrinsic and extrinsic parameters of the camera, the 3D reconstruction module 402 reconstructs the 3D point cloud of the target using the image grid coordinate diagrams under different viewing angles and the position coordinates of the camera, and the size calculation module 403 measures the size of the target in the 3D space, wherein for irregularly shaped targets, the size of the bounding box is calculated for approximate representation;

[0064] S4. Three-dimensional modeling: After the point cloud conversion module 501 performs denoising, filtering and smoothing on the three-dimensional point cloud, it converts the three-dimensional point cloud into a three-dimensional network model using Poisson reconstruction or Delaunay triangulation method. The optimization processing module 502 maps the texture information in the camera image to the three-dimensional mesh model to obtain the three-dimensional model of the corresponding image. The model verification module 503 described later exports the plane image according to the three-dimensional model, obtains the comparison image, compares it with the camera image, and marks the missing and excess parts.

[0065] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0066] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A photogrammetric three-dimensional modeling method, characterized in that: The specific steps include: S1. Camera parameter acquisition: obtaining the camera's posture information when taking photos and the camera's position coordinates when taking videos, and storing the camera's photographed images; S2, target positioning: establish plane coordinates with the pixel points of the image as coordinates, obtain the image grid coordinate map, screen the feature sub-characteristics with repeated characteristics, and use the feature descriptors to match the feature points, obtain the corresponding feature points of the same target in different images, and eliminate the mismatched points. According to the matching results, extract several two-dimensional plane coordinates of the same target from the image grid coordinate map; S3. Target size analysis: A camera imaging model is established based on the camera's intrinsic and extrinsic parameters. The three-dimensional point cloud of the target is reconstructed using the image grid coordinate diagrams at different viewing angles and the camera's position coordinates. The size of the target is measured in three-dimensional space. For targets with irregular shapes, the size of the bounding box is calculated for approximate representation. S4, 3D modeling: After denoising, filtering and smoothing the 3D point cloud, the 3D point cloud is converted into a 3D network model, and the texture information in the camera image is mapped to the 3D mesh model to obtain the 3D model of the corresponding image.

2. A photogrammetric 3D modeling system, characterized in that: The invention comprises a video-based three-dimensional modeling platform (1), wherein the video-based three-dimensional modeling platform (1) comprises a basic analysis unit (2), a target positioning unit (3), a dimension analysis unit (4) and a three-dimensional modeling unit (5), wherein the basic analysis unit (2) is connected to the target positioning unit (3), the target positioning unit (3) is connected to the dimension analysis unit (4), and the dimension analysis unit (4) is connected to the three-dimensional modeling unit (5).

3. A photogrammetric 3D modeling system according to claim 2, characterized in that: The basic analysis unit (2) comprises a position coordinate recording module (201) and a posture recording module (202), and the position coordinate recording module (201) is connected to the posture recording module (202).

4. A photogrammetric 3D modeling system according to claim 3, characterized in that: The position coordinate recording module (201) is used by the user to record the position coordinates of the camera when shooting, and stores the camera photographed images; The posture recording module (202) is used to obtain posture information of the camera when taking photos, wherein the posture information includes a pitch angle, a yaw angle and a roll angle.

5. A photogrammetric 3D modeling system according to claim 4, characterized in that: The target positioning unit (3) comprises a coordinate simulation module (301), a target setting module (302) and a similar marking module (303); the coordinate simulation module (301) is connected to the target setting module (302), and the target setting module (302) is connected to the similar marking module (303).

6. A photogrammetric 3D modeling system according to claim 5, characterized in that: The coordinate simulation module (301) is used to establish plane coordinates using the pixel points of the image as coordinates to obtain an image grid coordinate diagram; The target setting module (302) is used to screen feature sub-elements with repeated characteristics, and use feature descriptors to match feature points, obtain feature points of the same target corresponding to different images, and eliminate mismatched points; The same type marking module (303) is used to extract a number of two-dimensional plane coordinates of the same target from the image grid coordinate diagram according to the matching result.

7. A photogrammetric 3D modeling system according to claim 6, characterized in that: The size analysis unit (4) comprises a camera imaging modeling module (401), a three-dimensional reconstruction module (402) and a size calculation module (403); the camera imaging modeling module (401) is connected to the three-dimensional reconstruction module (402), and the three-dimensional reconstruction module (402) is connected to the size calculation module (403).

8. The photogrammetric 3D modeling system according to claim 7, characterized in that: The camera imaging modeling module (401) is used to establish a camera imaging model according to the camera's internal and external parameters; The three-dimensional reconstruction module (402) is used to reconstruct a three-dimensional point cloud of the target using the image grid coordinate diagrams at different viewing angles and the position coordinates of the camera; The size calculation module (403) is used to measure the size of the target in three-dimensional space, wherein for targets with irregular shapes, the size of the bounding box is calculated for approximate representation.

9. A photogrammetric 3D modeling system according to claim 8, characterized in that: The three-dimensional modeling unit (5) comprises a point cloud conversion module (501), an optimization processing module (502) and a model verification module (503), wherein the point cloud conversion module (501) is connected to the optimization processing module (502), and the optimization processing module (502) is connected to the model verification module (503).

10. The photogrammetric 3D modeling system according to claim 9, characterized in that: The point cloud conversion module (501) is used to perform denoising, filtering and smoothing on the three-dimensional point cloud, and then convert the three-dimensional point cloud into a three-dimensional network model; The optimization processing module (502) is used to map the texture information in the camera image to the three-dimensional mesh model to obtain the three-dimensional model of the corresponding image; The model verification module (503) is used to derive a plane image based on the three-dimensional model, obtain a comparison image, compare it with the camera image, and mark missing and extra parts.

Images