CN110019901A - Three-dimensional model search device, searching system, search method and computer readable storage medium - Google Patents
Three-dimensional model search device, searching system, search method and computer readable storage medium Download PDFInfo
- Publication number
- CN110019901A CN110019901A CN201710824563.6A CN201710824563A CN110019901A CN 110019901 A CN110019901 A CN 110019901A CN 201710824563 A CN201710824563 A CN 201710824563A CN 110019901 A CN110019901 A CN 110019901A
- Authority
- CN
- China
- Prior art keywords
- model
- dimensional
- feature data
- dimensional image
- unit
- 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 55
- 238000012545 processing Methods 0.000 claims abstract description 58
- 238000002845 discoloration Methods 0.000 claims abstract description 22
- 238000012937 correction Methods 0.000 claims abstract description 15
- 238000009877 rendering Methods 0.000 claims description 22
- 230000010287 polarization Effects 0.000 claims description 17
- 230000008859 change Effects 0.000 claims description 16
- 230000005484 gravity Effects 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 238000000926 separation method Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 5
- 230000009467 reduction Effects 0.000 claims description 4
- 230000008901 benefit Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 13
- 238000013461 design Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 5
- 238000007689 inspection Methods 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 1
- 235000007926 Craterellus fallax Nutrition 0.000 description 1
- 240000007175 Datura inoxia Species 0.000 description 1
- 125000000174 L-prolyl group Chemical group [H]N1C([H])([H])C([H])([H])C([H])([H])[C@@]1([H])C(*)=O 0.000 description 1
- 239000000729 antidote Substances 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000007717 exclusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Classifications
-
- 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
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/005—General purpose rendering architectures
Abstract
It includes coordinate correcting unit that the present invention, which provides a kind of three-dimensional model search device, carries out coordinate correction to the first model;Two dimensional image generation unit, for obtaining the standard two-dimensional image for completing the first model of coordinates correction;Two dimensional image processing unit carries out constituency to the standard two-dimensional image of the first model and discoloration is handled, obtains the fisrt feature data for describing first model;Feature acquiring unit, for obtaining second feature data;Fisrt feature data and multiple second feature data are compared respectively and are calculated the similarity value of fisrt feature data and each second feature data by characteristic matching unit;And output unit, multiple second models and corresponding similarity value are exported as search result;Second model is threedimensional model.The present invention also provides a kind of method for searching three-dimension model, three-dimensional model searching system and method for searching three-dimension model provided by the invention have the advantages of being quickly obtained search result, better user experience.
Description
[technical field]
The present invention relates to the first model index field, in particular to a kind of three-dimensional model search device, searching system, retrieval
Method and computer readable storage medium.
[background technique]
Currently, threedimensional model using more and more extensive, Design of Industrial Product, virtual reality, 3d gaming, education, shadow
Threedimensional model is widely used in the industries such as apparent motion picture.Threedimensional model with the presence of number in terms of million on the internet, and daily all
There is a large amount of threedimensional model to generate and propagate, thus there is the urgent needs that retrieval analysis is carried out to threedimensional model.
Existing method for searching three-dimension model can substantially be divided into two major classes: search method based on keyword and based on interior
The search method of appearance, being referred to based on the search method of keyword has text to describe the feature of threedimensional model, such as threedimensional model
Size, material, color etc., this method is suitable for the little situation of early stage threedimensional model amount, with the urgency of threedimensional model quantity
Increase severely length, and the search method based on keyword does not adapt to retrieval and requires, the research of more and more method for searching three-dimension model
Content-based retrieval method is turned to, existing content-based retrieval method is broadly divided into three classes: based on feature vector
Method, the method based on topological structure and the method based on two dimensional image.
In traditional content-based retrieval method, some method processing modes are complicated, and data volume is big, cannot more quickly
Ground retrieves as a result, some methods are then that feature is fairly simple, and the information of model describes not accurate enough, and retrieval precision is not high, leads
Cause user experience bad.
[summary of the invention]
To overcome above-mentioned traditional method for searching three-dimension model that cannot relatively rapid retrieve as a result, user experience is bad
The problem of, the present invention provides a kind of three-dimensional model search device, searching system, search method and computer readable storage medium.
The present invention provides a kind of three-dimensional model search device, is used to retrieve the first model of input, with matching
Corresponding second model, the three-dimensional model search device include: coordinate correcting unit, carry out coordinate to the first model and rectify
Just, so that the center of gravity of first model and change in coordinate axis direction are the same as the scheduled origin of three-dimensional model search device and change in coordinate axis direction one
It causes;Two dimensional image generation unit, for obtaining the standard two-dimensional image for completing the first model of coordinates correction;Two dimensional image processing
Unit carries out constituency to the standard two-dimensional image of the first model and discoloration is handled, obtains the fisrt feature for describing first model
Data;Feature acquiring unit, for obtaining second feature data corresponding with multiple second models;Characteristic matching unit,
Fisrt feature data and multiple second feature data are compared respectively and are calculated fisrt feature data and each second spy by it
Levy the similarity value of data;And output unit, multiple second models and corresponding similarity value are exported as retrieval knot
Fruit;Second model is threedimensional model.
Preferably, the two dimensional image processing unit includes selecting module, when first model forms multiple standards two
When tieing up image, the selecting module is formed by choose in multiple standard two-dimensional images in the first model has incidence relation extremely
Few three standard two-dimensional images, and two dimensional image processing is carried out to be described to selected at least three standard two-dimensional images
The fisrt feature data of first model.
Preferably, the two dimensional image processing unit further comprises rendering module, and the rendering module is to constituency module
Two dimensional image after frame choosing carries out the two dimensional image that rendering is converted into square.
Preferably, the two dimensional image processing unit further comprises reducing module, and the diminution module is to the rendering
Two dimensional image afterwards carries out reduction operation.
Preferably, the two dimensional image processing unit further comprises polarization module, and the polarization module is to discoloration
Treated, and gray scale image polarizes the two dimensional image for bianry image by separation of scheduled grayscale value.
The present invention also provides a kind of three-dimensional model searching system, the three-dimensional model searching system is using three-dimensional mould above-mentioned
Type retrieves device.
Preferably, the three-dimensional model searching system further comprises 3 d model library, is stored in the 3 d model library
Multiple second models and/or the corresponding second feature data with second model.
Preferably, the three-dimensional model searching system further comprises typing unit, and the typing unit is in threedimensional model
When inventory contains multiple second models and second feature data, typing unit is by the first model fisrt feature data corresponding with its
It is entered into 3 d model library;For the typing unit when 3 d model library is only stored with the second model, typing unit is by first
Model is entered into 3 d model library.
Preferably, the three-dimensional model searching system further comprises converting unit, and the converting unit is respectively and three-dimensional
Model library and the connection of three-dimensional model search device, when 3 d model library is only stored with the second model, the converting unit is used
In the second feature data that the second model conversion in 3 d model library is formed to description second model.
The present invention also provides a kind of method for searching three-dimension model, for being retrieved to the first model of input, described three
Dimension module search method is the following steps are included: S101: is corrected to the coordinate of the first model of input so that first model
Center of gravity it is consistent with reference axis with the scheduled origin of three-dimensional model search device with change in coordinate axis direction;S102: the first model is obtained
Standard two-dimensional image;S103: standard two-dimensional image or association standard two-dimensional image to the first model of acquisition carry out constituency
Processing selects the region with fisrt feature data with frame;S104: the two dimensional image of frame choosing carries out discoloration and handles to be described
The fisrt feature data of first model;S105: second feature data corresponding with multiple second models are obtained;S106:
Fisrt feature data and second feature data are compared respectively, and calculate fisrt feature data and multiple second feature data
Similarity;And S107: multiple second models and corresponding similarity value are exported as search result.
The present invention also provides a kind of computer-readable storage mediums, are used to store the computer journey of three-dimensional model search
For sequence to retrieve to the first model of input, the computer program executes a kind of method for searching three-dimension model above-mentioned.
Compared with existing design, the three-dimensional model search device of offer provided by the present invention is had the advantage that
Coordinate correcting unit, two dimensional image generation unit, two dimensional image processing are passed sequentially through by the first model of input
The processing of unit can effectively reduce the data volume handled during the first model index, so that retrieving more quick, user's body
It tests preferably.
Three-dimensional model searching system, method for searching three-dimension model and computer-readable storage medium provided by the invention
With with three-dimensional model search device beneficial effect having the same provided by the present invention, this will not be repeated here.
[Detailed description of the invention]
Fig. 1 is a kind of schematic block diagram of three-dimensional model searching system of first embodiment of the invention.
Fig. 2 is the specific module schematic block diagram of two dimensional image processing unit shown in Fig. 1.
Fig. 3 is the schematic diagram that three two dimensional images are obtained from threedimensional model.
Fig. 4 is the schematic diagram that three two dimensional images of acquisition are carried out with Rendering operations.
Fig. 5 is the schematic diagram that two dimensional image is carried out to reduction operation.
Fig. 6 is a kind of schematic block diagram of variant embodiment of two dimensional image processing unit shown in Fig. 2.
Fig. 7 is the schematic diagram of the bianry image formed after gray scale image is polarized.
Fig. 8 is to indicate that gray scale image carries out the schematic diagram of several polarization process with grayscale value.
Fig. 9 is a kind of variant embodiment of three-dimensional model searching system shown in FIG. 1.
Figure 10 is a kind of schematic block diagram of three-dimensional model searching system of second embodiment of the invention.
Figure 11 is the specific module schematic block diagram of two dimensional image processing unit shown in Figure 10.
Figure 12 is a kind of schematic block diagram of method for searching three-dimension model of third embodiment of the invention.
Figure 13 is a kind of schematic block diagram of method for searching three-dimension model of fourth embodiment of the invention.
[specific embodiment]
In order to make the purpose of the present invention, technical solution and advantage are more clearly understood, with reference to the accompanying drawings and embodiments, right
The present invention is further described.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and do not have to
It is of the invention in limiting.
Referring to Fig. 1, the first embodiment of the present invention provides a kind of three-dimensional model searching system 100, three-dimensional model search
System 100 includes three-dimensional model search device 10, input unit 20 and 3 d model library 30.Three-dimensional model search device 10 divides
Not and input unit 20 and the connection of 3 d model library 30, user's input model in input unit 20, which can be with
For threedimensional model or two dimensional image, preferably threedimensional model, after three-dimensional model search device 10 is handled the model of input
Generate the characteristic for describing the model, and the corresponding spy of threedimensional model that will be stored in this feature data and 3 d model library 30
Sign data are compared, and finally export search result, identical so as to be taken at input model in three-dimensional modeling data storehouse
Or similar threedimensional model.
Defining the model inputted from input unit 20 is the first model, and the first model can be threedimensional model and be also possible to two
Tie up image.
The corresponding characteristic of first model is fisrt feature data.The threedimensional model for being stored in 3 d model library 30 is
Second model, the characteristic of corresponding second model are second feature data.It is appreciated that the second model and second feature number
According to number be it is multiple.
Three-dimensional model search device 10 includes coordinate correcting unit 101, two dimensional image generation unit 102, at two dimensional image
Manage unit 103, feature acquiring unit 104, characteristic matching unit 105 and output unit 106.First model of input successively passes through
Cross coordinate correcting unit 101, two dimensional image generation unit 102, two dimensional image processing unit 103, feature acquiring unit 104 with
And the processing of characteristic matching unit 105, finally search result is exported in output unit 106.
It should be noted that output unit 106 can be display terminal, which supports user using more
Kind mode consults the second model retrieved.Specifically, user can be translated or be revolved by way of mouse drag
The second model for turning to retrieve suitable position or viewing angle into output unit 106 carry out the second model of threedimensional model
It checks or user can look into the second model by clicking the thumbnail at each visual angle of the second model retrieved
It sees.With the design method, user can check that mode checks have to the second model retrieved according to what oneself was wanted
Preferable practicability.
Coordinate correcting unit 101 is used to carry out coordinate correction to the first model of input, so that the center of gravity of first model
It is consistent with the scheduled origin of three-dimensional model search device 10, and make the change in coordinate axis direction and model index dress of first model
It is consistent to set 10 scheduled reference axis, which is further processed the first model convenient for the later period.
Input the first model can be one, it is two or more.
When the first model is threedimensional model, which be can be using the threedimensional model gone out designed by software, should
Software can be the Three-dimensional Design Softwares such as Pro/E, AutoCAD, 3DSMAX, Solid Works or UG, but be not limited to these
Three-dimensional Design Software.When the first model of input is threedimensional model, about the correction of threedimensional model coordinate, in the prior art,
The method that some pairs of threedimensional models carry out rotational correction, such as PCA (principal component has been proposed
Analysis) coordinate antidote.Each threedimensional model inputted using 101 Duis of coordinate correcting unit carries out coordinates correction.
When the first model of input is two dimensional image, that is, when what is inputted is picture or photo, coordinate correcting unit 101
The adjustment of placed angle is carried out to the every picture or photo inputted, so that the center of gravity of picture and change in coordinate axis direction and model are examined
The scheduled reference axis of rope device 10 is consistent.In the present embodiment so that edge of photo or picture is parallel to the horizontal plane or vertically i.e.
It can.
Two dimensional image generation unit 102 is used to obtain the standard two-dimensional image for the first model for completing coordinates correction.
Specifically, when the first model of input is threedimensional model, two dimensional image generation unit 102 will be from scheduled direction
First model is projected, the first model after coordinate is corrected generates main view, rearview, left view, right view, vertical view
Figure and bottom view six open standard two-dimensional image.In some specific embodiments, using the positive losing side from scheduled each reference axis
It is projected to the first model, the first model after coordinate is corrected generates main view, rearview, left view, right view, bows
View and bottom view six open standard two-dimensional image.In other embodiments, two dimensional image generation unit 102 is not limited in three-dimensional
The positive negative direction of each reference axis of 10 intrinsic coordinates system of model searching device projects the first model, is also possible to three
Dimension module is retrieved the scheduled aphylactic map projection direction of device 10 and is projected to the first model after coordinate correction, and the first model is raw
At multiple standard two-dimensional images, which can be referred to as association mark for describing the same threedimensional model
Quasi- two dimensional image.
When the first model of input is two dimensional image, two dimensional image generation unit 102 will be through coordinate correcting unit 101
Picture or photo after correction regenerate standard two-dimensional image.
When the two dimensional image of input is multiple, i.e., three or three or more, preferably three.Each two dimensional image is corresponding
Standard two-dimensional image is obtained, which can be referred to as to be associated with for describing the same threedimensional model
Standard two-dimensional image.
Preferably, using two dimensional image as when the first mode input, at least three visual angles vertical each other is selected to be shown
View as two dimensional image input, such as main view (or rearview), left view (or right view) and top view (or bottom view)
As input, and corresponds to and obtain its standard two-dimensional image.
Two dimensional image processing unit 103 is used for the one or more standards two generated of two dimensional image generation unit 102
It ties up image and carries out image preprocessing, to generate the fisrt feature data for describing first model.
Feature acquiring unit 104 is for obtaining second feature data corresponding with multiple second models.
Characteristic matching unit 105 by fisrt feature data and multiple second feature data for comparing and calculating first
The similarity value of characteristic and each second feature data.
Output unit 106 is used to export multiple second models and similarity value corresponding with second model of retrieval,
And it is ranked up from large to small or from small to large by similarity value.
Referring to Fig. 2, specifically, two dimensional image processing unit 103 includes selecting module 1031, constituency module 1032, wash with watercolours
It contaminates module 1033, reduce module 1034 and discoloration module 1035.The one or more standard two-dimensional image that first model generates
By selecting module 1031, constituency module 1032, rendering module 1033, the processing for reducing module 1034 and discoloration module 1035
The fisrt feature data for describing first model are generated afterwards.
Referring to Fig. 3, when the first model is two dimensional image, typically entering multiple X-Y schemes to improve retrieval precision
Picture, and the corresponding standard two-dimensional image for obtaining multiple two dimensional images.Selecting module 1031 is used for multiple standard two-dimensionals in generation
Chosen at least three associated standard two-dimensional images in image, preferably three, and by three standard two-dimensional figures of selection
Intended manner of the incidence relation of picture as subsequent first model three-view diagram image selection.
Preferably, when the first model is threedimensional model, three standard two-dimensional images of selection are vertical each other with three
The view that visual angle is shown, such as main view (or rearview), left view (or right view) and top view (or bottom view) conduct
Optimal selection mode, for the first model with symmetrical feature, which can preferably, maximumlly be opened up
Show the fisrt feature data of the first model, improves retrieval precision.Further, the design method is relevant due to only choosing three
The image of relationship, and do not have to all handle plurality of pictures, the treating capacity of data can be effectively reduced, the speed of retrieval is accelerated
Degree, when being especially stored with a large amount of second models in 3 d model library 30, can retrieve faster corresponding second model with
Obtain preferable user experience.It, can also be using to generation for the precision for further increasing retrieval in some other embodiment
Multiple standard two-dimensional images all handled so that the fisrt feature data more horn of plenty of first model is described with increase inspection
Suo Jingdu.Constituency module 1032 is used to select the region with fisrt feature data in the standard two-dimensional image center of selection, has
Body constituency operation using the marquee tool automatically in every standard two-dimensional image delimit a region, the region to formation two
The boundary for tieing up the first model of profile carries out frame choosing, does not have the region of feature to exclude periphery.The design method exclusion criteria two
There is no the region of fisrt feature data in dimension image, reduce the data volume of operation, the speed of retrieval can be accelerated.Other one
In a little embodiments, constituency operates the mode that can select color-match, to having fisrt feature data to have in standard two-dimensional image
Color region is selected, which further eliminates the white area for selecting frame circle to select, and has further reduced the number of operation
According to amount, so that search result process is more quick.It is appreciated that the constituency operation of constituency module 1032 can also be in selecting module
It is carried out before or after 1031 selection two dimensional images to be treated.
Referring to Fig. 4, rendering module 1033 is used to that the two dimensional image after the choosing of 1032 frame of constituency module to be rendered and be marked
Standardization conversion, so that the two dimensional image of selection is converted to an equal amount of square two dimensional image, the purpose of the design method exists
In facilitate the later period to the reduction operation of two dimensional image.
Referring to Fig. 5, reducing the two dimensional image contracting that module 1034 is the square formed after the rendering of rendering module 1033
The small gray scale image at 8x8 pixel or 16x16 pixel.The data for further having compressed image by reducing the processing of module 1034
Treating capacity improves retrieval rate.It, can also be with it is understood that two dimensional image is not limited to be reduced into 8x8 pixel or 16x16 pixel
It is other sizes, depending on the precision to be retrieved.In general, the pixel that image includes is more, contained data volume
More, retrieval precision is higher, and image includes that pixel is fewer, and contained data volume is fewer, and retrieval precision is lower.
Discoloration module 1035 is to carry out discoloration processing to the region in the two dimensional image after diminution with fisrt feature data,
So that two dimensional image forms gray scale image.This operation simplify the data processing step needed in the first model index to color,
Reduce data processing amount, accelerates retrieval rate.
Fig. 6-8 is please referred to, in some other embodiment, two dimensional image processing unit 103 further comprises polarization module
1036, the gray scale image that polarization module 1036 obtains after being handled discoloration as separation using selected some grayscale value a carries out pole
Change obtains bianry image.Specifically, using the grayscale value a selected in the grayscale value of 0-255 as separation, when in pixel
It is 0 by the grayscale value polarization of the pixel when grayscale value is greater than selected grayscale value a, when the grayscale value in pixel is less than choosing
It is 1 by the grayscale value polarization of the pixel when fixed grayscale value a.It such as with grayscale value 127 is separation by three gray scale images
It is polarised, pixel polarization of the grayscale value greater than 127 is 0, and pixel polarization of the grayscale value less than 127 is 1, then pole
Whole gray scale image is to form the binary map for carrying out feature description by 0 and 1 after change.With the design method, gray scale image passes through pole
Bianry image is formed after change processing, the feature of the first model is described using binary map, further reduces later retrieval process
The treating capacity of middle data can effectively accelerate the speed of retrieval.
Grayscale value a as separation can be selected in the following way: total with the whole grayscale value of two dimensional image
Grayscale value a of the average value of sum as separation.Thus, polarization module 1036 further comprise to the grayscale value of two dimensional image into
Row is averaged.With the design method, image is after polarization, and regardless of brightness, the data of the bianry image formed are not
It can change, eliminate influence of the brightness to search result, effectively improve the precision of retrieval.Or grayscale value a can also
To be determined according to the precision of customer demand and retrieval.
Feature acquiring unit 104 includes two kinds of sides obtaining second feature data corresponding with multiple second models
Formula: one is to obtain the second feature data that prestore namely the second feature data corresponding second model stores together with its
In 3 d model library 30, for three-dimensional model search device 10 during retrieval, feature acquiring unit 104 is directly from three-dimensional mould
Second feature data are directly acquired in type library 30 to be compared;Secondly to obtain the second feature data by data conversion,
The second model is only stored i.e. in 3 d model library, does not store second feature data corresponding with the second model, threedimensional model inspection
Rope device 10 during retrieval, 104 indirect gain of feature acquiring unit from the second model by data it is transformed come the
Two characteristics.Data in specific second of acquisition modes convert the content being referred in second embodiment of the invention.
Referring to Fig. 9, three-dimensional model searching system 100 further comprises typing unit in some other embodiment
107, typing unit 107 is used for after three-dimensional model searching system 100 completes retrieval, specifically can be in characteristic matching unit
After 105 complete characteristic matching, typing unit 107 is automatically by the first model and/or corresponding first spy with first model
Sign data inputting is stored into 3 d model library 30.Specifically, corresponding with the second model when being prestored in 3 d model library 30
When second feature data, typing unit 107 automatically records the first model and fisrt feature data corresponding with first model
Enter into 3 d model library 30 one kind for storing and being automatically converted to the second model and the second spy corresponding with second model
Levy data;When not prestoring second feature data corresponding with the second model in 3 d model library 30, namely only it is stored with
When two models, the first model is entered into automatically and the second mould is stored and be automatically converted in 3 d model library 30 by typing unit 107
One kind of type.With the design method, the amount of storage of 3 d model library 30 can be automatically updated in the setting of typing unit 107, no
It needs manually to carry out typing, is highly convenient for operating.It is appreciated that typing unit 107 is not limited to be used in three-dimensional model search system
After system 100 completes retrieval, just automatically by the first model and fisrt feature data inputting corresponding with first model to three-dimensional
In model library 30, before the characteristic matching unit 105 of three-dimensional model searching system 100 is matched, typing unit 107 is completed
The typing of first model is also possible.It should be noted that if typing unit 107 is set to the progress of characteristic matching unit 105
Before matching, then the first model inputted can become most similar search result and show in output unit 106.
Referring to Fig. 10, the second embodiment of the present invention is to provide a kind of three-dimensional model searching system 200, threedimensional model inspection
The place that cable system 200 is different from three-dimensional model search device 100 is three-dimensional model searching system 200 in addition to including three-dimensional mould
Type searching system 100 still further comprises converting unit 50a, and converting unit 50a and 3 d model library 30 connect, the conversion list
First 50a is used to carry out conversion processing to the second model stored in 3 d model library 30 special to form the second of the second model of description
Data are levied, which is not necessarily to add storage region in 3 d model library 30 to store corresponding second spy of each second model
Data are levied, further, operation is with more flexibility during which makes three-dimensional model search.
Specifically, converting unit 50a includes coordinate correcting unit 501, two dimensional image generation unit 502 and two dimensional image
Processing unit 503.The second model in 3 d model library 30 successively passes through coordinate correcting unit 501, two dimensional image generation unit
502 and the processing of two dimensional image processing unit 503 form the second feature data of the second model of description, then by the second feature
The fisrt feature data for the first model that data and needs are retrieved, which are matched, obtains search result.Due to 3 d model library 30
In the second feature data of the second model be in-time generatin, because storing second feature data without separately setting storage region.
Figure 11 is please referred to, detailed, two dimensional image processing unit 503 includes selecting module 5031, constituency module 5032, wash with watercolours
It contaminates module 5033, reduce module 5034, discoloration module 5035 and polarization module 5036, selecting module 5031, constituency module
5032, operation and the aforementioned selection mould of rendering module 5033, diminution module 5034, discoloration module 5035 and the module 5036 that polarizes
Block 1031, rendering module 1033, reduces module 1034, discoloration module 1035 and polarization module 1036 at constituency module 1032
Operation is consistent, and this will not be repeated here.Different places are, selection of the selecting module 5031 in multiple two dimensional images to generation
It does not carry out three two dimensional images of selection not instead of in an established manner in the process, can arbitrarily select three two dimensional images.This sets
Meter mode can increase the flexibility of operation, all not set especially for three two dimensional images of set selection mode selection
When meter will be put, the two dimensional image by converting other visual angles increases the feature of the second model of description, can increase retrieval
Accuracy.
Figure 12 is please referred to, the third embodiment of the present invention is to provide a kind of method for searching three-dimension model S10, for input
The first model retrieved, method for searching three-dimension model S10 the following steps are included:
S101: correcting the coordinate of the first model of input so that the center of gravity and change in coordinate axis direction of first model are same
The scheduled origin of three-dimensional model search device is consistent with reference axis.
S102: the standard two-dimensional image of the first model is obtained.
Specifically, the first model of input can be threedimensional model or two dimensional image, and the first model inputted can be
One or more, therefore, the standard two-dimensional image of the first model of acquisition can be one or more.
Preferably, when the first model of input is threedimensional model, to the threedimensional model after coordinates correction according to predetermined
Projecting direction is projected, multiple standard two-dimensional images are obtained, and chooses wherein at least three as association standard two-dimensional image;
Multiple standard two-dimensionals are obtained when the first model of input is two dimensional image, and by inputting multiple two dimensional images
Image, and wherein at least three are chosen as association standard two-dimensional image.
S103: standard two-dimensional image or association standard two-dimensional image to the first model of acquisition carry out constituency and handle with frame
Select the region with fisrt feature data.
S104: the two dimensional image of frame choosing carries out discoloration processing to obtain the fisrt feature data for describing first model.
S105: second feature data corresponding with multiple second models are obtained;
S106: fisrt feature data and second feature data are compared respectively, and calculate fisrt feature data and more
The similarity of a second feature data;And
S107: multiple second models and corresponding similarity value are exported as search result.
In some other embodiment, after step s 103 further include:
S1032: standardization rendering carries out rendering processing to two dimensional image, and to two after rendering after the processing of constituency
Image standardized processing is tieed up, different size of two dimensional image is converted to the square of unified size, to accelerate data processing speed
Rate.
S1033: reducing, and after standardization rendering, diminution processing is carried out to two dimensional image, to be further reduced data processing
Amount.Specifically, reducing the two dimensional image for forming 8x8 pixel or 16x16 pixel.It is appreciated that the X-Y scheme formed after reducing
As being not limited to 8x8 pixel or 16x16 pixel, the image for being reduced into other sizes is also possible.
After step s 104 further include S1041: polarization is polarised to form two the gray scale image after discoloration
It is worth image.
The method for searching three-dimension model provided in the present embodiment, the first model of input generate after a series of processing
The first model of description fisrt feature data volume it is less, can soon with the second model corresponding for being stored in model library
Two characteristics are matched, and search result can be relatively rapid obtained, and user experience is preferable.
In step S104, obtaining second feature data corresponding with multiple second models includes two ways: its
One is to obtain the second feature data that prestore namely the second feature data corresponding second model prestores with its, retrieval
Second feature data can be directly acquired in the process to be compared;Secondly to obtain the second feature data by data conversion,
Namely second feature data corresponding with the second model do not prestore, and need to get the second model by data conversion.
In some other embodiment, method for searching three-dimension model further includes steps of special by the first of acquisition
Sign data and before or after second feature data are compared, by the first model of input and or with first model it is corresponding
Fisrt feature data storage.Specifically, when with its, corresponding second model is to prestore to second feature data, typing unit
Automatically the first model and fisrt feature data inputting corresponding with first model are stored;When no pre- in 3 d model library
When depositing second feature data corresponding with the second model namely the corresponding second feature data of the second model do not prestore, and need
When second model is got by data conversion, typing unit automatically stores the first model typing.
Figure 13 is please referred to, the fourth embodiment of the present invention is to provide a kind of method for searching three-dimension model S20, threedimensional model inspection
The place that Suo Fangfa S20 is different from method for searching three-dimension model S10 in 3rd embodiment is:
Method for searching three-dimension model S20 other than including the steps that method for searching three-dimension model S10, further comprise with
Lower step S201-S202:
S201: the second model is transferred from 3 d model library;And
S202: the second model is converted to form second feature data by data.
Following step is specifically included specifically, the second model is converted by data and to form the step S202 of second feature data
Rapid S2021-S2023:
S2021: center of gravity and change in coordinate axis direction so that second model are corrected to the coordinate of the second model of acquisition
It is consistent with reference axis with the scheduled origin of three-dimensional model search device;
S2022: the second model is projected according to pre-determined projection directions, obtains multiple standard two-dimensional images;
S2023: the standard two-dimensional image or association standard two-dimensional image of the second model are obtained.
Specifically, coordinates correction is carried out to the second model in database, which can be with the first model
Method for correcting coordinate it is identical.The second model after coordinates correction is projected according to pre-determined projection directions, obtains multiple marks
Quasi- two dimensional image, and wherein at least three are chosen as association standard two-dimensional image.
S2024: to the standard two-dimensional image or association standard two-dimensional image for obtaining the second model, discoloration processing is carried out to obtain
The second feature data of second model must be described.
Specifically, the standard two-dimensional image to the second model of acquisition or association standard two-dimensional image carry out discoloration processing with
Obtain the second feature data for describing second model, and second feature data be stored in 3 d model library 30, so as to
Fisrt feature data are compared and are called.
In some other embodiment, method for searching three-dimension model S20 is further included steps of in step S2023
It later further include step
S20231: standard two-dimensional image to the second model of acquisition or association standard two-dimensional image carry out constituency processing with
Frame selects the region with second feature data.
S20232: rendering carries out the comparison that rendering is processed so that two dimensional image to two dimensional image after the processing of constituency
Feature becomes apparent.
S20233: reducing, and after rendering processing, diminution processing is carried out to two dimensional image, to be further reduced data processing
Amount.Specifically, reducing the two dimensional image for forming 8x8 pixel or 16x16 pixel.It is appreciated that the X-Y scheme formed after reducing
As being not limited to 8x8 pixel or 16x16 pixel, the image for being reduced into other sizes is also possible.
Further include step S20241 after step S2024:
S20241: polarization is polarised to form bianry image the gray scale image after discoloration.
The fifth embodiment of the present invention provides a kind of computer-readable storage medium, is used to store three-dimensional model search
Computer program to be retrieved to the first model of input, computer program execute following steps:
S101: correcting the coordinate of the first model of input so that the center of gravity and change in coordinate axis direction of first model are same
The scheduled origin of three-dimensional model search device is consistent with reference axis.
S102: the standard two-dimensional image or association standard two-dimensional image of the first model are obtained.
S103: standard two-dimensional image or association standard two-dimensional image to the first model of acquisition carry out constituency and handle with frame
Select the region with fisrt feature data.
S104: the two dimensional image of frame choosing carries out discoloration processing to obtain the fisrt feature data for describing first model.
S105: second feature data corresponding with multiple second models are obtained;
S106: fisrt feature data and second feature data are compared respectively, and calculate fisrt feature data and more
The similarity of a second feature data;And
S107: multiple second models and corresponding similarity value are exported as search result.
The form of general computer-readable medium includes: floppy disk (floppy disk), flexible disc (flexible
Disk), hard disk, tape, remaining any magnetic medium, CD-ROM, remaining any optical medium, punched card (punch
Cards), paper tape (paper tape), remaining any physical medium of pattern with hole, random access memory (RAM),
Programmable read only memory (PROM), erasable programmable read-only memory (EPROM), the read-only storage of quick flashing erasable programmable
Device (FLASH-EPROM), remaining any memory chip or cassette (cartridge) or it is any remaining computer can be allowed to read
Medium.Instruction can further be sent or receive by a transmission medium." transmission medium " this term may include any tangible
Or invisible medium, it can be used to store, encode or carry the instruction for being used to execute to machine, and include that number or simulation are logical
Believe the intangible medium of signal or the communication of remaining promotion above-metioned instruction.Transmission medium includes coaxial cable, copper wire and optical fiber,
Contain the conducting wire of the bus for transmitting a computer data signal.
Compared with existing design, the three-dimensional model search device of offer provided by the present invention is had the advantage that
Coordinate correcting unit, two dimensional image generation unit, two dimensional image processing are passed sequentially through by the first model of input
The processing of unit can effectively reduce the data volume handled during the first model index, so that retrieving more quick, user's body
It tests preferably.
Three-dimensional model searching system, method for searching three-dimension model and computer-readable storage medium provided by the invention
With with three-dimensional model search device beneficial effect having the same provided by the present invention, this will not be repeated here.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in original of the invention
Made any modification within then, equivalent replacement and improvement etc. should all be comprising within protection scope of the present invention.
Claims (11)
1. a kind of three-dimensional model search device is used to retrieve the first model of input, to match corresponding second mould
Type, which is characterized in that the three-dimensional model search device includes:
Coordinate correcting unit carries out coordinate correction to the first model, so that the center of gravity of first model and change in coordinate axis direction are same
The scheduled origin of three-dimensional model search device is consistent with change in coordinate axis direction;
Two dimensional image generation unit, for obtaining the standard two-dimensional image for completing the first model of coordinates correction;
Two dimensional image processing unit carries out constituency to the standard two-dimensional image of the first model and discoloration is handled, obtain description this
The fisrt feature data of one model;
Feature acquiring unit, for obtaining second feature data corresponding with multiple second models;
Fisrt feature data and multiple second feature data are compared respectively and calculate fisrt feature by characteristic matching unit
The similarity value of data and each second feature data;And
Output unit exports multiple second models and corresponding similarity value as search result;Second model is
Threedimensional model.
2. three-dimensional model search device as described in claim 1, it is characterised in that: the two dimensional image processing unit includes choosing
Module is selected, when first model forms multiple standard two-dimensional images, the selecting module is formed by more in the first model
At least three standard two-dimensional images chosen in standard two-dimensional image and there is incidence relation are opened, and selected at least three are marked
Quasi- two dimensional image carries out two dimensional image processing to obtain the fisrt feature data for describing first model.
3. three-dimensional model search device as described in claim 1, it is characterised in that: the two dimensional image processing unit is further
Including rendering module, the two dimensional image after the rendering module selects constituency module frame carries out the two dimension that rendering is converted into square
Image.
4. three-dimensional model search device as claimed in claim 3, it is characterised in that: the two dimensional image processing unit is further
Including reducing module, the module that reduces is to the two dimensional image progress reduction operation after the rendering.
5. three-dimensional model search device as described in claim 1, it is characterised in that: the two dimensional image processing unit is further
Including the module that polarizes, the polarization module is using scheduled grayscale value to be separation by institute to discoloration treated gray scale image
Stating two dimensional image polarization is bianry image.
6. a kind of three-dimensional model searching system, it is characterised in that: the three-dimensional model searching system uses such as claim 1-5
Described in any item three-dimensional model search devices.
7. three-dimensional model searching system as claimed in claim 6, it is characterised in that: the three-dimensional model searching system is further
Including 3 d model library, multiple second models are stored in the 3 d model library and/or corresponding with second model
Two characteristics.
8. three-dimensional model searching system as claimed in claim 7, it is characterised in that: the three-dimensional model searching system is further
Including typing unit, the typing unit is when 3 d model library is stored with multiple second models and second feature data, typing
Unit is by the first model fisrt feature data inputting corresponding with its into 3 d model library;The typing unit is in threedimensional model
When library is only stored with the second model, the first model is entered into 3 d model library by typing unit.
9. three-dimensional model searching system as claimed in claim 8, it is characterised in that: the three-dimensional model searching system is further
Including converting unit, the converting unit is connected with 3 d model library and three-dimensional model search device respectively, in threedimensional model
When library is only stored with the second model, the converting unit be used to form the second model conversion in 3 d model library description this
The second feature data of two models.
10. a kind of method for searching three-dimension model, for being retrieved to the first model of input, which is characterized in that the three-dimensional
Model retrieval method the following steps are included:
S101: correcting the coordinate of the first model of input so that the center of gravity and change in coordinate axis direction of first model are the same as three-dimensional
The scheduled origin of model searching device is consistent with reference axis;
S102: the first model criteria two dimensional image is obtained;
S103: standard two-dimensional image or association standard two-dimensional image to the first model of acquisition are carried out constituency processing and are selected with frame
Region with fisrt feature data;
S104: the two dimensional image of frame choosing carries out discoloration processing to obtain the fisrt feature data for describing first model;
S105: second feature data corresponding with multiple second models are obtained;
S106: fisrt feature data and second feature data are compared respectively, and calculate fisrt feature data and multiple
The similarity of two characteristics;And
S107: multiple second models and corresponding similarity value are exported as search result.
11. a kind of computer-readable storage medium is used to store the computer program of three-dimensional model search to input
First model is retrieved, it is characterised in that: the computer program executes following steps:
S101: correcting the coordinate of the first model of input so that the center of gravity and change in coordinate axis direction of first model are the same as three-dimensional
The scheduled origin of model searching device is consistent with reference axis;
S102: the standard two-dimensional image of the first model is obtained;
S103: standard two-dimensional image or association standard two-dimensional image to the first model of acquisition are carried out constituency processing and are selected with frame
Region with fisrt feature data;
S104: the two dimensional image of frame choosing carries out discoloration processing to obtain the fisrt feature data for describing first model;
S105: second feature data corresponding with multiple second models are obtained;
S106: fisrt feature data and second feature data are compared respectively, and calculate fisrt feature data and multiple
The similarity of two characteristics;And
S107: multiple second models and corresponding similarity value are exported as search result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710824563.6A CN110019901A (en) | 2017-09-13 | 2017-09-13 | Three-dimensional model search device, searching system, search method and computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710824563.6A CN110019901A (en) | 2017-09-13 | 2017-09-13 | Three-dimensional model search device, searching system, search method and computer readable storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110019901A true CN110019901A (en) | 2019-07-16 |
Family
ID=67186294
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710824563.6A Pending CN110019901A (en) | 2017-09-13 | 2017-09-13 | Three-dimensional model search device, searching system, search method and computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110019901A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111161399A (en) * | 2019-12-10 | 2020-05-15 | 盎锐(深圳)信息科技有限公司 | Data processing method and component for generating three-dimensional model based on two-dimensional image |
CN112100231A (en) * | 2020-07-17 | 2020-12-18 | 四川长宁天然气开发有限责任公司 | Correlation method and system for shale gas ground engineering entity information and digital model |
CN117745955A (en) * | 2024-02-20 | 2024-03-22 | 北京飞渡科技股份有限公司 | Method and device for generating urban building scene based on building base vector data |
CN111161399B (en) * | 2019-12-10 | 2024-04-19 | 上海青燕和示科技有限公司 | Data processing method and assembly for generating three-dimensional model based on two-dimensional image |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101350016A (en) * | 2007-07-20 | 2009-01-21 | 富士通株式会社 | Device and method for searching three-dimensional model |
CN101458714A (en) * | 2008-12-30 | 2009-06-17 | 清华大学 | Three-dimensional model search method based on precision geodesic |
CN102375831A (en) * | 2010-08-13 | 2012-03-14 | 富士通株式会社 | Three-dimensional model search device and method thereof and model base generation device and method thereof |
JP2015201151A (en) * | 2014-04-04 | 2015-11-12 | 国立大学法人豊橋技術科学大学 | Three-dimensional model retrieval system, and three-dimensional model retrieval method |
-
2017
- 2017-09-13 CN CN201710824563.6A patent/CN110019901A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101350016A (en) * | 2007-07-20 | 2009-01-21 | 富士通株式会社 | Device and method for searching three-dimensional model |
CN101458714A (en) * | 2008-12-30 | 2009-06-17 | 清华大学 | Three-dimensional model search method based on precision geodesic |
CN102375831A (en) * | 2010-08-13 | 2012-03-14 | 富士通株式会社 | Three-dimensional model search device and method thereof and model base generation device and method thereof |
JP2015201151A (en) * | 2014-04-04 | 2015-11-12 | 国立大学法人豊橋技術科学大学 | Three-dimensional model retrieval system, and three-dimensional model retrieval method |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111161399A (en) * | 2019-12-10 | 2020-05-15 | 盎锐(深圳)信息科技有限公司 | Data processing method and component for generating three-dimensional model based on two-dimensional image |
CN111161399B (en) * | 2019-12-10 | 2024-04-19 | 上海青燕和示科技有限公司 | Data processing method and assembly for generating three-dimensional model based on two-dimensional image |
CN112100231A (en) * | 2020-07-17 | 2020-12-18 | 四川长宁天然气开发有限责任公司 | Correlation method and system for shale gas ground engineering entity information and digital model |
CN112100231B (en) * | 2020-07-17 | 2023-10-13 | 四川长宁天然气开发有限责任公司 | Association method and system of shale gas ground engineering entity information and digital model |
CN117745955A (en) * | 2024-02-20 | 2024-03-22 | 北京飞渡科技股份有限公司 | Method and device for generating urban building scene based on building base vector data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108509848B (en) | The real-time detection method and system of three-dimension object | |
US10049308B1 (en) | Synthesizing training data | |
CN109870983B (en) | Method and device for processing tray stack image and system for warehousing goods picking | |
CN108121997A (en) | Use the object classification in the image data of machine learning model | |
US9349076B1 (en) | Template-based target object detection in an image | |
CN109960742B (en) | Local information searching method and device | |
CN109977983B (en) | Method and device for obtaining training image | |
CN109948497A (en) | A kind of object detecting method, device and electronic equipment | |
CN111680678B (en) | Target area identification method, device, equipment and readable storage medium | |
CN111462120A (en) | Defect detection method, device, medium and equipment based on semantic segmentation model | |
Biasotti et al. | SHREC’14 track: Retrieval and classification on textured 3D models | |
US10134149B2 (en) | Image processing | |
CN104915673A (en) | Object classification method and system based on bag of visual word model | |
AU2018202767A1 (en) | Data structure and algorithm for tag less search and svg retrieval | |
JP2002288687A (en) | Device and method for calculating feature amount | |
CN112927353A (en) | Three-dimensional scene reconstruction method based on two-dimensional target detection and model alignment, storage medium and terminal | |
CN111507357A (en) | Defect detection semantic segmentation model modeling method, device, medium and equipment | |
CN109740674A (en) | A kind of image processing method, device, equipment and storage medium | |
CN108229505A (en) | Image classification method based on FISHER multistage dictionary learnings | |
CN110147460B (en) | Three-dimensional model retrieval method and device based on convolutional neural network and multi-view map | |
CN115115825A (en) | Method and device for detecting object in image, computer equipment and storage medium | |
CN115984662A (en) | Multi-mode data pre-training and recognition method, device, equipment and medium | |
CN110019901A (en) | Three-dimensional model search device, searching system, search method and computer readable storage medium | |
CN110110829A (en) | A kind of two dimensional code processing method and processing device | |
CN113570615A (en) | Image processing method based on deep learning, electronic equipment and storage medium |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190716 |