CN107993234A - A kind of extracting method and device for cheating region - Google Patents
A kind of extracting method and device for cheating region Download PDFInfo
- Publication number
- CN107993234A CN107993234A CN201610952339.0A CN201610952339A CN107993234A CN 107993234 A CN107993234 A CN 107993234A CN 201610952339 A CN201610952339 A CN 201610952339A CN 107993234 A CN107993234 A CN 107993234A
- Authority
- CN
- China
- Prior art keywords
- area
- testing image
- segmentation threshold
- default segmentation
- region
- 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
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention belongs to vision technique field, there is provided a kind of extracting method and device for cheating region, this method include：Obtain and segmentation threshold corresponding first area in testing image is each preset in multiple default segmentation thresholds；The first area is rotated according to first area angle of inclination relative to horizontal direction, and stretches postrotational first area；According to the circle similarity for obtaining the first area after stretching；The first area that circle similarity is less than after the stretching of default similarity threshold is removed, and determines that the first area after residual stretch is second area；The flat ratio of second area is obtained, and characterizes the number that the second area of the same area occurs under multiple default segmentation thresholds；If the number that the flat ratio of second area and the second area occur under the multiple default segmentation threshold meets the first preset condition, the second area is then determined as hole region and merges the hole region of characterization the same area, the accuracy rate in extraction wild environment Zhong Keng regions can be improved by the present invention.
Description
Technical field
The invention belongs to vision technique field, more particularly to a kind of extracting method and device for cheating region.
Background technology
At present, the existing extraction algorithm to cheating region is concentrated mainly in the detection to circular pit region, specifically for
The detection in wild environment Zhong Keng regions is fewer.The common detection algorithm to circular pit region has Hough transform method, feature to carry
Follow the example of, fitting process etc..However, these algorithms require that the profile in detected region has certain regularity, this is intuitively seen
The burr in region cannot excessively, distortion cannot be excessively serious, and the burr for being detected region is excessive, excessively serious distortion is exactly open country
Feature possessed by external environment hole.Therefore existing algorithm is difficult to show good identification when cheating region in face of wild environment
Effect, the accuracy rate in extraction wild environment Zhong Keng regions is than relatively low.
Therefore, it is necessary to a kind of new technical solution is proposed, to solve above-mentioned technical problem.
The content of the invention
In consideration of it, the embodiment of the present invention provides a kind of extracting method and device for cheating region, to improve extraction wild environment
The accuracy rate in Zhong Keng regions.
The first aspect of the embodiment of the present invention, there is provided a kind of extracting method for cheating region, the described method includes：
Obtain the gray value of each pixel in testing image；
Mapping is treated to described according to the gray value of each pixel and multiple default segmentation thresholds in the testing image respectively
As being split, and it is corresponding in the testing image to obtain each default segmentation threshold in the multiple default segmentation threshold
First area；
Obtaining each default segmentation threshold, corresponding first area is relative to horizontal direction in the testing image
Angle of inclination, and corresponding first area is rotated according to the angle of inclination, so that the first area and horizontal direction parallel；
According to the postrotational each default segmentation threshold of preset rules stretching corresponding the in the testing image
One region；
Obtain the barycenter of each default first area of the segmentation threshold in the testing image after corresponding stretching
Positional information, and obtain according to the positional information of the barycenter circle similarity of the first area after corresponding stretching；
Circle is removed from each default first area of the segmentation threshold in the testing image after corresponding stretching
Similarity is less than the first area after the stretching of default similarity threshold, and determines that each default segmentation threshold is treated described
First area in altimetric image after remaining corresponding stretching is right in the testing image for each default segmentation threshold
The second area answered；
The flat ratio of each default segmentation threshold corresponding second area in the testing image is obtained, and it is described
Multiple default segmentation thresholds characterize the second area of the same area described in the testing image in corresponding second area
The number occurred under multiple default segmentation thresholds；
If each default segmentation threshold in the testing image flat ratio of corresponding second area and this second
The number that region occurs under the multiple default segmentation threshold meets preset condition, it is determined that the second area is hole region；
The multiple default segmentation threshold is characterized into the more of the same area in the testing image in corresponding hole region
A hole region merging technique is hole region, to extract the hole region in the testing image.
The second aspect of the embodiment of the present invention, there is provided a kind of extraction element for cheating region, described device include：
Gray value acquisition module, for obtaining the gray value of each pixel in testing image；
Image segmentation module, for according to the gray value of each pixel and multiple default segmentation thresholds in the testing image
The testing image is split respectively, and obtains and segmentation threshold is each preset in the multiple default segmentation threshold described
Corresponding first area in testing image；
Angle acquisition module, for obtaining each default segmentation threshold corresponding firstth area in the testing image
Domain angle of inclination relative to horizontal direction, and corresponding first area is rotated according to the angle of inclination, so that firstth area
Domain and horizontal direction parallel；
Region stretching module, for being treated according to the postrotational each default segmentation threshold of preset rules stretching described
Corresponding first area in altimetric image；
Barycenter acquisition module, for obtaining each default segmentation threshold in the testing image after corresponding stretching
First area barycenter positional information, and the first area after corresponding stretching is obtained according to the positional information of the barycenter
Circle similarity；
Region removes module, for from each default segmentation threshold in the testing image after corresponding stretching
The first area that circle similarity is less than after the stretching of default similarity threshold is removed in first area, and is determined described each default
First area of the segmentation threshold in the testing image after remaining corresponding stretching exists for each default segmentation threshold
Corresponding second area in the testing image；
Number acquisition module, for obtaining each default segmentation threshold corresponding secondth area in the testing image
The flat ratio in domain, and the multiple default segmentation threshold characterize the same area in the testing image in corresponding second area
The number that occurs under the multiple default segmentation threshold of second area；
Area determination module, if for each default segmentation threshold corresponding second area in the testing image
Flat ratio and the number that occurs under the multiple default segmentation threshold of the second area meet preset condition, it is determined that this
Two regions are hole region；
Region merging technique module, for by the multiple default segmentation threshold in the testing image it is corresponding hole region in
The multiple hole region merging techniques for characterizing the same area are hole region, to extract the hole region in the testing image.
Existing beneficial effect is the embodiment of the present invention compared with prior art：The embodiment of the present invention is multiple pre- by obtaining
If each presetting segmentation threshold corresponding first area in testing image in segmentation threshold, rotated according to the angle of inclination of calculating
Corresponding first area, and the postrotational first area is stretched, it is corresponding in testing image to obtain each default segmentation threshold
Stretching after first area barycenter positional information, and obtain according to the positional information of the barycenter the after corresponding stretching
The circle similarity in one region, removes the first area being less than after the stretching of default similarity threshold, and determines remaining firstth area
Domain is second area, obtains each default segmentation threshold flat ratio of corresponding second area and secondth area in testing image
The number that domain occurs under multiple default segmentation thresholds, in the flat ratio of second area and the second area in multiple default segmentations
The number occurred under threshold value meets the first preset condition, just determines the second area for hole region.The embodiment of the present invention can be right
Irregular hole region in wild environment is detected and extracts, and removes the non-hole region in wild environment, and it is wild to improve extraction
The accuracy rate in external environment Zhong Keng regions.
Brief description of the drawings
To describe the technical solutions in the embodiments of the present invention more clearly, below will be to embodiment or description of the prior art
Needed in attached drawing be briefly described, it should be apparent that, drawings in the following description be only the present invention some
Embodiment, for those of ordinary skill in the art, without having to pay creative labor, can also be according to these
Attached drawing obtains other attached drawings.
Fig. 1 is that the extracting method in the hole region that the embodiment of the present invention one provides realizes flow chart；
Fig. 2 a are the exemplary plots divided using five rectangles to the 3rd region；Fig. 2 b are the 3 after being divided in Fig. 2 a
The exemplary plot of the gray-value variation gradient in region；
Fig. 3 is the actual figure of circumgyration stretch to first area；
Fig. 4 is to characterize the exemplary plot that multiple hole regions of the same area overlap；
Fig. 5 is the composition schematic diagram of the extraction element in hole region provided by Embodiment 2 of the present invention.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to the accompanying drawings and embodiments, it is right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Embodiment one：
Fig. 1 shows the flow of realizing of the extracting method in the hole region that the embodiment of the present invention one provides, described to realize flow
Details are as follows：
Step S101, obtains the gray value of each pixel in testing image.
In embodiments of the present invention, since the testing image directly acquired by video camera is coloured image, and exist very
More interference, there is also many tiny objects (such as gravel) in external environment out of office in addition will also result in image disruption.Therefore,
Before the gray value of each pixel in obtaining the testing image, the testing image is pre-processed.Pretreatment is as follows：
These interference are usually mutation noise in external environment out of office, without any rule, nonlinear filter can be selected and remove, common
Have median filter and two-sided filter in nonlinear filter, median filter processing speed faster, the place of two-sided filter
It is more preferable to manage effect.Testing image is transformed to gray level image by coloured image after noise remove.It is to be measured in order to further highlight
The feature in relatively large region in image, can eliminate small-sized black hole using the closed operation in Morphological scale-space.
Step S102, according to the gray value of each pixel and multiple default segmentation thresholds in the testing image respectively to institute
State testing image to be split, and obtain and segmentation threshold is each preset in the multiple default segmentation threshold in the testing image
In corresponding first area.
In embodiments of the present invention, the multiple default segmentation threshold is a default segmentation threshold sequence, presets segmentation
Threshold value is pre-set gray value, and the initial value and stop value of the sequence can carry out sets itself according to being actually needed.Compared with
Good, when testing image is substantially ground, initial value could be provided as the gray value of all pixels in the testing image
The half of average value, stop value can be that initial value adds 50.When it is non-ground that testing image is most of, initial value can be set
50 are set to, stop value can be that initial value adds 50.The initial value and stop value of above-mentioned multiple default segmentation thresholds are through excessive
What the experiment of amount obtained, it can substantially meet the process demand of all wild environments.
In embodiments of the present invention, will be treated as segmentation threshold (i.e. multiple default segmentation thresholds) using a series of gray values
Altimetric image is handled testing image into row threshold division, and using the feature after these Threshold segmentations, can thus be obtained
Feature can not be obtained to testing image progress single treatment.
Optionally, it is described right respectively according to the gray value of each pixel and multiple default segmentation thresholds in the testing image
The testing image is split, and obtains each default segmentation threshold in the multiple default segmentation threshold and treat mapping described
Corresponding first area includes as in：
Pixel of the gray value more than the first default segmentation threshold is partitioned into from all pixels of the testing image, and really
The fixed pixel is described first default segmentation threshold corresponding first pixel in the testing image；
If in the testing image there are the multiple adjacent described first default segmentation thresholds in the testing image it is right
The first pixel answered, then by the multiple the described first adjacent default segmentation threshold corresponding first in the testing image
Combination of pixels forms described first default segmentation threshold corresponding 3rd region in the testing image；
Obtain the gray-value variation ladder in described first default segmentation threshold corresponding 3rd region in the testing image
Degree, and gray-value variation gradient is removed in corresponding 3rd region in the testing image from the described first default segmentation threshold
The 3rd region of the second preset condition is unsatisfactory for, and determines that the described first default segmentation threshold is remaining in the testing image
3rd region is described first default segmentation threshold corresponding first area in the testing image；
Pixel of the gray value more than the second default segmentation threshold is partitioned into from all pixels of the testing image, and really
The fixed pixel is described first default segmentation threshold corresponding first pixel in the testing image；
If in the testing image there are the multiple adjacent described second default segmentation thresholds in the testing image it is right
The first pixel answered, then by the multiple the described second adjacent default segmentation threshold corresponding first in the testing image
Combination of pixels forms described first default segmentation threshold corresponding 3rd region in the testing image；
Obtain the gray-value variation ladder in described second default segmentation threshold corresponding 3rd region in the testing image
Degree, and gray-value variation gradient is removed in corresponding 3rd region in the testing image from the described second default segmentation threshold
The 3rd region of the second preset condition is unsatisfactory for, and determines that the described second default segmentation threshold is remaining in the testing image
3rd region is described second default segmentation threshold corresponding first area in the testing image；
And so on, until having traveled through the multiple default segmentation threshold.
In embodiments of the present invention, the first preset condition is being unsatisfactory for according to the removal of the gray-value variation gradient in the 3rd region
The 3rd region before, can combine video camera parameter and requirement of engineering a part is unsatisfactory for require region to be removed,
The limitation of this subregion size can be defined by experience, it is assumed that it is 15 pictures that region, which is limited to minimum widith, here
Element, is highly 4 pixels, then removes width and be less than 15 pixels, highly the 3rd region less than 4 pixels.The gray value
Variable gradient is the 3rd region gray-value variation situation a part of inside edge to region.The hole region of usual wild environment
Gray-value variation from periphery to center is from high to low.It can be made a return journey in practice using this property unless cheating region, so
And also there are a kind of special circumstances, that is, the gray value for cheating center is higher than the gray value in other regions in cheating.Therefore utilizing gray value
Change can only consider from the edge in hole region to an internal part when removing hole region, and ignore the situation at hole center, can keep away
Exempt to miss the possibility compared with shallow hole.
Specifically, being calculated to simplify, the 3rd region can be divided and carry out gray value change using five rectangles
The calculating of change.It is the exemplary plot divided using five rectangles to the 3rd region as shown in Figure 2 a, in fig. 2 a by the 3rd area
Domain carries out section from midline, is from top to bottom 10 zonules by the 3rd region division, i.e. 10 zonules are from the 3rd area
To center, then from center to edge, above-mentioned 10 zonule non-overlapping copies, center is the midline in Fig. 2 a at the edge in domain.It is logical
Cross total gray value that formula (1) calculates each zonule：
Wherein sumkFor total gray value of k-th of zonule, rowStart is the initial row of k-th of zonule, rowEnd
For the end line of k-th of zonule, colStart is the starting row of k-th of zonule, and colEnd is the end of k-th of cell
Row, X (i, j) are behavior i in k-th zonule, the gray value for the pixel being classified as at j positions.
The area of each zonule is calculated by formula (2)：
Nk=((colEnd-colStart) * (rowEnd-rowStart))k (2)
Wherein, NkFor the area of k-th of zonule.
The average gray value of each zonule is calculated by formula (3)：
averagek=sumk/Nk (3)
Wherein, averagekFor the average gray value of k-th of zonule.
Fig. 2 b are the gray-value variation gradients in the 3rd region after being divided in Fig. 2 a.Wherein, 11 points on abscissa from
Left-to-right to represent 10 zonules from top to bottom and center line in the 3rd region, ordinate represents average gray value, and dotted line represents
The gray-value variation situation of each pixel of midline during to three region divisions, solid line represent the flat of each zonule after division
Equal gray-value variation situation.When from edge, inwardly the average gray value of continuous three zonules is all reduced and (such as opened from top edge
The gradient of average gray value to the average gray value of second zonule of first zonule begun is negative, second zonule
The gradient of average gray value to the average gray value of the 3rd zonule be also negative；Or the tenth since lower edge is small
The gradient of the average gray value in region to the average gray value of the 9th zonule is negative, the average gray value of the 9th zonule
Gradient to the average gray value of the 8th zonule is also negative) when, it is believed that the 3rd region is doubtful hole region, otherwise to be non-
Cheat region.From the solid line in Fig. 2 b, the average gray value of the average gray value of the 9th zonule to the 8th zonule
Gradient first for be after negative just, it is undesirable, therefore corresponding 3rd regions of Fig. 2 b are non-hole region.
Exemplary, multiple default segmentation thresholds are first from treating from 51 default segmentation thresholds of gray value 50 to 100
The pixel that gray value is more than 50 is partitioned into all pixels of altimetric image, if there are multiple adjacent gray scales in the testing image
Pixel of the value more than 50, then combination of pixels formation first default segmentation threshold 50 of the plurality of adjacent gray value more than 50 is at this
Corresponding 3rd region in testing image, there may be multiple non-conterminous 3rd regions, each 3rd area in testing image
Domain all includes the pixel that multiple adjacent gray values are more than 50, and removes gray-value variation gradient not from multiple 3rd regions
Meet the 3rd region of the second preset condition；Then the picture that gray value is more than 51 is partitioned into from all pixels of testing image
Element, to be more than if there are the pixel that multiple adjacent gray values are more than 51, the plurality of adjacent gray value in the testing image
51 combination of pixels forms the second default corresponding 3rd region in the testing image of segmentation threshold 51, can in testing image
Can be there are multiple non-conterminous 3rd regions, each 3rd region includes the pixel that multiple adjacent gray values are more than 51,
And the 3rd region that gray-value variation gradient is unsatisfactory for the second preset condition is removed from multiple 3rd regions；Then from treating mapping
The pixel that gray value is more than 52 is partitioned into all pixels of picture, if there are multiple adjacent gray values are big in the testing image
In 52 pixel, then combination of pixels formation threeth default segmentation threshold 52 of the plurality of adjacent gray value more than 52 is to be measured at this
Corresponding 3rd region in image, there may be multiple non-conterminous 3rd regions, each 3rd region in testing image
Include the pixel that multiple adjacent gray values are more than 52, and remove gray-value variation gradient from multiple 3rd regions and be discontented with
3rd region of the second preset condition of foot；And so on, until having traveled through 51 default segmentation thresholds.
Step S103, obtain each default segmentation threshold in the testing image corresponding first area relative to
The angle of inclination of horizontal direction, and corresponding first area is rotated according to the angle of inclination, so that the first area and level
Direction is parallel.
In embodiments of the present invention, first area angle of inclination relative to horizontal direction refers to by the first area
In the straight line of particle, across that line angle of inclination relative to horizontal direction of the length maximum of first area.
Optionally, the acquisition each default segmentation threshold corresponding first area in the testing image is opposite
Include in the angle of inclination of horizontal direction, and according to the angle of inclination corresponding first area of rotation：
Obtain the position letter of each default segmentation threshold barycenter of corresponding first area in the testing image
Breath；
Slope will be changed to preset slope intervals in the range of predetermined angle by the straight line of the barycenter, and obtained each
Length of the corresponding straight line by the barycenter of slope in first area corresponding with the barycenter；
From the length of the corresponding straight line by the barycenter of each slope of acquisition in corresponding first area, look into
The length of maximum is looked for, and is obtained and the corresponding straight line by the barycenter of maximum length；
The slope with the corresponding line correspondences by the barycenter of maximum length is obtained, to obtain and the barycenter
Corresponding first area angle of inclination relative to horizontal direction；
First area corresponding with the barycenter is rotated according to the angle of inclination.
Specifically, the position of the barycenter of first area can be calculated by formula (4)：
Wherein, M and N represents the scope of the transverse and longitudinal coordinate of first area, x respectivelyiAnd yiPixel in first area is represented respectively
Transverse and longitudinal coordinate value.
After the positional information of barycenter of first area is got, can by by barycenter straight line according between default slope
Value is carried out from predetermined angle scope every (such as 0.1), and calculates the length for the line that each slope falls in first area, and
The length of maximum is found, that is, finds the corresponding straight line of length of maximum, the firstth area is rotated according to the slope of the line correspondences
Domain.Preferably, it is less than the width in hole region due to cheating the height in region as viewed from the perspective of video camera, therefore the predetermined angle model
Enclose forArrive
Step S104, according to the postrotational each default segmentation threshold of preset rules stretching in the testing image
Corresponding first area.
Optionally, it is described to stretch the postrotational each default segmentation threshold in the testing image according to preset rules
In corresponding first area include：
Obtain the height of postrotational each default segmentation threshold corresponding first area in the testing image
And width；
The height of postrotational each default segmentation threshold corresponding first area in the testing image is stretched,
So that the height and width of postrotational each default segmentation threshold corresponding first area in the testing image
It is identical.
In embodiments of the present invention, height and width of the postrotational first area in testing image are obtained, by first
High elongation of the region in testing image to the first area being of same size in testing image (because from video camera
The height of postrotational first area can not possibly be more than the width in the region from the point of view of angle) so that the shape of the first area is drawn
It is similar to extend circle.Due to stretching postrotational first area simply by the high elongation of first area extremely with being of same size, not
It is that each putting on the profile to the first area stretches, postrotational first area can not be stretched as justifying, can only
It is similar to justify.If Fig. 3 is the actual figure of circumgyration stretch to first area.
Step S105, obtains each default firstth area of the segmentation threshold in the testing image after corresponding stretching
The positional information of the barycenter in domain, and it is similar according to the circle of the first area after the corresponding stretching of the positional information of barycenter acquisition
Degree.
Specifically, according to m point (such as 18 points) is taken at equal intervals on the profile of first area after the stretch, will be above-mentioned
Position of the m point in testing image is respectively defined as (x1,y1), (x2,y2) ..., (xm,ym), counted by formula (5) and (6)
Each point is calculated to the average distance mean of barycenter:
di=sqrt ((xi-x0)+(yi-y0)) (5)
According to formula (7), (8) and (9) calculate the circle similarity similarity of the first area after stretching：
disi=abs (di-mean) (7)
Similarity=mean/meanl (9)
Wherein, sqrt is square root function, and abs is ABS function.
In embodiments of the present invention, from the firstth area after the stretching during circle similarity of the first area after calculating stretching
According to series of points is taken at equal intervals on the profile in domain, done so as to greatly reduce break edge to what circle similarity measure was brought
Disturb, afterwards using the average distance mean and edge at edge to barycenter to the average distance between the new circle produced by mean
Relation between meanl judges round similarity, and compared with conventional method, this considerably reduce calculation amount, improves to non-rule
Then round accuracy of judgement degree.
Step S106, from each default first area of the segmentation threshold in the testing image after corresponding stretching
The middle first area for removing circle similarity and being less than after the stretching of default similarity threshold, and determine each default segmentation threshold
First area in the testing image after remaining corresponding stretching is each default segmentation threshold described to be measured
Corresponding second area in image.
Step S107, obtain each default segmentation threshold in the testing image corresponding second area it is flat
Rate, and the multiple default segmentation threshold characterize secondth area of the same area in the testing image in corresponding second area
The number that domain occurs under the multiple default segmentation threshold.
Optionally, it is described to obtain the multiple default segmentation threshold table in corresponding second area in the testing image
The number that the second area of sign the same area occurs under the multiple default segmentation threshold includes：
Obtain the positional information of first default segmentation threshold barycenter of corresponding second area in the testing image；
Obtain the positional information of second default segmentation threshold corresponding second area in the testing image；
If the corresponding second area covering described first in the testing image of the described second default segmentation threshold is default
The barycenter of segmentation threshold corresponding second area in the testing image, it is determined that the described second default segmentation threshold is described
Corresponding second area and described first default segmentation threshold corresponding second area in the testing image in testing image
For the same area, and the number that the corresponding second area of the same area occurs under the multiple default segmentation threshold is added 1；
Obtain the position letter of described second default segmentation threshold barycenter of corresponding second area in the testing image
Breath；
Obtain the positional information of the 3rd default segmentation threshold corresponding second area in the testing image；
If the corresponding second area covering described second in the testing image of the described 3rd default segmentation threshold is default
The barycenter of segmentation threshold corresponding second area in the testing image, it is determined that the described 3rd default segmentation threshold is described
Corresponding second area and described second default segmentation threshold corresponding second area in the testing image in testing image
For the same area, and the number that the corresponding second area of the same area occurs under the multiple default segmentation threshold is added 1；
And so on, until having traveled through the multiple default segmentation threshold.
By taking gray value 50,51,52 3 default segmentation thresholds as an example, if default segmentation threshold 51 is right in testing image
The barycenter for the default corresponding second area in testing image of segmentation threshold 50 of second area covering answered, it is determined that default segmentation
Threshold value 51 corresponding second area and default segmentation threshold 50 corresponding second area in testing image in testing image is
The same area, the number that the corresponding second area of the same area occurs in the case where these three preset segmentation threshold are 1；It is if default
The corresponding second area covering in testing image of segmentation threshold 52 presets segmentation threshold 51 corresponding second in testing image
The barycenter in region, it is determined that default segmentation threshold 53 corresponding second area and default segmentation threshold 51 in testing image are being treated
Corresponding second area is the same area in altimetric image, due to presetting the corresponding second area in testing image of segmentation threshold 51
Corresponding second area is the same area in testing image with default segmentation threshold 50, then above three presets segmentation threshold
50th, 51,52 in testing image corresponding second area be the same area, the corresponding second area of the same area this three
The number occurred under a default segmentation threshold is 2.
Step S108, if the flat ratio of each default segmentation threshold corresponding second area in the testing image
The number occurred with the second area under the multiple default segmentation threshold meets the first preset condition, it is determined that secondth area
Domain is hole region.
In embodiments of the present invention, the flat ratio of second area for height of the second area in the testing image and
The ratio of width.
In embodiments of the present invention, due to video camera shooting angle and the distance difference apart from subject, doubtful hole region
(i.e. second area) can show the different ellipse of flat ratio, elliptical flat ratio more remote apart from video camera in testing image
It is higher, otherwise lower apart from the nearlyer flat ratio of video camera, correspondingly, the higher region of flat ratio is in multiple default segmentations
The number (number occurred under multiple default segmentation thresholds) being detected under threshold value is more, the lower region of flat ratio
The number being detected under multiple default segmentation thresholds is fewer.Therefore region quilt under multiple default segmentation thresholds can be utilized
Relation between the number and flat ratio that detect further judges doubtful hole region.Assuming that video camera is vertically shot
When cheating region, hole region occurrence number under multiple default segmentation thresholds is 10, and flat ratio at this time is 1, video camera distance hole
When region is farthest, the number that hole region occurs under multiple default segmentation thresholds is 1, and flat ratio (is assumed to take the photograph for 0.26 at this time
The minimum widith in the hole region that camera is capable of detecting when is 15 pixels, and minimum constructive height is 4 pixels, and flat ratio is at this time), function y=12.1621*x-2.1621 is so can be obtained by, wherein y represents hole region in multiple default segmentations
The number occurred under threshold value, x represent the flat ratio in the hole region.If the flat ratio of the second area in step S108 and more
The number occurred under a default segmentation threshold meets above-mentioned formula, it is determined that the second area is hole region, if being unsatisfactory for above-mentioned
Formula, it is determined that the second area is non-hole region.
Step S109, by the multiple default segmentation threshold in the testing image it is corresponding hole region in characterize it is same
Multiple hole region merging techniques in region are hole region, to extract the hole region in the testing image.
The present invention implement in, due to multiple default segmentation thresholds in testing image it is corresponding hole region in characterize it is same
Multiple hole regions in region are there may be lap, if Fig. 4 is showing of characterizing that multiple hole regions of the same area overlap
Illustration.Multiple hole regions of overlapping characterization the same area are merged, are hole region by multiple hole region merging techniques, should
Cheating region includes the Zone Full in multiple hole regions of characterization the same area, such as first overlapping cases in Fig. 4, region 1
Inclusion region 2, the hole region after merging are region 1.It is a hole by the multiple hole region merging techniques that will characterize same hole region
Region, can avoid extracting the repetition for cheating region.
It should be noted that if the quantity in the hole region of characterization the same area is 1, then without merging, the hole is directly extracted
Region.
The embodiment of the present invention is feasible by verification experimental verification, cheats picture and 100 width using 100 web and is surveyed without hole picture
Examination, test environment matlab2014, machine are configured to CPU：Core i5-5200U, 2.2GHz.RAM is 4GB, DDR3L, is grasped
It is windows7 (32bits) as system, result of the test such as table 1.
1 result of the test of table
| Project | As a result |
| Detection success rate with hole picture | 92% |
| False alarm rate with hole picture | 18% |
| Without the false alarm rate of hole picture | 4% |
| Average operating time with hole picture | 7.596 the second |
| Without the average operating time of hole picture | 5.304 the second |
It should be noted that the embodiment of the present invention is to establish circulation, time of circulation according to the multiple default segmentation threshold
Number is identical with the quantity of default segmentation threshold, and circulation is exactly the processing procedure under a default segmentation threshold each time, every
Doubtful hole area is removed according to the circle similarity of gray-value variation gradient, circumgyration stretch first area, first area in one cycle
Domain, after circulation terminates, occurs secondary according to the flat ratio of second area and the second area under multiple default segmentation thresholds
Several relations removes possible non-hole region.
The embodiment of the present invention can be detected and extract to the irregular hole region in wild environment, remove wild environment
In non-hole region, improve the accuracy rate in extraction wild environment Zhong Keng regions.
Embodiment two：
Fig. 5 shows the composition schematic diagram of the extraction element in hole region provided by Embodiment 2 of the present invention, for the ease of saying
It is bright, illustrate only with the relevant part of the embodiment of the present invention, details are as follows：
Gray value acquisition module 51, for obtaining the gray value of each pixel in testing image；
Image segmentation module 52, for according to the gray value of each pixel in the testing image and multiple default segmentation thresholds
Value respectively splits the testing image, and obtains and segmentation threshold is each preset in the multiple default segmentation threshold in institute
State corresponding first area in testing image；
Angle acquisition module 53, for obtaining each default segmentation threshold corresponding first in the testing image
Region angle of inclination relative to horizontal direction, and corresponding first area is rotated according to the angle of inclination so that this first
Region and horizontal direction parallel；
Region stretching module 54, for stretching the postrotational each default segmentation threshold described according to preset rules
Corresponding first area in testing image；
Barycenter acquisition module 55, for obtaining each default segmentation threshold corresponding stretching in the testing image
The positional information of the barycenter of first area afterwards, and the first area after corresponding stretching is obtained according to the positional information of the barycenter
Circle similarity；
Region removes module 56, for from each default segmentation threshold in the testing image after corresponding stretching
First area in remove circle similarity and be less than the first area after the stretching of default similarity threshold, and determine described each pre-
If first area of the segmentation threshold in the testing image after remaining corresponding stretching is each default segmentation threshold
The corresponding second area in the testing image；
Number acquisition module 57, for obtaining each default segmentation threshold corresponding second in the testing image
The flat ratio in region, and the multiple default segmentation threshold characterize same area in the testing image in corresponding second area
The number that the second area in domain occurs under the multiple default segmentation threshold；
Area determination module 58, if for each default segmentation threshold corresponding secondth area in the testing image
The number that the flat ratio in domain and the second area occur under the multiple default segmentation threshold meets the first preset condition, then really
The fixed second area is hole region；
Region merging technique module 59, for by the multiple default segmentation threshold in the testing image corresponding hole region
Multiple hole region merging techniques of middle characterization the same area are hole region, to extract the hole region in the testing image.
Optionally, described image segmentation module 52 includes：
First cutting unit, for being partitioned into gray value from all pixels of the testing image more than first default point
The pixel of threshold value is cut, and determines the pixel for described first default segmentation threshold corresponding first picture in the testing image
Element；
First assembled unit, if to exist for there are the multiple adjacent described first default segmentation thresholds in the testing image
Corresponding first pixel in the testing image, then by the multiple the described first adjacent default segmentation threshold described to be measured
Corresponding first combination of pixels forms described first default segmentation threshold corresponding 3rd area in the testing image in image
Domain；
First acquisition unit, for obtaining described first default segmentation threshold corresponding 3rd area in the testing image
The gray-value variation gradient in domain, and from the described first default segmentation threshold in the testing image in corresponding 3rd region
Except gray value variable gradient is unsatisfactory for the 3rd region of the second preset condition, and determine the described first default segmentation threshold described
Remaining 3rd region is described first default segmentation threshold corresponding first area in the testing image in testing image；
Second cutting unit, for being partitioned into gray value from all pixels of the testing image more than second default point
The pixel of threshold value is cut, and determines the pixel for described first default segmentation threshold corresponding first picture in the testing image
Element；
Second assembled unit, if to exist for there are the multiple adjacent described second default segmentation thresholds in the testing image
Corresponding first pixel in the testing image, then by the multiple the described second adjacent default segmentation threshold described to be measured
Corresponding first combination of pixels forms described first default segmentation threshold corresponding 3rd area in the testing image in image
Domain；
Second acquisition unit, for obtaining described second default segmentation threshold corresponding 3rd area in the testing image
The gray-value variation gradient in domain, and from the described second default segmentation threshold in the testing image in corresponding 3rd region
Except gray value variable gradient is unsatisfactory for the 3rd region of the second preset condition, and determine the described second default segmentation threshold described
Remaining 3rd region is described second default segmentation threshold corresponding first area in the testing image in testing image；
Traversal Unit, be used for and so on, until having traveled through the multiple default segmentation threshold.
Optionally, the angle acquisition module 53 includes：
Barycenter acquiring unit, for obtaining each default segmentation threshold corresponding firstth area in the testing image
The positional information of the barycenter in domain；
Slope changes unit, for will be changed by the straight line of the barycenter with presetting slope intervals in the range of predetermined angle
Variable slope, and obtain length of the corresponding straight line by the barycenter of each slope in the corresponding first area of the barycenter
Degree；
Length searching unit, for the corresponding straight line by the barycenter of each slope from acquisition corresponding first
In length in region, the length of maximum is searched, and obtain and the corresponding straight line by the barycenter of maximum length；
Slope acquiring unit, it is oblique with the corresponding line correspondences by the barycenter of maximum length for obtaining
Rate, to obtain first area angle of inclination relative to horizontal direction corresponding with the barycenter；
Region rotary unit, for rotating first area corresponding with the barycenter according to the angle of inclination.
Optionally, the region stretching module 54 includes：
High width acquiring unit, corresponds to for obtaining the postrotational each default segmentation threshold in the testing image
First area height and width；
High elongation unit, corresponds to for stretching the postrotational each default segmentation threshold in the testing image
First area height so that the postrotational each default segmentation threshold is corresponding first in the testing image
The height in region and it is of same size.
Optionally, the number acquisition module 57 includes：
First acquisition unit, for obtaining first default segmentation threshold corresponding second area in the testing image
The positional information of barycenter；
Second acquisition unit, for obtaining second default segmentation threshold corresponding second area in the testing image
Positional information；
First determination unit, if for described second default segmentation threshold corresponding second area in the testing image
Cover the barycenter of described first default segmentation threshold corresponding second area in the testing image, it is determined that described second is pre-
If segmentation threshold corresponding second area in the testing image presets segmentation threshold in the testing image with described first
In corresponding second area be the same area, and by the corresponding second area of the same area in the multiple default segmentation threshold
The number of lower appearance adds 1；
3rd acquiring unit, for obtaining described second default segmentation threshold corresponding secondth area in the testing image
The positional information of the barycenter in domain；
4th acquiring unit, for obtaining the 3rd default segmentation threshold corresponding second area in the testing image
Positional information；
Second determination unit, if for the described 3rd default segmentation threshold corresponding second area in the testing image
Cover the barycenter of described second default segmentation threshold corresponding second area in the testing image, it is determined that the described 3rd is pre-
If segmentation threshold corresponding second area in the testing image presets segmentation threshold in the testing image with described second
In corresponding second area be the same area, and by the corresponding second area of the same area under the multiple default segmentation threshold
The number of appearance adds 1；
Traversal Unit, be used for and so on, until having traveled through the multiple default segmentation threshold.
The extraction element in hole region provided in an embodiment of the present invention can be used in foregoing corresponding embodiment of the method one,
Details are referring to the description of above-described embodiment one, and details are not described herein.
The technical staff in the field can be understood that, for convenience and simplicity of description, only with above-mentioned each function
The division progress of module, can be as needed and by above-mentioned function distribution by different function moulds for example, in practical application
Block is completed, i.e. the internal structure of described device is divided into different function modules, and above-mentioned function module can both use hardware
Form is realized, can also be realized in the form of software.In addition, the specific name of each function module is also only to facilitate mutually
Difference, is not limited to the protection domain of the application.
In conclusion the embodiment of the present invention can be detected and extract to the irregular hole region in wild environment, go
Except the non-hole region in wild environment, the accuracy rate for extracting wild environment Zhong Keng regions is improved.
Those of ordinary skill in the art are further appreciated that all or part of step realized in above-described embodiment method is can
To instruct relevant hardware to complete by program, the program can be stored in a computer read/write memory medium
In, the storage medium, including ROM/RAM, disk, CD etc..
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
All any modification, equivalent and improvement made within refreshing and principle etc., should all be included in the protection scope of the present invention.
Claims (10)
- A kind of 1. extracting method for cheating region, it is characterised in that the described method includes：Obtain the gray value of each pixel in testing image；According to the gray value of each pixel in the testing image and multiple default segmentation thresholds respectively to the testing image into Row segmentation, and obtain and segmentation threshold is each preset in the multiple default segmentation threshold corresponding first in the testing image Region；Obtain each default segmentation threshold corresponding first area is relative to horizontal direction in the testing image and incline Rake angle, and corresponding first area is rotated according to the angle of inclination, so that the first area and horizontal direction parallel；According to postrotational each default segmentation threshold corresponding firstth area in the testing image of preset rules stretching Domain；Obtain the position of the barycenter of each default first area of the segmentation threshold in the testing image after corresponding stretching Confidence ceases, and the circle similarity of the first area after corresponding stretching is obtained according to the positional information of the barycenter；It is similar that circle is removed from each default first area of the segmentation threshold in the testing image after corresponding stretching Degree is less than the first area after the stretching of default similarity threshold, and determines that each default segmentation threshold treats mapping described First area as in after remaining corresponding stretching is corresponding in the testing image for each default segmentation threshold Second area；The flat ratio of each default segmentation threshold corresponding second area in the testing image is obtained, and it is the multiple Default segmentation threshold characterizes the second area of the same area the multiple in the testing image in corresponding second area The number occurred under default segmentation threshold；If described each default the segmentation threshold flat ratio of corresponding second area and the second area in the testing image The number occurred under the multiple default segmentation threshold meets the first preset condition, it is determined that the second area is hole region；The multiple default segmentation threshold is characterized to multiple holes of the same area in the testing image in corresponding hole region Region merging technique is hole region, to extract the hole region in the testing image.
- 2. the according to the method described in claim 1, it is characterized in that, gray scale according to each pixel in the testing image Value and multiple default segmentation thresholds respectively split the testing image, and obtain every in the multiple default segmentation threshold A default segmentation threshold corresponding first area in the testing image includes：Pixel of the gray value more than the first default segmentation threshold is partitioned into from all pixels of the testing image, and determines to be somebody's turn to do Pixel is described first default segmentation threshold corresponding first pixel in the testing image；If in the testing image there are the multiple adjacent described first default segmentation thresholds in the testing image it is corresponding First pixel, then by the multiple described first adjacent default segmentation threshold corresponding first pixel in the testing image Combination forms described first default segmentation threshold corresponding 3rd region in the testing image；The gray-value variation gradient in described first default segmentation threshold corresponding 3rd region in the testing image is obtained, and Gray-value variation gradient is removed in corresponding 3rd region in the testing image from the described first default segmentation threshold to be discontented with 3rd region of the second preset condition of foot, and determine the described first default segmentation threshold the remaining 3rd in the testing image Region is described first default segmentation threshold corresponding first area in the testing image；Pixel of the gray value more than the second default segmentation threshold is partitioned into from all pixels of the testing image, and determines to be somebody's turn to do Pixel is described first default segmentation threshold corresponding first pixel in the testing image；If in the testing image there are the multiple adjacent described second default segmentation thresholds in the testing image it is corresponding First pixel, then by the multiple described second adjacent default segmentation threshold corresponding first pixel in the testing image Combination forms described first default segmentation threshold corresponding 3rd region in the testing image；The gray-value variation gradient in described second default segmentation threshold corresponding 3rd region in the testing image is obtained, and Gray-value variation gradient is removed in corresponding 3rd region in the testing image from the described second default segmentation threshold to be discontented with 3rd region of the second preset condition of foot, and determine the described second default segmentation threshold the remaining 3rd in the testing image Region is described second default segmentation threshold corresponding first area in the testing image；And so on, until having traveled through the multiple default segmentation threshold.
- 3. according to the method described in claim 1, it is characterized in that, the acquisition each default segmentation threshold is treated described Corresponding first area angle of inclination relative to horizontal direction in altimetric image, and rotate corresponding first according to the angle of inclination Region includes：Obtain the positional information of each default segmentation threshold barycenter of corresponding first area in the testing image；Slope will be changed to preset slope intervals in the range of predetermined angle by the straight line of the barycenter, and obtain each slope Length of the corresponding straight line by the barycenter in first area corresponding with the barycenter；From the length of the corresponding straight line by the barycenter of each slope of acquisition in corresponding first area, search most Big length, and obtain and the corresponding straight line by the barycenter of maximum length；The slope with the corresponding line correspondences by the barycenter of maximum length is obtained, it is corresponding with the barycenter to obtain First area angle of inclination relative to horizontal direction；First area corresponding with the barycenter is rotated according to the angle of inclination.
- It is 4. according to the method described in claim 1, it is characterized in that, described postrotational described each according to preset rules stretching Default segmentation threshold corresponding first area in the testing image includes：Obtain the height and width of postrotational each default segmentation threshold corresponding first area in the testing image Degree；The height of postrotational each default segmentation threshold corresponding first area in the testing image is stretched, so that The postrotational each default segmentation threshold height of corresponding first area and is of same size in the testing image.
- 5. method according to any one of claims 1 to 4, it is characterised in that described to obtain the multiple default segmentation threshold It is worth in the testing image in corresponding second area and characterizes the second area of the same area in the multiple default segmentation threshold The lower number occurred of value includes：Obtain the positional information of first default segmentation threshold barycenter of corresponding second area in the testing image；Obtain the positional information of second default segmentation threshold corresponding second area in the testing image；If the described second default segmentation threshold corresponding described first default segmentation of second area covering in the testing image The barycenter of threshold value corresponding second area in the testing image, it is determined that the described second default segmentation threshold is described to be measured Corresponding second area is same with described first default segmentation threshold corresponding second area in the testing image in image One region, and the number that the corresponding second area of the same area occurs under the multiple default segmentation threshold is added 1；Obtain the positional information of described second default segmentation threshold barycenter of corresponding second area in the testing image；Obtain the positional information of the 3rd default segmentation threshold corresponding second area in the testing image；If the described 3rd default segmentation threshold corresponding described second default segmentation of second area covering in the testing image The barycenter of threshold value corresponding second area in the testing image, it is determined that the described 3rd default segmentation threshold is described to be measured Corresponding second area is same with described second default segmentation threshold corresponding second area in the testing image in image One region, and the number that the corresponding second area of the same area occurs under the multiple default segmentation threshold is added 1；And so on, until having traveled through the multiple default segmentation threshold.
- 6. a kind of extraction element for cheating region, it is characterised in that described device includes：Gray value acquisition module, for obtaining the gray value of each pixel in testing image；Image segmentation module, for being distinguished according to the gray value of each pixel in the testing image and multiple default segmentation thresholds The testing image is split, and obtains and segmentation threshold is each preset in the multiple default segmentation threshold described to be measured Corresponding first area in image；Angle acquisition module, for obtaining each default segmentation threshold corresponding first area phase in the testing image Rotate corresponding first area for the angle of inclination of horizontal direction, and according to the angle of inclination so that the first area with Horizontal direction parallel；Region stretching module, for treating mapping described according to the postrotational each default segmentation threshold of preset rules stretching The corresponding first area as in；Barycenter acquisition module, for obtaining each default segmentation threshold the in the testing image after corresponding stretching The positional information of the barycenter in one region, and according to the circle phase of the first area after the corresponding stretching of the positional information of barycenter acquisition Like degree；Region removes module, for each presetting first of segmentation threshold in the testing image after corresponding stretching from described The first area that circle similarity is less than after the stretching of default similarity threshold is removed in region, and determines each default segmentation First area of the threshold value in the testing image after remaining corresponding stretching is each default segmentation threshold described Corresponding second area in testing image；Number acquisition module, for obtaining each default segmentation threshold corresponding second area in the testing image Flat ratio, and the multiple default segmentation threshold characterize the of the same area in the testing image in corresponding second area The number that two regions occur under the multiple default segmentation threshold；Area determination module, if for each default segmentation threshold in the testing image corresponding second area it is flat The number that flat rate and the second area occur under the multiple default segmentation threshold meets the first preset condition, it is determined that this Two regions are hole region；Region merging technique module, for the multiple default segmentation threshold to be characterized in the testing image in corresponding hole region Multiple hole region merging techniques of the same area are hole region, to extract the hole region in the testing image.
- 7. device according to claim 6, it is characterised in that described image segmentation module includes：First cutting unit, for being partitioned into gray value from all pixels of the testing image more than the first default segmentation threshold The pixel of value, and determine the pixel for described first default segmentation threshold corresponding first pixel in the testing image；First assembled unit, if in the testing image there are the multiple adjacent described first default segmentation thresholds described Corresponding first pixel in testing image, then by the multiple the described first adjacent default segmentation threshold in the testing image In corresponding first combination of pixels form described first default segmentation threshold corresponding 3rd region in the testing image；First acquisition unit, for obtaining described first default segmentation threshold corresponding 3rd region in the testing image Gray-value variation gradient, and from the described first default segmentation threshold ash disposal is gone in corresponding 3rd region in the testing image Angle value variable gradient is unsatisfactory for the 3rd region of the second preset condition, and determines the described first default segmentation threshold described to be measured Remaining 3rd region is described first default segmentation threshold corresponding first area in the testing image in image；Second cutting unit, for being partitioned into gray value from all pixels of the testing image more than the second default segmentation threshold The pixel of value, and determine the pixel for described first default segmentation threshold corresponding first pixel in the testing image；Second assembled unit, if in the testing image there are the multiple adjacent described second default segmentation thresholds described Corresponding first pixel in testing image, then by the multiple the described second adjacent default segmentation threshold in the testing image In corresponding first combination of pixels form described first default segmentation threshold corresponding 3rd region in the testing image；Second acquisition unit, for obtaining described second default segmentation threshold corresponding 3rd region in the testing image Gray-value variation gradient, and from the described second default segmentation threshold ash disposal is gone in corresponding 3rd region in the testing image Angle value variable gradient is unsatisfactory for the 3rd region of the second preset condition, and determines the described second default segmentation threshold described to be measured Remaining 3rd region is described second default segmentation threshold corresponding first area in the testing image in image；Traversal Unit, be used for and so on, until having traveled through the multiple default segmentation threshold.
- 8. device according to claim 6, it is characterised in that the angle acquisition module includes：Barycenter acquiring unit, for obtaining each default segmentation threshold corresponding first area in the testing image The positional information of barycenter；Slope changes unit, for will be changed tiltedly with presetting slope intervals in the range of predetermined angle by the straight line of the barycenter Rate, and obtain length of the corresponding straight line by the barycenter of each slope in the corresponding first area of the barycenter；Length searching unit, for the corresponding straight line by the barycenter of each slope from acquisition in corresponding first area In length in, search maximum length, and obtain and the corresponding straight line by the barycenter of maximum length；Slope acquiring unit, for obtaining the slope with the corresponding line correspondences by the barycenter of maximum length, with Obtain first area angle of inclination relative to horizontal direction corresponding with the barycenter；Region rotary unit, for rotating first area corresponding with the barycenter according to the angle of inclination.
- 9. device according to claim 6, it is characterised in that the region stretching module includes：High width acquiring unit, for obtaining the postrotational each default segmentation threshold corresponding the in the testing image The height and width in one region；High elongation unit, for stretching the postrotational each default segmentation threshold corresponding the in the testing image The height in one region, so that postrotational each default segmentation threshold corresponding first area in the testing image Height and be of same size.
- 10. according to claim 6 to 9 any one of them device, it is characterised in that the number acquisition module includes：First acquisition unit, for obtaining the barycenter of first default segmentation threshold corresponding second area in the testing image Positional information；Second acquisition unit, for obtaining the position of second default segmentation threshold corresponding second area in the testing image Information；First determination unit, if the corresponding second area covering in the testing image for the described second default segmentation threshold The barycenter of described first default segmentation threshold corresponding second area in the testing image, it is determined that described second default point It is right in the testing image to cut threshold value corresponding second area and described first default segmentation threshold in the testing image The second area answered is the same area, and the corresponding second area of the same area is gone out under the multiple default segmentation threshold Existing number adds 1；3rd acquiring unit, for obtaining described second default segmentation threshold corresponding second area in the testing image The positional information of barycenter；4th acquiring unit, for obtaining the position of the 3rd default segmentation threshold corresponding second area in the testing image Information；Second determination unit, if the corresponding second area covering in the testing image for the described 3rd default segmentation threshold The barycenter of described second default segmentation threshold corresponding second area in the testing image, it is determined that the described 3rd default point It is right in the testing image to cut threshold value corresponding second area and described second default segmentation threshold in the testing image The second area answered is the same area, and the corresponding second area of the same area is occurred under the multiple default segmentation threshold Number add 1；Traversal Unit, be used for and so on, until having traveled through the multiple default segmentation threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610952339.0A CN107993234A (en) | 2016-10-26 | 2016-10-26 | A kind of extracting method and device for cheating region |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610952339.0A CN107993234A (en) | 2016-10-26 | 2016-10-26 | A kind of extracting method and device for cheating region |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107993234A true CN107993234A (en) | 2018-05-04 |
Family
ID=62029481
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610952339.0A Pending CN107993234A (en) | 2016-10-26 | 2016-10-26 | A kind of extracting method and device for cheating region |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107993234A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102920477A (en) * | 2012-03-05 | 2013-02-13 | 杭州弘恩医疗科技有限公司 | Device and method for determining target region boundary of medical image |
| CN102944226A (en) * | 2012-12-03 | 2013-02-27 | 哈尔滨工业大学 | Meteor crater detecting method based on bright and dark area pairing |
| CN103093463A (en) * | 2013-01-15 | 2013-05-08 | 南京航空航天大学 | Meteor crater detecting method based on gray level image |
| CN104537673A (en) * | 2015-01-09 | 2015-04-22 | 西安电子科技大学 | Infrared image segmentation method based on multiple threshold values and self-adaptation fuzzy clustering |
| CN104637057A (en) * | 2015-02-04 | 2015-05-20 | 昆明理工大学 | Grayscale-gradient entropy multi-threshold fast division method based on genetic algorithm |
-
2016
- 2016-10-26 CN CN201610952339.0A patent/CN107993234A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102920477A (en) * | 2012-03-05 | 2013-02-13 | 杭州弘恩医疗科技有限公司 | Device and method for determining target region boundary of medical image |
| CN102944226A (en) * | 2012-12-03 | 2013-02-27 | 哈尔滨工业大学 | Meteor crater detecting method based on bright and dark area pairing |
| CN103093463A (en) * | 2013-01-15 | 2013-05-08 | 南京航空航天大学 | Meteor crater detecting method based on gray level image |
| CN104537673A (en) * | 2015-01-09 | 2015-04-22 | 西安电子科技大学 | Infrared image segmentation method based on multiple threshold values and self-adaptation fuzzy clustering |
| CN104637057A (en) * | 2015-02-04 | 2015-05-20 | 昆明理工大学 | Grayscale-gradient entropy multi-threshold fast division method based on genetic algorithm |
Non-Patent Citations (4)
| Title |
|---|
| 孟令江: ""基于立体视觉的野外环境坑区域识别研究"", 《HTTP://IR.SIA.CN/HANDLE/173321/19675》 * |
| 孟令江等: ""野外环境下图像中坑区域的提取方法"", 《计算机应用》 * |
| 方青松等: ""基于光斑室内移动机器人的定位导航技术"", 《微型机与应用》 * |
| 郭烈等: ""基于二维最大熵阈值分割的坑识别方法"", 《计算机工程与应用》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wei et al. | Counterfeit iris detection based on texture analysis | |
| Gopalakrishnan et al. | Salient region detection by modeling distributions of color and orientation | |
| CN101246544B (en) | Iris positioning method based on boundary point search and minimum kernel value similarity region edge detection | |
| Kumar et al. | A fast DCT based method for copy move forgery detection | |
| CN104464079B (en) | Multiple Currencies face amount recognition methods based on template characteristic point and topological structure thereof | |
| CN103678504B (en) | Similarity-based breast image matching image searching method and system | |
| CN107092871B (en) | Remote sensing image building detection method based on multiple dimensioned multiple features fusion | |
| CN105225523B (en) | A kind of parking space state detection method and device | |
| CN105160330B (en) | A kind of automobile logo identification method and vehicle-logo recognition system | |
| Short et al. | Latent fingerprint segmentation using ridge template correlation | |
| Ma et al. | Text detection in natural images based on multi-scale edge detetion and classification | |
| CN107330027A (en) | A kind of Weakly supervised depth station caption detection method | |
| CN103945228A (en) | Video intra-frame copy-move tampering detection method based on space and time relevance | |
| CN104268595B (en) | general object detection method and system | |
| CN109635669A (en) | Image classification method, the training method of device and disaggregated model, device | |
| CN108446690A (en) | A kind of human face in-vivo detection method based on various visual angles behavioral characteristics | |
| Rangkuti et al. | Batik image retrieval based on similarity of shape and texture characteristics | |
| Ardizzone et al. | Visual saliency by keypoints distribution analysis | |
| CN111950568A (en) | Embroidery stitch pattern feature point extraction and matching method | |
| CN107993233A (en) | A kind of localization method and device for cheating region | |
| WO2022178945A1 (en) | Face swapping forgery detection method and system based on three-dimensional facial shape, and device | |
| CN108629771B (en) | A kind of blind evaluation method of picture quality with scale robustness | |
| CN113496215A (en) | Method and device for detecting human face of living body and electronic equipment | |
| CN107993234A (en) | A kind of extracting method and device for cheating region | |
| CN110288616A (en) | A method of based on dividing shape and RPCA to divide hard exudate in eye fundus image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180504 |