CN112270742A - Method for model retrieval based on object natural characteristic points - Google Patents
Method for model retrieval based on object natural characteristic points Download PDFInfo
- Publication number
- CN112270742A CN112270742A CN202011181296.3A CN202011181296A CN112270742A CN 112270742 A CN112270742 A CN 112270742A CN 202011181296 A CN202011181296 A CN 202011181296A CN 112270742 A CN112270742 A CN 112270742A
- Authority
- CN
- China
- Prior art keywords
- image
- background
- dimensional model
- dimensional
- data
- 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
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000005070 sampling Methods 0.000 claims abstract description 4
- 230000000877 morphologic effect Effects 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 2
- 239000000463 material Substances 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 238000009877 rendering Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/04—Texture mapping
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/155—Segmentation; Edge detection involving morphological operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Graphics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the field of three-dimensional modeling, and discloses a method for model retrieval based on object natural characteristic points, which comprises the following steps: image acquisition: acquiring an object image with a simple background; background denoising: removing the background of the object in the object image; removing the surface of the pattern: removing the surface of the object and only leaving the point line data of the object; and (3) side line point taking: sampling and collecting the point line of the object to generate a two-dimensional array; generating two-dimensional model data: setting different identification positions at four corners of the two-dimensional array to generate two-dimensional model data; model retrieval: and inputting the two-dimensional model data into a model library for model retrieval by taking the two-dimensional model data as a query condition, and acquiring three-dimensional model data of an object similar to the object. By the method, the three-dimensional model of the object close to the object can be quickly obtained through the plane image of the object. The three-dimensional model is used as the input of the three-dimensional model of the target object, rapid modeling can be realized based on priori knowledge, and modeling efficiency is improved.
Description
Technical Field
The invention relates to the field of digital modeling, in particular to a method for model retrieval based on object natural characteristic points.
Background
In recent years, 3D digital modeling is widely applied to industrial internet, and 3D models can visually show the real conditions of mines, plants, industrial equipment, materials and the like, thereby well supporting the functions of a plurality of industrial informatization systems such as monitoring command, asset management, production management and the like.
However, 3D models, or digital factories, need to be built for 3D models of real objects.
In the existing 3D modeling method, a plurality of pictures, 360-degree images, and materials such as a size chart of an object are introduced into 3Dmax or other tools through tool software such as 3Dmax, and are converted to obtain a model prototype, and then a final 3D model is obtained through manual editing.
The 3D model can be constructed mainly through three modes of three-dimensional manufacturing software, scanning instrument generation and plane image rendering, and a 3D model file mainly comprises two parts of 'grid (material, animation') and 'map'. The traditional 3D modeling process can be summarized as four steps as a whole:
first step, modeling tool
There are many excellent modeling software on the market, such as 3DMAX, ArcGIS, Maya, AutoCAD, etc., which usually provide some basic geometric elements, such as cubes, spheres, etc., and then construct a complex geometric scene through a series of geometric operations (translation, rotation, stretching, etc.).
Second, effect editing
Generally, a common 3D model is made, considering the best results with the least number of surfaces. And adding proper material textures to the model, unfolding UV, assisting in various pasters such as colors and normal lines, creating lighting effects, enabling the model to become more precise and more vivid, setting animation, and finally rendering and exporting.
Thirdly, visualization and publishing
After the model and the map material are exported, the last step is the visual display and formal release of the model, and the last step is the most important step for generating the value of the model after the model is created.
The fourth step, model processing
Modelers may search many tools to perform surface reduction, UV unfolding, baking, format conversion, and other processes on the 3D model, so as to make the model smaller, not destroy the effect of the modeler, and quickly display the model.
The traditional 3D modeling process consumes a great amount of labor and time, and how to perform rapid and automatic 3D modeling of the object can greatly reduce the cost of 3D modeling.
Disclosure of Invention
In view of the above-mentioned drawbacks of the prior art, an object of the present invention is to provide a method for rapidly obtaining a pre-modeling of a 3D model of an object, so as to speed up modeling and reduce the cost of 3D modeling.
In order to achieve the above object, the present invention provides a method for model retrieval based on object natural feature points, comprising:
image acquisition: acquiring an object image with a simple background;
background denoising: removing the background of the object in the object image;
removing the surface of the pattern: removing the surface of the object and only leaving the point line data of the object;
and (3) side line point taking: sampling and collecting the point line of the object to generate a two-dimensional array;
generating two-dimensional model data: setting different identification positions at four corners of the two-dimensional array to generate two-dimensional model data;
model retrieval: and inputting the two-dimensional model data as a query condition into a model library for model retrieval to obtain three-dimensional model data of the approximate object.
Further, the image acquisition specifically includes:
setting a plane background;
arranging a target object in front of a background, wherein the target object and the background have a certain distance interval;
and setting the camera to be vertical to the background, and acquiring the image of the target object with the background.
Further, the background denoising specifically includes:
converting the object image into a gray scale image;
carrying out self-adaptive threshold processing on the gray level image through a second-order Gaussian filter function;
performing morphological closing operation on the binary gray level image to remove tiny irrelevant elements in the image;
and acquiring the outline of the object and removing the background according to the result of the morphological closing operation of the gray-scale image.
Further, the graphic removal mask body includes: and detecting geometric corner points and edges in the gray-scale image by a corner point detection algorithm.
Further, the edge line point fetching specifically includes: and storing all detected corner points and edges in an m x n two-dimensional array, wherein m is the width of the image, n is the height of the image, the array index is the pixel coordinate of the corresponding geometric corner point, and the value corresponding to the array index is whether the geometric corner point exists, so that the characteristic data of the image, namely the model data of a certain view angle of the object, is established.
Further, the generating the two-dimensional model data specifically includes: and respectively setting different identification bits at each corner of the two-dimensional array obtained by edge line point taking.
Further, the method also comprises graph splitting between the graph surface removing and the side line point taking, wherein the graph splitting is to split the object into a plurality of components according to the morphological characteristics of the object, determine the geometric corner points and the side lines of each component, store the detected angles in different two-dimensional arrays according to the components, and establish the characteristic data of images of different components.
By the method for carrying out model retrieval based on the natural characteristic points of the object, the two-dimensional model data of a certain visual angle of the object can be automatically and quickly established by acquiring the natural characteristic points such as geometric corner points, side lines and the like in the image of the object, the three-dimensional model data of the object similar to the two-dimensional model data is acquired by inquiring the model database by taking the two-dimensional model data as the inquiry condition, and the 3D modeling of the target object is carried out on the basis of the three-dimensional model data, so that the 3D modeling efficiency can be improved, and the 3D modeling cost can be reduced.
Drawings
FIG. 1 is a flow chart of a method of model retrieval based on natural feature points of an object of the present invention;
Detailed Description
To further illustrate the various embodiments, the invention provides the accompanying drawings. The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the embodiments. Those skilled in the art will appreciate still other possible embodiments and advantages of the present invention with reference to these figures. Elements in the figures are not drawn to scale and like reference numerals are generally used to indicate like elements.
The invention will now be further described with reference to the accompanying drawings and detailed description.
Example 1
As shown in fig. 1, the present invention provides a method for model retrieval based on natural feature points of an object, which acquires an image of the object with a simple background, automatically acquires two-dimensional model data of the object at a certain viewing angle, and acquires three-dimensional model data approximate to the object by using the two-dimensional model data as a query condition. The method comprises the following steps:
1. image acquisition: an image of an object with a simple background is acquired.
Setting a plane background;
placing a target object in front of a background, wherein the target object and the background have a certain distance interval;
and setting the camera to be vertical to the background, and acquiring the image of the target object with the background.
By the method, the background is easy to identify and remove, and the outline of the object is extracted.
2. Background denoising: and removing the background of the object in the object image. The method specifically comprises the following steps:
(1) converting the object image into a grey-scale map;
(2) carrying out self-adaptive threshold processing on the gray level image through a second-order Gaussian filter function;
(3) performing morphological closing operation on the binary gray level image to remove tiny irrelevant elements in the image;
(4) and acquiring the outline of the object and removing the background according to the result of the morphological closing operation of the gray-scale image.
The morphological closing operation can generally smooth the contour of the object, close narrower discontinuities and elongated ravines, eliminate small voids, fill breaks in the contour line, and facilitate object extraction and background removal.
Since the distance between the background and the camera is consistent and the same or similar gray scale is provided, the background and the object can be conveniently separated through the adaptive threshold processing, and the contour of the object can be obtained.
3. And (4) removing the surface of the object by the pattern, and only leaving the point line data of the object.
And detecting geometric corner points and edges in the gray-scale image by a corner point detection algorithm.
The edge line is a set of pixel points of points where the first-order reciprocal (i.e., the gradient of the gray level) jumps.
The geometric corner points may be of several types: (1) the pixel point corresponding to the local maximum of the first derivative (i.e., the gradient of the gray level); (2) the intersection point of two or more sidelines; (3) points in the image where both gradient values and rates of change in gradient direction are high; (4) the first derivative is maximum at the geometric angle point and the second derivative is zero, indicating the direction in which the edge variation of the object is discontinuous.
4. And (3) side line point taking: sampling and collecting the point line of the object to generate a two-dimensional array.
And storing all the detected geometric corner points and edges in an m x n two-dimensional array, wherein m is the width of the image, n is the height of the image, the array index is the pixel coordinate of the corresponding geometric corner point, and the value corresponding to the array index is whether the geometric corner point exists, so that the characteristic data of the image, namely the image characteristic data of a certain view angle of the object, is established.
5. Generating two-dimensional model data: and setting different identification positions at four corners of the two-dimensional array to generate two-dimensional model data. Different identification positions are arranged at four corners of the two-dimensional array, so that the uniqueness of the two-dimensional model data of the planar image in the database can be ensured after the image is rotated.
6. Model retrieval: and inputting the two-dimensional model data into a model library for model retrieval by taking the two-dimensional model data as a query condition, and acquiring three-dimensional model data of an object similar to the object. The three-dimensional model can be used as the input of the three-dimensional model modeling of the target object, so that the rapid modeling can be realized based on the priori knowledge, and the modeling efficiency is improved.
7. And (3) splitting a graph: and graph splitting can be further included between the graph surface removing and the edge line point taking. The graph splitting is to select the geometric corner points and the side lines of the object through manual operation according to the morphological characteristics of the object, and split the object into a plurality of parts. And then, by taking points through side lines, respectively storing the geometric corner points and the side line data of the object in different two-dimensional arrays according to the parts, and establishing the feature data of different part images. This process is an optional process.
By the method for searching the model based on the natural characteristic points of the object, the three-dimensional model of the object close to the object can be quickly obtained by acquiring the natural characteristic points such as geometric corner points, side lines and the like in the image of the object. On the basis, the retrieved three-dimensional model can be used as the input of the three-dimensional model modeling of the target object, so that the rapid modeling can be realized based on the prior knowledge, and the modeling efficiency is improved.
While the invention has been particularly shown and described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (6)
1. A method for model retrieval based on object natural feature points is characterized by comprising the following steps:
image acquisition: acquiring an object image with a simple background;
background denoising: removing the background of the object in the object image;
removing the surface of the pattern: removing the surface of the object, and only leaving the geometric corner point and the side line data of the object;
and (3) side line point taking: sampling and collecting geometric corner points and side lines of the object to generate a two-dimensional array;
generating two-dimensional model data: setting different identification positions at four corners of the two-dimensional array to generate two-dimensional model data;
model retrieval: and inputting the two-dimensional model data into a model library for model retrieval by taking the two-dimensional model data as a query condition, and acquiring three-dimensional model data of an object similar to the object.
2. The method according to claim 1, wherein the image acquisition comprises in particular:
setting a plane background;
arranging a target object in front of a background, wherein the target object and the background have a certain distance interval;
and setting the camera to be vertical to the background, and acquiring the image of the target object with the background.
3. The method of claim 1, wherein the background noise reduction specifically comprises:
converting the object image into a gray scale image;
carrying out self-adaptive threshold processing on the gray level image through a second-order Gaussian filter function;
performing morphological closing operation on the binary gray level image to remove tiny irrelevant elements in the image;
and acquiring the outline of the object and removing the background according to the result of the morphological closing operation of the gray-scale image.
4. The method of claim 1, wherein the graphical facade is specifically: and detecting geometric corner points and side lines in the gray-scale image by a corner point detection algorithm to obtain geometric corner point and side line data.
5. The method of claim 1, wherein the edge fetching specifically comprises: and storing all the detected geometric corner points and edges in an m x n two-dimensional array, wherein m is the width of the image, n is the height of the image, the array index is the pixel coordinate of the corresponding geometric corner point, and the value corresponding to the array index is whether the geometric corner point exists or not, so as to establish the characteristic data of the image.
6. The method of claim 1, further comprising a graph splitting between the graph surface removing and the edge line fetching, wherein the graph splitting is to split the object into a plurality of parts by selecting geometric corner points and edge lines of the object according to morphological features of the object.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011181296.3A CN112270742A (en) | 2020-10-29 | 2020-10-29 | Method for model retrieval based on object natural characteristic points |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011181296.3A CN112270742A (en) | 2020-10-29 | 2020-10-29 | Method for model retrieval based on object natural characteristic points |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112270742A true CN112270742A (en) | 2021-01-26 |
Family
ID=74344942
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011181296.3A Pending CN112270742A (en) | 2020-10-29 | 2020-10-29 | Method for model retrieval based on object natural characteristic points |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112270742A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113112608A (en) * | 2021-04-20 | 2021-07-13 | 厦门汇利伟业科技有限公司 | Method for automatically establishing three-dimensional model from object graph |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000075856A1 (en) * | 1999-06-07 | 2000-12-14 | Metrologic Instruments, Inc. | Unitary package identification and dimensioning system employing ladar-based scanning methods |
| WO2019157924A1 (en) * | 2018-02-13 | 2019-08-22 | 视辰信息科技(上海)有限公司 | Real-time detection method and system for three-dimensional object |
-
2020
- 2020-10-29 CN CN202011181296.3A patent/CN112270742A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000075856A1 (en) * | 1999-06-07 | 2000-12-14 | Metrologic Instruments, Inc. | Unitary package identification and dimensioning system employing ladar-based scanning methods |
| WO2019157924A1 (en) * | 2018-02-13 | 2019-08-22 | 视辰信息科技(上海)有限公司 | Real-time detection method and system for three-dimensional object |
Non-Patent Citations (3)
| Title |
|---|
| 康锋: "倾斜摄影技术在城市三维建模中的应用探讨", 《集成电路与嵌入式系统》, vol. 20, no. 230, 1 February 2020 (2020-02-01), pages 40 - 42 * |
| 樊亚春: "基于形状检索的场景图像三维建模", 《高技术通讯》, vol. 23, no. 8, 31 August 2013 (2013-08-31), pages 781 - 788 * |
| 阮孟贵: "基于图像三维模型重建的研究", 《中国优秀硕士学位论文全文数据库（信息科技辑）》, no. 2012, 15 February 2012 (2012-02-15), pages 138 - 2236 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113112608A (en) * | 2021-04-20 | 2021-07-13 | 厦门汇利伟业科技有限公司 | Method for automatically establishing three-dimensional model from object graph |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10510148B2 (en) | Systems and methods for block based edgel detection with false edge elimination | |
| CN105427317B (en) | A kind of method suitable for various visual angles automatization registration multistation ground laser point cloud data | |
| Xu et al. | Reconstruction of scaffolds from a photogrammetric point cloud of construction sites using a novel 3D local feature descriptor | |
| CN112836734A (en) | Heterogeneous data fusion method and device and storage medium | |
| CN108921939A (en) | A kind of method for reconstructing three-dimensional scene based on picture | |
| CN113112608A (en) | Method for automatically establishing three-dimensional model from object graph | |
| CN107944459A (en) | A kind of RGB D object identification methods | |
| Yogeswaran et al. | 3d surface analysis for automated detection of deformations on automotive body panels | |
| CN107134008A (en) | A kind of method and system of the dynamic object identification based under three-dimensional reconstruction | |
| Pound et al. | A patch-based approach to 3D plant shoot phenotyping | |
| CN112307553A (en) | Method for extracting and simplifying three-dimensional road model | |
| CN111489416A (en) | Tunnel axis fitting method and application in calculation of over-under excavation square measure | |
| CN114022542A (en) | Three-dimensional reconstruction-based 3D database manufacturing method | |
| Poullis | Large-scale urban reconstruction with tensor clustering and global boundary refinement | |
| CN113487633A (en) | Point cloud contour extraction method and device, computer equipment and storage medium | |
| CN111612886B (en) | Indoor three-dimensional model generation method and system | |
| Pacheco et al. | Reconstruction of high resolution 3D objects from incomplete images and 3D information | |
| CN114821571A (en) | Point cloud processing method for power cable identification and reconstruction | |
| Yin et al. | [Retracted] Virtual Reconstruction Method of Regional 3D Image Based on Visual Transmission Effect | |
| CN113628170B (en) | Laser line extraction method and system based on deep learning | |
| CN112270742A (en) | Method for model retrieval based on object natural characteristic points | |
| CN116452604B (en) | Complex substation scene segmentation method, device and storage medium | |
| CN113111741A (en) | Assembly state identification method based on three-dimensional feature points | |
| CN116958485A (en) | Visual field analysis method and device | |
| CN112802101A (en) | Hierarchical template matching method based on multi-dimensional pyramid |
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 |