CN110163954A - Three-dimensional house type model generating method, device, equipment and storage medium - Google Patents
Three-dimensional house type model generating method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN110163954A CN110163954A CN201910292777.2A CN201910292777A CN110163954A CN 110163954 A CN110163954 A CN 110163954A CN 201910292777 A CN201910292777 A CN 201910292777A CN 110163954 A CN110163954 A CN 110163954A
- Authority
- CN
- China
- Prior art keywords
- wall
- floor plan
- house type
- pixel
- type model
- 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 61
- 238000012549 training Methods 0.000 claims description 43
- 238000013136 deep learning model Methods 0.000 claims description 32
- 238000010586 diagram Methods 0.000 claims description 31
- 238000006243 chemical reaction Methods 0.000 abstract description 3
- 238000004891 communication Methods 0.000 description 6
- 238000010276 construction Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000003321 amplification Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000003199 nucleic acid amplification method Methods 0.000 description 2
- 208000001613 Gambling Diseases 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000001351 cycling effect Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
Classifications
-
- 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
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/04—Architectural design, interior design
Abstract
The present invention relates to modeling technique fields, a kind of three-dimensional house type model generating method, device, equipment and storage medium are provided, it the described method comprises the following steps: when receiving two-dimentional floor plan, the key point of the two-dimentional floor plan is obtained, the key point is turning point, tie point, bifurcation and/or the endpoint occurred in the two-dimentional floor plan；It determines the corresponding area to be tested of the key point, the pixel of preset condition will be met in the area to be tested labeled as wall pixel；The wall in the two-dimentional floor plan is determined according to the key point and the wall pixel based on breadth-first search；According to the wall, the two-dimentional floor plan is generated into three-dimensional house type model.The present invention gradually identifies wall by starting point of key point around, finally according to wall, constructs three-dimensional house type model, realize the conversion of two-dimentional floor plan to three-dimensional house type model, facilitate the detailed understanding house type of user by first extracting the key point in two-dimentional floor plan.
Description
Technical field
The present invention relates to modeling technique field more particularly to a kind of three-dimensional house type model generating method, device, equipment and meters
Calculation machine readable storage medium storing program for executing.
Background technique
In real estate industry, client is not in the case where going to scene to check, it is desirable to when understanding house type, generally by pin
Personnel's two-dimentional house type drawing of granting is sold, or viewing design specialist uses mold in kind, the three-dimensional house type mould built by manpower
Type.The former is due to being in the plane, direct feel spatially cannot to be brought to client；The latter due to technical threshold height,
Output is small, lacks the means for automatically generating three-dimensional house type model.In similar patent, such as 201710673386.6 users by
It is clicked in drawing, gradually determines module position in house type, this mode still needs manpower to participate in, hence it is evident that intelligence not enough
Can, and 201510156080.4 are to handle photo by traditional graph processing means not obtaining although saving manpower
It popularizes and applies in market to accurate location, therefore not.Therefore, it is badly in need of one kind at present and had not only saved expert's manpower, but also can be quickly
The method for accurately generating house type threedimensional model, and then spread to the pain spot demand that market solves client.
Summary of the invention
The main purpose of the present invention is to provide a kind of three-dimensional house type model generating method, device, equipment and computers can
Read storage medium, it is intended to it is more complicated to solve existing three-dimensional house type model foundation, and establishes slow-footed technical problem.
To achieve the above object, the present invention provides a kind of three-dimensional house type model generating method, and the three-dimensional house type model is raw
At method the following steps are included:
When receiving two-dimentional floor plan, the key point of the two-dimentional floor plan is obtained, the key point is the two dimension
Turning point, tie point, bifurcation and/or the endpoint occurred in floor plan；
It determines the corresponding area to be tested of the key point, the pixel of preset condition will be met in the area to be tested
Labeled as wall pixel；
The two-dimentional house type is determined according to the key point and the wall pixel based on breadth-first search
Wall in figure；
According to the wall, the two-dimentional floor plan is generated into three-dimensional house type model.
Optionally, the corresponding area to be tested of the determination key point, it is default by meeting in the area to be tested
The pixel of condition be labeled as wall pixel the step of include:
Using the key point as input terminal, using the input terminal as the center of circle, presetted pixel value is radius, is constructed to be detected
Region；
Rgb (RGB color standard) value of the first pixel in the area to be tested is obtained, and according to the rgb value, really
The rgb median of the fixed rgb value；
It determines in the area to be tested, if there are rgb value unlabelled second pictures equal with the rgb median
Vegetarian refreshments；
If it exists, second pixel is labeled as wall pixel, and using the wall pixel as output end；
Using the output end as the input terminal, and recycle described in execution using the input terminal as the center of circle, presetted pixel
Second pixel is labeled as wall pixel, and will to if it exists by the step of value is radius, constructs area to be tested
There is no rgb value and corresponding rgb until determining in the area to be tested in the step of wall pixel is as output end
The equal unlabelled pixel of median.
Optionally, the key point is described to be based on breadth-first search including at least two, according to the key point
With the wall pixel, the step of determining the wall in the two-dimentional floor plan, includes:
Based on breadth-first search, key point all in the two-dimentional floor plan is traversed, successively determines two passes
The wall path constituted between key point with the presence or absence of the wall pixel；
If it exists, it is determined that described two key points and the wall path are wall.
Optionally, described when receiving two-dimentional floor plan, obtain the key point of the two-dimentional floor plan, the key point
For occur in the two-dimentional floor plan turning point, tie point, bifurcation and/or endpoint the step of include:
When receiving two-dimentional floor plan, the two-dimentional floor plan is inputted in trained deep learning model, is generated
Crucial point diagram, the deep learning model are obtained based on the crucial point diagram training of the two-dimentional floor plan marked and corresponding mark；
Based on the crucial point diagram, corresponding key point is extracted, wherein the key point is to go out in the two-dimentional floor plan
Existing turning point, tie point, bifurcation and/or endpoint.
Optionally, described when receiving two-dimentional floor plan, based on the two-dimentional floor plan, obtain corresponding key point
Before step, the method also includes:
It obtains multiple and different to the two-dimentional floor plan of training, constructs the training set for training the deep learning model；
Obtain each crucial point diagram to mark in the two-dimentional floor plan of training；
It, wait train two-dimentional floor plan as the input of the deep learning model, corresponding will will be marked in the training set
Output of the crucial point diagram as the deep learning model, training obtains the deep learning model.
Optionally, described according to the wall, include: by the step of two-dimentional floor plan generation three-dimensional house type model
The corresponding default three-dimensional wall of the wall is obtained, and determines first of the wall in the two-dimentional floor plan
Position；
According to the first position, determine the default three-dimensional wall in the corresponding second position of default three-dimensional platform；
The default three-dimensional wall is placed on the second position, three-dimensional house type model is constructed.
Optionally, the three-dimensional house type model generating method further include:
The three-dimensional house type model is shown in corresponding display interface；
It when receiving the scaling instruction based on the display interface, is instructed according to the scaling, to the three-dimensional house type
Model carries out scaling.
In addition, to achieve the above object, the present invention also provides a kind of three-dimensional house type model generating means, the three-dimensional house type
Model generating means include:
Module is obtained, for when receiving two-dimentional floor plan, obtaining the key point of the two-dimentional floor plan, the key
Point is turning point, tie point, bifurcation and/or the endpoint occurred in the two-dimentional floor plan；
Mark module, it is pre- by meeting in the area to be tested for determining the corresponding area to be tested of the key point
If the pixel of condition is labeled as wall pixel；
Determining module, according to the key point and the wall pixel, is determined for being based on breadth-first search
Wall in the two dimension floor plan；
Generation module, for according to the wall, the two-dimentional floor plan to be generated three-dimensional house type model.
In addition, to achieve the above object, the present invention also provides a kind of three-dimensional house type model generation device, the three-dimensional house type
Model generation device includes processor, memory and is stored in can be executed on the memory and by the processor three
House type model generator is tieed up, wherein realizing when the three-dimensional house type model generator is executed by the processor as above-mentioned
Three-dimensional house type model generating method the step of.
In addition, to achieve the above object, it is described computer-readable the present invention also provides a kind of computer readable storage medium
Three-dimensional house type model generator is stored on storage medium, wherein the three-dimensional house type model generator is executed by processor
When, it realizes such as the step of above-mentioned three-dimensional house type model generating method.
The present invention provides a kind of three-dimensional house type model generating method and obtains the two dimension when receiving two-dimentional floor plan
The key point of floor plan, the key point are turning point, tie point, bifurcation and/or the end occurred in the two-dimentional floor plan
Point；It determines the corresponding area to be tested of the key point, the pixel that preset condition is met in the area to be tested is marked
For wall pixel；The two dimension is determined according to the key point and the wall pixel based on breadth-first search
Wall in floor plan；According to the wall, the two-dimentional floor plan is generated into three-dimensional house type model.The present invention is by first extracting
Key point in two-dimentional floor plan gradually identifies pixel by starting point of key point around, by qualified pixel
Labeled as wall, finally according to wall, three-dimensional house type model is constructed, realizes the conversion of two-dimentional floor plan to three-dimensional house type model,
Facilitate the detailed understanding house type of user.
Detailed description of the invention
Fig. 1 is the hardware structural diagram of three-dimensional house type model generation device involved in the embodiment of the present invention；
Fig. 2 is the flow diagram of the three-dimensional house type model generating method first embodiment of the present invention；
Fig. 3 is the flow diagram of the three-dimensional house type model generating method second embodiment of the present invention；
Fig. 4 is the functional block diagram of the three-dimensional house type model generating means first embodiment of the present invention.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
The present embodiments relate to three-dimensional house type model generating method be mainly used in three-dimensional house type model generation device,
The three-dimensional house type model generation device can be the equipment that PC, portable computer, mobile terminal etc. have display and processing function.
Referring to Fig.1, Fig. 1 is that the hardware configuration of three-dimensional house type model generation device involved in the embodiment of the present invention shows
It is intended to.In the embodiment of the present invention, three-dimensional house type model generation device may include processor 1001 (such as CPU), communication bus
1002, user interface 1003, network interface 1004, memory 1005.Wherein, communication bus 1002 for realizing these components it
Between connection communication；User interface 1003 may include display screen (Display), input unit such as keyboard (Keyboard)；
Network interface 1004 optionally may include standard wireline interface and wireless interface (such as WI-FI interface)；Memory 1005 can
To be high speed RAM memory, it is also possible to stable memory (non-volatile memory), such as magnetic disk storage, deposits
Reservoir 1005 optionally can also be the storage device independently of aforementioned processor 1001.
Three-dimensional house type model is generated it will be understood by those skilled in the art that hardware configuration shown in Fig. 1 is not constituted
The restriction of equipment may include perhaps combining certain components or different component cloth than illustrating more or fewer components
It sets.
With continued reference to Fig. 1, the memory 1005 in Fig. 1 as a kind of computer readable storage medium may include operation system
System, network communication module and three-dimensional house type model generator.
In Fig. 1, network communication module is mainly used for connecting server, carries out data communication with server；And processor
1001 can call the three-dimensional house type model generator stored in memory 1005, and execute provided in an embodiment of the present invention three
Tie up house type model generating method.
The embodiment of the invention provides a kind of three-dimensional house type model generating methods.
It is the flow diagram of the three-dimensional house type model generating method first embodiment of the present invention referring to Fig. 2, Fig. 2.
In the present embodiment, it is described three-dimensional house type model generating method the following steps are included:
Step S10 obtains the key point of the two-dimentional floor plan when receiving two-dimentional floor plan, and the key point is
Turning point, tie point, bifurcation and/or the endpoint occurred in the two dimension floor plan；
Step S20 determines the corresponding area to be tested of the key point, will meet preset condition in the area to be tested
Pixel be labeled as wall pixel；
Step S30 is based on breadth-first search, according to the key point and the wall pixel, determine described in
Wall in two-dimentional floor plan；
The two-dimentional floor plan is generated three-dimensional house type model according to the wall by step S40.
The present embodiment is not since wall has apparent identification feature, so pass through the key in the two-dimentional floor plan of first extraction
Point gradually identifies pixel using key point as starting point around, and qualified pixel is labeled as wall, is finally obtained
Corresponding three-dimensional wall is taken, three-dimensional house type model is constructed.
Each step will be described in detail below:
Step S10 obtains the key point of the two-dimentional floor plan when receiving two-dimentional floor plan, and the key point is
Turning point, tie point, bifurcation and/or the endpoint occurred in the two dimension floor plan.
In the present embodiment, equipment, such as mobile phone is can be used in user, is scanned shooting to two-dimentional floor plan, or receive
The two-dimentional floor plan that another terminal transmission comes locates the two dimension floor plan when equipment receives two-dimentional floor plan in advance
Reason, wherein if the two dimension floor plan is the two-dimentional floor plan of scanning shoot, pretreatment includes background deletion and distortion correction
Deng, it is possible to understand that, band has powerful connections the two-dimentional floor plan of scanning shoot in most cases, and due to camera, or
It so that picture is generated distortion etc., these factors or can mostly will affect identification to two-dimentional floor plan less；If the two dimension family
Type figure is the two-dimentional floor plan that another terminal sends over, then pretreatment includes the finishing of resolution adjustment and picture size,
It should be understood that itself resolution ratio of different equipment is different, then the picture size of itself is also possible to not
The same, therefore need to pre-process two-dimentional floor plan, to obtain more clean two-dimentional floor plan, wherein after pretreatment
Two-dimentional floor plan include wall, house type component and markup information etc., house type component refers in addition to wall and markup information
Element, specifically include the soft dress such as sofa, TV, bed, dining table.It should be understood that the dress firmly such as built-in wardrobe and lavatory also belongs to family
Type component, i.e. house type component include hard dress and soft dress.
When equipment receives two-dimentional floor plan, it is based on deep learning model, is obtained from the two dimension floor plan corresponding
Kpoint (KeyPoint) key point, wherein the kpoint key point be the turning point occurred in two-dimentional floor plan, tie point,
Bifurcation and/or endpoint etc..It should be understood that wall is indicated by heavy black line in two-dimentional floor plan, such as when two walls are mutual
When connection, two line segment intersections are shown as in two-dimentional floor plan, there are tie point, i.e. the corresponding two sections of black line segments of two walls are total
There are three endpoint, one of endpoint, which shares, is used as tie point, and above-mentioned tie point and endpoint all can serve as kpoint key point.
Before this, equipment needs carry out the deep learning based on kpoint key point, to lead in two-dimentional floor plan
It crosses the convolutional neural networks such as densenet deep learning and extracts kpoint key point.
Specifically, step S10 includes:
The two-dimentional floor plan is inputted trained deep learning model when receiving two-dimentional floor plan by step S11
In, generate crucial point diagram, crucial point diagram instruction of the deep learning model based on the two-dimentional floor plan marked and corresponding mark
It gets.
The present embodiment inputs trained deep learning model when equipment receives two-dimentional floor plan, by two-dimentional floor plan
In, generate crucial point diagram, wherein the deep learning model is the key point based on the two-dimentional floor plan marked and corresponding mark
Figure training obtains.
Before the step S10, the method also includes:
Step a, obtains multiple and different to the two-dimentional floor plan of training, constructs the instruction for training the deep learning model
Practice collection；
Equipment obtain it is multiple and different to the two-dimentional floor plan of training, based on waiting for the two-dimentional floor plan of training by these, building
For training the training set of the deep learning model.It should be noted that equipment treats trained X-Y scheme during building
It is handled in advance, makes its size and uniform format, this is for the ease of batch processing.
Step b obtains each crucial point diagram to mark in the two-dimentional floor plan of training；
Equipment obtains user to each crucial point diagram to the two-dimentional floor plan mark of training, which is that user is prior
The result being labeled in two-dimentional floor plan.
Step c, using in the training set to the two-dimentional floor plan of training as the input of the deep learning model, will be right
Output of the crucial point diagram that should be marked as the deep learning model, training obtain the deep learning model.
, wait train two-dimentional floor plan as the input of deep learning model, the key point of mark will will be corresponded in training set
Scheme the output as deep learning model, training obtains deep learning model.
Need to illustrate when, for the accuracy for ensuring deep learning model, the two-dimentional floor plan in training set wants enough,
The present embodiment does not limit the quantity in training set to the two-dimentional floor plan of training, general to use million two-dimentional floor plans as one
Training set, in practical applications, more to the two-dimentional floor plan of training in training set, the crucial point diagram of deep learning model output
It is more accurate.
Step S12 extracts corresponding key point based on the crucial point diagram, wherein the key point is the two-dimentional family
Turning point, tie point, bifurcation and/or the endpoint occurred on type figure.
After obtaining crucial point diagram, equipment extracts the kpoint key point in crucial point diagram, it is possible to understand that, the key point
Figure is made of kpoint key point, wherein kpoint key point is turning point, tie point, the bifurcated occurred in two-dimentional floor plan
Point and/or endpoint etc..
Step S20 determines the corresponding area to be tested of the key point, will meet preset condition in the area to be tested
Pixel be labeled as wall pixel.
In the present embodiment, equipment determines corresponding area to be tested according to kpoint key point, it is possible to understand that, one
In Zhang Erwei floor plan, kpoint key point has multiple, therefore according to kpoint key point, determines that corresponding area to be tested also has
Multiple, i.e., each kpoint key point corresponds to an area to be tested.
Then, equipment will meet the pixel of preset condition labeled as wall pixel in area to be tested, it is possible to understand that
, in two-dimentional floor plan, since the color between wall, house type component and floor or background is different, therefore not
With the pixel on object, rgb value is different；Pixel only on same object, rgb value is just identical, such as wall
The rgb value of two pixels on body is identical.
Specifically, step S20 includes:
Step S21, using the key point as input terminal, using the input terminal as the center of circle, presetted pixel value is radius, structure
Build area to be tested.
Equipment is using kpoint key point as input terminal, using input terminal as the center of circle, presetted pixel value be radius, building to
Detection zone, wherein presetted pixel value the present embodiment is preferably 5 pixels, as having 18 key points in two-dimentional floor plan, then with 18
A key point constructs 18 area to be tested respectively using 5 pixels as radius for the center of circle.
Step S22 obtains the rgb value of the first pixel in the area to be tested, and according to the rgb value, determines institute
State the rgb median of rgb value.
Equipment obtains the rgb value of the first pixel in area to be tested, and (it is green that rgb:r, which is red (Red), g,
(Green), b is blue (Blue), rgb value refers to its brightness, is indicated from 0,1,2 ... until 255 with integer), wherein
First pixel refers to pixel all in area to be tested.
According to the corresponding rgb value of the first pixel, the rgb median of all rgb values is determined.
As having 5 pixels in a certain area to be tested, this corresponding rgb value of 5 pixels be respectively as follows: first point (0,
0,0), second point (0,0,0), thirdly (0,0,0), the 4th point (255,235,205) and the 5th point (255,235,205), then
The rgb median of the rgb value of this 5 pixels is (0,0,0), and it is black for representing the corresponding rgb median of the area to be tested
Color.
It should be noted that the reason of taking median here is, rgb value is by integer representation, therefore the present embodiment is not sought
Average value, but seek the corresponding rgb median of the area to be tested.
Step S23 is determined in the area to be tested, if there are unmarked equal with the rgb median of rgb value
The second pixel.
The corresponding rgb value of first pixel is compared with determining rgb median one by one, determines the area to be tested
In, if there are rgb value second pixels equal with rgb median, and second pixel is to be not labeled as wall
Pixel.
It should be noted that in actual process, since there may be a little color difference for two-dimentional floor plan, Europe can be passed through
Formula distance calculation formula, calculates the Euclidean distance in a certain pixel and the center of circle, if obtaining Euclidean distance less than 10, also thinks the picture
The rgb value of vegetarian refreshments is equal to the rgb median in the area to be tested.
Step S24, and if it exists, second pixel is labeled as wall pixel, and the wall pixel is made
For output end.
If it is determined that there are the second pixels that rgb value is equal to the rgb median that area to be tested gambling is won in area to be tested
Point, and second pixel is not labeled, then the second pixel is labeled as wall pixel, and the wall pixel is made
For output end, wherein the second pixel refers to that in the first pixel, rgb value is equal to the pixel of rgb median, quantity
Less than or equal to the quantity of the first pixel.
Step S25 using the output end as the input terminal, and is recycled described in execution using the input terminal as the center of circle,
Second pixel is labeled as wall pixel to if it exists by the step of presetted pixel value is radius, constructs area to be tested
Point, and using the wall pixel as the step of output end, until determining in the area to be tested, do not exist rgb value with
The equal unlabelled pixel of corresponding rgb median.
Hereafter, the step of executing above-mentioned S21-S24 is recycled, that is, is existed as the input terminal of subsequent cycle using this output end
In subsequent cycle, this output end, wall pixel will be used as input terminal, using the wall pixel as the center of circle, presetted pixel
Value is radius, constructs new area to be tested again, and obtain the corresponding rgb of all pixels point in the new area to be tested
Value, determines the rgb median of the rgb value of all pixels point in new area to be tested, and in these pixels, it is determined whether
There are the rgb value unlabelled pixels equal with rgb median, and if it exists, these pixels are also then labeled as wall picture
Vegetarian refreshments, and using the wall pixel of these new labels as output end, while it being re-used as the input terminal of subsequent cycle, it then proceedes to
Above-mentioned steps are recycled, until determining in area to be tested there is no the rgb value unlabelled picture equal with corresponding rgb median
Vegetarian refreshments.
It should be understood that in cyclic process, since current input terminal is the output end of a upper circulation, and area to be detected
The radius in domain is fixed, so, when being determined area to be tested as the center of circle using output end, which is necessarily also wrapped
The input terminal for having included a upper circulation, since the input terminal of a upper circulation has been marked as wall pixel, therefore in previous cycle
It is not repeated to mark.
That is the present embodiment constantly judges that the pixel for closing on direction is using kpoint key point as starting point vertically and horizontally
No is wall, if it is determined that pixel on certain direction is wall, which is labeled as wall pixel, and by current picture
After vegetarian refreshments is labeled as wall pixel, using the direction as extending direction, continues to judge and mark, until traversing in two-dimentional floor plan
All kpoint key points, kpoint key point and extension path, collectively form the wall in two-dimentional floor plan at this time.
It should be understood that once determining failure in one direction, that is, determine the rgb value of the pixel in a direction not
Equal to the rgb median in area to be tested where the pixel, then illustrate that the pixel is not wall, system will be no longer to this
Continue to extend judgement in direction.
Step S30 is based on breadth-first search, according to the key point and the wall pixel, determine described in
Wall in two-dimentional floor plan.
In the present embodiment, equipment traverses key point all in two-dimentional floor plan, determines according to breadth-first search
Marked all wall pixels, and then determine the wall in the two dimension floor plan, it is possible to understand that, wall pixel represents
Be a point of wall can determine two-dimentional floor plan therefore after wall pixel all in two-dimentional floor plan has been determined
In wall.
Specifically, step S30 includes:
Step S31 is based on breadth-first search, traverses key point all in the two-dimentional floor plan, successively really
The wall path constituted between fixed two key points with the presence or absence of the wall pixel；
Equipment traverses key point all in two-dimentional floor plan, is with two key points according to breadth-first search
Target determines the wall path constituted between the two key points with the presence or absence of wall pixel.
It should be understood that key point is endpoint or tie point of wall etc., i.e. in a wall, include at least two keys
Point, two key points of same wall, certainly exists wall path between the two, so by judging the pass in two-dimentional floor plan
Key point, two-by-two between the wall path that is constituted with the presence or absence of wall pixel, and then determine when whether the first two key point be same
The key point of wall, if so, when the first two key point and the wall path being linked to be between them are wall；If it is not,
It is not same wall that then explanation, which works as the first two key point, and wall path is not present in the line between them.
Step S32, and if it exists, then determine described two key points and the wall path is wall.
If it is determined that when there are the wall paths that wall pixel is constituted between the first two key point, it is determined that when the first two key
Point and wall path are wall.
Based on same method, wall all in two dimension floor plan is determined.
The two-dimentional floor plan is generated three-dimensional house type model according to the wall by step S40.
In the present embodiment, after the wall in two-dimentional floor plan has been determined, equipment generates two-dimentional floor plan according to wall
Three-dimensional house type model.
Specifically, step S40 includes:
Step S41 obtains the corresponding default three-dimensional wall of the wall, and determines the wall in the two-dimentional floor plan
In first position.
Equipment obtains the corresponding default three-dimensional wall of wall, and determines first position of the wall in two-dimentional floor plan,
In, default three-dimensional wall is built in advance, also, establishes the corresponding relationship of wall and three-dimensional wall in advance, and wall is being determined
After body, the default three-dimensional wall of acquisition can be corresponded to, and the size of three-dimensional wall then can be according to the mark in two-dimentional floor plan
Information is determined.Specifically, it should be understood that wall in two-dimentional floor plan, be indicated with black line segment, therefore
After the black line segment for representing wall in two-dimentional floor plan has been determined, further determine that this represents the length of the black line segment of wall,
Then can be according to the length of black line segment, in default three-dimensional library, the default three-dimensional wall of the corresponding length of acquisition, and the default three-dimensional
The width and height of wall then can uniformly be preset as a fixed value, this is because in real building course, a house type
Wall be usually contour and wide, therefore equipment can obtain corresponding default three-dimensional wall according to the wall in two-dimentional floor plan
Body.
Step S42 determines the default three-dimensional wall in default three-dimensional platform corresponding second according to the first position
Position.
Likewise, equipment stores default three-dimensional platform in advance, the default three-dimensional platform is corresponding with two-dimentional floor plan.Root
According to first position of the wall in two-dimentional floor plan, that is, it can determine default three-dimensional wall corresponding second on default three-dimensional platform
Position.
The default three-dimensional wall is placed on the second position, constructs three-dimensional house type model by step S43.
According to default three-dimensional wall, and the second position determined, default three-dimensional wall is placed on the second position, structure
Build three-dimensional house type model.
It should be noted that have existing for door and window as being on some walls, therefore the step further include: in two-dimentional floor plan
It is middle to identify the icon for representing door and window, door and window is opened up in the corresponding three-dimensional wall of the icon.Meanwhile the nerve by constructing in advance
Network model identifies the house type component in two-dimentional floor plan, obtains the position and direction of house type component, and obtains the family
The corresponding three-dimensional house type component of type component.
By three-dimensional wall, the three-dimensional wall with door and window and three-dimensional house type component, three-dimensional is combined into a manner of playing with building blocks
House type model.
The present embodiment provides a kind of three-dimensional house type model generating methods, i.e., when receiving two-dimentional floor plan, based on described
Two-dimentional floor plan obtains corresponding key point；Determine the corresponding area to be tested of the key point, it will be in the area to be tested
Meet the pixel of preset condition labeled as wall pixel；Based on breadth-first search, according to the key point and institute
Wall pixel is stated, determines the wall in the two-dimentional floor plan；According to the wall, the two-dimentional floor plan is generated three-dimensional
House type model.The present invention is gradually identified by first extracting the key point in two-dimentional floor plan using key point as starting point around
Qualified pixel is labeled as wall by pixel, finally according to wall, is constructed three-dimensional house type model, is realized two-dimentional family
Type figure facilitates the detailed understanding house type of user to the conversion of three-dimensional house type model.
Further, the second embodiment of the three-dimensional house type model generating method of the present invention is proposed based on first embodiment.Three
The difference for tieing up the second embodiment and the first embodiment of three-dimensional house type model generating method of house type model generating method is, joins
According to Fig. 3, three-dimensional house type model generating method further include:
The three-dimensional house type model is shown in corresponding display interface by step S50；
Step S60 is instructed, to described when receiving the scaling instruction based on the display interface according to the scaling
Three-dimensional house type model carries out scaling.
In the present embodiment, after generating three-dimensional house type model, three-dimensional house type model can be shown in corresponding display circle
Face, and when receiving scaling instruction, scaling is carried out to the three-dimensional house type model, so that user will appreciate that the thin of the house type
Section.
Each step will be described in detail below:
The three-dimensional house type model is shown in corresponding display interface by step S50.
In the present embodiment, after generating three-dimensional house type model, three-dimensional house type model can be shown in display circle of this equipment
Face generates json file such as the display interface of mobile phone, or by the three-dimensional house type model, is transferred to corresponding receiving device, with
For the receiving device by the three-dimensional house type model be shown or 3d print, can also by the three-dimensional house type model projection to pair
On the screen answered.
Step S60 is instructed, to described when receiving the scaling instruction based on the display interface according to the scaling
Three-dimensional house type model carries out scaling.
In the present embodiment, in the amplification that, when display interface is shown, receivable user assigns by three-dimensional house type model, reduce
Or rotation instruction, and according to the instruction, it corresponds to and executes amplification, diminution or rotation process, enable users to recognize current family
Each corner of type.
The present embodiment can receive what user assigned on the basis of the three-dimensional house type model after generating three-dimensional house type model
Scaling instruction, and the corresponding scaling that executes operates, it is raw to improve three-dimensional house type model for the current house type of the understanding for keeping user more careful
At the intelligence of method.
In addition, the embodiment of the present invention also provides a kind of three-dimensional house type model generating means.
It is the functional block diagram of the three-dimensional house type model generating means first embodiment of the present invention referring to Fig. 4, Fig. 4.
In the present embodiment, the three-dimensional house type model generating means include:
Module 10 is obtained, for when receiving two-dimentional floor plan, obtaining the key point of the two-dimentional floor plan, the pass
Key point is turning point, tie point, bifurcation and/or the endpoint occurred in the two-dimentional floor plan；
Mark module 20 will meet for determining the corresponding area to be tested of the key point in the area to be tested
The pixel of preset condition is labeled as wall pixel；
Determining module 30, for being based on breadth-first search, according to the key point and the wall pixel, really
Wall in the fixed two-dimentional floor plan；
Generation module 40, for according to the wall, the two-dimentional floor plan to be generated three-dimensional house type model.
Further, the mark module 20 specifically includes:
First construction unit, for using the key point as input terminal, using the input terminal as the center of circle, presetted pixel value
For radius, area to be tested is constructed；
First determination unit, for obtaining the rgb value of the first pixel in the area to be tested, and according to the rgb
Value, determines the rgb median of institute's rgb value；
Second determination unit, for determining in the area to be tested, if there are rgb values and the rgb median phase
Deng unlabelled second pixel；
Marking unit, for if it exists, will second pixel labeled as wall pixel, and by the wall pixel
Point is used as output end；
Cycling element is used for using the output end as the input terminal, and is recycled described in execution and be with the input terminal
Second pixel is labeled as wall to if it exists by the step of center of circle, presetted pixel value is radius, constructs area to be tested
Body image vegetarian refreshments, and using the wall pixel as the step of output end, until determining in the area to be tested, do not exist
The rgb value unlabelled pixel equal with corresponding rgb median.
Further, the determining module 30 specifically includes:
Third determination unit traverses key all in the two-dimentional floor plan for being based on breadth-first search
Point successively determines the wall path constituted between two key points with the presence or absence of the wall pixel；
4th determination unit, for if it exists, it is determined that described two key points and the wall path are wall.
Further, the acquisition module 10 specifically includes:
Generation unit, for when receiving two-dimentional floor plan, the two-dimentional floor plan to be inputted trained depth
It practises in model, generates crucial point diagram, key of the deep learning model based on the two-dimentional floor plan marked and corresponding mark
Point diagram training obtains；
Extraction unit, for extracting corresponding key point, wherein the key point is described based on the crucial point diagram
Turning point, tie point, bifurcation and/or the endpoint occurred in two-dimentional floor plan.
Further, the deep learning model specifically includes:
Second construction unit, it is multiple and different to the two-dimentional floor plan of training for obtaining, it constructs for training the depth
The training set of learning model；
First acquisition unit, for obtaining each crucial point diagram to mark in the two-dimentional floor plan of training；
Training unit, for using in the training set to the two-dimentional floor plan of training as the defeated of the deep learning model
Enter, using the crucial point diagram of corresponding mark as the output of the deep learning model, training obtains the deep learning model.
Further, the generation module 40 specifically includes:
Second acquisition unit for obtaining the corresponding default three-dimensional wall of the wall, and determines the wall described
First position in two-dimentional floor plan；
5th determination unit, for determining the default three-dimensional wall in default three-dimensional platform according to the first position
The corresponding second position；
Third construction unit constructs three-dimensional house type for the default three-dimensional wall to be placed on the second position
Model.
Further, the three-dimensional house type model generating means further include:
Display module, for the three-dimensional house type model to be shown in corresponding display interface；
Scaling module is right for being instructed according to the scaling when receiving the scaling instruction based on the display interface
The three-dimensional house type model carries out scaling.
Wherein, modules and unit and above-mentioned three-dimensional house type model generation side in above-mentioned three-dimensional house type model generating means
Each step is corresponding in method embodiment, and function and realization process no longer repeat one by one here.
In addition, the embodiment of the present invention also provides a kind of computer readable storage medium.
Three-dimensional house type model generator is stored on computer readable storage medium of the present invention, wherein the three-dimensional house type
When model generator is executed by processor, realize such as the step of above-mentioned three-dimensional house type model generating method.
Wherein, three-dimensional house type model generator, which is performed realized method, can refer to three-dimensional house type model of the invention
Each embodiment of generation method, details are not described herein again.
It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, method, article or the system that include a series of elements not only include those elements, and
And further include other elements that are not explicitly listed, or further include for this process, method, article or system institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do
There is also other identical elements in the process, method of element, article or system.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words does the prior art
The part contributed out can be embodied in the form of software products, which is stored in one as described above
In storage medium (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that terminal device (it can be mobile phone,
Computer, server, air conditioner or network equipment etc.) execute method described in each embodiment of the present invention.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of three-dimensional house type model generating method, which is characterized in that the three-dimensional house type model generating method includes following step
It is rapid:
When receiving two-dimentional floor plan, the key point of the two-dimentional floor plan is obtained, the key point is the two-dimentional house type
Turning point, tie point, bifurcation and/or the endpoint occurred on figure；
It determines the corresponding area to be tested of the key point, the pixel that preset condition is met in the area to be tested is marked
For wall pixel；
It is determined in the two-dimentional floor plan based on breadth-first search according to the key point and the wall pixel
Wall；
According to the wall, the two-dimentional floor plan is generated into three-dimensional house type model.
2. three-dimensional house type model generating method as described in claim 1, which is characterized in that the determination key point is corresponding
Area to be tested, the step of pixel of preset condition is labeled as wall pixel packet will be met in the area to be tested
It includes:
Using the key point as input terminal, using the input terminal as the center of circle, presetted pixel value is radius, constructs area to be detected
Domain；
The rgb value of the first pixel in the area to be tested is obtained, and according to the rgb value, determines the rgb of the rgb value
Median；
It determines in the area to be tested, if there are rgb value unlabelled second pixels equal with the rgb median
Point；
If it exists, second pixel is labeled as wall pixel, and using the wall pixel as output end；
It using the output end as the input terminal, and recycles described in execution using the input terminal as the center of circle, presetted pixel value is
Second pixel is labeled as wall pixel, and will be described to if it exists by the step of radius, building area to be tested
There is no rgb value and position in corresponding rgb until determining in the area to be tested in the step of wall pixel is as output end
The equal unlabelled pixel of number.
3. three-dimensional house type model generating method as described in claim 1, which is characterized in that the key point includes at least two
It is a, it is described to be based on breadth-first search, according to the key point and the wall pixel, determine the two-dimentional floor plan
In wall the step of include:
Based on breadth-first search, key point all in the two-dimentional floor plan is traversed, successively determines two key points
Between with the presence or absence of the wall pixel constitute wall path；
If it exists, it is determined that described two key points and the wall path are wall.
4. three-dimensional house type model generating method as described in claim 1, which is characterized in that described to receive two-dimentional floor plan
When, the key point of the two-dimentional floor plan is obtained, the key point is the turning point occurred in the two-dimentional floor plan, connection
The step of point, bifurcation and/or endpoint includes:
When receiving two-dimentional floor plan, the two-dimentional floor plan is inputted in trained deep learning model, is generated crucial
Point diagram, the deep learning model are obtained based on the crucial point diagram training of the two-dimentional floor plan marked and corresponding mark；
Based on the crucial point diagram, corresponding key point is extracted, wherein occur in the key point two-dimentional floor plan
Turning point, tie point, bifurcation and/or endpoint.
5. three-dimensional house type model generating method as claimed in claim 4, which is characterized in that described to receive two-dimentional floor plan
When, based on the two-dimentional floor plan, before the step of obtaining corresponding key point, the method also includes:
It obtains multiple and different to the two-dimentional floor plan of training, constructs the training set for training the deep learning model；
Obtain each crucial point diagram to mark in the two-dimentional floor plan of training；
, wait train two-dimentional floor plan as the input of the deep learning model, the pass of mark will will be corresponded in the training set
Output of the key point diagram as the deep learning model, training obtain the deep learning model.
6. three-dimensional house type model generating method as described in claim 1, which is characterized in that it is described according to the wall, by institute
Stating the step of two-dimentional floor plan generates three-dimensional house type model includes:
The corresponding default three-dimensional wall of the wall is obtained, and determines first of the wall in the two-dimentional floor plan
It sets；
According to the first position, determine the default three-dimensional wall in the corresponding second position of default three-dimensional platform；
The default three-dimensional wall is placed on the second position, three-dimensional house type model is constructed.
7. the three-dimensional house type model generating method as described in claim 1 to 6 any one, which is characterized in that described according to institute
After the step of stating wall, the two-dimentional floor plan is generated three-dimensional house type model, the three-dimensional house type model generating method is also
Include:
The three-dimensional house type model is shown in corresponding display interface；
It when receiving the scaling instruction based on the display interface, is instructed according to the scaling, to the three-dimensional house type model
Carry out scaling.
8. a kind of three-dimensional house type model generating means, which is characterized in that the three-dimensional house type model generating means include:
Module is obtained, for when receiving two-dimentional floor plan, obtaining the key point of the two-dimentional floor plan, the key point is
Turning point, tie point, bifurcation and/or the endpoint occurred in the two dimension floor plan；
Mark module will meet default item for determining the corresponding area to be tested of the key point in the area to be tested
The pixel of part is labeled as wall pixel；
Determining module, for being based on breadth-first search, according to the key point and the wall pixel, determine described in
Wall in two-dimentional floor plan；
Generation module, for according to the wall, the two-dimentional floor plan to be generated three-dimensional house type model.
9. a kind of three-dimensional house type model generation device, which is characterized in that the three-dimensional house type model generation device include processor,
Memory and it is stored in the three-dimensional house type model generator that can be executed on the memory and by the processor, wherein
When the three-dimensional house type model generator is executed by the processor, three as described in any one of claims 1 to 7 are realized
The step of tieing up house type model generating method.
10. a kind of computer readable storage medium, which is characterized in that be stored with three-dimensional family on the computer readable storage medium
Type model generator, wherein realizing such as claim 1 to 7 when the three-dimensional house type model generator is executed by processor
Any one of described in three-dimensional house type model generating method the step of.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910292777.2A CN110163954A (en) | 2019-04-12 | 2019-04-12 | Three-dimensional house type model generating method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910292777.2A CN110163954A (en) | 2019-04-12 | 2019-04-12 | Three-dimensional house type model generating method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110163954A true CN110163954A (en) | 2019-08-23 |
Family
ID=67639037
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910292777.2A Pending CN110163954A (en) | 2019-04-12 | 2019-04-12 | Three-dimensional house type model generating method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110163954A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110706230A (en) * | 2019-10-29 | 2020-01-17 | 国网黑龙江省电力有限公司电力科学研究院 | Tower abnormity automatic detection method based on prior information |
| CN111008484A (en) * | 2019-12-24 | 2020-04-14 | 深圳市蓝图速配科技有限公司 | Processing method for converting two-dimensional house type graph into three-dimensional scene graph |
| CN111027118A (en) * | 2019-11-19 | 2020-04-17 | 广东博智林机器人有限公司 | Actual measurement actual quantity task point searching and task dispatching method and system |
| CN111080518A (en) * | 2019-11-27 | 2020-04-28 | 贝壳技术有限公司 | Point cloud image splicing method and device and storage medium |
| CN113239785A (en) * | 2021-05-11 | 2021-08-10 | 百安居信息技术(上海)有限公司 | Method, system, storage medium and electronic device for automatically identifying house type based on convolutional neural network |
| CN113240768A (en) * | 2019-12-18 | 2021-08-10 | 北京城市网邻信息技术有限公司 | House type graph processing method, system, device and computer readable storage medium |
| CN113642092A (en) * | 2021-10-18 | 2021-11-12 | 西南交通大学 | Building space path capturing method |
| CN114092642A (en) * | 2021-11-18 | 2022-02-25 | 北京字节跳动网络技术有限公司 | Three-dimensional house type model generation method, device and equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150077282A1 (en) * | 2013-09-17 | 2015-03-19 | Farrokh Mohamadi | Real-time, two dimensional (2-d) tracking of first responders with identification inside premises |
| CN105279787A (en) * | 2015-04-03 | 2016-01-27 | 北京明兰网络科技有限公司 | Method for generating three-dimensional (3D) building model based on photographed house type image identification |
| CN108133116A (en) * | 2018-01-15 | 2018-06-08 | 北京居然设计家家居连锁集团有限公司 | CAD floor plan wall thickness automatic identifying methods and system |
| CN108717726A (en) * | 2018-05-11 | 2018-10-30 | 北京家印互动科技有限公司 | Three-dimensional house type model generating method and device |
| CN108717188A (en) * | 2018-05-29 | 2018-10-30 | 成都理工大学 | Quick wall compensation method suitable for MIMO through-wall radar imagings |
-
2019
- 2019-04-12 CN CN201910292777.2A patent/CN110163954A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150077282A1 (en) * | 2013-09-17 | 2015-03-19 | Farrokh Mohamadi | Real-time, two dimensional (2-d) tracking of first responders with identification inside premises |
| CN105279787A (en) * | 2015-04-03 | 2016-01-27 | 北京明兰网络科技有限公司 | Method for generating three-dimensional (3D) building model based on photographed house type image identification |
| CN108133116A (en) * | 2018-01-15 | 2018-06-08 | 北京居然设计家家居连锁集团有限公司 | CAD floor plan wall thickness automatic identifying methods and system |
| CN108717726A (en) * | 2018-05-11 | 2018-10-30 | 北京家印互动科技有限公司 | Three-dimensional house type model generating method and device |
| CN108717188A (en) * | 2018-05-29 | 2018-10-30 | 成都理工大学 | Quick wall compensation method suitable for MIMO through-wall radar imagings |
Non-Patent Citations (1)
| Title |
|---|
| 朱俊芳: "基于结构构件识别的户型图三维重建算法研究", 中国优秀硕士学位论文全文数据库信息科技辑, no. 07, pages 138 - 639 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110706230A (en) * | 2019-10-29 | 2020-01-17 | 国网黑龙江省电力有限公司电力科学研究院 | Tower abnormity automatic detection method based on prior information |
| CN111027118A (en) * | 2019-11-19 | 2020-04-17 | 广东博智林机器人有限公司 | Actual measurement actual quantity task point searching and task dispatching method and system |
| CN111027118B (en) * | 2019-11-19 | 2024-01-19 | 广东博智林机器人有限公司 | Actual measurement real-quantity task point searching and task dispatching method and system |
| CN111080518A (en) * | 2019-11-27 | 2020-04-28 | 贝壳技术有限公司 | Point cloud image splicing method and device and storage medium |
| CN111080518B (en) * | 2019-11-27 | 2023-10-20 | 贝壳技术有限公司 | Point cloud image stitching method and device and storage medium |
| CN113240768A (en) * | 2019-12-18 | 2021-08-10 | 北京城市网邻信息技术有限公司 | House type graph processing method, system, device and computer readable storage medium |
| CN113240768B (en) * | 2019-12-18 | 2022-03-15 | 北京城市网邻信息技术有限公司 | House type graph processing method, system, device and computer readable storage medium |
| CN111008484A (en) * | 2019-12-24 | 2020-04-14 | 深圳市蓝图速配科技有限公司 | Processing method for converting two-dimensional house type graph into three-dimensional scene graph |
| CN113239785A (en) * | 2021-05-11 | 2021-08-10 | 百安居信息技术(上海)有限公司 | Method, system, storage medium and electronic device for automatically identifying house type based on convolutional neural network |
| CN113642092A (en) * | 2021-10-18 | 2021-11-12 | 西南交通大学 | Building space path capturing method |
| CN114092642A (en) * | 2021-11-18 | 2022-02-25 | 北京字节跳动网络技术有限公司 | Three-dimensional house type model generation method, device and equipment |
| CN114092642B (en) * | 2021-11-18 | 2024-01-26 | 抖音视界有限公司 | Three-dimensional house type model generation method, device and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110163954A (en) | Three-dimensional house type model generating method, device, equipment and storage medium | |
| Li et al. | PDR-Net: Perception-inspired single image dehazing network with refinement | |
| CN109508678B (en) | Training method of face detection model, and detection method and device of face key points | |
| CN102184221B (en) | Real-time video abstract generation method based on user preferences | |
| Zollhöfer et al. | Automatic reconstruction of personalized avatars from 3D face scans | |
| CN104601964B (en) | Pedestrian target tracking and system in non-overlapping across the video camera room of the ken | |
| WO2019210555A1 (en) | People counting method and device based on deep neural network and storage medium | |
| KR20190018483A (en) | Internet-based facial beautification system | |
| CN110399825A (en) | Facial expression moving method, device, storage medium and computer equipment | |
| CN109446981A (en) | A kind of face's In vivo detection, identity identifying method and device | |
| CN103425964B (en) | Image processing equipment and image processing method | |
| CN106384383A (en) | RGB-D and SLAM scene reconfiguration method based on FAST and FREAK feature matching algorithm | |
| CN106101561A (en) | Camera focusing detection method and device | |
| CN109359527B (en) | Hair region extraction method and system based on neural network | |
| CN111814620A (en) | Face image quality evaluation model establishing method, optimization method, medium and device | |
| CN110334605A (en) | Gesture identification method, device, storage medium and equipment neural network based | |
| CN103299342A (en) | Method and apparatus for providing a mechanism for gesture recognition | |
| CN109410316A (en) | Method, tracking, relevant apparatus and the storage medium of the three-dimensional reconstruction of object | |
| CN110136244A (en) | Three-dimensional house type model generating method, device, equipment and storage medium | |
| CN107247920A (en) | Interaction control method, device and computer-readable recording medium | |
| CN112001274A (en) | Crowd density determination method, device, storage medium and processor | |
| CN108961227A (en) | A kind of image quality evaluating method based on airspace and transform domain multiple features fusion | |
| CN109685829A (en) | Eye-controlling focus method, apparatus and electronic equipment based on image | |
| Vidanapathirana et al. | Plan2scene: Converting floorplans to 3d scenes | |
| CN109740527B (en) | Image processing method in video frame |
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 |