CN108345888A - A kind of connected domain extracting method and device - Google Patents
A kind of connected domain extracting method and device Download PDFInfo
- Publication number
- CN108345888A CN108345888A CN201810141802.2A CN201810141802A CN108345888A CN 108345888 A CN108345888 A CN 108345888A CN 201810141802 A CN201810141802 A CN 201810141802A CN 108345888 A CN108345888 A CN 108345888A
- Authority
- CN
- China
- Prior art keywords
- value
- mark value
- pixel
- mark
- pixel point
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of connected domain extracting method and device, the method includes：Image to be extracted is converted into bianry image；Each pixel in bianry image is scanned successively, for the pixel scanned, according to the mark value of the first gray value of current pixel point, the first object pixel and second target pixel points adjacent with current pixel point, determine the mark value of current pixel point, and if it is determined that the first mark value and the second mark value of the first object pixel and second target pixel points adjacent with the current pixel point not etc. and be not 0, first mark value and the second mark value is right as equivalence, it is recorded in code table of equal value；It is right according to equivalence present in equivalent table, determine equivalent sequence；According to each target label value for including in the equivalent sequence, the pixel that mark value in the bianry image is the target label value is determined as the connected domain in the image to be extracted.Provide a kind of succinct, efficient connected domain extraction scheme.
Description
Technical field
The present invention relates to image identification technical field more particularly to a kind of connected domain extracting method and devices.
Background technology
Connected domain extraction is often the first step of image recognition, is the important component part of image recognition technology, connected domain
Extraction can shoot for the feature of science fiction movies and provide accurate positioning, can be used for the image recognition and target of security protection industry
Information extraction is even used in radar system obtain the parameters such as the position of moving target, phase, angle and speed.
Such as：Connected domain extraction can be used to carry out Car license recognition, text identification, subtitle recognition, vehicle tracking, pedestrian's intrusion detection
Deng.
Existing connected domain extracting method can be mainly divided into two major classes, and the first kind is in sequential scan bianry image first
Each pixel carries out the label of connected domain, mark value is assigned for each pixel, according still further to backward, sequence, the sequence of backward
Scanning bianry image mark value is optimized, wherein according to backward, sequence, backward sequential scan bianry image to mark value
It optimizes as the current pixel point for scanning, according to the mark value of current pixel point, the pixel on the current pixel point left side
Mark value and current pixel point top pixel mark value in minimum mark value, to the mark value of current pixel point into
Row update, and then keep the mark value of pixel in same connected domain identical, to achieve the purpose that mark connected domain.Second class
For the thought based on seed mediated growth method, i.e., by first specifying a base reference point or region, then again with the mark point or area
Based on domain, the extending marking one by one by way of recurrence or storehouse, until all adjacent points have all been labeled.
Although above-mentioned two classes method may be used to connected domain detection, but all can not easily extract the spy of connected domain
Sign.Wherein first kind algorithm, which needs to take multiple scan image, can just determine connected domain, in processing containing compared with multiply connected domain
Image when speed can be very slow.Second class algorithm would generally consider to realize using recursive mode, real using recursive fashion
Now, when some connected domain is larger, necessarily there is very big depth of recursion.Under extreme case, as N*M sizes image in
For all pixels point all in the same connected domain, depth of recursion is up to N*M, and the efficiency for handling larger image will can not be endured
's.And if replacing recurrence by the way of storehouse, although recurrence can be solved the problems, such as, such method is either with point
Based on extension or extended with behavior base, can not all avoid each pixel from being required for all being marked by multiple access
Note, and such method is when handling larger connected domain, as all pixels point is in the same connected domain, efficiency of algorithm in image
It is very low.
Invention content
A kind of connected domain extracting method of present invention offer and device account for solve to exist in the prior art connected domain extraction
The problem that process resource is more, efficiency is low.
The invention discloses a kind of connected domain extracting method, the method includes：
Image to be extracted is converted into bianry image；
Each pixel in the bianry image is scanned successively, for the pixel scanned, according to current pixel point
The mark value of first gray value, the first object pixel and second target pixel points adjacent with current pixel point determines current
The mark value of pixel, and if it is determined that the first object pixel and second target pixel points adjacent with the current pixel point
The first mark value and the second mark value not etc. and be not 0, using first mark value and second mark value as etc.
Valence pair is recorded in code table of equal value；
It is right according to equivalence present in the equivalent table, determine equivalent sequence；
It is the mesh by mark value in the bianry image in the equivalent sequence according to each target label value for including
The pixel of mark mark value is determined as the connected domain in the image to be extracted.
Further, described to scan each pixel in the bianry image successively and include：
According to sequence from top to bottom from left to right, each pixel in the bianry image is scanned；Or,
According to sequence from left to right from top to bottom, each pixel in the bianry image is scanned；Or,
According to sequence from top to bottom from right to left, each pixel in the bianry image is scanned；Or,
According to sequence from right to left from top to bottom, each pixel in the bianry image is scanned.
Further, if according to sequence from top to bottom from left to right, each pixel in the bianry image is scanned,
It is described for the pixel scanned, according to the first gray value of current pixel point, the first object adjacent with current pixel point
The mark value of pixel and the second target pixel points determines that the mark value of current pixel point includes：
For the pixel scanned, judge whether the first gray value of current pixel point is 0；
If the first gray value of current pixel point is 0, the mark value of current pixel point is determined as 0；
If the first gray value of current pixel point not be 0, judge the current pixel point left side first object pixel and
Whether the first mark value and the second mark value of the second target pixel points of top are 0；If so, by the mark of current pixel point
Note value is determined as fixed maximum mark value and adds 1, if not, according in first mark value and second mark value most
Small non-zero mark value, determines the mark value of current pixel point.
Further, described that first mark value and second mark value is right as equivalence, equivalent code is recorded
Before in table, the method further includes：
It is right with the presence or absence of the equivalence comprising first mark value and second mark value in the code table of equal value to judge；
If not, carrying out subsequent step.
Further, after the mark value of the determining current pixel point, the method further includes：
According to the coordinate of the mark value of current pixel point and current pixel point, current pixel point in the image to be extracted
In the second gray value, preset gray threshold, the corresponding connected domain attribute value of the mark value is updated；
It is described according to each target label value for including in the equivalent sequence, be institute by mark value in the bianry image
The pixel for stating target label value is determined as after the connected domain in the image to be extracted, and the method further includes：
According to each target label value for including in the corresponding connected domain attribute value of each mark value and the equivalent sequence,
Determine the connected domain attribute value of the connected domain in the image to be extracted.
Further, the connected domain attribute value includes：
The starting point of connected domain, terminating point, area, weights are cumulative, abscissa x weighted accumulations, ordinate y weighted accumulations.
The invention discloses a kind of connected domain extraction element, described device includes：
Conversion module, for image to be extracted to be converted to bianry image；
Code table generation module is marked, for scanning each pixel in the bianry image successively, for the picture scanned
Vegetarian refreshments, according to the first gray value of current pixel point, the first object pixel adjacent with current pixel point and the second target picture
The mark value of vegetarian refreshments, determines the mark value of current pixel point, and if it is determined that the first object adjacent with the current pixel point
The first mark value and the second mark value of pixel and the second target pixel points are not etc. and 0, by first mark value
It is right as equivalence with second mark value, it is recorded in code table of equal value；
First determining module, it is right according to equivalence present in the equivalent table to be used for, and determines equivalent sequence；
Second determining module, for according to each target label value for including in the equivalent sequence, by the binary map
Mark value is that the pixel of the target label value is determined as the connected domain in the image to be extracted as in.
Further, the label code table generation module is specifically used for according to sequence from top to bottom from left to right, scanning
Each pixel in the bianry image；Or, according to sequence from left to right from top to bottom, scan each in the bianry image
Pixel；Or, according to sequence from top to bottom from right to left, each pixel in the bianry image is scanned；Or, according to from the right side
To left sequence from top to bottom, each pixel in the bianry image is scanned.
Further, the label code table generation module, if be specifically used for according to sequence from top to bottom from left to right,
Each pixel in the bianry image is scanned, for the pixel scanned, judges that the first gray value of current pixel point is
No is 0；If the first gray value of current pixel point is 0, the mark value of current pixel point is determined as 0；If current pixel
First gray value of point is not 0, judges the first object pixel on the current pixel point left side and the second target pixel points of top
The first mark value and the second mark value whether be 0；If so, by the mark value of current pixel point be determined as it is fixed most
Big mark value adds 1, if not, according to non-zero mark value minimum in first mark value and second mark value, determines
The mark value of current pixel point.
Further, described device further includes：
Judgment module, for judging in the code table of equal value with the presence or absence of comprising first mark value and second mark
The equivalence of note value is right；If it is judged that be no, carry out subsequently using first mark value and second mark value as etc.
The step in code table of equal value is recorded in valence pair.
Further, described device further includes：
Update module, for according to the coordinate of the mark value of current pixel point and current pixel point, current pixel point in institute
The second gray value, the preset gray threshold in image to be extracted are stated, the corresponding connected domain attribute value of the mark value is carried out
Update；
Third determining module, for including according in the corresponding connected domain attribute value of each mark value and the equivalent sequence
Each target label value, determine the connected domain attribute value of the connected domain in the image to be extracted.
The invention discloses a kind of connected domain extracting method and device, the method includes：Image to be extracted is converted to
Bianry image；Each pixel in the bianry image is scanned successively, for the pixel scanned, according to current pixel point
The mark value of first gray value, the first object pixel and second target pixel points adjacent with current pixel point determines current
The mark value of pixel, and if it is determined that the first object pixel and second target pixel points adjacent with the current pixel point
The first mark value and the second mark value not etc. and be not 0, using first mark value and second mark value as etc.
Valence pair is recorded in code table of equal value；It is right according to equivalence present in the equivalent table, determine equivalent sequence；According to the equivalence
The each target label value for including in sequence determines the pixel that mark value in the bianry image is the target label value
For the connected domain in the image to be extracted.Due in embodiments of the present invention, each pixel in scanning bianry image successively
When point, if it is determined that the first mark value and second of the adjacent first object pixel and the second target pixel points of current pixel point
Mark value is not equal and is not 0, and the first mark value and the second mark value is right as equivalence, is recorded in code table of equal value, and root
According to equivalence present in code table of equal value to determining equivalent sequence, according to each target label value for including in equivalent sequence, by two
Mark value is that the pixel of target label value is determined as the connected domain in image to be extracted in value image, it is only necessary to run-down, just
It can be by different mark values be corresponded in bianry image but the pixel in same connected domain is determined as connection in image to be extracted
Domain simplifies the process of connected domain extraction, avoids and be repeatedly scanned to bianry image, reduces occupied by connected domain extraction
Process resource, improve connected domain extraction efficiency.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
Obtain other attached drawings according to these attached drawings.
Fig. 1 is a kind of connected domain extraction process schematic diagram that the embodiment of the present invention 1 provides；
Fig. 2 is a kind of scanning window schematic diagram that the embodiment of the present invention 2 provides；
Fig. 3 is a kind of connected domain extraction element structural schematic diagram that the embodiment of the present invention 5 provides.
Specific implementation mode
To make the objectives, technical solutions, and advantages of the present invention clearer, make into one below in conjunction with the attached drawing present invention
Step ground detailed description, it is clear that described embodiment is only a part of the embodiment of the present invention, rather than whole implementation
Example.Based on the embodiments of the present invention, obtained by those of ordinary skill in the art without making creative efforts
Every other embodiment, shall fall within the protection scope of the present invention.
Embodiment 1：
Fig. 1 is a kind of connected domain extraction process schematic diagram provided in an embodiment of the present invention, which includes：
S101：Image to be extracted is converted into bianry image.
Connected domain extracting method provided in an embodiment of the present invention is applied to electronic equipment, which can adopt for image
Collect equipment, such as video camera, camera, can also be the equipment such as mobile phone, PC (PC), tablet computer.
Bianry image refers to each pixel in image only there are two types of value or tonal gradation state, usually " 0 " or
" 1 ", bianry image connected component labeling be from the dot matrix image being only made of " 0 " (background dot) and " 1 " (target point), will be adjacent
Target point extract.The purpose is to find connected domain all in image, and all pixels that will belong to same connected domain
Point is marked with unique mark value.Wherein electronic equipment can be based on electricity when carrying out connected component labeling to bianry image
Field Programmable Gate Array (Field-Programmable Gate Array, FPGA) in sub- equipment is realized.
Specifically, according to preset gray threshold, image to be extracted is converted into bianry image.
S102：Each pixel in the bianry image is scanned successively, for the pixel scanned, according to current pixel
First gray value of point, the mark value of the first object pixel and second target pixel points adjacent with current pixel point, determine
The mark value of current pixel point, and if it is determined that the first object pixel and second target picture adjacent with the current pixel point
The first mark value and the second mark value of vegetarian refreshments not etc. and be not 0, will first mark value and second mark value works
It is of equal value right, is recorded in code table of equal value.
In embodiments of the present invention, can be according to positive sequence, sequence such as from top to bottom from left to right can also be according to inverse
Sequence scans each pixel in bianry image such as according to sequence from top to bottom from right to left, to each pixel into rower
Note, for the pixel scanned, the pixel on the current pixel point left side is made wherein if scanning bianry image according to positive sequence
For the first object pixel adjacent with current pixel point, using the pixel of current pixel point top as with current pixel point phase
The second adjacent target pixel points；If scanning bianry image according to backward, for the pixel scanned, by the current pixel point right side
The pixel on side as the first object pixel adjacent with current pixel point, using the following pixel of current pixel point as with
The second adjacent object pixel of current pixel point, wherein if the current pixel point of scanning is in the edge of bianry image, it can not
When determining the first object pixel adjacent with current pixel point and/or the second target pixel points, first object pixel is determined
And/or second target pixel points mark value be 0, and then determine current pixel point mark value.
Specifically, each pixel in bianry image is scanned successively, for the pixel scanned, if current pixel
First gray value of point is 0, then the mark value of current pixel point is determined as 0, if the first gray value of current pixel point is non-
0 and the first mark value of first object pixel and the second mark value of the second target pixel points be 0, by current pixel point
Mark value be determined as fixed maximum mark value and add 1, if the first gray value of current pixel point is non-zero, and first object
First mark value of pixel and the second mark value of the second target pixel points are not 0, are marked according to the first mark value and second
Minimum non-zero mark value, determines the mark value of current pixel point in note value.
In addition, if the first mark value of the current adjacent first object pixel of picture point and the second target pixel points and the
Two mark values differ, and are not 0, and the first mark value and the second mark value is right as equivalence, are recorded in code table of equal value.Example
Such as：First mark value of the adjacent first object pixel of current pixel point is that the second mark value of the 3, second target pixel points is
2, (3,2) are right as equivalence, it is recorded in code table of equal value.
S103：It is right according to equivalence present in the equivalent table, determine equivalent sequence.
S104：According to each target label value for including in the equivalent sequence, it is by mark value in the bianry image
The pixel of the target label value is determined as the connected domain in the image to be extracted.
Specifically, in embodiments of the present invention, it is right according to equivalence present in equivalent table, determine the equivalence between mark value
Relationship determines equivalent sequence, and then according to each target mark for including in equivalent sequence according to the equivalence relation between mark value
Note value determines during label is in bianry image that the pixel for the target label value for including is the company in image to be extracted
Logical domain.Such as：Equivalence present in equivalent table is to for (5,3), (3,2), it is determined that 5-3-2 is an equivalent sequence, determines two
The pixel that mark value is 5 or 3 or 2 in value image is the connected domain in image to be extracted.
In addition, if according to of equal value right present in equivalent table, multiple equivalent sequences are determined, for each sequence of equal value
Row determine in bianry image that label be the pixel for the target label for including in valence sequence is the connection in image to be extracted
Domain.Such as：Equivalence present in equivalent table is to for (5,3), (3,2), (4,1), it is determined that 5-3-2 is equivalent sequence, a 4-1
For an equivalent sequence, the pixel that mark value is 5 or 3 or 2 in determining bianry image respectively is the connection in image to be extracted
The pixel that mark value is 4 or 1 in domain, bianry image is the connected domain in image to be extracted.
After determining the corresponding connected domain of equivalent sequence, then illustrate the corresponding two-value of remaining each mark value in bianry image
As soon as the pixel in image is the connected domain in image to be extracted, and then can determine that out image to be extracted by single pass
In all connected domain.Such as：Equivalent sequence is 5-3-2, and mark value present in bianry image includes 1,2,3,4,5,6, really
Determine after the pixel that mark value in bianry image is 5 or 3 or 2 is connected domain in image to be extracted, to determine that bianry image is got the bid
The pixel that note value is 1 is a connected domain in image to be extracted, determines that the pixel that mark value is 4 in bianry image is to wait for
A connected domain in extraction image determines that mark value is 6 in bianry image pixel is a company in image to be extracted
Logical domain.
Preferably, for the ease of the determination of connected domain, electronic equipment can configure a twoport for storing mark value
Random access memory (Random Access Memory, RAM), is indicated with U1, is that each mark value distributes one in U1
Storage address, if mark value 1 corresponds to the storage address 1 of U1, the storage address 2 ... of the correspondence of mark value 2 U1, in bianry image
When each pixel is scanned, if the first gray value of current pixel point is non-zero and the first label of first object pixel
Second mark value of value and the second target pixel points is 0, and the mark value of current pixel point is determined as fixed maximum mark
Note value adds 1, such as：Current maximum mark value is 5, it is determined that the mark value of current pixel point is 6, in the storage address 6 of U1
Add mark value 6；If the first mark value of the current adjacent first object pixel of picture point and the second target pixel points and the
Two mark values differ, and are not 0, then depositing for larger mark value in the first mark value and the second mark value will be stored in U1
The mark value stored in storage address is updated to mark value smaller in the first mark value and the second mark value, such as：First label
Value is that the 5, second mark value is 3, and the first mark value is more than the second mark value, will be stored in U1 in mark value " 5 " storage address 5
The mark value " 5 " of storage is updated to mark value " 3 ", will if follow-up detect that the first mark value be the 2, second mark value is 3 again
The mark value " 3 " being stored in U1 in the storage address 3 and storage address 5 of mark value " 3 " is updated to mark value " 2 ", until
Scanning is completed, and determines that the corresponding mark value of the identical storage address of mark value is stored in U1 corresponds to same connected domain.
Due in embodiments of the present invention, in each pixel in scanning bianry image successively, if it is determined that current picture
The first mark value and the second mark value of the adjacent first object pixel of vegetarian refreshments and the second target pixel points not etc. and be not
0, the first mark value and the second mark value is right as equivalence, it is recorded in code table of equal value, and according to present in code table of equal value
Mark value in bianry image is mesh in equivalent sequence according to each target label value for including to determining equivalent sequence by equivalence
The pixel of mark mark value is determined as the connected domain in image to be extracted, it is only necessary to which run-down can will correspond in bianry image
Different mark values but it is determined as the connected domain in image to be extracted in the pixel of same connected domain, simplifies connected domain extraction
Process, avoid and repeatedly bianry image be scanned, reduce connected domain and extract occupied process resource, the company of improving
The efficiency of logical domain extraction.
Embodiment 2：
For the ease of the pixel in bianry image is marked, on the basis of the above embodiments, of the invention real
It applies in example, it is described to scan each pixel in the bianry image successively and include：
According to sequence from top to bottom from left to right, each pixel in the bianry image is scanned；Or,
According to sequence from left to right from top to bottom, each pixel in the bianry image is scanned；Or,
According to sequence from top to bottom from right to left, each pixel in the bianry image is scanned；Or,
According to sequence from right to left from top to bottom, each pixel in the bianry image is scanned.
If according to sequence from top to bottom from left to right, each pixel in the bianry image is scanned, it is described to be directed to
The pixel scanned, according to the first gray value of current pixel point, the first object pixel adjacent with current pixel point and
The mark value of second target pixel points determines that the mark value of current pixel point includes：
For the pixel scanned, judge whether the first gray value of current pixel point is 0；
If the first gray value of current pixel point is 0, the mark value of current pixel point is determined as 0；
If the first gray value of current pixel point not be 0, judge the current pixel point left side first object pixel and
Whether the first mark value and the second mark value of the second target pixel points of top are 0；If so, by the mark of current pixel point
Note value is determined as fixed maximum mark value and adds 1, if not, according in first mark value and second mark value most
Small non-zero mark value, determines the mark value of current pixel point.
In embodiments of the present invention, can be according to positive sequence, sequence such as from top to bottom from left to right or from left to right from upper
Sequence under, is scanned bianry image, can also according to backward, sequence such as from top to bottom from right to left or from the right side to
Left sequence from top to bottom is scanned bianry image.
Specifically, the scanning window in the regions 2*2 as shown in Figure 2 may be used, according to from left to right suitable from top to bottom
Sequence scans each pixel in bianry image, and wherein E (i, j) indicates current pixel point, and D (i, j-1) is indicated and current pixel point
Adjacent first object pixel, B (i-1, j) indicate that second target pixel points adjacent with current pixel point, wherein i expressions are worked as
Abscissa of the preceding pixel point in bianry image, y identify ordinate of the current pixel point in bianry image.Specifically, adopting
When being scanned to each pixel in bianry image with the scanning window of 2*2, tagging scheme can be expressed as, if E
Gray value is 0, determines that the mark value of E is 0；If the gray value of E is not that the mark value of 0, B and D is 0, determine that the mark value of E is
Fixed maximum mark value adds 1；If it is 0 that the gray value of E, which is not the mark value of 0, B, the mark value of D is not 0, determine E's
Mark value is identical as the mark value of D；If the mark value that the mark value that the gray value of E is not 0, B is not 0, D is 0, determine E's
Mark value is identical as the mark value of B；If the mark value that the mark value that the gray value of E is not 0, B is not 0, D is not 0, by B's
Minimum mark value, is determined as the mark value of E in mark value and the mark value of D.
Embodiment 3：
In order to avoid identical equivalence is right, repetition is recorded in code table of equal value, on the basis of the various embodiments described above,
It is described that first mark value and second mark value is right as equivalence in the embodiment of the present invention, code table of equal value is recorded
In before, the method further includes：
It is right with the presence or absence of the equivalence comprising first mark value and second mark value in the code table of equal value to judge；
If not, carrying out subsequent step.
Specifically, if the first label of the first object pixel and second target pixel points adjacent with current pixel point
Value and the second mark value do not wait and are not 0, judge in code table of equal value with the presence or absence of comprising the first mark value and the second mark value
Equivalence it is right, if it is, illustrate that the equivalence comprising the first mark value and the second mark value is right in code table of equal value, without by first
The equivalence of second mark value of mark value is to being recorded in code table of equal value, if it is not, then illustrating in code table of equal value comprising the first mark
The equivalence of note value and the second mark value is right, needs the equivalence by the second mark value of the first mark value to code table of equal value is recorded
In.
Such as：If the first mark value of the first object pixel and second target pixel points adjacent with current pixel point
It is not waited with the second mark value and is not 0, it is 3 that the first mark value, which is the 5, second mark value, the first mark value and the second mark value
It is of equal value of equal value to for (5,3), there is equivalence to (5,3) in code table of equal value, being not necessarily to for (5,3), include in code table of equal value
(5,3) are recorded in code table of equal value in equivalence again.
Embodiment 4：
Connection Domain Properties are known for the ease of user, on the basis of the various embodiments described above, in the embodiment of the present invention
In, after the mark value of the determining current pixel point, the method further includes：
According to the coordinate of the mark value of current pixel point and current pixel point, current pixel point in the image to be extracted
In the second gray value, preset gray threshold, the corresponding connected domain attribute value of the mark value is updated；
It is described according to each target label value for including in the equivalent sequence, be institute by mark value in the bianry image
The pixel for stating target label value is determined as after the connected domain in the image to be extracted, and the method further includes：
According to each target label value for including in the corresponding connected domain attribute value of each mark value and the equivalent sequence,
Determine the connected domain attribute value of the connected domain in the image to be extracted.
Preferably, connected domain attribute value includes：
The starting point of connected domain, terminating point, area, weights are cumulative, the weighting of abscissa (x) weighted accumulation, ordinate (y) is tired
Add.
Specifically, for the pixel scanned, after the mark value for determining current pixel point, existed according to current pixel point
The difference of the second gray value and preset gray threshold in image to be extracted, as the weights of current pixel point, by current picture
The abscissa of vegetarian refreshments and the product of weights are made as the x weights of current pixel point, by the product of the ordinate of current pixel point and weights
For the y weights of current pixel point, according to the coordinate of current pixel point, weights, x weights and y weights to the corresponding company of the mark value
The starting point in logical domain, terminating point, area, weights are cumulative, abscissa x weighted accumulations, ordinate y weighted accumulations are updated, i.e.,
Compared with by the abscissa of the abscissa of current pixel point starting point corresponding with the mark value, by the ordinate of current pixel point with
The ordinate of the corresponding starting point of the mark value compares, and chooses minimum abscissa and ordinate to the corresponding starting of the mark value
The coordinate of point is updated；Compared with by the abscissa of the abscissa of current pixel point terminating point corresponding with the mark value, it will work as
Compared with the ordinate of the ordinate terminating point corresponding with the mark value of preceding pixel point, maximum abscissa and ordinate pair are chosen
The coordinate of the corresponding terminating point of the mark value is updated；The corresponding area of the mark value is added 1；By the corresponding power of the mark value
The cumulative weights plus current pixel point of value；The corresponding x weighted accumulations of the mark value are added to the x weights of current pixel point；It will
The corresponding y weighted accumulations of the mark value add the y weights of current pixel point.After the mark value for wherein determining current pixel point, such as
Fruit do not exist the corresponding connected domain attribute value of the mark value, determine the corresponding updated connected domain of the mark value starting point and
The coordinate of terminating point is that the coordinate of current pixel point, area 1, weights accumulated value are that the weights of current pixel point, x weightings are tired
Add as the x weights of current pixel point, the y weights that y weighted accumulations are current pixel point.
It is target label value by mark value in bianry image in equivalent sequence according to each target label value for including
After pixel is determined as the connected domain in image to be extracted, the abscissa of the corresponding starting point of each target label value is compared
Ordinate compared with, the corresponding starting point of each target label value is compared, and chooses minimum abscissa and ordinate as should
The abscissa and ordinate of the starting point of connected domain in image to be extracted；Each target scale value is remembered to the cross of corresponding terminating point
Coordinate is compared, the ordinate of the corresponding terminating point of each target label value is compared, and is chosen minimum abscissa and is indulged
Abscissa and ordinate of the coordinate as the terminating point of the connected domain in the image to be extracted；Each target label value is corresponding
Area is added the area as the connected domain in the image to be extracted；The corresponding weights of each target label value are added to be used as and are somebody's turn to do
The weights of connected domain in image to be extracted；The corresponding x weighted accumulations of each target label value are added and are used as the figure to be extracted
The x weighted accumulations of connected domain as in；The corresponding y weighted accumulations of each target label value are added as in the image to be extracted
Connected domain y weighted accumulations.
Preferably, for the ease of being connected to the determination of Domain Properties, electronic equipment can also configure one for storing connected domain
The dual port RAM of attribute value, is indicated with U2, is that each mark value distributes a storage address in U2, as mark value 1 corresponds to U2
Storage address 1, mark value 2 correspond to the storage address 2 ... of U2, and each storage address is corresponding for storing the storage address in U2
The corresponding connected domain attribute value of mark value.Electronic equipment can also configure one for storing the double of final connected domain attribute value
Mouth RAM, is indicated with U3, and electronic equipment is according to the mark value that each address stores in U1, the label with each address storage in U2
It is worth corresponding connected domain attribute value, it is right according to the equivalence stored in U1, equivalent sequence is determined, in same equivalent sequence
The corresponding connection Domain Properties of mark value merge, and determine final connected domain attribute value, are stored by U3.Preferably, in order to
The acquisition that the data of same mark value are corresponded in U1 and U2 is corresponded to for same mark value in U1 and U2 convenient for electronic equipment
Identical storage address, such as：Mark value 1 corresponds to the storage address 1 in storage address 1 and U2 in U1.
If electronic equipment handles continuous multiple image, for every frame image, start during frame blanking area
The data in U1 and U2 are arranged, final connected domain attribute value are stored by U3, the connected domain attribute value of a frame image has extracted
At.Next frame image come can start temporarily calculate next frame image connected domain attribute value.Preferably, can will be calculated
Various connected domain combinations of attributes are sequentially output at every 32 bit (bit) as an attribute.After end of output, final connected domain category
Property start from storage address 1 reset.
Embodiment 5：
Fig. 3 is a kind of connected domain extraction element structural schematic diagram provided in an embodiment of the present invention, which includes：
Conversion module 31, for image to be extracted to be converted to bianry image；
Code table generation module 32 is marked, for scanning each pixel in the bianry image successively, for what is scanned
Pixel, according to the first gray value of current pixel point, the first object pixel adjacent with current pixel point and the second target
The mark value of pixel, determines the mark value of current pixel point, and if it is determined that first mesh adjacent with the current pixel point
The first mark value and the second mark value for marking pixel and the second target pixel points be not equal and be not 0, and described first is marked
Value and second mark value are right as equivalence, are recorded in code table of equal value；
First determining module 33, it is right according to equivalence present in the equivalent table to be used for, and determines equivalent sequence；
Second determining module 34, for according to each target label value for including in the equivalent sequence, by the two-value
Mark value is that the pixel of the target label value is determined as the connected domain in the image to be extracted in image.
The label code table generation module 32 is specifically used for according to sequence from top to bottom from left to right, scanning described two
It is worth each pixel in image；Or, according to sequence from left to right from top to bottom, each pixel in the bianry image is scanned
Point；Or, according to sequence from top to bottom from right to left, each pixel in the bianry image is scanned；Or, according to from right to left
Sequence from top to bottom scans each pixel in the bianry image.
If the label code table generation module 32 scans institute specifically for according to sequence from top to bottom from left to right
Each pixel in bianry image is stated, for the pixel scanned, judges whether the first gray value of current pixel point is 0；
If the first gray value of current pixel point is 0, the mark value of current pixel point is determined as 0；If the of current pixel point
One gray value is not 0, judges the first of the first object pixel on the current pixel point left side and the second target pixel points of top
Whether mark value and the second mark value are 0；If so, the mark value of current pixel point is determined as fixed maximum mark
Value plus 1, if not, according to non-zero mark value minimum in first mark value and second mark value, determines current picture
The mark value of vegetarian refreshments.
Described device further includes：
Judgment module 35, for judging in the code table of equal value with the presence or absence of comprising first mark value and described second
The equivalence of mark value is right；If it is judged that be no, carry out subsequently using first mark value and second mark value as
It is of equal value right, the step in code table of equal value is recorded.
Described device further includes：
Update module 36, for being existed according to the mark value and the coordinate of current pixel point, current pixel point of current pixel point
The second gray value, preset gray threshold in the image to be extracted, to the corresponding connected domain attribute value of the mark value into
Row update；
Third determining module 37, for being wrapped according in the corresponding connected domain attribute value of each mark value and the equivalent sequence
The each target label value contained, determines the connected domain attribute value of the connected domain in the image to be extracted.
The invention discloses a kind of connected domain extracting method and device, the method includes：Image to be extracted is converted to
Bianry image；Each pixel in the bianry image is scanned successively, for the pixel scanned, according to current pixel point
The mark value of first gray value, the first object pixel and second target pixel points adjacent with current pixel point determines current
The mark value of pixel, and if it is determined that the first object pixel and second target pixel points adjacent with the current pixel point
The first mark value and the second mark value not etc. and be not 0, using first mark value and second mark value as etc.
Valence pair is recorded in code table of equal value；It is right according to equivalence present in the equivalent table, determine equivalent sequence；According to the equivalence
The each target label value for including in sequence determines the pixel that mark value in the bianry image is the target label value
For the connected domain in the image to be extracted.Due in embodiments of the present invention, each pixel in scanning bianry image successively
When point, if it is determined that the first mark value and second of the adjacent first object pixel and the second target pixel points of current pixel point
Mark value is not equal and is not 0, and the first mark value and the second mark value is right as equivalence, is recorded in code table of equal value, and root
According to equivalence present in code table of equal value to determining equivalent sequence, according to each target label value for including in equivalent sequence, by two
Mark value is that the pixel of target label value is determined as the connected domain in image to be extracted in value image, it is only necessary to run-down, just
It can be by different mark values be corresponded in bianry image but the pixel in same connected domain is determined as connection in image to be extracted
Domain simplifies the process of connected domain extraction, avoids and be repeatedly scanned to bianry image, reduces occupied by connected domain extraction
Process resource, improve connected domain extraction efficiency.
For systems/devices embodiment, since it is substantially similar to the method embodiment, so the comparison of description is simple
Single, the relevent part can refer to the partial explaination of embodiments of method.
It should be understood by those skilled in the art that, embodiments herein can be provided as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the application
Apply the form of example.Moreover, the application can be used in one or more wherein include computer usable program code computer
The computer program production implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
The form of product.
The application is with reference to method, the flow of equipment (system) and computer program product according to the embodiment of the present application
Figure and/or block diagram describe.It should be understood that can be realized by computer program instructions every first-class in flowchart and/or the block diagram
The combination of flow and/or box in journey and/or box and flowchart and/or the block diagram.These computer programs can be provided
Instruct the processor of all-purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine so that the instruction executed by computer or the processor of other programmable data processing devices is generated for real
The device for the function of being specified in present one flow of flow chart or one box of multiple flows and/or block diagram or multiple boxes.
These computer program instructions, which may also be stored in, can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that instruction generation stored in the computer readable memory includes referring to
Enable the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one box of block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device so that count
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, in computer or
The instruction executed on other programmable devices is provided for realizing in one flow of flow chart or multiple flows and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Although the preferred embodiment of the application has been described, created once a person skilled in the art knows basic
Property concept, then additional changes and modifications can be made to these embodiments.So it includes excellent that the following claims are intended to be interpreted as
It selects embodiment and falls into all change and modification of the application range.
Obviously, various changes and modifications can be made to the invention without departing from essence of the invention by those skilled in the art
God and range.In this way, if these modifications and changes of the present invention belongs to the range of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to include these modifications and variations.
Claims (11)
1. a kind of connected domain extracting method, which is characterized in that the method includes：
Image to be extracted is converted into bianry image；
Each pixel in the bianry image is scanned successively, for the pixel scanned, according to the first of current pixel point
The mark value of gray value, the first object pixel and second target pixel points adjacent with current pixel point, determines current pixel
The mark value of point, and if it is determined that the first object pixel and second target pixel points adjacent with the current pixel point the
One mark value and the second mark value not etc. and be not 0, will first mark value and second mark value as of equal value
It is right, it is recorded in code table of equal value；
It is right according to equivalence present in the equivalent table, determine equivalent sequence；
It is the target mark by mark value in the bianry image in the equivalent sequence according to each target label value for including
The pixel of note value is determined as the connected domain in the image to be extracted.
2. the method as described in claim 1, which is characterized in that described to scan each pixel packet in the bianry image successively
It includes：
According to sequence from top to bottom from left to right, each pixel in the bianry image is scanned；Or,
According to sequence from left to right from top to bottom, each pixel in the bianry image is scanned；Or,
According to sequence from top to bottom from right to left, each pixel in the bianry image is scanned；Or,
According to sequence from right to left from top to bottom, each pixel in the bianry image is scanned.
3. method as claimed in claim 2, which is characterized in that if according to sequence from top to bottom from left to right, scan institute
State each pixel in bianry image, it is described for the pixel scanned, according to the first gray value of current pixel point, with work as
The mark value of preceding pixel point adjacent first object pixel and the second target pixel points, determines the mark value packet of current pixel point
It includes：
For the pixel scanned, judge whether the first gray value of current pixel point is 0；
If the first gray value of current pixel point is 0, the mark value of current pixel point is determined as 0；
If the first gray value of current pixel point is not 0, first object pixel and the top on the current pixel point left side are judged
The second target pixel points the first mark value and the second mark value whether be 0；If so, by the mark value of current pixel point
It is determined as fixed maximum mark value and adds 1, if not, according to minimum in first mark value and second mark value
Non-zero mark value determines the mark value of current pixel point.
4. the method as described in claim 1, which is characterized in that described to make first mark value and second mark value
To be of equal value right, before being recorded in code table of equal value, the method further includes：
It is right with the presence or absence of the equivalence comprising first mark value and second mark value in the code table of equal value to judge；
If not, carrying out subsequent step.
5. the method as described in claim 1, which is characterized in that after the mark value of the determining current pixel point, the side
Method further includes：
According to the coordinate of the mark value of current pixel point and current pixel point, current pixel point in the image to be extracted
Second gray value, preset gray threshold are updated the corresponding connected domain attribute value of the mark value；
It is described according to each target label value for including in the equivalent sequence, be the mesh by mark value in the bianry image
The pixel of mark mark value is determined as after the connected domain in the image to be extracted, and the method further includes：
According to each target label value for including in the corresponding connected domain attribute value of each mark value and the equivalent sequence, determine
The connected domain attribute value of connected domain in the image to be extracted.
6. method as claimed in claim 5, which is characterized in that the connected domain attribute value includes：
The starting point of connected domain, terminating point, area, weights are cumulative, abscissa x weighted accumulations, ordinate y weighted accumulations.
7. a kind of connected domain extraction element, which is characterized in that described device includes：
Conversion module, for image to be extracted to be converted to bianry image；
Code table generation module is marked, for scanning each pixel in the bianry image successively, for the pixel scanned,
According to the first gray value of current pixel point, the first object pixel adjacent with current pixel point and the second target pixel points
Mark value, determines the mark value of current pixel point, and if it is determined that the first object pixel adjacent with the current pixel point
Not etc. and be not 0 with the first mark value of the second target pixel points and the second mark value, by first mark value and described
Second mark value is right as equivalence, is recorded in code table of equal value；
First determining module, it is right according to equivalence present in the equivalent table to be used for, and determines equivalent sequence；
Second determining module will be in the bianry image for according to each target label value for including in the equivalent sequence
Mark value is that the pixel of the target label value is determined as the connected domain in the image to be extracted.
8. device as claimed in claim 7, which is characterized in that the label code table generation module is specifically used for according to from upper
Sequence from left to right is arrived down, each pixel in the bianry image is scanned；Or, according to from left to right from top to bottom suitable
Sequence scans each pixel in the bianry image；Or, according to sequence from top to bottom from right to left, the binary map is scanned
Each pixel as in；Or, according to sequence from right to left from top to bottom, each pixel in the bianry image is scanned.
9. device as claimed in claim 8, which is characterized in that the label code table generation module, if specifically for according to
Sequence from left to right from top to bottom scans each pixel in the bianry image, for the pixel scanned, judges to work as
Whether the first gray value of preceding pixel point is 0；If the first gray value of current pixel point is 0, by the label of current pixel point
Value is determined as 0；If the first gray value of current pixel point is not 0, the first object pixel on the current pixel point left side is judged
Whether it is 0 with the first mark value of the second target pixel points of top and the second mark value；If so, by current pixel point
Mark value is determined as fixed maximum mark value and adds 1, if not, according in first mark value and second mark value
Minimum non-zero mark value, determines the mark value of current pixel point.
10. device as claimed in claim 7, which is characterized in that described device further includes：
Judgment module, for judging in the code table of equal value with the presence or absence of comprising first mark value and second mark value
Equivalence it is right；If it is judged that being no, it is subsequently right as equivalence using first mark value and second mark value to carry out,
The step in code table of equal value is recorded.
11. device as claimed in claim 7, which is characterized in that described device further includes：
Update module, for being waited for described according to the mark value and the coordinate of current pixel point, current pixel point of current pixel point
The second gray value, the preset gray threshold in image are extracted, the corresponding connected domain attribute value of the mark value is updated；
Third determining module, for every according to include in the corresponding connected domain attribute value of each mark value and the equivalent sequence
A target label value determines the connected domain attribute value of the connected domain in the image to be extracted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810141802.2A CN108345888A (en) | 2018-02-11 | 2018-02-11 | A kind of connected domain extracting method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810141802.2A CN108345888A (en) | 2018-02-11 | 2018-02-11 | A kind of connected domain extracting method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108345888A true CN108345888A (en) | 2018-07-31 |
Family
ID=62960170
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810141802.2A Pending CN108345888A (en) | 2018-02-11 | 2018-02-11 | A kind of connected domain extracting method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108345888A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109671081A (en) * | 2018-12-25 | 2019-04-23 | 凌云光技术集团有限责任公司 | A kind of bad cluster statistical method and device based on FPGA look-up table |
| CN109850518A (en) * | 2018-11-12 | 2019-06-07 | 太原理工大学 | A kind of real-time mining adhesive tape early warning tearing detection method based on infrared image |
| CN110517282A (en) * | 2019-08-07 | 2019-11-29 | 哈尔滨工业大学 | A kind of bianry image connected component labeling method |
| CN111213178A (en) * | 2019-03-29 | 2020-05-29 | 深圳市大疆创新科技有限公司 | Connected domain processing method, data processing device and computer readable storage medium |
| CN111833398A (en) * | 2019-04-16 | 2020-10-27 | 杭州海康威视数字技术股份有限公司 | Method and device for marking pixel points in image |
| CN111985508A (en) * | 2020-09-01 | 2020-11-24 | 安徽萤瞳科技有限公司 | Target connected domain shape analysis method suitable for linear array CCD |
| CN112017218A (en) * | 2020-09-09 | 2020-12-01 | 杭州海康威视数字技术股份有限公司 | Image registration method and device, electronic equipment and storage medium |
| CN112668676A (en) * | 2020-12-12 | 2021-04-16 | 浙江大华技术股份有限公司 | Package security inspection image processing method, electronic device and storage medium |
| CN112686919A (en) * | 2020-12-29 | 2021-04-20 | 上海闻泰信息技术有限公司 | Object boundary line determining method and device, electronic equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101727654A (en) * | 2009-08-06 | 2010-06-09 | 北京理工大学 | Method realized by parallel pipeline for performing real-time marking and identification on connected domains of point targets |
| CN104112123A (en) * | 2014-07-09 | 2014-10-22 | 宁波摩视光电科技有限公司 | Defect characteristic extraction and identification method of AOI system used for bullet apparent defect detection |
| CN104866856A (en) * | 2015-05-17 | 2015-08-26 | 西南石油大学 | Imaging log image solution cave information picking method based on connected domain equivalence pair processing |
| CN105427287A (en) * | 2015-11-06 | 2016-03-23 | 中国电子科技集团公司第二十八研究所 | Projection transformation-based connected region marking method |
| US20170103536A1 (en) * | 2015-10-13 | 2017-04-13 | Fujitsu Limited | Counting apparatus and method for moving objects |
| CN106932408A (en) * | 2017-03-20 | 2017-07-07 | 山东农业大学 | A kind of paper cocooning frame pollution silk cocoon detection method based on image procossing |
-
2018
- 2018-02-11 CN CN201810141802.2A patent/CN108345888A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101727654A (en) * | 2009-08-06 | 2010-06-09 | 北京理工大学 | Method realized by parallel pipeline for performing real-time marking and identification on connected domains of point targets |
| CN104112123A (en) * | 2014-07-09 | 2014-10-22 | 宁波摩视光电科技有限公司 | Defect characteristic extraction and identification method of AOI system used for bullet apparent defect detection |
| CN104866856A (en) * | 2015-05-17 | 2015-08-26 | 西南石油大学 | Imaging log image solution cave information picking method based on connected domain equivalence pair processing |
| US20170103536A1 (en) * | 2015-10-13 | 2017-04-13 | Fujitsu Limited | Counting apparatus and method for moving objects |
| CN105427287A (en) * | 2015-11-06 | 2016-03-23 | 中国电子科技集团公司第二十八研究所 | Projection transformation-based connected region marking method |
| CN106932408A (en) * | 2017-03-20 | 2017-07-07 | 山东农业大学 | A kind of paper cocooning frame pollution silk cocoon detection method based on image procossing |
Non-Patent Citations (1)
| Title |
|---|
| 左敏 等: ""基于等价对的图像连通域标记算法"", 《计算机仿真》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109850518A (en) * | 2018-11-12 | 2019-06-07 | 太原理工大学 | A kind of real-time mining adhesive tape early warning tearing detection method based on infrared image |
| CN109850518B (en) * | 2018-11-12 | 2022-01-28 | 太原理工大学 | Real-time mining adhesive tape early warning tearing detection method based on infrared image |
| CN109671081A (en) * | 2018-12-25 | 2019-04-23 | 凌云光技术集团有限责任公司 | A kind of bad cluster statistical method and device based on FPGA look-up table |
| CN111213178A (en) * | 2019-03-29 | 2020-05-29 | 深圳市大疆创新科技有限公司 | Connected domain processing method, data processing device and computer readable storage medium |
| CN111833398A (en) * | 2019-04-16 | 2020-10-27 | 杭州海康威视数字技术股份有限公司 | Method and device for marking pixel points in image |
| CN111833398B (en) * | 2019-04-16 | 2023-09-08 | 杭州海康威视数字技术股份有限公司 | Pixel point marking method and device in image |
| CN110517282A (en) * | 2019-08-07 | 2019-11-29 | 哈尔滨工业大学 | A kind of bianry image connected component labeling method |
| CN111985508A (en) * | 2020-09-01 | 2020-11-24 | 安徽萤瞳科技有限公司 | Target connected domain shape analysis method suitable for linear array CCD |
| CN111985508B (en) * | 2020-09-01 | 2023-09-19 | 安徽萤瞳科技有限公司 | Target connected domain shape analysis method suitable for linear array CCD |
| CN112017218A (en) * | 2020-09-09 | 2020-12-01 | 杭州海康威视数字技术股份有限公司 | Image registration method and device, electronic equipment and storage medium |
| CN112668676A (en) * | 2020-12-12 | 2021-04-16 | 浙江大华技术股份有限公司 | Package security inspection image processing method, electronic device and storage medium |
| CN112686919A (en) * | 2020-12-29 | 2021-04-20 | 上海闻泰信息技术有限公司 | Object boundary line determining method and device, electronic equipment and storage medium |
| CN112686919B (en) * | 2020-12-29 | 2022-10-21 | 上海闻泰信息技术有限公司 | Object boundary line determining method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108345888A (en) | A kind of connected domain extracting method and device | |
| CN108830280B (en) | Small target detection method based on regional nomination | |
| CN110570433B (en) | Image semantic segmentation model construction method and device based on generation countermeasure network | |
| CN110443140B (en) | Text positioning method, device, computer equipment and storage medium | |
| EP2605186B1 (en) | Method and apparatus for recognizing a character based on a photographed image | |
| CN110942071A (en) | License plate recognition method based on license plate classification and LSTM | |
| CN110399760A (en) | A kind of batch two dimensional code localization method, device, electronic equipment and storage medium | |
| CN111507337A (en) | License plate recognition method based on hybrid neural network | |
| CN111598076B (en) | Method and device for detecting and processing date in label image | |
| CN114519858A (en) | Document image recognition method and device, storage medium and electronic equipment | |
| CN108710881B (en) | Neural network model, candidate target area generation method and model training method | |
| CN110210467B (en) | Formula positioning method of text image, image processing device and storage medium | |
| CN114529909A (en) | Sample data set generation method and device and electronic equipment | |
| CN106709393B (en) | QR two-dimensional code binarization method and system | |
| CN104036272A (en) | Text detection method and electronic device | |
| CN112101113B (en) | Lightweight unmanned aerial vehicle image small target detection method | |
| US8705134B2 (en) | Method of processing an image to clarify text in the image | |
| CN111079749B (en) | End-to-end commodity price tag character recognition method and system with gesture correction | |
| CN112132780A (en) | Reinforcing steel bar quantity detection method and system based on deep neural network | |
| CN111199228A (en) | License plate positioning method and device | |
| CN114511862B (en) | Form identification method and device and electronic equipment | |
| US9798932B2 (en) | Video extraction method and device | |
| CN111191658A (en) | Texture description method and image classification method based on generalized local binary pattern | |
| CN115063679A (en) | Pavement quality assessment method based on deep learning | |
| US8908986B1 (en) | Systems and methods for selecting ink colors |
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 | ||
| CB02 | Change of applicant information |
Address after: C10, No. 1199 Bin'an Road, Binjiang District, Hangzhou City, Zhejiang Province Applicant after: Zhejiang Huarui Technology Co.,Ltd. Address before: C10, No. 1199 Bin'an Road, Binjiang District, Hangzhou City, Zhejiang Province Applicant before: ZHEJIANG HUARAY TECHNOLOGY Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180731 |
|
| RJ01 | Rejection of invention patent application after publication |