CN105844615A - Method of identifying regular rectangular area and apparatus thereof - Google Patents
Method of identifying regular rectangular area and apparatus thereof Download PDFInfo
- Publication number
- CN105844615A CN105844615A CN201610153207.1A CN201610153207A CN105844615A CN 105844615 A CN105844615 A CN 105844615A CN 201610153207 A CN201610153207 A CN 201610153207A CN 105844615 A CN105844615 A CN 105844615A
- Authority
- CN
- China
- Prior art keywords
- edge
- widthwise
- group
- longitudinal edge
- longitudinal
- 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.)
- Granted
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a method of identifying a regular rectangular area and an apparatus thereof. The method comprises the following steps of carrying out edge point detection operation on an image containing the regular rectangular area so as to acquire edge points in the image; according to distribution of the acquired edge points, acquiring a transverse edge set including at least two transverse edges and a longitudinal edge set including at least two longitudinal edges; and constructing at least one transverse edge pair satisfying a predetermined condition from the transverse edge set. Each transverse edge pair is verified, the transverse edge pair which passes the verification is the effective transverse edge pair and the rectangular area determined by the effective transverse edge pair is the regular rectangular area. In the technical scheme in the invention, the regular rectangular area in the image can be accurately identified.
Description
Technical field
It relates to image identification technical field, particularly relate to method and the dress of recognition rule rectangular area
Put.
Background technology
Developing rapidly recently as automatic test, various automated test tools emerge in an endless stream.Mesh
Before have a lot of Web User Interface automated test frame on the market, such as WatiN,
Selinimu, WebDriver, also Coded UI in VS2010 etc..These frameworks can be known
Not and operate the control in Web, the operations such as analog subscriber inputs, click, it is achieved Web automatization surveys
Examination.
But the automatic test stability being currently based on user interface is the highest, the figure that correlation technique is used
As identification technology mainly uses Feature Points Matching mode, processing speed is relatively slow, and for special
During demand region segmentation, the most only segmentation regular rectangular shape region, it is impossible to reach good segmentation effect.
Additionally, in other much application, there is also the demand of recognition rule rectangular area.
Summary of the invention
The disclosure provides the method and device of a kind of recognition rule rectangular area, using the teaching of the invention it is possible to provide higher identification
Accuracy.
One side according to the disclosure, it is provided that a kind of method of recognition rule rectangular area, including: to bag
Image containing regular rectangular shape region carries out endpoint detections computing and obtains the marginal point in image；According to obtaining
The distribution of the described marginal point taken obtains including the widthwise edge group of at least two widthwise edges, and/or according to obtaining
The distribution of the described marginal point taken obtains including the longitudinal edge group of at least two longitudinal edges；From described widthwise edge
In group, structure meets at least one widthwise edge pair of predetermined condition, and/or constructs from described longitudinal edge group
Meet at least one longitudinal edge pair of predetermined condition；To each widthwise edge to verifying, by checking
Widthwise edge is to for effective widthwise edge pair, and the described effective widthwise edge rectangular area to determining is described rule square
Shape region；Or to limit longitudinally in each to verifying, by the longitudinal edge of checking to for effective longitudinal edge
Right, the described effective longitudinal edge rectangular area to determining is described regular rectangular shape region.
According to some embodiments, the image comprising regular rectangular shape region is carried out endpoint detections computing and obtains
The marginal point taken in image includes: utilize preset algorithm image to carry out endpoint detections computing and acquisition figure
Marginal point in Xiang, wherein said preset algorithm includes Laplacian operator, Canny operator, Sobel
Operator, Prewitt operator, Kirsch operator, compass operator or Roberts Cross operator.
According to some embodiments, the image comprising regular rectangular shape region is carried out endpoint detections computing and obtains
The marginal point taken in image includes: image is carried out smothing filtering；Image after smothing filtering is transported
Calculate；The extreme point of detection calculations is to obtain the marginal point in image.
According to some embodiments, image is carried out smothing filtering and includes: use Gaussian function that image is carried out
Smothing filtering.
According to some embodiments, obtain including at least two laterally according to the distribution of the described marginal point obtained
The widthwise edge group on limit includes: carry out being grouped to obtain multiple widthwise edge according to ordinate value by described marginal point
Edge point group, for each transverse edge point group, if the continuous boundary point that described transverse edge point group comprises
Number is more than first threshold, then added as a widthwise edge by the line segment that described transverse edge point group is constituted
Enter described widthwise edge group.
According to some embodiments, obtain including at least two longitudinal directions according to the distribution of the described marginal point obtained
The longitudinal edge group on limit includes: carry out being grouped to obtain multiple longitudinal edge according to abscissa value by described marginal point
Edge point group, for each longitudinal edge point group, if the continuous boundary point that described longitudinal edge point group comprises
Number is more than Second Threshold, then added as a longitudinal edge by the line segment that described longitudinal edge point group is constituted
Enter described longitudinal edge group.
According to some embodiments, from described widthwise edge group, to meet at least one of predetermined condition horizontal for structure
While to including: for wantonly two widthwise edges in described widthwise edge group, if it is determined that described two widthwise edges
Initial abscissa value difference and terminate abscissa value difference be respectively less than the 3rd threshold value, and described two horizontal strokes
To the difference of the ordinate value on limit in the 4th threshold range, then using described two widthwise edges as one laterally
Limit pair.
According to some embodiments, from described longitudinal edge group, structure meets at least one longitudinal direction of predetermined condition
While to including: for wantonly two longitudinal edges in described longitudinal edge group, if it is determined that described two longitudinal edges
The difference of initial ordinate value and terminate the difference of ordinate value and be respectively less than the 5th threshold value, and described two are indulged
To the difference of the abscissa value on limit in the 6th threshold range, then using described two longitudinal edges as a longitudinal direction
Limit pair.
According to some embodiments, each widthwise edge is included carrying out checking: find and described widthwise edge pair
Corresponding longitudinal edge pair, if there is corresponding longitudinal edge pair, the most described widthwise edge is to for effective widthwise edge
Right.
According to some embodiments, limit longitudinally in each is included carrying out checking: find and described longitudinal edge pair
Corresponding widthwise edge pair, if there is corresponding widthwise edge pair, the most described longitudinal edge is to for effective longitudinal edge
Right.
According to a further aspect in the invention, it is provided that the device of a kind of recognition rule rectangular area, including limit
Edge point acquisition module, for carrying out endpoint detections computing to the image comprising regular rectangular shape region and obtaining
Marginal point in image；Limit group acquisition module, for being wrapped according to the distribution of the described marginal point obtained
Include the widthwise edge group of at least two widthwise edges, and/or wrapped according to the distribution of the described marginal point obtained
Include the longitudinal edge group of at least two longitudinal edges；While to constructing module, for constructing from described widthwise edge group
Meet at least one widthwise edge pair of predetermined condition, and/or structure meets predetermined from described longitudinal edge group
At least one longitudinal edge pair of condition；Authentication module, is used for each widthwise edge verifying, passes through
The widthwise edge of checking is to for effective widthwise edge pair, and the described effective widthwise edge rectangular area to determining is described
Regular rectangular shape region；Or to limit longitudinally in each to verifying, by the longitudinal edge of checking to for effectively
Longitudinal edge pair, the described effective longitudinal edge rectangular area to determining is described regular rectangular shape region.
According to some embodiments, described marginal point acquisition module is configured to utilize preset algorithm to carry out image
Endpoint detections computing and obtain the marginal point in image, wherein said preset algorithm includes Laplacian
Operator, Canny operator, Sobel operator, Prewitt operator, Kirsch operator, compass operator or
Roberts Cross operator.
According to some embodiments, described marginal point acquisition module is configured that and image is carried out smothing filtering；
Image after smothing filtering is carried out computing；The extreme point of detection calculations is to obtain the marginal point in image.
According to some embodiments, described marginal point acquisition module is configured to use Gaussian function to carry out image
Smothing filtering.
According to some embodiments, described limit group acquisition module is configured that described marginal point according to vertical coordinate
Value carries out being grouped to obtain multiple transverse edge point group, for each transverse edge point group, if described horizontal stroke
The continuous boundary comprised to group of edge points is counted out more than first threshold, then by described transverse edge point group structure
The line segment become adds described widthwise edge group as a widthwise edge；And/or by described marginal point according to horizontal stroke
Coordinate figure carries out being grouped to obtain multiple longitudinal edge point group, for each longitudinal edge point group, if institute
State the continuous boundary that longitudinal edge point group comprises and count out more than Second Threshold, then by described longitudinal edge point
The line segment that group is constituted adds described longitudinal edge group as a longitudinal edge.
According to some embodiments, constructing module was configured that for appointing in described widthwise edge group by described limit
Article two, widthwise edge, if it is determined that the difference of the initial abscissa value of described two widthwise edges and termination abscissa
The difference of value is respectively less than the 3rd threshold value, and the difference of the ordinate value of described two articles of widthwise edges is at the 4th threshold range
In, then using described two widthwise edges as a widthwise edge pair；And/or in described longitudinal edge group
Wantonly two longitudinal edges, if it is determined that the difference of the initial ordinate value of described two longitudinal edges and the vertical seat of termination
The difference of scale value is respectively less than the 5th threshold value, and the difference of the abscissa value of described two articles of longitudinal edges is at the 6th threshold value model
In enclosing, then using described two longitudinal edges as a longitudinal edge pair.
According to some embodiments, described authentication module is configured that to find and indulges corresponding with described widthwise edge
To limit pair, if there is corresponding longitudinal edge pair, it is determined that described widthwise edge is to for effective widthwise edge pair；
Or find with described longitudinal edge corresponding widthwise edge pair, if there is corresponding widthwise edge pair, the most really
Fixed described longitudinal edge is to for effective longitudinal edge pair.
Technical scheme according to some example embodiment of the disclosure, it is possible to accurately identify the regular square in image
Shape region.
It should be appreciated that it is only exemplary and explanatory that above general description and details hereinafter describe
, the disclosure can not be limited.
Accompanying drawing explanation
Accompanying drawing herein is merged in description and constitutes the part of this specification, it is shown that meet this
Bright embodiment, and for explaining the principle of the present invention together with description.
Fig. 1 illustrates the method flow diagram of the recognition rule rectangular area according to an illustrative embodiments；
Fig. 2 illustrates the method flow diagram of the recognition rule rectangular area according to another exemplary embodiment；
Fig. 3 illustrates the method flow diagram of the recognition rule rectangular area according to another exemplary embodiment；
Fig. 4 illustrate according to the present invention one example embodiment to the widthwise edge process to verifying；
Fig. 5 illustrate according to another example embodiment of the present invention to the longitudinal edge process to verifying；
Fig. 6 illustrate according to another example embodiment of the present invention to the widthwise edge process to verifying；
Fig. 7 illustrates the block diagram of the device of the recognition rule rectangular area according to an illustrative embodiments.
Detailed description of the invention
It is described more fully with example embodiment referring now to accompanying drawing.But, example embodiment can be with many
The form of kind is implemented, and is not understood as limited to embodiment set forth herein；On the contrary, it is provided that these are implemented
Example makes the present invention will fully and completely, and the design of example embodiment is conveyed to this area all sidedly
Technical staff.The most identical reference represents same or similar part, thus by omission to it
Repeated description.
Additionally, described feature, structure or characteristic can be combined in any suitable manner one or
In more embodiments.In the following description, it is provided that many details thus provide the reality to the present invention
Execute fully understanding of example.It will be appreciated, however, by one skilled in the art that the technology of the present invention can be put into practice
Scheme and do not have in described specific detail one or more, or other method, group can be used
Unit, device, etc..In other cases, be not shown in detail or describe known features, method, device,
Realize or operation is to avoid fuzzy each aspect of the present invention.
Block diagram shown in accompanying drawing is only functional entity, it is not necessary to must be with physically separate entity
Corresponding.I.e., it is possible to use software form to realize these functional entitys, or at one or more softwares
The module of hardening realizes the part of these functional entitys or functional entity, or at heterogeneous networks and/or
Processor device and/or microcontroller device realize these functional entitys.
Rim detection is the basic problem in image procossing and computer vision, and the purpose of rim detection is mark
Know the obvious point of brightness flop in digital picture.Notable change in image attributes generally reflects attribute
Critical event and change.Different images gray scale is different, and boundary typically has obvious edge, utilizes this
Feature can split image.But for specific demand region segmentation, the most only split regular rectangular shape
Region, only relies on edge detection algorithm can not reach good segmentation effect.Present inventive concept provides
A kind of image-recognizing method, can on the basis of edge detection algorithm the rectangular area of recognition rule.
Fig. 1 illustrates the method flow diagram of the recognition rule rectangular area according to an exemplary embodiment.This reality
Execute example to be applicable to identify the rule button in image or there is the regular icon etc. of limbus, but this reality
The application executing example is not limited to this.
As it is shown in figure 1, at S110, the image comprising regular rectangular shape region is carried out endpoint detections computing
And obtain the marginal point in image.
The operation that image carries out the marginal point in endpoint detections computing acquisition image can use multiple side
Method, the such as available Second-order Operator preset or single order operator carry out endpoint detections computing acquisition to image
Marginal point in image.Second-order Operator can be Laplacian operator and Canny operator etc., one
Rank operator can be Sobel operator, Prewitt operator, Kirsch operator, compass operator and
Roberts Cross operator etc..Such as, Laplacian operator utilizes second dervative information, have each to
The same sex, i.e. unrelated with change in coordinate axis direction, after coordinate axes rotates, gradient result is constant.Make image through two
After the differential of rank, produce a precipitous zero cross point in edge, can determine whether limit according to this zero cross point
Edge.The most such as, Canny edge detection operator is that John F.Canny developed in 1986
One multistage edge detection algorithm, signal to noise ratio is estimated with location product, is obtained optimization and approach by it
Operator.
According to an embodiment, obtain marginal point and can use mode described below, but the invention is not restricted to
This.
First image is carried out smothing filtering.Filter function can be to select Gauss according to human visual system
Function, but the invention is not restricted to this.
Then, the image after smothing filtering is carried out computing, and the extreme point of detection calculations.To use
As a example by Laplacian operator obtains the marginal point in image, to the smoothed image after smothing filtering
Carrying out Laplace's operation, the extreme point of detection Laplace's operation is to obtain the marginal point in image.
That is, using the zero cross point of second dervative as extreme point, and the bigger peak value of corresponding first derivative, thus
Obtain the marginal point in image.
In S120, according to the distribution of the marginal point obtained, obtain including the horizontal stroke of at least two widthwise edges
To limit group.
Various method can be used to obtain widthwise edge group.Such as, can be by acquired marginal point according to vertical seat
Scale value carries out being grouped to obtain multiple transverse edge point group, for each transverse edge point group, if this is horizontal
The continuous boundary that group of edge points comprises is counted out more than a first threshold, then this transverse edge point group constituted
Line segment be included into widthwise edge group as a widthwise edge.Depending on this first threshold can be according to concrete application,
Such as can use empirical value or statistical value etc. (lower same).
According to some embodiments, the marginal point drawn S110 respectively is according to vertical coordinate (Y-axis coordinate) value
It is grouped, using marginal point identical for ordinate value (Y value) as one group of transverse edge point group undetermined.
Then, for each group of transverse edge undetermined some group, travel through the abscissa value (X-axis coordinate) of each point,
If abscissa value (X value) is continuously (such as: X value is 2,3,4,5 ... 24,25 continued presences)
Some number more than a first threshold or in the range of a codomain, (bound of codomain can be according to specifically should
With based on experience value or statistical value determines), then it is assumed that this group abscissa value (X value) continuous print point composition one
Bar widthwise edge.Obtain all qualified widthwise edges according to this.
In S130, from described widthwise edge group, structure meets at least one widthwise edge of predetermined condition
Right.Widthwise edge is to can be as two horizontal edges to determine a rectangular area.Meet the widthwise edge pair of predetermined condition
Can be as two horizontal edges to determine that one meets intended rectangular area.
Various method construct widthwise edge pair can be used.Such as, in widthwise edge group wantonly two laterally
Limit, if it is determined that the difference of the initial abscissa value of two widthwise edges and the difference of termination abscissa value are respectively less than
3rd threshold value, and the ordinate value of two articles of widthwise edges differ from the 4th threshold range, then by two widthwise edges
As a widthwise edge pair.Depending on these threshold values or threshold range can be according to concrete application, such as, can use
Empirical value or statistical value etc..
According to some embodiments, in widthwise edge group, take a pair widthwise edge, sit initial for their X-axis
Scale value subtracts each other, and takes absolute value；Their X-axis is terminated coordinate figure subtract each other, take absolute value.
If it follows that the two of back absolute values are all within a threshold value (such as, the 3rd threshold value),
Then continue to subtract each other this two limits coordinate figure in Y-axis, take absolute value, a step after entrance；Otherwise return
Return back.
Then, if (such as, the 4th threshold value in the range of a codomain of the coordinate difference in back Y-axis
Scope), then it is assumed that this can make up to be confirmed or the regular rectangular shape region of checking, a square to widthwise edge
The starting point in shape region is the X in two widthwise edges, the minima (x in Y-coordinate valuemin, ymin),
Terminating point is then X, the maximum (x in Y-coordinate valuemax, ymax).The rectangular area of composition can represent
For Rect (xmin, ymin, xmax-xmin, ymax-ymin).Otherwise return the first step.Repeat into
Row said process, until completing all paired of all widthwise edges and processing.
Rectangular area Rect (xmin, ymin, xmax-xmin, ymax-ymin) be meant that, with
“xmin" as X-axis origin coordinates, with " ymin" as Y-axis origin coordinates, with
“xmax-xmin" as rectangle width, with " ymax-ymin" as the rectangle model of rectangular elevation
Enclose.
At S140, to each widthwise edge to verifying, by the widthwise edge of checking to for the most laterally
Limit pair, effective widthwise edge rectangular area to determining is the expection regular rectangular shape region identified.
Can adopt in various manners to each widthwise edge to verifying.Some according to the present invention are implemented
Example, to each widthwise edge is that find can be with widthwise edge to be verified to right to the mode verified
The longitudinal edge pair answered, i.e. finding can be with the widthwise edge to be verified longitudinal edge to constituting enclosing square
Right.If there is such longitudinal edge pair, widthwise edge the most to be verified is to for effective widthwise edge pair, its structure
The rectangular area become is effective rectangular area.Such rectangular area can a corresponding rectangular buttons or rectangle
Icon, but the invention is not restricted to this.
According to the scheme of example embodiment of the present invention, utilize endpoint detections, between widthwise edge
Position relationship determines regular rectangular shape region, it is possible to accurately identify the regular rectangular shape region in image.It addition,
Can have simple and accurate by rectangular area being verified the accuracy improving identification further
The high advantage of rate.
Fig. 2 illustrates the method flow diagram of the recognition rule rectangular area according to another exemplary embodiment.
The present embodiment is applicable to identify the rule button in image or have the regular icon etc. of limbus, but
The application of the present embodiment is not limited to this.
The method of the recognition rule rectangular area according to another exemplary embodiment as shown in Figure 2 and Fig. 1
Shown method is similar to, and differing only according to the method shown in Fig. 2 is by obtaining longitudinal edge group and indulging
To limit to carrying out region recognition.
As in figure 2 it is shown, at S210, the image comprising regular rectangular shape region is carried out endpoint detections computing
And obtain the marginal point in image.This is similar with the process described in S110, and here is omitted.
In S220, according to the distribution of the marginal point obtained, obtain including the vertical of at least two longitudinal edges
To limit group.
Similar with S120, various method can be used to obtain longitudinal edge group.Such as, can be by acquired limit
Edge point carries out being grouped to obtain multiple longitudinal edge point group according to abscissa value, if longitudinal edge point group comprises
Continuous boundary count out more than a Second Threshold, then the line segment this longitudinal edge point group constituted is as one
Bar longitudinal edge.Obtain all qualified longitudinal edges according to this, obtain longitudinal edge group.This Second Threshold can
It is different from pre-determined first threshold.
According to some embodiments, the marginal point drawn S210 respectively is according to abscissa (X-axis coordinate) value
It is grouped, using marginal point identical for abscissa value (X value) as one group of longitudinal edge point group undetermined.
Then, for each group of longitudinal edge undetermined some group, travel through the ordinate value (Y-axis coordinate) of each point,
If ordinate value (Y value) is continuously (such as: Y value is 2,3,4,5 ... 24,25 continued presences)
Some number more than a Second Threshold or in the range of a codomain, (bound of codomain can be according to specifically should
With based on experience value or statistical value determines), then it is assumed that this group ordinate value (Y value) continuous print point constitute one
Bar longitudinal edge.Obtain all qualified longitudinal edges according to this.
In S230, from described longitudinal edge group, structure meets at least one longitudinal edge of predetermined condition
Right.Longitudinal edge is to can be as two longitudinal edges to determine a rectangular area.Meet the longitudinal edge pair of predetermined condition
Can be as two longitudinal edges to determine that one meets intended rectangular area.
Similar with S130, various method construct longitudinal edge pair can be used.Such as, in longitudinal edge group
Wantonly two longitudinal edges, if it is determined that the difference of the initial ordinate value of two longitudinal edges and terminate vertical coordinate
The difference of value is respectively less than the 5th threshold value, and the difference of the abscissa value of two articles of longitudinal edges is in the 6th threshold range,
Then using two longitudinal edges as a longitudinal edge pair.3rd threshold value can be identical with aforementioned 5th threshold value or not
With, the 4th threshold range can be identical or different with aforementioned 6th threshold range.These threshold values or threshold range
Depending on can be according to concrete application, such as, can use empirical value or statistical value etc..
According to some embodiments, in longitudinal edge group, take a pair longitudinal edge, sit initial for their Y-axis
Scale value subtracts each other, and takes absolute value；Their Y-axis is terminated coordinate figure subtract each other, take absolute value.
If it follows that the two of back absolute values are all within a threshold value (such as, the 5th threshold value),
Then continue to subtract each other this two limits coordinate figure in X-axis, take absolute value, a step after entrance；Otherwise return
Return back.
Then, if (such as, the 6th threshold value in the range of a codomain of the coordinate difference in back X-axis
Scope), then it is assumed that this can make up to be confirmed or the regular rectangular shape region of checking, a square to longitudinal edge
The starting point in shape region is the Y in two longitudinal edges, the minima (x in X-coordinate valuemin, ymin),
Terminating point is then Y, the maximum (x in X-coordinate valuemax, ymax).The rectangular area of composition can represent
For Rect (xmin, ymin, xmax-xmin, ymax-ymin).Otherwise return the first step.Repeat into
Row said process, until completing all paired of all longitudinal edges and processing.
At S240, to limit longitudinally in each to verifying, by the longitudinal edge of checking to for the most longitudinally
Limit pair, effective longitudinal edge rectangular area to determining is the expection regular rectangular shape region identified.
Can adopt in various manners to limit longitudinally in each to verifying.Some according to the present invention are implemented
Example, to limit is longitudinally in each that find can be with longitudinal edge to be verified to right to the mode verified
The widthwise edge pair answered, i.e. finding can be with the longitudinal edge to be verified widthwise edge to constituting enclosing square
Right.If there is such widthwise edge pair, longitudinal edge the most to be verified is to for effective longitudinal edge pair, its structure
The rectangular area become is effective rectangular area.Such rectangular area can a corresponding rectangular buttons or rectangle
Icon, but the invention is not restricted to this.
According to the scheme of example embodiment of the present invention, utilize endpoint detections, between longitudinal edge
Position relationship determines regular rectangular shape region, it is possible to improve identification by verifying rectangular area
Accuracy, has the advantages such as simple, easy-to-use and accuracy rate is high.
Fig. 3 illustrates the method flow diagram of the recognition rule rectangular area according to another exemplary embodiment.
The present embodiment is applicable to identify the rule button in image or have the regular icon etc. of limbus, but
The application of the present embodiment is not limited to this.
The method of the recognition rule rectangular area according to another exemplary embodiment as shown in Figure 3 and Fig. 1
Similar with the method shown in Fig. 2, differ only according to the method shown in Fig. 3 be by obtaining laterally and
Longitudinal edge group and horizontal and vertical limit, to carrying out region recognition, have i.e. been used in combination shown in Fig. 1 and Fig. 2
Method and process, here is omitted.
Fig. 4 illustrate according to the present invention one example embodiment to the widthwise edge process to verifying.
As shown in Figure 4, at S410, at least one longitudinal edge obtained according to aforementioned process to carrying out time
Go through.
At S420, choose a longitudinal edge pair, for the longitudinal edge pair currently chosen, utilize longitudinal edge pair
Minimum and maximum abscissa respectively with minimum and maximum abscissa Xmin, Xmax of widthwise edge pair to be tested
Subtract each other, can take absolute value, if it is determined that absolute value is all within predetermined threshold, then enter S430；Otherwise
Return S410, continue to travel through at least one longitudinal edge pair.Without finding the longitudinal edge pair meeting condition,
Widthwise edge the most to be tested is to not over checking, it is impossible to be used for determining a rectangular area.
At S430, for the longitudinal edge pair currently chosen, utilize the minimum and maximum vertical coordinate of longitudinal edge pair
Minimum and maximum vertical coordinate Ymin, Ymax with widthwise edge pair to be tested subtract each other respectively, take absolute value, as
Really absolute value is all within predetermined threshold, then it is assumed that widthwise edge to be tested is to for effective widthwise edge pair, and it is true
Fixed rectangular area is the expection regular rectangular shape region identified.Otherwise return S410, continue traversal at least
One longitudinal edge pair.Without finding the longitudinal edge pair meeting condition, widthwise edge the most to be tested is to not having
Have by checking, it is impossible to be used for determining a rectangular area.
Fig. 5 illustrate according to another example embodiment of the present invention to the longitudinal edge process to verifying.
Process shown in Fig. 5 is similar with the process shown in Fig. 4, and differing only in is to utilize widthwise edge to right
Longitudinal edge is to verifying.
As it is shown in figure 5, at S510, at least one widthwise edge obtained according to aforementioned process is traveled through.
At S520, choose a widthwise edge pair, for the widthwise edge pair currently chosen, utilize widthwise edge pair
Minimum and maximum abscissa respectively with minimum and maximum abscissa Xmin, Xmax of longitudinal edge pair to be tested
Subtract each other, take absolute value, if absolute value is all within predetermined threshold, then enter S530；Otherwise return S510,
Continue to travel through at least one widthwise edge pair.Without finding the widthwise edge pair meeting condition, the most to be tested
Longitudinal edge to not over checking, it is impossible to be used for determining a rectangular area.
At S530, for the widthwise edge pair currently chosen, utilize the minimum and maximum vertical coordinate of widthwise edge pair
Minimum and maximum vertical coordinate Ymin, Ymax with longitudinal edge pair to be tested subtract each other respectively, take absolute value, as
Really absolute value is all within predetermined threshold, then it is assumed that longitudinal edge to be tested is to for effective longitudinal edge pair, and it is true
Fixed rectangular area is the expection regular rectangular shape region identified.Otherwise return S510, continue traversal at least
One widthwise edge pair.Without finding the widthwise edge pair meeting condition, longitudinal edge the most to be tested is to not having
Have by checking, it is impossible to be used for determining a rectangular area.
Fig. 6 illustrate according to another example embodiment of the present invention to the widthwise edge process to verifying.
Process shown in Fig. 6 and the difference of the process shown in Fig. 4 are to make utilization according to previously mentioned
The longitudinal edge group obtained to widthwise edge to verifying.
As shown in Figure 6, at S610, the longitudinal edge group obtained according to aforementioned process is traveled through.
At S620, choose a longitudinal edge, for the longitudinal edge currently chosen, utilize the abscissa of longitudinal edge
Subtract each other with the minimum abscissa Xmin of widthwise edge pair to be tested, take absolute value, if absolute value is at predetermined threshold
Within value, then enter S630；Otherwise return S610, continue traversal longitudinal edge group.Full without finding
The longitudinal edge of foot condition, widthwise edge the most to be tested is to not over checking, it is impossible to be used for determining a rectangle
Region.
At S630, for the longitudinal edge currently chosen, utilize the minimum and maximum vertical coordinate of longitudinal edge respectively
Subtract each other with minimum and maximum vertical coordinate Ymin, Ymax of widthwise edge pair to be tested, take absolute value, if absolutely
To value all within predetermined threshold, then it is assumed that this longitudinal edge is an effective longitudinal edge, enter S640；No
Then return S610, continue traversal longitudinal edge group.Without finding the longitudinal edge meeting condition, the most to be checked
The widthwise edge tested is to not over checking, it is impossible to be used for determining a rectangular area.
At S640, the longitudinal edge group obtained according to aforementioned process is traveled through.
At S650, choose a longitudinal edge, for the longitudinal edge currently chosen, utilize the abscissa of longitudinal edge
Subtract each other with the maximum abscissa Xmax of widthwise edge pair to be tested, take absolute value.If absolute value is at predetermined threshold
Within value, then enter S660；Otherwise return S640, continue traversal longitudinal edge group.Full without finding
The longitudinal edge of foot condition, widthwise edge the most to be tested is to not over checking, it is impossible to be used for determining a rectangle
Region.
At S660, for the longitudinal edge currently chosen, utilize the minimum and maximum vertical coordinate of longitudinal edge respectively
Subtract each other with minimum and maximum vertical coordinate Ymin, Ymax of widthwise edge pair to be tested, take absolute value, if absolutely
To value all within predetermined threshold, then it is assumed that this longitudinal edge is another effective longitudinal edge.At this moment, can recognize
For widthwise edge to be tested to for effective widthwise edge pair, its rectangular area determined is the expection rule identified
Rectangular area.Otherwise return S640, continue traversal longitudinal edge group.The vertical of condition is met without finding
Xiang Bian, widthwise edge the most to be tested is to not over checking, it is impossible to be used for determining a rectangular area.
It is likewise possible to utilize widthwise edge group to longitudinal edge to verifying, here is omitted.
It can be readily appreciated that based on present disclosure and teaching, it is also possible to use other various similar
Mode is verified.
Fig. 7 illustrates the block diagram of the device of the recognition rule rectangular area according to an illustrative embodiments.
As it is shown in fig. 7, include edge according to the device 700 of the recognition rule rectangular area of present embodiment
Point acquisition module 710, limit group acquisition module 720, limit are to constructing module 730 and authentication module 740.
The image that marginal point acquisition module 710 can be used for comprising regular rectangular shape region carries out endpoint detections
Computing and obtain the marginal point in image.
According to example embodiment, marginal point acquisition module 710 can be configured to: image is carried out smothing filtering；
Image after smothing filtering is carried out computing；The extreme point of detection calculations is to obtain the marginal point in image.
The distribution that limit group acquisition module 720 can be used for according to the described marginal point obtained obtains including at least two
The widthwise edge group of bar widthwise edge.Limit group acquisition module 720 also can be according to the distribution of the described marginal point obtained
Obtain including the longitudinal edge group of at least two longitudinal edges.
According to example embodiment, limit group acquisition module 720 can be configured to: by described marginal point according to vertical seat
Scale value carries out being grouped to obtain multiple transverse edge point group, for each transverse edge point group, if described
The continuous boundary that transverse edge point group comprises is counted out more than first threshold, then by described transverse edge point group
The line segment constituted adds described widthwise edge group as a widthwise edge.Limit group acquisition module 720 also can be joined
It is set to: carry out being grouped to obtain multiple longitudinal edge point group according to abscissa value by described marginal point, for
Each longitudinal edge point group, if the continuous boundary that described longitudinal edge point group comprises is counted out more than second
Threshold value, then add described longitudinal edge using the line segment that described longitudinal edge point group is constituted as a longitudinal edge
Group.
While constructing module 730 be can be used for structure from described widthwise edge group meet at least the one of predetermined condition
Individual widthwise edge pair.Predetermined condition is met while constructing module 730 also can be constructed from described longitudinal edge group
At least one longitudinal edge pair.
According to example embodiment, while constructing module 730 be can be configured to: in described widthwise edge group
Wantonly two widthwise edges, if it is determined that the difference of the initial abscissa value of described two widthwise edges and terminate horizontal seat
The difference of scale value is respectively less than the 3rd threshold value, and the difference of the ordinate value of described two articles of widthwise edges is at the 4th threshold value model
In enclosing, then using described two widthwise edges as a widthwise edge pair.While constructing module 730 be can also configure
For: for wantonly two longitudinal edges in described longitudinal edge group, if it is determined that initiateing of described two longitudinal edges
The difference of ordinate value and the difference of ordinate value of terminating are respectively less than the 5th threshold value, and described two longitudinal edges
The difference of abscissa value is in the 6th threshold range, then using described two longitudinal edges as a longitudinal edge pair.
Authentication module 740 can be used for each widthwise edge verifying, by checking widthwise edge to for
Effectively widthwise edge pair, the described effective widthwise edge rectangular area to determining is described regular rectangular shape region.Test
Card module 740 also can be to limit longitudinally in each to verifying, by the longitudinal edge of checking to for the most longitudinally
Limit pair, the described effective longitudinal edge rectangular area to determining is described regular rectangular shape region.
According to example embodiment, authentication module 740 is configured that to find and indulges corresponding with described widthwise edge
To limit pair, if there is corresponding longitudinal edge pair, it is determined that described widthwise edge is to for effective widthwise edge pair.
Authentication module 740 may be additionally configured to: finds with described longitudinal edge corresponding widthwise edge pair, if there is
Corresponding widthwise edge pair, it is determined that described longitudinal edge is to for effective longitudinal edge pair.
The device 700 of recognition rule rectangular area according to embodiments of the present invention can realize previously described
According to the method for the present application design, repeat no more with aforementioned similar explanation.
By above detailed description, those skilled in the art is it can be readily appreciated that according to embodiments of the present invention
And device can have one or more of the following advantages.
According to some embodiments, the solution of the present invention can accurately identify rule on the basis of edge detection algorithm
Rectangular area then, is applicable to identify the rule button in image or have the regular icon of limbus
Deng.
According to some embodiments, it is possible to accurately identify the regular rectangular shape region in image.
According to some embodiments, can be by rectangular area being verified the standard improving identification further
Really property, has the advantages such as simple and accuracy rate is high.
By the description of above embodiment, those skilled in the art is it can be readily appreciated that the embodiment of the present invention
Method and corresponding module can by software or partial software hardening by the way of realize.Therefore, this
The technical scheme of bright embodiment can embody with the form of software product, and this software product can store
In a non-volatile memory medium (can be CD-ROM, USB flash disk, portable hard drive etc.), if including
Dry instruction with so that calculating equipment (can be personal computer, server, mobile terminal or
Person's network equipment etc.) perform method according to embodiments of the present invention.
It will be understood by those skilled in the art that accompanying drawing is the schematic diagram of example embodiment, the mould in accompanying drawing
Block or flow process are not necessarily implemented necessary to the present invention, therefore cannot be used for limiting the protection of the present invention
Scope.
It will be appreciated by those skilled in the art that above-mentioned each module can be distributed in device according to the description of embodiment
In, it is also possible to carry out respective change and be disposed other than in one or more devices of the present embodiment.Above-mentioned reality
The module executing example can merge into a module, it is also possible to is further split into multiple submodule.
More than it is particularly shown and described the exemplary embodiment of the present invention.It should be understood that the present invention is not
Be limited to institute's inventive embodiment, on the contrary, it is intended to contain be included in claims spirit and
In the range of various amendments and equivalent arrangements.
Claims (14)
1. the method for a recognition rule rectangular area, it is characterised in that including:
The edge image comprising regular rectangular shape region being carried out endpoint detections computing and obtain in image
Point；
Distribution according to the described marginal point obtained obtains including the widthwise edge group of at least two widthwise edges, and
/ or obtain including the longitudinal edge group of at least two longitudinal edges according to the distribution of the described marginal point obtained；
From described widthwise edge group, structure meets at least one widthwise edge pair of predetermined condition, and/or from institute
State structure in longitudinal edge group and meet at least one longitudinal edge pair of predetermined condition；
To each widthwise edge to verifying, by the widthwise edge of checking to for effective widthwise edge pair, described
The effectively widthwise edge rectangular area to determining is described regular rectangular shape region；Or to limit longitudinally in each to entering
Row checking, by the longitudinal edge of checking to for effective longitudinal edge pair, the described effective longitudinal edge square to determining
Shape region is described regular rectangular shape region.
Method the most according to claim 1, it is characterised in that to comprising regular rectangular shape region
The marginal point that image carries out endpoint detections computing and obtains in image includes: utilize preset algorithm to image
Carrying out endpoint detections computing and obtain the marginal point in image, wherein said preset algorithm includes
Laplacian operator, Canny operator, Sobel operator, Prewitt operator, Kirsch operator, sieve
Calculate son or Roberts Cross operator.
Method the most according to claim 1, it is characterised in that to comprising regular rectangular shape region
The marginal point that image carries out endpoint detections computing and obtains in image includes:
Image is carried out smothing filtering；
Image after smothing filtering is carried out computing；
The extreme point of detection calculations is to obtain the marginal point in image.
Method the most according to claim 3, it is characterised in that image is carried out smothing filtering bag
Include: use Gaussian function that image is carried out smothing filtering.
Method the most according to claim 1, it is characterised in that:
Distribution according to the described marginal point obtained obtains including the widthwise edge group bag of at least two widthwise edges
Include: carry out being grouped to obtain multiple transverse edge point group, for often according to ordinate value by described marginal point
One transverse edge point group, if the continuous boundary that described transverse edge point group comprises is counted out more than the first threshold
Value, then add described widthwise edge using the line segment that described transverse edge point group is constituted as a widthwise edge
Group；And/or
Distribution according to the described marginal point obtained obtains including the longitudinal edge group bag of at least two longitudinal edges
Include: carry out being grouped to obtain multiple longitudinal edge point group, for often according to abscissa value by described marginal point
One longitudinal edge point group, if the continuous boundary that described longitudinal edge point group comprises is counted out more than the second threshold
Value, then add described longitudinal edge using the line segment that described longitudinal edge point group is constituted as a longitudinal edge
Group.
Method the most according to claim 1, it is characterised in that:
From described widthwise edge group, structure meets at least one widthwise edge of predetermined condition to including: for institute
State wantonly two widthwise edges in widthwise edge group, if it is determined that the initial abscissa value of described two widthwise edges
Difference and terminate the difference of abscissa value and be respectively less than the 3rd threshold value, and the ordinate value of described two widthwise edges
Difference is in the 4th threshold range, then using described two widthwise edges as a widthwise edge pair；And/or
From described longitudinal edge group, structure meets at least one longitudinal edge of predetermined condition to including: for institute
State wantonly two longitudinal edges in longitudinal edge group, if it is determined that the initial ordinate value of described two longitudinal edges
Difference and terminate the difference of ordinate value and be respectively less than the 5th threshold value, and the abscissa value of described two longitudinal edges
Difference is in the 6th threshold range, then using described two longitudinal edges as a longitudinal edge pair.
Method the most according to claim 1, it is characterised in that:
Each widthwise edge is included carrying out checking: find with described widthwise edge corresponding longitudinal edge pair,
If there is corresponding longitudinal edge pair, the most described widthwise edge is to for effective widthwise edge pair；Or
Limit longitudinally in each is included carrying out checking: find with described longitudinal edge corresponding widthwise edge pair,
If there is corresponding widthwise edge pair, the most described longitudinal edge is to for effective longitudinal edge pair.
8. the device of a recognition rule rectangular area, it is characterised in that including:
Marginal point acquisition module, for carrying out endpoint detections computing to the image comprising regular rectangular shape region
And obtain the marginal point in image；
Limit group acquisition module, for obtaining including at least two horizontal strokes according to the distribution of the described marginal point obtained
To the widthwise edge group on limit, and/or obtain including that at least two are indulged according to the distribution of the described marginal point obtained
To the longitudinal edge group on limit；
While to constructing module, meet at least one horizontal stroke of predetermined condition for structure from described widthwise edge group
To limit pair, and/or structure meets at least one longitudinal edge pair of predetermined condition from described longitudinal edge group；
Authentication module, is used for each widthwise edge verifying, by the widthwise edge of checking to for effectively
Widthwise edge pair, the described effective widthwise edge rectangular area to determining is described regular rectangular shape region；Or it is right
Limit is to verifying longitudinally in each, by the longitudinal edge of checking to for effective longitudinal edge pair, described the most vertical
It is described regular rectangular shape region to the rectangular area determined to limit.
Device the most according to claim 8, it is characterised in that described marginal point acquisition module is joined
The marginal point being set to utilize preset algorithm that image is carried out endpoint detections computing and obtain in image, wherein
Described preset algorithm includes that Laplacian operator, Canny operator, Sobel operator, Prewitt calculate
Son, Kirsch operator, compass operator or Roberts Cross operator.
Device the most according to claim 9, it is characterised in that described marginal point acquisition module is joined
It is set to:
Image is carried out smothing filtering；
Image after smothing filtering is carried out computing；
The extreme point of detection calculations is to obtain the marginal point in image.
11. devices according to claim 10, it is characterised in that described marginal point acquisition module
It is configured to use Gaussian function that image is carried out smothing filtering.
12. devices according to claim 8, it is characterised in that described limit group acquisition module configures
For:
Carry out being grouped to obtain multiple transverse edge point group, for often according to ordinate value by described marginal point
One transverse edge point group, if the continuous boundary that described transverse edge point group comprises is counted out more than the first threshold
Value, then add described widthwise edge using the line segment that described transverse edge point group is constituted as a widthwise edge
Group；And/or
Carry out being grouped to obtain multiple longitudinal edge point group, for often according to abscissa value by described marginal point
One longitudinal edge point group, if the continuous boundary that described longitudinal edge point group comprises is counted out more than the second threshold
Value, then add described longitudinal edge using the line segment that described longitudinal edge point group is constituted as a longitudinal edge
Group.
13. devices according to claim 8, it is characterised in that constructing module is configured by described limit
For:
For wantonly two widthwise edges in described widthwise edge group, if it is determined that initiateing of described two widthwise edges
The difference of abscissa value and the difference of abscissa value of terminating are respectively less than the 3rd threshold value, and described two widthwise edges
The difference of ordinate value is in the 4th threshold range, then using described two widthwise edges as a widthwise edge pair；
And/or
For wantonly two longitudinal edges in described longitudinal edge group, if it is determined that initiateing of described two longitudinal edges
The difference of ordinate value and the difference of ordinate value of terminating are respectively less than the 5th threshold value, and described two longitudinal edges
The difference of abscissa value is in the 6th threshold range, then using described two longitudinal edges as a longitudinal edge pair.
14. devices according to claim 13, it is characterised in that described authentication module configures
For:
Find with described widthwise edge corresponding longitudinal edge pair, if there is corresponding longitudinal edge pair, the most really
Fixed described widthwise edge is to for effective widthwise edge pair；Or
Find with described longitudinal edge corresponding widthwise edge pair, if there is corresponding widthwise edge pair, the most really
Fixed described longitudinal edge is to for effective longitudinal edge pair.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610153207.1A CN105844615B (en) | 2016-03-17 | 2016-03-17 | The method and device of recognition rule rectangular area |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610153207.1A CN105844615B (en) | 2016-03-17 | 2016-03-17 | The method and device of recognition rule rectangular area |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105844615A true CN105844615A (en) | 2016-08-10 |
| CN105844615B CN105844615B (en) | 2018-10-26 |
Family
ID=56588281
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610153207.1A Active CN105844615B (en) | 2016-03-17 | 2016-03-17 | The method and device of recognition rule rectangular area |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105844615B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108805072A (en) * | 2018-06-06 | 2018-11-13 | 西安纸贵互联网科技有限公司 | A kind of qualitative recognition method and device of external diagnosis reagent |
| CN108830314A (en) * | 2018-06-06 | 2018-11-16 | 西安纸贵互联网科技有限公司 | A kind of sxemiquantitative recognition methods of external diagnosis reagent and device |
| CN113485705A (en) * | 2021-06-30 | 2021-10-08 | 深圳软牛科技有限公司 | QML Rectangle component-based frame selection method, device, equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101534370A (en) * | 2008-03-11 | 2009-09-16 | 株式会社Pfu | Image processor, image processing method, and computer program product |
| CN102034254A (en) * | 2010-09-29 | 2011-04-27 | 广东威创视讯科技股份有限公司 | Method for recognizing geometric figure |
| CN103593860A (en) * | 2013-10-27 | 2014-02-19 | 沈阳建筑大学 | Cutting method for any irregular blank |
| CN104714666A (en) * | 2013-12-12 | 2015-06-17 | 鸿合科技有限公司 | Intelligent pen and stroke recognition processing method thereof |
| CN104952072A (en) * | 2015-06-16 | 2015-09-30 | 华中科技大学 | Rectangle detection method based on genetic algorithm |
| US20160044196A1 (en) * | 2014-08-08 | 2016-02-11 | Kyocera Document Solutions Inc. | Image processing apparatus |
-
2016
- 2016-03-17 CN CN201610153207.1A patent/CN105844615B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101534370A (en) * | 2008-03-11 | 2009-09-16 | 株式会社Pfu | Image processor, image processing method, and computer program product |
| CN102034254A (en) * | 2010-09-29 | 2011-04-27 | 广东威创视讯科技股份有限公司 | Method for recognizing geometric figure |
| CN103593860A (en) * | 2013-10-27 | 2014-02-19 | 沈阳建筑大学 | Cutting method for any irregular blank |
| CN104714666A (en) * | 2013-12-12 | 2015-06-17 | 鸿合科技有限公司 | Intelligent pen and stroke recognition processing method thereof |
| US20160044196A1 (en) * | 2014-08-08 | 2016-02-11 | Kyocera Document Solutions Inc. | Image processing apparatus |
| CN104952072A (en) * | 2015-06-16 | 2015-09-30 | 华中科技大学 | Rectangle detection method based on genetic algorithm |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108805072A (en) * | 2018-06-06 | 2018-11-13 | 西安纸贵互联网科技有限公司 | A kind of qualitative recognition method and device of external diagnosis reagent |
| CN108830314A (en) * | 2018-06-06 | 2018-11-16 | 西安纸贵互联网科技有限公司 | A kind of sxemiquantitative recognition methods of external diagnosis reagent and device |
| CN113485705A (en) * | 2021-06-30 | 2021-10-08 | 深圳软牛科技有限公司 | QML Rectangle component-based frame selection method, device, equipment and storage medium |
| CN113485705B (en) * | 2021-06-30 | 2023-11-21 | 深圳软牛科技有限公司 | QML (quality control language) component-based frame selection method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105844615B (en) | 2018-10-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110503108A (en) | Architecture against regulations recognition methods and device, storage medium, computer equipment | |
| CN109376631B (en) | Loop detection method and device based on neural network | |
| CN109284674A (en) | A kind of method and device of determining lane line | |
| Min et al. | Support vector machine approach to fall recognition based on simplified expression of human skeleton action and fast detection of start key frame using torso angle | |
| EP2528035B1 (en) | Apparatus and method for detecting a vertex of an image | |
| CN103477352A (en) | Gesture recognition using depth images | |
| CN109640066B (en) | Method and device for generating high-precision dense depth image | |
| CN103119609B (en) | A kind of method and apparatus determining video foreground main image area | |
| CN109660783A (en) | Virtual reality parallax correction | |
| CN103020993A (en) | Visual saliency detection method by fusing dual-channel color contrasts | |
| CN105844615A (en) | Method of identifying regular rectangular area and apparatus thereof | |
| CN103488356A (en) | Infrared camera three-dimensional imaging-based touch recognition method | |
| US20200074149A1 (en) | Method and apparatus for obtaining painting | |
| CN113240678A (en) | Plane information detection method and system | |
| CN110415285A (en) | Image processing method, device and electronic equipment | |
| CN109833003B (en) | Sweeping robot and method for dividing region thereof | |
| CN102866885B (en) | Method and device for confirming clicking position in webpage | |
| Pasquale et al. | Enabling depth-driven visual attention on the icub humanoid robot: Instructions for use and new perspectives | |
| CN111768005A (en) | Training method and device for lightweight detection model, electronic equipment and storage medium | |
| JP5478520B2 (en) | People counting device, people counting method, program | |
| CN111784660B (en) | Method and system for analyzing frontal face degree of face image | |
| WO2015079054A1 (en) | Estimating gaze from un-calibrated eye measurement points | |
| Ranade et al. | Novel single view constraints for manhattan 3d line reconstruction | |
| CN107066982A (en) | The recognition methods of human face characteristic point and device | |
| WO2016001920A1 (en) | A method of perceiving 3d structure from a pair of images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |