CN110866945A - Method for generating three-dimensional tree by automatic identification of oblique photography model - Google Patents
Method for generating three-dimensional tree by automatic identification of oblique photography model Download PDFInfo
- Publication number
- CN110866945A CN110866945A CN201911105576.3A CN201911105576A CN110866945A CN 110866945 A CN110866945 A CN 110866945A CN 201911105576 A CN201911105576 A CN 201911105576A CN 110866945 A CN110866945 A CN 110866945A
- Authority
- CN
- China
- Prior art keywords
- tree
- dimensional
- model
- oblique photography
- generating
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
- G06T7/55—Depth or shape recovery from multiple images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention provides a method for generating a three-dimensional tree by automatically identifying an oblique photography model, wherein input data in the method comprise an oblique photography photo, a space-three result, point cloud data and oblique photography result data, output data comprise an xml file storing position and posture information of all simplex tree models and a model of each tree, and processing steps comprise extraction and identification of the tree, space matching and parameterization generation. The invention can automatically generate the target three-dimensional tree model based on the prior oblique photography data result of the user without any manual intervention, the result effect of the XML + OSGB output after automatic processing can be completely comparable with the effect of games and even CG animation, and the real target data can be obtained, so that the subsequent application becomes possible, including the management, inquiry, analysis, editing and the like of trees.
Description
Technical Field
The invention relates to a method for generating a three-dimensional tree by automatically identifying an oblique photography model.
Background
At present, automatic object processing cannot be realized on a tree automatically generated by oblique photography, most oblique photography modeling methods are based on aerial photographs, the tree model is incomplete due to the fact that objects in the photographs are shielded, for the tree, the photographs are difficult to obtain the complete details of the tree, information including tree branches and the like is lost, the existing oblique photography model automatically generated based on a photograph reconstruction method is an integral surface model (commonly called a layer of bark), generally, a triangular mesh with textures is used, and such data results can only be used for viewing and cannot be applied, such as query, statistics, editing and the like.
At present, aiming at a building model, semi-automatic extraction and objectification can be carried out through some software and tools, but aiming at vegetation, particularly trees, the objectification is not carried out in an automatic processing mode, only manual cutting can be carried out, and the efficiency is low and the cost is high.
The tree model automatically generated by oblique photography has very poor effect, generally, one tree is a hemispherical epidermis model, the effect is completely unacceptable once the tree is viewed from a high altitude of hundreds of meters, and if the tree is observed from a viewpoint by people on the ground, the conditions of serious influence on the effect, such as deformation, leak, no texture and the like, can be found, so that the tree model generated by oblique photography can be completely deleted in many practical projects.
Disclosure of Invention
The invention provides a method for generating a three-dimensional tree by automatically identifying a tilted photography model, which solves the problem of how to realize a complete tree model for the automatically generated tree by tilted photography, and the technical scheme is as follows:
a method for generating a three-dimensional tree by automatic identification of a tilted photography model comprises the following steps:
s1: acquiring aerial triangulation results, point cloud data and three-dimensional model data in an OSGB format according to the picture content and the picture external parameters of oblique photography;
s2: extracting tree objects of trees in the oblique photography picture by using the point cloud data;
s3: finding out a photo corresponding to the tree object and generating a corresponding sub-image;
s4: performing tree species identification on all the subimages;
s5: completing space matching, and determining to obtain specific data of each tree by using three-dimensional model data;
s6: and placing the tree species model in the step S4 at the spatial position corresponding to the tree object.
Further, in step S2, the extraction of the tree object includes two steps: 1) performing feature extraction and feature screening on input point cloud data, and then performing point cloud classification, thereby screening out point cloud data related to a tree and removing the point cloud data unrelated to the tree; 2) and performing two-dimensional projection on the three-dimensional point cloud data, reducing the dimension to a two-dimensional plane space, and performing image processing in a rasterization mode to obtain the point cloud data of the tree object.
Further, in step 2), the image processing includes pixel segmentation, shape analysis and instantiation.
Further, in step S3, the aerial triangulation result is used to filter all the photos once, and the photos that reflect the most tree features are screened out as the photos corresponding to the tree object, and the pixel range of the tree object on the photos is inversely calculated according to the bounding box of the point cloud and the internal and external parameters of the camera, so as to generate the sub-images of the tree object.
Further, in step S4, each tree species provides not less than 30 training samples, and the sub-images are detected to identify the tree type and image area in each sub-image.
Further, in step S5, the spatial matching includes the following steps:
1) calculating the three-dimensional space range of each tree according to the image space range of the tree species identified in the sub-images and the internal reference and the external reference of the corresponding photo;
2) obtaining a corresponding triangular mesh model in a three-dimensional space by using the image space range and the three-dimensional model data of oblique photography;
3) and calculating the position and the size of the tree according to the grid model of the tree in the three-dimensional space, and finally obtaining the position and the height of each tree and the radius of the crown projected on the two-dimensional plane.
In the step 2), a bounding box is calculated according to the range of each tree by the triangular mesh model, and the area of the bounding box projected to the near cutting surface of the viewing cone is calculated according to the position relation between the viewing cone and the bounding box.
Further, in step S6, the tree species model automatically generates an instance object of a tree by using the objectification result and the parameter-driven tree growing system according to the specific data of each tree.
Further, in step S6, when the instance object of the tree is implemented, the operations adopted include: 1) matching in a rotary zooming mode; 2) generation is driven parametrically by the plant growth system.
Further, in step S6, the tree is placed at the corresponding spatial position, and the XML file and the osgb file of the tree model are output.
The method for automatically identifying and generating the three-dimensional tree by the oblique photography model can automatically generate the objectified three-dimensional tree model based on the existing oblique photography data result of a user without any manual intervention. The result effect of the XML + OSGB output after the automatic processing by the method of the invention can completely match with the effect of games and even CG animation. The method can obtain real objectified data after automatic processing, so that subsequent application becomes possible, including management, query, analysis, editing and the like of the tree.
Drawings
FIG. 1 is a schematic flow diagram of a method for generating a three-dimensional tree for automatic identification of the oblique photography model;
FIG. 2 is a schematic flow chart of screening input point cloud data for point clouds associated with trees;
FIG. 3 is a schematic flow chart of processing three-dimensional point cloud data to obtain point cloud data of a targeted tree;
FIG. 4 is a schematic illustration of the generation of a sub-image on a photograph;
FIG. 5 is a schematic illustration of the determination of the location and size of a tree from a mesh model of the tree in three-dimensional space;
FIG. 6 is a schematic diagram of a mesh model of a tree in three-dimensional space projected onto a two-dimensional plane;
fig. 7 is a diagram illustrating an effect of opening one of the tree osgb files by the osdviewer;
FIG. 8 is a schematic diagram of the effect of displaying the data result automatically generated by the method.
Detailed Description
The invention provides a method for automatically identifying and generating a three-dimensional tree by a tilted photography model, which adopts a brand-new processing flow, can fully automatically identify, extract and generate an objectified three-dimensional tree model based on a tilted photography photo, an aerial triangulation result and final result three-dimensional model data (a data result generated by tilted photography processing, namely OSGB model data), and meets the requirement of later-stage application.
The processing flow of the method is shown in fig. 1, and the input data includes the following four types:
1) the picture format can adopt JPG format or tiff format; the photo resolution does not exceed 20000 x 20000.
2) The aerial triangulation results, derived from the output results of the oblique photography process, serve to calibrate the external parameters of all the photographs, i.e., the camera position and pose at the time each photograph was taken.
3) LAS point cloud data, LAS refers to point cloud data in the LAS file format, which typically includes location and color information of the point cloud. All references to point cloud data in a document refer to this data type and are consistent.
4) The OSGB model, i.e. the final resultant three-dimensional model data, is generally in the OSGB file format, which can be exported by the oblique photography software currently in the market. The format in the invention is also the osgb file with the pyramid structure which is common to oblique photography.
After input data pass through the processing module, the output data are xml files storing xml position posture information of all the simplex tree models; and an OSGB model for each tree, the OSGB model being in an OSGB file format, the map also being contained inside the OSGB file.
The black box between the input data and the output data is the core processing module of the invention, and the specific operation flow of the processing module is as follows:
s1: tree extraction and identification
1) The method comprises the following steps of (1) extracting a tree object by utilizing dense point cloud data output by oblique photography, wherein the specific process comprises the following two steps:
a, point cloud classification: as shown in fig. 1, feature extraction and feature screening are performed on input point cloud data, and then point cloud classification is performed. Therefore, point cloud related to the tree is screened out, and point cloud data unrelated to the tree is eliminated. The above steps are common processing modes in the art, and the present invention is not described in detail.
B, objectification: as shown in fig. 2, after the three-dimensional point cloud data is two-dimensionally projected (that is, the third component z in the (x, y, z) coordinate of the three-dimensional point is removed to become the (x, y) plane coordinate), and reduced to the two-dimensional plane space, image processing can be performed in a rasterization manner, including pixel segmentation, shape analysis and final instantiation, so as to obtain the point cloud data of the objectified tree. The pixel segmentation, the shape analysis and the instantiation of the two-dimensional image space are common processing modes, and the description of the invention is omitted.
2) Finding out a photo corresponding to the tree object and generating a sub-image;
extracting a tree object through point cloud data;
and filtering the photos once by using the aerial triangulation result to screen out the photos which have great contribution to the tree, namely the photos which can embody the most characteristics of the tree object.
As shown in fig. 4, the pixel range on the photo is back calculated according to the bounding box of the point cloud and the intrinsic parameters (abbreviated as "intrinsic parameters") and extrinsic parameters (abbreviated as "extrinsic parameters") of the camera, and a sub-image is generated. The sub-image enables the subsequent identification method to be more targeted, and the efficiency and the identification accuracy can be greatly improved.
The camera intrinsic parameters are parameters related to the characteristics of the camera itself, such as the focal length, pixel size, etc. of the camera; the camera-out parameters are parameters in a world coordinate system, such as the position, rotation direction, etc. of the camera.
3) And performing tree species identification on all the sub-images by using a deep learning method.
The original oblique photographs are first used to train the trees that may be involved in the scene in which the photographs were taken, typically 30 training samples per tree. And detecting all the sub-images screened in the last step by a machine learning method, identifying the tree type and more accurate image area in each sub-image, and adopting a Faster R-CNN method for improving the performance.
The Faster R-CNN method is a deep learning algorithm which is commonly used in the industry, and is an efficient implementation of the R-CNN algorithm. For a picture, about 2000 candidate regions are generated by the R-CNN based on a selective search method, then each candidate region is reset to be in a fixed size and sent into a CNN model, and finally a feature vector is obtained. This feature vector is then fed into a multi-class SVM classifier to predict the probability values of objects contained in the candidate region belonging to each class. Each class trains an SVM classifier, and the probability of the class is deduced from the feature vector. In order to improve the positioning accuracy, the R-CNN trains a boundary frame regression model finally, and the accurate position of the frame is corrected through the boundary frame regression model.
S2: spatial matching
1) The three-dimensional spatial extent of each tree, typically a frustum of a tree, is calculated from the image space extent of the tree species identified in the previous sub-image, and the internal and external references of the corresponding photograph.
2) The corresponding triangular mesh model in the three-dimensional space is obtained by utilizing the image space range and the three-dimensional model data (OSGB model data) of oblique photography, only the geometric information of the model is needed here, and no texture information is needed.
As shown in fig. 5, one bounding box is calculated from the range of each tree, and the area of the bounding box projected onto the near-clipping plane of the viewing pyramid is calculated from the positional relationship between the viewing pyramid and the bounding box.
3) Completing the objectification, namely calculating the position and size of the tree according to the grid model of the tree in the three-dimensional space, and finally obtaining the position (x, y, z) of each tree, the height H of the tree, and the radius R of the crown projected on the two-dimensional plane, as shown in fig. 6.
S3: parametric generation
1) First, a proper tree model is automatically matched from a plant system library according to tree species information.
And selecting a proper tree species model base according to the area related to the specific project, covering all the tree species which are trained in the early stage, and directly matching in the model base after automatic identification.
2) And automatically generating an example object of the tree by using the information of the objectification, such as the height, the horizontal radius and the like of the tree through a parameter-driven tree growing system, and directly scaling the matrix if the application requirement is not high. Wherein the tree growing system may automatically generate an instance of a tree in a parameter-driven manner, the parameters comprising: tree type, tree height, horizontal radius of crown.
The following are two specific descriptions of the instantiation of the tree:
a, directly matching by adopting a rotation scaling mode, firstly carrying out a standardization operation on trees in a model library to unify the trees into a standard scale, such as 1M, and then respectively calculating the translation rotation and the scaling of each example, wherein the specific calculation mode is as follows:
the (x, y, z) in the translation matrix is the position of the tree computed in the previous spatial matching process.
X and y in the scaling matrix are the radius R of the projection of the crown to the two-dimensional plane in the previous spatial matching, and z is the height H of the tree obtained by spatial matching.
Rotation (rotate);
the axis of rotation is the Z-axis, which refers to the vector (0,0,1), and the angle of rotation is a random value, ensuring that the tree orientation is random for each instance.
B is generated by parameterization drive of a plant growth system, the generation can be realized by means of speedTree software, and trees generated by parameter automatic drive do not need rotation and scaling transformation, and only translation matrix transformation is needed.
3) And placing the tree at a corresponding spatial position by using the objectification result, and outputting an XML file and an osgb file of the tree model.
The final exported XML file contains the following information:
ModelName: exported osgb model name
Location X/location Y/location Z: location information of model
Matrix3 rotational scaling Matrix of model is a 3 x3 Float Matrix
The following is an example:
fig. 7 is an effect of opening one of the tree species osgb files with the osdviewer.
FIG. 8 is a schematic diagram of the effect of displaying the data result automatically generated by the method.
The method has the greatest advantage that the target three-dimensional tree model is generated based on the existing oblique photography data result of the user in a full-automatic mode without any manual intervention. At present, hundreds of cities in China acquire and generate oblique photography data through unmanned planes or manned planes, although the generation method of oblique photography is automatic, in order to achieve the purpose, a large amount of manual processing is required to be used for objectification and model correction, the time and the cost are very high, a full-automatic objectification and optimization method is urgently needed, and the problem in the industry is solved.
As mentioned above, the effect of vegetation, especially trees, in the oblique photography data generated by photographs is very poor, and in many cases, the vegetation is directly abandoned, and the effect of the XML + OSGB result output after the automatic processing by the method of the present invention can completely match with the effect of games and even CG animations.
For the problem that the model generated by oblique photography is a layer of skin and the structure is seriously lacked, the method can automatically process to obtain truly targeted data, so that the subsequent application becomes possible, including the management, query, analysis, editing and the like of trees.
Claims (10)
1. A method for generating a three-dimensional tree by automatic identification of a tilted photography model comprises the following steps:
s1: acquiring aerial triangulation results, point cloud data and three-dimensional model data in an OSGB format according to the picture content and the picture external parameters of oblique photography;
s2: extracting tree objects of trees in the oblique photography picture by using the point cloud data;
s3: finding out a photo corresponding to the tree object and generating a corresponding sub-image;
s4: performing tree species identification on all the subimages;
s5: completing space matching, and determining to obtain specific data of each tree by using three-dimensional model data;
s6: and placing the tree species model in the step S4 at the spatial position corresponding to the tree object.
2. The method for generating three-dimensional trees with automatic identification of oblique photography model according to claim 1, characterized in that: in step S2, the extraction of the tree object includes two steps: 1) performing feature extraction and feature screening on input point cloud data, and then performing point cloud classification, thereby screening out point cloud data related to a tree and removing the point cloud data unrelated to the tree; 2) and performing two-dimensional projection on the three-dimensional point cloud data, reducing the dimension to a two-dimensional plane space, and performing image processing in a rasterization mode to obtain the point cloud data of the tree object.
3. The method for generating three-dimensional trees with automatic identification of oblique photography model according to claim 2, characterized in that: in step 2), image processing includes pixel segmentation, shape analysis, and instantiation.
4. The method for generating three-dimensional trees with automatic identification of oblique photography model according to claim 1, characterized in that: in step S3, the aerial triangulation result is used to filter all the photos once, and the photos that show the most tree features are screened out as the photos corresponding to the tree object, and the pixel range of the tree object on the photos is inversely calculated according to the bounding box of the point cloud and the internal and external parameters of the camera, so as to generate the sub-images of the tree object.
5. The method for generating three-dimensional trees with automatic identification of oblique photography model according to claim 1, characterized in that: in step S4, each tree species provides not less than 30 training samples, and the sub-images are detected to identify the tree type and image area in each sub-image.
6. The method for generating three-dimensional trees with automatic identification of oblique photography model according to claim 1, characterized in that: in step S5, the spatial matching includes the steps of:
1) calculating the three-dimensional space range of each tree according to the image space range of the tree species identified in the sub-images and the internal reference and the external reference of the corresponding photo;
2) obtaining a corresponding triangular mesh model in a three-dimensional space by using the image space range and the three-dimensional model data of oblique photography;
3) and calculating the position and the size of the tree according to the grid model of the tree in the three-dimensional space, and finally obtaining the position and the height of each tree and the radius of the crown projected on the two-dimensional plane.
7. The method for generating three-dimensional trees with automatic identification of oblique photography model of claim 6, characterized in that: in the step 2), the triangular mesh model calculates a bounding box according to the range of each tree, and calculates the area of the bounding box projected to the near cutting surface of the viewing cone according to the position relation between the viewing cone and the bounding box.
8. The method for generating three-dimensional trees with automatic identification of oblique photography model according to claim 1, characterized in that: in step S6, the tree species model automatically generates an instance object of a tree by using the objectification result and the parameter-driven tree growing system according to the specific data of each tree.
9. The method for generating three-dimensional trees with automatic identification of oblique photography model of claim 8, characterized in that: in step S6, when the instance object of the tree is implemented, the operations adopted include: 1) matching in a rotary zooming mode; 2) generation is driven parametrically by the plant growth system.
10. The method for generating three-dimensional trees with automatic identification of oblique photography model according to claim 1, characterized in that: in step S6, the tree is placed at the corresponding spatial position, and the XML file and the osgb file of the tree model are output.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911105576.3A CN110866945A (en) | 2019-11-13 | 2019-11-13 | Method for generating three-dimensional tree by automatic identification of oblique photography model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911105576.3A CN110866945A (en) | 2019-11-13 | 2019-11-13 | Method for generating three-dimensional tree by automatic identification of oblique photography model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110866945A true CN110866945A (en) | 2020-03-06 |
Family
ID=69654095
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911105576.3A Pending CN110866945A (en) | 2019-11-13 | 2019-11-13 | Method for generating three-dimensional tree by automatic identification of oblique photography model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110866945A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111783552A (en) * | 2020-06-09 | 2020-10-16 | 当家移动绿色互联网技术集团有限公司 | Live-action three-dimensional model monomer method and device, storage medium and electronic equipment |
| CN114186603A (en) * | 2021-09-28 | 2022-03-15 | 北京博能科技股份有限公司 | Multi-source model single fusion method and system, storage medium and terminal |
| CN115861549A (en) * | 2023-02-27 | 2023-03-28 | 航天宏图信息技术股份有限公司 | Method and device for automatically replacing tree model in live-action three-dimensional data |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104077806A (en) * | 2014-07-10 | 2014-10-01 | 天津中科遥感信息技术有限公司 | Automatic separate extraction method based on city building three-dimensional model |
| CN104361628A (en) * | 2014-11-27 | 2015-02-18 | 南宁市界围工程咨询有限公司 | Three-dimensional real scene modeling system based on aviation oblique photograph measurement |
| KR20160060403A (en) * | 2014-11-20 | 2016-05-30 | 삼성전자주식회사 | Method and apparatus for calibrating image |
| CN109711410A (en) * | 2018-11-20 | 2019-05-03 | 北方工业大学 | Three-dimensional object rapid segmentation and identification method, device and system |
| CN110136259A (en) * | 2019-05-24 | 2019-08-16 | 唐山工业职业技术学院 | A kind of dimensional Modeling Technology based on oblique photograph auxiliary BIM and GIS |
| CN110310355A (en) * | 2019-06-21 | 2019-10-08 | 永州电力勘测设计院有限公司 | Oblique photograph model monomerization approach based on multitexture mapping |
| JP2019185449A (en) * | 2018-04-11 | 2019-10-24 | 国立大学法人信州大学 | Forest resource information calculation method and forest resource information calculation device |
-
2019
- 2019-11-13 CN CN201911105576.3A patent/CN110866945A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104077806A (en) * | 2014-07-10 | 2014-10-01 | 天津中科遥感信息技术有限公司 | Automatic separate extraction method based on city building three-dimensional model |
| KR20160060403A (en) * | 2014-11-20 | 2016-05-30 | 삼성전자주식회사 | Method and apparatus for calibrating image |
| CN104361628A (en) * | 2014-11-27 | 2015-02-18 | 南宁市界围工程咨询有限公司 | Three-dimensional real scene modeling system based on aviation oblique photograph measurement |
| JP2019185449A (en) * | 2018-04-11 | 2019-10-24 | 国立大学法人信州大学 | Forest resource information calculation method and forest resource information calculation device |
| CN109711410A (en) * | 2018-11-20 | 2019-05-03 | 北方工业大学 | Three-dimensional object rapid segmentation and identification method, device and system |
| CN110136259A (en) * | 2019-05-24 | 2019-08-16 | 唐山工业职业技术学院 | A kind of dimensional Modeling Technology based on oblique photograph auxiliary BIM and GIS |
| CN110310355A (en) * | 2019-06-21 | 2019-10-08 | 永州电力勘测设计院有限公司 | Oblique photograph model monomerization approach based on multitexture mapping |
Non-Patent Citations (3)
| Title |
|---|
| 杜靖: "基于倾斜摄影影像的真三维单体化模型精细建模方法研究", 《中国优秀博硕士学位论文全文数据库(硕士)基础科学辑》 * |
| 林磊等: "《AutoCAD 2004、3ds max 6、PHOTOSHOP CS建筑效果图制作教程 建筑设计师之路》", 31 July 2004, 电子科技大学出版社 * |
| 陈杭: "一种基于多重纹理映射的倾斜摄影模型单体化技术", 《测绘与空间地理信息》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111783552A (en) * | 2020-06-09 | 2020-10-16 | 当家移动绿色互联网技术集团有限公司 | Live-action three-dimensional model monomer method and device, storage medium and electronic equipment |
| CN111783552B (en) * | 2020-06-09 | 2024-06-25 | 北京五一视界数字孪生科技股份有限公司 | Live three-dimensional model singulation method and device, storage medium and electronic equipment |
| CN114186603A (en) * | 2021-09-28 | 2022-03-15 | 北京博能科技股份有限公司 | Multi-source model single fusion method and system, storage medium and terminal |
| CN114186603B (en) * | 2021-09-28 | 2022-07-01 | 北京博能科技股份有限公司 | Multi-source model single fusion method and system, storage medium and terminal |
| CN115861549A (en) * | 2023-02-27 | 2023-03-28 | 航天宏图信息技术股份有限公司 | Method and device for automatically replacing tree model in live-action three-dimensional data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2011362799B2 (en) | 3D streets | |
| US10521694B2 (en) | 3D building extraction apparatus, method and system | |
| CN110866945A (en) | Method for generating three-dimensional tree by automatic identification of oblique photography model | |
| Santos et al. | 3D plant modeling: localization, mapping and segmentation for plant phenotyping using a single hand-held camera | |
| CN104134234A (en) | Full-automatic three-dimensional scene construction method based on single image | |
| CN112307553B (en) | Method for extracting and simplifying three-dimensional road model | |
| CN112200854B (en) | Leaf vegetable three-dimensional phenotype measuring method based on video image | |
| CN112818925A (en) | Urban building and crown identification method | |
| Frommholz et al. | Extracting semantically annotated 3D building models with textures from oblique aerial imagery | |
| WO2024088071A1 (en) | Three-dimensional scene reconstruction method and apparatus, device and storage medium | |
| CN110070608B (en) | Method for automatically deleting three-dimensional reconstruction redundant points based on images | |
| CN114972646A (en) | Method and system for extracting and modifying independent ground objects of live-action three-dimensional model | |
| CN113378754B (en) | Bare soil monitoring method for construction site | |
| Frommholz et al. | Reconstructing buildings with discontinuities and roof overhangs from oblique aerial imagery | |
| CN116051980B (en) | Building identification method, system, electronic equipment and medium based on oblique photography | |
| Frommholz et al. | Inlining 3d reconstruction, multi-source texture mapping and semantic analysis using oblique aerial imagery | |
| CN114529689B (en) | Ceramic cup defect sample amplification method and system based on antagonistic neural network | |
| CN108921908B (en) | Surface light field acquisition method and device and electronic equipment | |
| Miles et al. | Computational photography | |
| Thomas et al. | A study on close-range photogrammetry in image based modelling and rendering (imbr) approaches and post-processing analysis | |
| Motayyeb et al. | Enhancing contrast of images to improve geometric accuracy of a UAV photogrammetry project | |
| KR102144268B1 (en) | A device for producing a 3D map for virtual reality using an image photographed by an unattended observation means and the method thereof | |
| Collazos et al. | A photogrammetric system for dendrometric feature estimation of individual trees | |
| CN115713700B (en) | Air-ground cooperative typical crop planting sample collection method | |
| US20240233261A1 (en) | Vector data projection and feature matching to determine three-dimensional structure |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210617 Address after: Room 14, 7 / F, building 5, Fourth District, 186 South Fourth Ring Road West, Fengtai District, Beijing, 100160 Applicant after: Beijing Feidu Technology Co.,Ltd. Address before: 311100 Room 502, building 2, Xizi international jinzuo, Nanyuan street, Yuhang District, Hangzhou City, Zhejiang Province Applicant before: Jingcheng Technology (Hangzhou) Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200306 |
|
| RJ01 | Rejection of invention patent application after publication |