CN107231551B - A kind of image detecting method and device - Google Patents
A kind of image detecting method and device Download PDFInfo
- Publication number
- CN107231551B CN107231551B CN201610178431.6A CN201610178431A CN107231551B CN 107231551 B CN107231551 B CN 107231551B CN 201610178431 A CN201610178431 A CN 201610178431A CN 107231551 B CN107231551 B CN 107231551B
- Authority
- CN
- China
- Prior art keywords
- gradient
- image
- detected
- feature vector
- amplitude
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N17/00—Diagnosis, testing or measuring for television systems or their details
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The embodiment provides a kind of image detecting method and devices, it is related to technical field of image processing, it solves in the prior art since high definition television is when carrying out clarity enhancing processing to the high-definition image got, it is the high-definition image obtained after image enhanced processing or true high-definition image that the high-definition image, which can not be told, so as to cause the high-definition image obtained after image enhanced processing occurred image distortion problems are shown in high definition television.This method comprises: by obtaining image to be detected, and extract the geometrical characteristic of image to be detected；At least two groups Gradient Features of the geometrical characteristic of image to be detected are determined by edge detection operator；Obtain the corresponding gradient amplitude figure of every group of Gradient Features of the geometrical characteristic of image to be detected；According to the feature vector of every width gradient amplitude figure of image to be detected and the precedence information of corresponding edge detection operator, determine whether image to be detected passes through scaling processing.The present invention is applied to image detection.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of image detecting methods and device.
Background technique
With the development at full speed of liquid crystal tv technology, high definition and ultra high-definition TV have become the market mainstream.However, point
The lower standard definition signal source of resolution is also still widely used in radio and television, Internet communication and multimedia recreation.When super
When high definition television receives the standard definition signal source of low resolution, the video signal source of these low resolution is had to pass through at image amplification
The high-definition image obtained after reason can be shown on ultra high-definition LCD screen.For example, the image of 480P resolution ratio is complete in 1080P
When being shown on high-definition liquid crystal screen, since the image resolution ratio of 480P is 720*480, the full HD liquid crystal display of 1080P
Resolution ratio is 1920*1080, needs to increase 720 pixels in horizontal direction by way of linear interpolation under normal conditions
1920 are added to, 480 pixels of vertical direction are increased to 1080, so that it is full of entire picture by pixel, into
And it realizes and shows the image of low resolution on full HD LCD screen.
However, the high-definition image either obtained after image enhanced processing or true high-definition image are being shown in
When on high-definition liquid crystal screen, it is required to high definition television and carries out clarity enhancing processing to it, carry out the image effect so that showing
More aesthetically pleasing nature.But since high definition television is when carrying out clarity enhancing processing to the high-definition image got, Wu Fafen
Discerning the high-definition image is the high-definition image obtained after image enhanced processing or true high-definition image, and can be according to phase
Same clarity adjustment amplitude carries out clarity adjustment to both high-definition images.Therefore, finer clear for needing to carry out
The high-definition image of clear degree enhancing processing obtained after image enhanced processing, will appear image when showing in high definition television
The phenomenon that distortion.
Summary of the invention
The embodiment of the present invention provides a kind of image detecting method and device, solves in the prior art due to high definition television
When carrying out clarity enhancing processing to the high-definition image got, can not tell the high-definition image is by image amplification
The high-definition image or true high-definition image obtained after reason, so as to cause the high definition figure obtained after image enhanced processing
As showing occurred image distortion problems in high definition television.
In order to achieve the above objectives, the embodiment of the present invention adopts the following technical scheme that
In a first aspect, providing a kind of image detecting method, comprising:
Image to be detected is obtained, and extracts the geometrical characteristic of described image to be detected；
At least two groups Gradient Features of the geometrical characteristic of described image to be detected are determined by edge detection operator, it is described every
The corresponding edge detection operator of group Gradient Features is different；
Obtain the corresponding gradient amplitude figure of every group of Gradient Features of the geometrical characteristic of described image to be detected；
According to the feature vector of every width gradient amplitude figure of described image to be detected and corresponding edge detection operator
Precedence information, determines whether described image to be detected passes through scaling processing.
On the other hand, a kind of image detection device is provided, comprising:
First obtains module, for obtaining image to be detected, and extracts the geometrical characteristic of described image to be detected；
First determining module, at least two of the geometrical characteristic for determining described image to be detected by edge detection operator
Group Gradient Features, the corresponding edge detection operator of every group of Gradient Features are different；
Second obtains module, the corresponding gradient of every group of Gradient Features of the geometrical characteristic for obtaining described image to be detected
Map of magnitudes；
Second determining module, for the feature vector and correspondence according to every width gradient amplitude figure of described image to be detected
Edge detection operator precedence information, determine whether described image to be detected passes through scaling processing.
The image detecting method and device that the embodiment of the present invention provides, by obtaining image to be detected, and are extracted to be checked
The geometrical characteristic of altimetric image, secondly, determining at least two groups gradient of the geometrical characteristic of image to be detected by edge detection operator
Feature, the corresponding edge detection operator of every group of Gradient Features are different；Then, every group of ladder of the geometrical characteristic of image to be detected is obtained
Spend the corresponding gradient amplitude figure of feature；Finally, according to the feature vector and correspondence of every width gradient amplitude figure of image to be detected
Edge detection operator precedence information, determine whether image to be detected passes through scaling processing.Compared with the existing technology, due to
Image by scaling processing using the poor definition between the obtained different gradient amplitude figures of different edge detection operators away from compared with
Greatly, therefore, several gradients obtained based on different edge detection operators that this programme passes through the geometrical characteristic to image to be detected
The feature vector of map of magnitudes is compared, to can determine whether the image to be detected carried out scaling processing.In this way, working as
When terminal carries out clarity processing to high-definition image, it can be amplified based on the testing result of the image to be detected to by image
The high-definition image and true high-definition image obtained after processing carries out corresponding clarity enhancing processing respectively, so that through
Crossing the high-definition image obtained after image enhanced processing is not in image fault when showing in high definition television.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be in embodiment or description of the prior art
Required attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is only some realities of the invention
Example is applied, it for those of ordinary skill in the art, without creative efforts, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is a kind of method flow diagram of image detecting method provided in an embodiment of the present invention；
Geometrical characteristic figure of the Fig. 2 by original image provided in an embodiment of the present invention and extraction original image；
Fig. 3 is integer Gaussian template figure provided in an embodiment of the present invention；
Fig. 4 is the original image of width landscape painting provided in an embodiment of the present invention and treated by two kinds of different operators
Image；
Fig. 5 is that Projection Character provided in an embodiment of the present invention converts schematic diagram；
Fig. 6 is a kind of structural schematic diagram of image detection device provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is 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.
The executing subject of image detecting method provided in an embodiment of the present invention can be image detection device, or for holding
The terminal of the above-mentioned image detecting method of row.Specifically, the terminal can for smart television, Intelligent television terminal, high definition set top box,
Tablet computer, laptop, Ultra-Mobile PC's (English: Ultra-mobile Personal Computer, letter
Claim: UMPC), net book, personal digital assistant (English: Personal Digital Assistant, referred to as: PDA) etc. terminals.
Wherein, image detection device can be the central processing unit (English: Central Processing Unit, letter in above-mentioned terminal
: CPU) it either can be referred to as control unit or functional module in above-mentioned terminal.
Image to be detected provided in an embodiment of the present invention can in signal source original SD image, high-definition image or
Full HD image and ultra high-definition image, when being also possible to standard definition signal source and being input to high definition terminal, high definition terminal is by amplification
The high-definition image obtained after reason, when can also be that high-definition signal source is input to SD terminal, SD terminal is after diminution is handled
Obtained SD image.It should be noted that it will be apparent to those skilled in the art that hereinafter mentioned " image to be detected "
It could alternatively be above-mentioned mentioned any one of image, meanwhile, for convenience of explanation, hereinafter mentioned " high definition "
It is the general designation of " high definition, full HD and ultra high-definition ".
The terms "and/or", only a kind of incidence relation for describing affiliated partner, indicates that there may be three kinds of passes
System, for example, A and/or B, can indicate: individualism A exists simultaneously A and B, these three situations of individualism B.In addition, herein
Middle character "/" typicallys represent the relationship that forward-backward correlation object is a kind of "or".
For the ease of clearly describing the technical solution of the embodiment of the present invention, in an embodiment of the present invention, use " the
One ", the printed words such as " second " distinguish function or the essentially identical identical entry of effect or similar item, and those skilled in the art can
To understand that the printed words such as " first ", " second " are not defined quantity and execution order.
The embodiment of the present invention provides a kind of image detecting method, as shown in Figure 1, this method comprises the following steps:
101, image detection device obtains image to be detected, and extracts the geometrical characteristic of image to be detected.
Wherein, the geometrical characteristic of image to be detected in the present embodiment is pixel grey scale occur in image to be detected sharply to become
The set of those of change pixel, the geometrical characteristic of the detection image includes but is not limited to: image edge information and image detail letter
Breath.Wherein, above-mentioned image edge information includes the Pixel Information for constituting image outline pixel, above-mentioned image detail information packet
The angle point information of image is included, and above-mentioned angle point refers to occur the violent pixel of brightness change or image border curve in image
The pixel of upper curvature maximum.
Illustratively, image detection device extracts the marginal information of image to be detected and detailed information and can be examined by edge
Measuring and calculating extracts.Specifically, edge detection operator is by taking information measure operator as an example here, the marginal information of obtained image
It is referred to Fig. 2 with detailed information, a is image to be detected, and b is by the extracted image geometry feature of information measure operator
Figure.
In addition, filtering out the interference in image and noise, and in order to eliminate information unrelated in image to be detected in order to extensive
Useful real information in multiple image to be detected, and enhance detectability for information about and simplify data to the maximum extent, from
And improve the reliability of feature extraction.Preferably, image can also be carried out to image to be detected before realizing step 101 to locate in advance
Reason operation.
Illustratively, image detection device can use low pass when carrying out the pretreatment operation to image to be detected
Filter carries out denoising to it.
For example, image detection device can realize denoising by way of Gauss weighting or Gaussian convolution, thus
Obtain the filter value of image to be detected.Wherein, the principle of Gauss weighting is that weighting is real twice respectively using two one-dimensional Gaussian kernels
Existing, the principle of Gaussian convolution is realized by convolution of a two-dimensional Gaussian kernel.
Example one is weighted by Gauss and realizes that pretreated details are provided below:
1) one-dimensional Gaussian integer template count is obtained
The one-dimensional Gaussian function of one discretization are as follows:
Wherein, σ is the constant greater than zero, indicates the degree of control noise reduction.
In order to which convenience of calculation is here by taking integer Gaussian template as an example.Illustratively, if the spectral window of the integer Gaussian template
When mouth figure is as shown in the figure a in Fig. 2, corresponding integer Gaussian mode domain is referred to the figure b in Fig. 2, specifically, in Fig. 2
Integer Gaussian template filter window figure for, detailed process that the corresponding one-dimensional integer Gaussian template of figure b in Fig. 2 generates
It is as follows:
Assuming that the size of integer Gaussian template is 2*N+1, so the variation range of x is [- N, N]；According to symmetry need
The data between [- N, 0] are generated, the minimum value g_min of N+1 floating number is sought；By all floating point values divided by most
The numerical value that the quotient of small value g_min obtains multiplied by 2 is the integer numerical value of the corresponding integer Gaussian template of each point of template.
2) filter value of image to be detected is obtained by weighted calculation twice according to obtained one-dimensional integer Gaussian template.With
For N=1, integer Gaussian template is after weighted calculation twice, filter value of the obtained image at (i, j) are as follows:
The process of one-dimensional Gaussian smoothing be before this (2*N+1) a data around horizontal (vertical direction) central pixel point with
(2*N+1) a template data carries out the value of convolution to replace original centerpoint value；After horizontal (vertical) directional smoothing, then
A template data of (2*N+1) a data and (2*N+1) carried out around vertical (horizontal direction) central pixel point carries out convolution
Value replaces original centerpoint value.
Example two realizes that pretreated details are provided below by Gaussian convolution:
1) two-dimensional integer Gaussian template numerical value is obtained
The two-dimensional Gaussian function of one discretization are as follows:Wherein, σ be greater than
Zero constant indicates the degree of control noise reduction.
In order to which convenience of calculation is here by taking integer Gaussian template as an example.Illustratively, if the spectral window of the integer Gaussian template
When mouth figure is as shown in the figure a in Fig. 3, corresponding integer Gaussian mode domain is referred to the figure b in Fig. 3, specifically, in Fig. 3
Integer Gaussian template filter window figure for, detailed process that the corresponding two-dimensional integer Gaussian template of figure b in Fig. 3 generates
It is as follows:
Assuming that the size of integer Gaussian template is (2*N+1) * (2*N+1)；So the variation range of x, y are [- N, N]；Root
Only need to generate x according to symmetry, data of the y between [- N, 0] seek the minimum value g_ of (N+1) * (N+1) a floating number
min；All floating point values is high for the corresponding integer of each point of template divided by the numerical value that the quotient of minimum value g_min obtains multiplied by 2
The integer numerical value of this template.
2) filter value of image to be detected is obtained by a convolutional calculation according to obtained two-dimensional integer Gaussian template.With
For N=1, two-dimensional integer Gaussian template is after a convolutional calculation, filter value of the obtained image at (i, j)
Are as follows:
Dimensional Gaussian smoothing process is (2*N+1) * (2*N+1) a data and (2*N+1) * (2*N around the pixel of center
+ 1) a template data carries out the value of convolution to replace original centerpoint value.
Preferably, in order to open brightness and chrominance separation, to be suitable for image processing process.To image to be detected into
The conversion that can also carry out color space before row pretreatment to image to be detected, i.e., by the color space of image to be detected
RGB is converted to YUV.
102, image detection device determines at least two groups ladder of the geometrical characteristic of image to be detected by edge detection operator
Spend feature.
Specifically, the Gradient Features of the geometrical characteristic of the image to be detected include gradient magnitude and gradient direction.The edge
Detective operators include but is not limited to: Roberts Cross operator, Prewitt operator, Sobel operator, Kirsch operator, compass
Operator, Canny operator, Laplacian operator and high pass operator.
Illustratively, image detection device carries out gradient spy by geometrical characteristic of the edge detection operator to image to be detected
Sign detection, calculates at least two groups Gradient Features of the geometrical characteristic of image to be detected.Specifically, every group of Gradient Features are corresponding
Edge detection operator is different.For example, being first gradient feature by the Gradient Features that Canny operator is calculated, by high pass
The Gradient Features that operator obtains are the second Gradient Features, and specific calculating process is as described below.
Example one, the gradient magnitude of the first gradient feature obtained by Canny operator and the specific steps of gradient direction
It is as described below:
1) according to the Canny operator matrix of acquisition obtain x to y to first-order partial derivative.
Specifically, obtaining Canny operator matrix template, the Canny operator matrix template are as follows:Then, to above-mentioned matrix sxAnd sySeek first-order partial derivative respectively, obtain x to y to
First-order partial derivative difference it is as follows:
Gx(i, j)=(I (i, j+1)-I (i, j)+I (i+1, j+1)-I (i+1, j))/2 (formula 5)
Gy(i, j)=(I (i, j)+I (i+1, j)+I (i, j+1)-I (i+1, j+1))/2 (formula 6)
2) according to x to y to first-order partial derivative calculate the gradient magnitude and gradient direction of first gradient feature
By above-mentioned formula 5 and formula 6, the mathematic(al) representation of available gradient magnitude and gradient direction are as follows:
θ (i, j)=arctan (Gy(i,j)/Gx(i, j)) (formula 8)
In addition, for convenience of calculation, after getting Gradient Features, can also to the gradient magnitude in the Gradient Features into
Row non-maxima suppression.
Example two, the gradient magnitude of the second Gradient Features obtained by high pass operator and the specific of gradient direction cross step
It is as described below:
1) according to the high pass operator matrix of acquisition obtain x to y to first-order partial derivative
Specifically, obtaining high pass operator matrix template, the high pass operator matrix template are as follows:Then, to above-mentioned matrix sxAnd syFirst-order partial derivative is sought respectively, is obtained
To x to y to first-order partial derivative difference it is as follows:
2) according to x to y to first-order partial derivative calculate the gradient magnitude and gradient direction of the second Gradient Features
By above-mentioned formula 9 and formula 10, the mathematic(al) representation of available gradient magnitude and gradient direction are as follows:
θ h (i, j)=arctan (Gy(i,j)/Gx(i, j)) (formula 12)
In addition, for convenience of calculation, after getting Gradient Features, can also to the gradient magnitude in the Gradient Features into
Row non-maxima suppression.
103, image detection device obtains the corresponding gradient amplitude of every group of Gradient Features of the geometrical characteristic of image to be detected
Figure.
In the present embodiment, image detection device is special based on geometry of the different edge detection operators to image to be detected
After sign carries out edge detection, the identical gradient amplitude figure with image to be detected size (M*N) will be obtained, i.e., in the present embodiment
Every width gradient amplitude figure of image to be detected is identical as the image size of image to be detected.Illustratively, it is mentioned with the figure a in Fig. 4
For the original image of the width landscape painting supplied, figure b is by canny operator treated image, and figure c is by high pass operator
Treated image.By the b and c in Fig. 4, this it appears that, the image obtained after the processing of canny operator is obviously than warp
Cross the image clearly obtained after the processing of high pass operator.Therefore different edge detection operators can be released to the several of identical image
What feature carries out edge detection, and the display effect of obtained gradient amplitude figure is different.
104, feature vector and corresponding side of the image detection device according to every width gradient amplitude figure of image to be detected
The precedence information of edge detective operators, determines whether image to be detected passes through scaling processing.
Wherein, the feature vector of above-mentioned gradient amplitude figure includes the pixel value of all elements in gradient amplitude figure.Example
Property, the precedence information of above-mentioned edge detection operator includes the priority level of the edge detection operator, wherein edge inspection
The priority level for calculating son is higher, then the clarity of the corresponding gradient amplitude figure of the edge detection operator is compared to priority etc.
The clarity of the low corresponding gradient amplitude figure of edge detection operator of grade is higher, specifically, the edge detection that priority level is high
The display effect of the corresponding gradient amplitude figure of the operator edge detection operator corresponding gradient amplitude figure low better than priority level,
That is the image outline point of the corresponding gradient amplitude figure of the high edge detection operator of priority level and the quantity of image angle point is greater than
The image outline point and image angle point and image angle point of the corresponding gradient amplitude figure of the low edge detection operator of priority level
Quantity.
The image detecting method that the embodiment of the present invention provides, by obtaining image to be detected, and extracts image to be detected
Geometrical characteristic, secondly, determining at least two groups Gradient Features of the geometrical characteristic of image to be detected by edge detection operator, often
The corresponding edge detection operator of group Gradient Features is different；Then, every group of Gradient Features of the geometrical characteristic of image to be detected are obtained
Corresponding gradient amplitude figure；Finally, according to the feature vector of every width gradient amplitude figure of image to be detected and corresponding edge
The precedence information of detective operators, determines whether image to be detected passes through scaling processing.Compared with the existing technology, due to by contracting
The image of processing is put using the poor definition between the obtained different gradient amplitude figures of different edge detection operators away from larger, because
This, several gradient amplitude figures obtained based on different edge detection operators that this programme passes through the geometrical characteristic to image to be detected
Feature vector be compared, to can determine whether the image to be detected carried out scaling processing.In this way, working as terminal pair
It, can be based on the testing result of the image to be detected, to after image enhanced processing when high-definition image carries out clarity processing
Obtained high-definition image and true high-definition image carries out corresponding clarity enhancing processing, so that by image respectively
The high-definition image obtained after enhanced processing is not in image fault when showing in high definition television.
Embodiment one:
Illustratively, when above-mentioned at least two groups Gradient Features include first gradient feature and the second Gradient Features, this
One Gradient Features correspond to first gradient map of magnitudes, and when which corresponds to the second gradient amplitude figure, step 104 can lead to
Following two implementation is crossed to realize:
Specifically, in the first implementation, step 104 includes the following steps:
104a1, image detection device are according to the pixel value of all elements in first gradient map of magnitudes and the second gradient amplitude
The pixel value of all elements in figure determines the element in first gradient map of magnitudes and the second gradient amplitude figure in same area
Pixel value meets the number of the element of the first predetermined condition.
First predetermined condition is precedence information and based on the corresponding edge detection operator of first gradient map of magnitudes
What the precedence information of the corresponding edge detection operator of two gradient amplitude figures obtained.
Illustratively, if the priority of corresponding first operator of first gradient map of magnitudes is corresponding higher than the second gradient amplitude figure
The second operator priority, then image detection device can be by same area in first gradient map of magnitudes and the second gradient amplitude figure
The pixel value of element is greater than the pixel value of element in the second gradient amplitude figure, and first gradient amplitude in interior first gradient map of magnitudes
The pixel of difference in figure in the pixel value of element and the second gradient amplitude figure between the pixel value of element less than the second predetermined threshold
Point screens, and determines whether image to be detected passes through scaling processing according to the number of these pixels；If first gradient width
The priority that degree schemes corresponding first operator is lower than the priority of corresponding second operator of the second gradient amplitude figure, then image detection
Device can be by the pixel of first gradient map of magnitudes and element in first gradient map of magnitudes in same area in the second gradient amplitude figure
It is worth the pixel value less than element in the second gradient amplitude figure, and the pixel value of element and the second gradient width in first gradient map of magnitudes
The absolute value of difference in degree figure between the pixel value of element is screened less than the pixel of the second predetermined threshold, and according to these
The number of pixel determines whether image to be detected passes through scaling processing.
104a2, image detection device are according to the member in first gradient map of magnitudes and the second gradient amplitude figure in same area
The number that the pixel value of element meets the element of the first predetermined condition determines whether image to be detected passes through scaling processing.
Illustratively, if the pixel value of the element in first gradient map of magnitudes and the second gradient amplitude figure in same area is full
When the number of the element of the first predetermined condition of foot is greater than the first predetermined threshold, then determine image to be detected for by scaling processing
Image；If the pixel value of the element in first gradient map of magnitudes and the second gradient amplitude figure in same area meets the first predetermined item
When the number of the element of part is less than or equal to the first predetermined threshold, then determine that image to be detected is the figure without scaling processing
Picture.
It should be noted that above-mentioned the first predetermined threshold and the second predetermined threshold is obtained according to experimental result
Empirical value.
Illustratively, by taking Canny operator and high pass operator as an example, first gradient map of magnitudes is denoted as GC here, by second
Gradient amplitude seal is GCD.
One specific example is enumerated to the first implementation, realizes that process is as follows:
1) it, counts at same position, GC value ratio GCD value is big, and the absolute value of difference is less than the number of TH, and specific procedure is such as
Shown in lower:
For (j=0；j<h；j++)
{
For (i=0；i<w；i++)
{if(GC(i,j)>GCD(i,j)&|GC(i,j)-GCD(i,j)|<TH)
{num_total++；
}
}
}
Wherein, w is picture traverse, and h is picture altitude, and TH is the second predetermined threshold, takes default value 100, can configure, with
The increase of image resolution ratio accordingly become larger.
2), num_total is normalized, obtains the corresponding confidence level parameter of element
Nout=num_total/ (w*h)；
3) do you, judge flag=Nout > THf? 1:0；Flag=1 indicates that input picture was scaled, and flag=0 indicates input
Image is not scaled processing.
Wherein, THf is the first predetermined threshold, is to be obtained by experiment, while exporting confidence level parameter Nout.Nout value
More being proximate to, a possibility that 1 expression image was amplified is bigger, and Nout value indicates image closer to original graph nearer it is to 0
Picture.
Embodiment two
Since image detection device is when the two-dimensional feature vector to gradient amplitude figure calculates, calculating process is excessively numerous
Trivial, therefore, the present embodiment realizes the spectral discrimination process of step 104 by providing another implementation, i.e., by ladder
Spend map of magnitudes two-dimensional feature vector carry out dimension-reduction treatment, thus directly according to the one-dimensional characteristic of the gradient amplitude figure after dimensionality reduction to
Amount is calculated, and calculating process is simplified.
Specifically, in the second implementation, step 104 includes the following steps:
104b1, image detection device carry out Projection Character transformation along X-axis and Y-axis respectively to every width gradient amplitude figure, obtain
The the first Y-axis feature vector and the first X-axis feature vector of first gradient map of magnitudes and the second Y-axis feature of the second gradient amplitude figure
Vector sum the second X-axis feature vector.
104b2, image detection device are according to the pictures of all elements in the first X-axis feature vector and the second X-axis feature vector
Element value determines that the pixel value of the element in the first X-axis feature vector and the second X-axis feature vector in same area meets second
The number Num1 of the element of predetermined condition, and according to the picture of all elements in the first Y-axis feature vector and the second Y-axis feature vector
Element value determines that the pixel value of the element in the first Y-axis feature vector and the second Y-axis feature vector in same area meets second
The number Num2 of the element of predetermined condition.
Second predetermined condition is precedence information and based on the corresponding edge detection operator of first gradient map of magnitudes
What the precedence information of the corresponding edge detection operator of two gradient amplitude figures obtained.
Illustratively, if the priority of corresponding first operator of first gradient map of magnitudes is corresponding higher than the second gradient amplitude figure
The second operator priority, then image detection device can be by same area in first gradient map of magnitudes and the second gradient amplitude figure
The pixel value of element in the corresponding first X-axis feature vector of interior first gradient map of magnitudes is corresponding greater than the second gradient amplitude figure
The second X-axis feature vector in element pixel value, and the pixel value of element and the second X-axis are special in the first X-axis feature vector
The pixel that difference between the pixel value of element in sign vector is less than third predetermined threshold screens, the point screened
Number be denoted as Num1；Simultaneously by the first gradient amplitude in first gradient map of magnitudes and the second gradient amplitude figure in same area
Scheme the element in corresponding first Y-axis feature vector pixel value be greater than the corresponding second Y-axis feature of the second gradient amplitude figure to
The pixel value of element in amount, and the pixel value of element and the element in the second Y-axis feature vector in the first Y-axis feature vector
Difference between pixel value is screened less than the pixel of the 4th predetermined threshold, and the number of the point screened is denoted as Num2,
And determine whether image to be detected passes through scaling processing according to Num1 and Num2.
If the priority of corresponding first operator of first gradient map of magnitudes corresponding lower than the second gradient amplitude figure second is calculated
The priority of son, then image detection device can be by first gradient map of magnitudes and first in same area in the second gradient amplitude figure
The pixel value of element in the corresponding first X-axis feature vector of gradient amplitude figure twoth X corresponding less than the second gradient amplitude figure
The pixel value of element in axis feature vector, and in the first X-axis feature vector in the pixel value of element and the second X-axis feature vector
Element pixel value between difference be less than third predetermined threshold pixel, screen, the number of the point screened
It is denoted as Num1；It is simultaneously that first gradient map of magnitudes is corresponding with the first gradient map of magnitudes in same area in the second gradient amplitude figure
The first Y-axis feature vector in element pixel value the second Y-axis feature vector corresponding less than the second gradient amplitude figure in
The pixel value of element, and in the first Y-axis feature vector the pixel value of element and the element in the second Y-axis feature vector pixel value
Between difference screened less than the pixel of the 4th predetermined threshold, the number of the point screened is denoted as Num2, and according to
Num1 and Num2 determines whether image to be detected passes through scaling processing.
104b3, image detection device determine to be checked according to the number Num1 and Num2 of the element for meeting the second predetermined condition
Whether altimetric image passes through scaling processing.
Illustratively, if the pixel value of the element in first gradient map of magnitudes and the second gradient amplitude figure in same area is full
When the number of the element of the second predetermined condition of foot is greater than five predetermined thresholds, then determine image to be detected for by scaling processing
Image；If the pixel value of the element in first gradient map of magnitudes and the second gradient amplitude figure in same area meets the second predetermined item
When the number of the element of part is less than or equal to five predetermined thresholds, then determine that image to be detected is the figure without scaling processing
Picture.
It should be noted that above-mentioned third predetermined threshold, the 4th predetermined threshold and the 5th predetermined threshold are bases
The empirical value that experimental result obtains.
Illustratively, if by taking two gradient amplitude figures as an example, and first gradient map of magnitudes is denoted as GC here, by the second ladder
Degree map of magnitudes is denoted as GCD.Specifically, Projection Character referring to Figure 5 converts schematic diagram, the gradient amplitude in step 104b1
Dimensionality reduction (i.e. Projection Character converts: image is projected respectively along x-axis or y-axis) process of figure is as described below.
(1) it is projected along x-axis to y-axis
GC projects to obtain along x-axis to y-axis:
GCD projects to obtain along x-axis to y-axis:
Wherein, sgcy (i) is i-th of element of transformed feature vector SGCY (the first Y-axis feature vector), sgcdy
It (i) is i-th of element of transformed feature vector SGCDY (the second Y-axis feature vector), w is input picture width, and i is to work as
The y-coordinate of preceding pixel.Assuming that the case where being 4Kx2k, SGCY and SGCDY are the vector of 2k dimension.
(2) it is projected along y-axis to x-axis
GC projects to obtain along y-axis to x-axis:
GCD projects to obtain along y-axis to x-axis:
Wherein, sgcx (i, j) is i-th of element of transformed feature vector SGCX (the first X-axis feature vector),
Sgcdx (i, j) is i-th of element of transformed feature vector SGCDX (the second X-axis feature vector), and h is input picture height
Degree, i is the x coordinate of current pixel.Assuming that the case where being 4Kx2k, vector SGCX and SGCDX are the vector of 4k dimension.
Illustratively, by taking Canny operator and high pass operator as an example, first gradient map of magnitudes is denoted as GC here, by second
Gradient amplitude seal is GCD, GC and GCD is carried out respectively x-axis and y-axis direction projection obtain the first X-axis feature vector SGCX,
First Y-axis feature vector SGCY, the second X-axis feature vector SGCDX and the second Y-axis feature vector SGCDY.
1) it, counts at SGCX and SGCDX same position, SGCX value ratio SGCDX value is big, and the absolute value of difference is less than THh
Number, specific procedure is as follows:
For (i=0；i<w；i++)
{if(sgcx(i)>sgcdx(i)&|sgcx(i)-sgcdx(i)|<THh)
{num_h++；
}
}
Wherein, w is picture traverse, and THh is third predetermined threshold, can take default value 100, can configure, with image
The increase of resolution ratio accordingly becomes larger.
2) it, counts at SGCY and SGCDY same position, SGCY value ratio SGCDY value is big, and the absolute value of difference is less than THv
Number, specific procedure is as follows:
For (i=0；i<h；i++)
{if(sgcy(i)>sgcdy(i)&|sgcy(i)-sgcdy(i)|<THv)
{num_v++；
}
}
Wherein, h is picture altitude, and the 4th predetermined threshold of THv can take default value 100, can configure, with image point
The increase of resolution accordingly becomes larger；
3), num_h and num_v are normalized
N0=num_h/w
N1=num_v/h
4) the corresponding confidence level parameter of element for the condition that meets, is calculated, is the larger value in N0 and N1
Nout=max (N0, N1)
5) do you, judge flag=Nout > THf? 1:0；Flag=1 indicates that input picture was scaled, and flag=0 indicates input
Image is not scaled processing.
Wherein, THf is the 5th predetermined threshold, which is to be obtained by experiment, while exporting confidence level parameter Nout.Nout
Value is bigger more being proximate to a possibility that 1 expression image was amplified, and Nout value indicates image closer to original nearer it is to 0
Image.
A kind of image detecting method that the embodiment of the present invention provides, by obtaining image to be detected, and is extracted to be detected
The geometrical characteristic of image, secondly, determining that at least two groups gradient of the geometrical characteristic of image to be detected is special by edge detection operator
Sign, the corresponding edge detection operator of every group of Gradient Features are different；Then, every group of gradient of the geometrical characteristic of image to be detected is obtained
The corresponding gradient amplitude figure of feature；Finally, according to the feature vector of every width gradient amplitude figure of image to be detected and corresponding
The precedence information of edge detection operator, determines whether image to be detected passes through scaling processing.This sample plan passes through to be checked
The feature vector of several gradient amplitude figures of the geometrical characteristic of altimetric image is compared, so that it is determined that whether going out the image to be detected
Carried out scaling processing.In this way, when terminal carries out clarity processing to high-definition image, it can be based on the inspection of the image to be detected
It surveys as a result, it is corresponding clear to carry out respectively to the high-definition image and true high-definition image obtained after image enhanced processing
Enhancing processing is spent, so that the high-definition image obtained after image enhanced processing is not in when showing in high definition television
Image fault.
The embodiment of the present invention provides a kind of image detection device, which examines for realizing above-mentioned image
Survey method, as shown in fig. 6, the image detection device 2 includes: the first acquisition module 21, the first determining module 22, second acquisition mould
Block 23 and the second determining module 24, in which:
First obtains module 21, for obtaining image to be detected, and extracts the geometrical characteristic of image to be detected.
First determining module 22, at least two groups of the geometrical characteristic for determining image to be detected by edge detection operator
Gradient Features, the corresponding edge detection operator of every group of Gradient Features are different.
Second obtains module 23, the corresponding gradient width of every group of Gradient Features of the geometrical characteristic for obtaining image to be detected
Degree figure.
Second determining module 24, for according to the feature vector of every width gradient amplitude figure of image to be detected and corresponding
The precedence information of edge detection operator, determines whether image to be detected passes through scaling processing.
Illustratively, above-mentioned Gradient Features include gradient magnitude and gradient direction, every width gradient width of the image to be detected
Degree figure is identical as the image size of the image to be detected.The feature vector of the gradient amplitude figure includes owning in the gradient amplitude figure
The pixel value of element, above-mentioned at least two groups Gradient Features include first gradient feature and the second Gradient Features, and the first gradient is special
Levy corresponding first gradient map of magnitudes, the corresponding second gradient amplitude figure of second Gradient Features.
Optionally, the second above-mentioned determining module 24 is specifically used for:
According to the pixel of all elements in the pixel value of all elements in first gradient map of magnitudes and the second gradient amplitude figure
It is predetermined to determine that the pixel value of the element in first gradient map of magnitudes and the second gradient amplitude figure in same area meets first for value
The number of the element of condition.
First predetermined condition is precedence information and based on the corresponding edge detection operator of first gradient map of magnitudes
What the precedence information of the corresponding edge detection operator of two gradient amplitude figures obtained；
Meet first according to the pixel value of the element in first gradient map of magnitudes and the second gradient amplitude figure in same area
The number of the element of predetermined condition determines whether image to be detected passes through scaling processing.
Optionally, above-mentioned second determining module 24 is according to same zone in first gradient map of magnitudes and the second gradient amplitude figure
The number that the pixel value of element in domain meets the element of the first predetermined condition determines whether image to be detected passes through scaling processing
When, it is specifically used for:
If it is pre- that the pixel value of the element in first gradient map of magnitudes and the second gradient amplitude figure in same area meets first
When the number of the element of fixed condition is greater than the first predetermined threshold, determine that image to be detected is the image Jing Guo scaling processing；
If it is pre- that the pixel value of the element in first gradient map of magnitudes and the second gradient amplitude figure in same area meets first
When the number of the element of fixed condition is less than or equal to the first predetermined threshold, determine that image to be detected is the figure without scaling processing
Picture.
Optionally, above-mentioned second determining module 24 is specifically used for:
Projection Character transformation is carried out along X-axis and Y-axis respectively to every width gradient amplitude figure, obtains the of first gradient map of magnitudes
The the second Y-axis feature vector and the second X-axis feature of one Y-axis feature vector and the first X-axis feature vector and the second gradient amplitude figure
Vector；
According to the pixel value of all elements in the first X-axis feature vector and the second X-axis feature vector, the first X-axis is determined
The pixel value of element in feature vector and the second X-axis feature vector in same area meets of the element of the second predetermined condition
Number Num1, and according to the pixel value of all elements in the first Y-axis feature vector and the second Y-axis feature vector, determine the first Y-axis
The pixel value of element in feature vector and the second Y-axis feature vector in same area meets of the element of the second predetermined condition
Number Num2.
Second predetermined condition is precedence information and based on the corresponding edge detection operator of first gradient map of magnitudes
What the precedence information of the corresponding edge detection operator of two gradient amplitude figures obtained；
Determine whether image to be detected passes through scaling processing according to Num1 and Num2.
A kind of image detection device that the embodiment of the present invention provides, the image detection device is by obtaining mapping to be checked
Picture, and the geometrical characteristic of image to be detected is extracted, secondly, determining the geometrical characteristic of image to be detected by edge detection operator
At least two groups Gradient Features, the corresponding edge detection operator of every group of Gradient Features are different；Then, the geometry of image to be detected is obtained
The corresponding gradient amplitude figure of every group of Gradient Features of feature；Finally, according to the feature of every width gradient amplitude figure of image to be detected
The precedence information of vector and corresponding edge detection operator, determines whether image to be detected passes through scaling processing.Relative to
The prior art, since the image Jing Guo scaling processing uses between the obtained different gradient amplitude figures of different edge detection operators
Poor definition is away from larger, and therefore, this programme is obtained by the geometrical characteristic to image to be detected based on different edge detection operators
To the feature vectors of several gradient amplitude figures be compared, to can determine whether the image to be detected carried out scaling
Processing.In this way, when terminal carries out clarity processing to high-definition image, it can be right based on the testing result of the image to be detected
The high-definition image and true high-definition image obtained after image enhanced processing, carries out respectively at corresponding clarity enhancing
Reason, so that the high-definition image obtained after image enhanced processing is not in that image loses when showing in high definition television
Very.
In several embodiments provided herein, it should be understood that disclosed terminal and method can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can be tied
Another system is closed or is desirably integrated into, or some features can be ignored or not executed.Another point, it is shown or discussed
Mutual coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING or logical of device or unit
Letter connection can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that the independent physics of each unit includes, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or the network equipment etc.) executes the portion of each embodiment the method for the present invention
Step by step.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (Read-Only Memory, abbreviation
ROM), random access memory (Random Access Memory, abbreviation RAM), magnetic or disk etc. are various can store
The medium of program code.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features；
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (6)
1. a kind of image detecting method characterized by comprising
Image to be detected is obtained, and extracts the geometrical characteristic of described image to be detected；
At least two groups Gradient Features of the geometrical characteristic of described image to be detected, every group of ladder are determined by edge detection operator
It is different to spend the corresponding edge detection operator of feature；
Obtain the corresponding gradient amplitude figure of every group of Gradient Features of the geometrical characteristic of described image to be detected；The gradient amplitude figure
Feature vector include all elements in the gradient amplitude figure pixel value；At least two groups Gradient Features include the first ladder
Feature and the second Gradient Features are spent, the first gradient feature corresponds to first gradient map of magnitudes, and second Gradient Features are corresponding
Second gradient amplitude figure；
According to the preferential of the feature vector of every width gradient amplitude figure of described image to be detected and corresponding edge detection operator
Grade information, determines whether described image to be detected passes through scaling processing, specifically includes: according to institute in the first gradient map of magnitudes
There is the pixel value of all elements in the pixel value and the second gradient amplitude figure of element, determines the first gradient map of magnitudes
Meet the number of the element of the first predetermined condition with the pixel value of the element in the second gradient amplitude figure in same area；Institute
Stating the first predetermined condition is the precedence information and described based on the corresponding edge detection operator of the first gradient map of magnitudes
What the precedence information of the corresponding edge detection operator of two gradient amplitude figures obtained；
Met according to the pixel value of the element in the first gradient map of magnitudes and the second gradient amplitude figure in same area
The number of the element of first predetermined condition determines whether described image to be detected passes through scaling processing.
2. the method according to claim 1, wherein described according to the first gradient map of magnitudes and described second
Described in the number that the pixel value of element in gradient amplitude figure in same area meets the element of first predetermined condition determines
Whether image to be detected specifically includes by scaling processing:
If the pixel value of the element in the first gradient map of magnitudes and the second gradient amplitude figure in same area meets institute
When stating the number of the element of the first predetermined condition greater than the first predetermined threshold, then determine described image to be detected for by scaling
The image of reason；
If the pixel value of the element in the first gradient map of magnitudes and the second gradient amplitude figure in same area meets institute
State the element of the first predetermined condition number be less than or equal to the first predetermined threshold when, then determine described image to be detected be without
Cross the image of scaling processing.
3. a kind of image detecting method characterized by comprising
Image to be detected is obtained, and extracts the geometrical characteristic of described image to be detected；
At least two groups Gradient Features of the geometrical characteristic of described image to be detected, every group of ladder are determined by edge detection operator
It is different to spend the corresponding edge detection operator of feature；
Obtain the corresponding gradient amplitude figure of every group of Gradient Features of the geometrical characteristic of described image to be detected；The gradient amplitude figure
Feature vector include all elements in the gradient amplitude figure pixel value；At least two groups Gradient Features include the first ladder
Feature and the second Gradient Features are spent, the first gradient feature corresponds to first gradient map of magnitudes, and second Gradient Features are corresponding
Second gradient amplitude figure；
According to the preferential of the feature vector of every width gradient amplitude figure of described image to be detected and corresponding edge detection operator
Grade information, determines whether described image to be detected passes through scaling processing, specifically includes: to every width gradient amplitude figure respectively along X-axis
With Y-axis carry out Projection Character transformation, obtain the first gradient map of magnitudes the first Y-axis feature vector and the first X-axis feature to
The the second Y-axis feature vector and the second X-axis feature vector of amount and the second gradient amplitude figure；
According to the pixel value of all elements in the first X-axis feature vector and the second X-axis feature vector, determine described
The pixel value of element in first X-axis feature vector and the second X-axis feature vector in same area meets the second predetermined item
The number Num1 of the element of part, and according to all elements in the first Y-axis feature vector and the second Y-axis feature vector
Pixel value determines the pixel of the element in the first Y-axis feature vector and the second Y-axis feature vector in same area
Value meets the number Num2 of the element of the second predetermined condition；Second predetermined condition is based on the first gradient map of magnitudes pair
The priority of the precedence information for the edge detection operator answered and the corresponding edge detection operator of the second gradient amplitude figure letter
What breath obtained；
Determine whether described image to be detected passes through scaling processing according to the Num1 and the Num2.
4. a kind of image detection device characterized by comprising
First obtains module, for obtaining image to be detected, and extracts the geometrical characteristic of described image to be detected；
First determining module, at least two groups ladder of the geometrical characteristic for determining described image to be detected by edge detection operator
Feature is spent, the corresponding edge detection operator of every group of Gradient Features is different；
Second obtains module, the corresponding gradient amplitude of every group of Gradient Features of the geometrical characteristic for obtaining described image to be detected
Figure；
Second determining module, for according to every width gradient amplitude figure of described image to be detected feature vector and corresponding side
The precedence information of edge detective operators, determines whether described image to be detected passes through scaling processing；
The feature vector of the gradient amplitude figure includes the pixel value of all elements in the gradient amplitude figure；At least two groups
Gradient Features include first gradient feature and the second Gradient Features, and the first gradient feature corresponds to first gradient map of magnitudes, institute
State the corresponding second gradient amplitude figure of the second Gradient Features；
Second determining module is specifically used for:
According to all elements in the pixel value of all elements in the first gradient map of magnitudes and the second gradient amplitude figure
Pixel value determines the pixel value of the element in the first gradient map of magnitudes and the second gradient amplitude figure in same area
Meet the number of the element of the first predetermined condition；First predetermined condition is based on the corresponding side of the first gradient map of magnitudes
The precedence information of the precedence information of edge detective operators and the corresponding edge detection operator of the second gradient amplitude figure obtains
's；
Met according to the pixel value of the element in the first gradient map of magnitudes and the second gradient amplitude figure in same area
The number of the element of first predetermined condition determines whether described image to be detected passes through scaling processing.
5. device according to claim 4, which is characterized in that second determining module is according to the first gradient width
The pixel value of element in degree figure and the second gradient amplitude figure in same area meets the element of first predetermined condition
Number when determining whether described image to be detected passes through scaling processing, be specifically used for:
If the pixel value of the element in the first gradient map of magnitudes and the second gradient amplitude figure in same area meets institute
When stating the number of the element of the first predetermined condition greater than the first predetermined threshold, determine described image to be detected for by scaling processing
Image；
If the pixel value of the element in the first gradient map of magnitudes and the second gradient amplitude figure in same area meets institute
State the element of the first predetermined condition number be less than or equal to the first predetermined threshold when, determine described image to be detected be without
The image of scaling processing.
6. a kind of image detection device characterized by comprising
First obtains module, for obtaining image to be detected, and extracts the geometrical characteristic of described image to be detected；
First determining module, at least two groups ladder of the geometrical characteristic for determining described image to be detected by edge detection operator
Feature is spent, the corresponding edge detection operator of every group of Gradient Features is different；
Second obtains module, the corresponding gradient amplitude of every group of Gradient Features of the geometrical characteristic for obtaining described image to be detected
Figure；
Second determining module, for according to every width gradient amplitude figure of described image to be detected feature vector and corresponding side
The precedence information of edge detective operators, determines whether described image to be detected passes through scaling processing；
The feature vector of the gradient amplitude figure includes the pixel value of all elements in the gradient amplitude figure；At least two groups
Gradient Features include first gradient feature and the second Gradient Features, and the first gradient feature corresponds to first gradient map of magnitudes, institute
State the corresponding second gradient amplitude figure of the second Gradient Features；
Second determining module is specifically used for:
Projection Character transformation is carried out along X-axis and Y-axis respectively to every width gradient amplitude figure, obtains the of the first gradient map of magnitudes
The the second Y-axis feature vector and the second X-axis of one Y-axis feature vector and the first X-axis feature vector and the second gradient amplitude figure
Feature vector；
According to the pixel value of all elements in the first X-axis feature vector and the second X-axis feature vector, determine described
The pixel value of element in first X-axis feature vector and the second X-axis feature vector in same area meets the second predetermined item
The number Num1 of the element of part, and according to all elements in the first Y-axis feature vector and the second Y-axis feature vector
Pixel value determines the pixel of the element in the first Y-axis feature vector and the second Y-axis feature vector in same area
Value meets the number Num2 of the element of the second predetermined condition；Second predetermined condition is based on the first gradient map of magnitudes pair
The priority of the precedence information for the edge detection operator answered and the corresponding edge detection operator of the second gradient amplitude figure letter
What breath obtained；
Determine whether described image to be detected passes through scaling processing according to the Num1 and the Num2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610178431.6A CN107231551B (en) | 2016-03-25 | 2016-03-25 | A kind of image detecting method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610178431.6A CN107231551B (en) | 2016-03-25 | 2016-03-25 | A kind of image detecting method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107231551A CN107231551A (en) | 2017-10-03 |
| CN107231551B true CN107231551B (en) | 2019-05-07 |
Family
ID=59932497
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610178431.6A Active CN107231551B (en) | 2016-03-25 | 2016-03-25 | A kind of image detecting method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107231551B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111260605B (en) * | 2019-04-11 | 2020-11-10 | 江苏盛凡信息服务有限公司 | Intelligent action execution method |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4860388B2 (en) * | 2005-09-27 | 2012-01-25 | 株式会社リコー | Image processing apparatus, image processing method, and computer-readable recording medium storing program for executing the method |
| CN101916442A (en) * | 2010-08-05 | 2010-12-15 | 大连理工大学 | Method for robustly positioning tampered region by utilizing GLCM characteristic |
| WO2015077946A1 (en) * | 2013-11-28 | 2015-06-04 | 中国科学院自动化研究所 | Method for positioning image tampering region based on dct coefficient |
| CN104182973A (en) * | 2014-08-11 | 2014-12-03 | 福州大学 | Image copying and pasting detection method based on circular description operator CSIFT (Colored scale invariant feature transform) |
| CN104700360B (en) * | 2015-04-01 | 2018-06-05 | 北京思朗科技有限责任公司 | Image-scaling method and system based on edge self-adaption |
-
2016
- 2016-03-25 CN CN201610178431.6A patent/CN107231551B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN107231551A (en) | 2017-10-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | Robust image and video dehazing with visual artifact suppression via gradient residual minimization | |
| Zhou et al. | Multi-scale weighted gradient-based fusion for multi-focus images | |
| US7623683B2 (en) | Combining multiple exposure images to increase dynamic range | |
| Lee et al. | Simultaneous localization, mapping and deblurring | |
| CN108093221B (en) | Suture line-based real-time video splicing method | |
| Aires et al. | Optical flow using color information: preliminary results | |
| US20060165311A1 (en) | Spatial standard observer | |
| CN106934806B (en) | It is a kind of based on text structure without with reference to figure fuzzy region dividing method out of focus | |
| CN112308797B (en) | Corner detection method and device, electronic equipment and readable storage medium | |
| CN110268712A (en) | Method and apparatus for handling image attributes figure | |
| US10475229B2 (en) | Information processing apparatus and information processing method | |
| JP2010266982A (en) | Image processing apparatus and method, as well as program | |
| CN110599455A (en) | Display screen defect detection network model, method and device, electronic equipment and storage medium | |
| US9299132B2 (en) | Automatically determining the size of a blur kernel | |
| WO2015198368A1 (en) | Image processing device and image processing method | |
| US8126275B2 (en) | Interest point detection | |
| CN107231551B (en) | A kind of image detecting method and device | |
| JP5928465B2 (en) | Degradation restoration system, degradation restoration method and program | |
| Zhai et al. | Progressive image restoration through hybrid graph Laplacian regularization | |
| Jiang et al. | Sensor-assisted image deblurring of consumer photos on smartphones | |
| JP2004133592A (en) | Image processor, image processing method and image processing program for magnifying image | |
| US20150131922A1 (en) | Method for removing noise of image signal and image processing device using the same | |
| Priyadharshini et al. | Research and analysis on segmentation and thresholding techniques | |
| ALAA et al. | Introcduction to image processing with Python | |
| Ou et al. | Color edge detection based on data fusion technology in presence of Gaussian noise |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 266100 Zhuzhou Road, Laoshan District, Shandong, No. 151, No. Patentee after: Hisense Video Technology Co.,Ltd. Address before: 266100 Zhuzhou Road, Laoshan District, Shandong, No. 151, No. Patentee before: HISENSE ELECTRIC Co.,Ltd. |