CN106097306A - Obtain the method for Image Edge-Detection operator, method for detecting image edge and device - Google Patents

Abstract

The invention discloses a kind of method of Image Edge-Detection operator, method for detecting image edge and device of obtaining, wherein, the method obtaining Image Edge-Detection operator includes: in the target image, mark off four neighborhoods including first object pixel；Described target image carries out bilinear interpolation process, and an optional pixel inserted in described four neighborhoods, as the second target pixel points, sets up described first object pixel and described second target pixel points position relationship in described four neighborhoods；According to described position relationship and bilinear interpolation formula, determine Image Edge-Detection functional relation；Image Edge-Detection operator is determined according to described Image Edge-Detection functional relation.The Image Edge-Detection operator using the method to obtain, is applied in the Image Edge-Detection of digital picture, it is possible to while reducing operand, improves noise resisting ability, and the suitability is more preferable.

Description

Obtain the method for Image Edge-Detection operator, method for detecting image edge and device

Technical field

The present invention relates to image edge processing technical field, particularly relate to a kind of side obtaining Image Edge-Detection operator Method, method for detecting image edge and device.

Background technology

The edge of image belongs to the radio-frequency component of image, is one of the basic feature of image.The target image that one width is concrete, After Image Edge-Detection processes, becoming a width edge-detected image, edge-detected image is identical with the size of target image, In edge-detected image, the marginal position of corresponding target image is usually highlighted, and remaining position is essentially black, thus Target image can effectively be distinguished by the edge using image with other object (such as other image, background etc.).Thus Understanding, find the edge of image, the many applications for image edge processing are extremely important, such as, and image segmentation, image The applications such as sharpening, graphical analysis and identification, when concrete operations, are required to first determine the marginal position of image.Existing skill In art, the method for Image Edge-Detection is generally used to determine the marginal position of image.

In digital picture, the radio-frequency component of image includes the part that in image, gray-value variation is violent, generally by gray scale Value changes violent position and is defined as the edge of digital picture.For digital picture, generally use various image borders Detective operators (such as differential operator, Sobel operator, Canny operator etc.) carries out Image Edge-Detection to it.Again due in numeral In image, each pixel being positioned on the edge of image, the comparison in difference of gray value is little, and change is mild, and is positioned at the limit of image The pixel at the edge of close image in pixel on edge, with image, both grey value differences are bigger, and change is violent, Based on this, the first derivative of the gray value of image and second dervative is generally used to calculate as the Image Edge-Detection of digital picture Son, namely differential operator, including first order differential operator and Second Order Differential Operator.

When using first order differential operator that digital picture is carried out Image Edge-Detection, noise is relatively big, and testing result is easily made an uproar Sound shadow is rung.When using Second Order Differential Operator that digital picture is carried out Image Edge-Detection, noise is less on the impact of testing result, But during detection, needing bigger data window, operand is bigger.Therefore, for the Image Edge-Detection of digital picture, one The suitability of rank differential operator and Second Order Differential Operator is the most poor, it is impossible to meets antinoise simultaneously and reduces the demand of operand.

So, in the technique of image edge detection of digital picture, the suitability of differential operator is poor, it is impossible to solve simultaneously Macrooperation amount and the problem of big noise.

Summary of the invention

The purpose of the embodiment of the present invention is that providing a kind of obtains the method for Image Edge-Detection operator, Image Edge-Detection Method and device, to solve in the technique of image edge detection of digital picture, the problem that the suitability of differential operator is poor.Pass through A kind of method of the present invention, it is possible to obtain Image Edge-Detection operator being relatively suitable for, is applied to this Image Edge-Detection operator Time in the technique of image edge detection of digital picture, there is preferable noise robustness, meanwhile, data window can be reduced, i.e. Reducing operand, the suitability is more preferable.

In order to solve above-mentioned technical problem, the embodiment of the invention discloses following technical scheme:

First aspect, embodiments provides a kind of method obtaining Image Edge-Detection operator, and the method includes:

In the target image, four neighborhoods including first object pixel are marked off；Described first object pixel For the pixel that line number index in described four neighborhoods and columns index are the most minimum；

Described target image is carried out bilinear interpolation process, an optional pixel conduct inserted in described four neighborhoods Second target pixel points, sets up described first object pixel and described second target pixel points position in described four neighborhoods Relation；

According to described position relationship and bilinear interpolation formula, determine Image Edge-Detection functional relation, described figure As edge indicator function relational expression is for calculating the gray value of described first object pixel and described second target pixel points The difference of gray value；

Image Edge-Detection operator is determined according to described Image Edge-Detection functional relation.

In conjunction with first aspect, in the first possible embodiment of first aspect, described described target image is entered Row bilinear interpolation processes, and an optional pixel inserted in described four neighborhoods is as the process of the second target pixel points, tool Body includes:

With default interpolation multiplying power, described target image is carried out bilinear interpolation process, it is thus achieved that inserting of described target image Value image；

In described interpolation image, select to insert the second target pixel points in described four neighborhoods, described second target picture Vegetarian refreshments is positioned at the position that in described four neighborhoods, row pixel second from the bottom and row pixel second from the bottom intersect.

In conjunction with the first possible embodiment of first aspect, the embodiment that the second in first aspect is possible In, described according to described position relationship and bilinear interpolation formula, determine the process of Image Edge-Detection functional relation, tool Body includes:

According to described position relationship, and bilinear interpolation formula, determine following about described first object pixel Gray value and the gray value of described second target pixel points between the first functional relation；

D (u, v)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+ 1)；

When determining each pixel in described target image as first object pixel, right in described interpolation image The second target pixel points answered；

Extract all described second target pixel points in described interpolation image, according to the first mesh described in described target image All described second target pixel points are rearranged by the queueing discipline of mark pixel, form reference picture；

According to described second target pixel points position coordinates in described reference picture and described first functional relationship Formula, determines between the gray value of the following gray value about described first object pixel and described second target pixel points Second functional relation；

D (i, j)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+ 1)；

According to described second functional relation, determine following Image Edge-Detection functional relation；

$\begin{array}{c}G(i,j)-D(i,j)=(v+u-uv)*G(i,j)+u(v-1)*G(i,j+1)+(u-1)v*G(i+1,j)-uv*\\ G(i+1,j+1)\end{array};$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents with reference to figure D The gray value of the pixel of the i-th row jth row in Xiang；G (i, j+1) represents the ash of the pixel of the i-th row jth+1 row in target image Angle value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target figure The gray value of the pixel of i+1 row jth+1 row in Xiang；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to the first corresponding mesh The fore-and-aft distance of mark pixel.

In conjunction with the embodiment that the second of first aspect is possible, at the third possible embodiment of first aspect In, the described process determining Image Edge-Detection operator according to described Image Edge-Detection functional relation, specifically include:

Described Image Edge-Detection functional relation is transformed to the Image Edge-Detection functional relationship of following convolution form Formula；

$G(i,j)-D(i,j)=\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right]\left[\begin{array}{cc}G(i,j)& G(i,j+1)\\ G(i+1,j)& G(i+1,j+1)\end{array}\right];$

Image Edge-Detection functional relation according to described convolution form determines following Image Edge-Detection operator；

$\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right];$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents with reference to figure D The gray value of the pixel of the i-th row jth row in Xiang；G (i, j+1) represents the ash of the pixel of the i-th row jth+1 row in target image Angle value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target figure The gray value of the pixel of i+1 row jth+1 row in Xiang；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to the first corresponding mesh The fore-and-aft distance of mark pixel.

Second aspect, embodiments provides the above-mentioned Image Edge-Detection operator of a kind of employing and enters image to be detected The method of row Image Edge-Detection, the method includes:

According to for the default interpolation multiplying power of image to be detected and described Image Edge-Detection operator, determine for described The concrete Image Edge-Detection operator of image to be detected；

One-row pixels point is increased, at its last string picture in the lower section of last column pixel of the image to be detected obtained The right side of vegetarian refreshments increases string pixel, obtains actually detected image；The one-row pixels point increased is by replicating image to be detected Row pixel second from the bottom obtain, the string pixel of increase by replicate image to be detected row pixel second from the bottom obtain ?；

Use described concrete Image Edge-Detection operator that described actually detected image is carried out Image Edge-Detection, thus obtain Obtain the gray value information of the edge-detected image of described image to be detected.

In conjunction with second aspect, in the first possible embodiment of second aspect, described basis is for mapping to be checked The default interpolation multiplying power of picture and described Image Edge-Detection operator, determine the concrete image border for described image to be detected The process of detective operators, specifically includes:

Obtain default interpolation multiplying power M for image to be detected；

According to described default interpolation multiplying power M, determine the value of u and v according to the following equation；

$u=v=\frac{M}{1+M};$

Value according to described u and v and described Image Edge-Detection operator, determine that described concrete Image Edge-Detection is calculated Son.

In conjunction with the first possible embodiment of second aspect, the embodiment that the second in second aspect is possible In, the described concrete Image Edge-Detection operator of described employing carries out Image Edge-Detection to described actually detected image, thus obtains Obtain the process of the gray value information of the edge-detected image of described image to be detected, specifically include:

Obtain the gray value data of described image to be detected, and determine institute according to the gray value data of described image to be detected State the gray value data of actually detected image；

According to the following equation, the gray value number of described concrete Image Edge-Detection operator and described actually detected image is used According to carrying out convolution, thus obtain the gray value information of the edge-detected image of described image to be detected；

Wherein, D represents the gray value of pixel in the edge-detected image of image to be detected, and I ' represents described actually detected The gray value data of 2 × 2 data windows in image, A represents described concrete Image Edge-Detection operator, | | represent absolute value fortune Calculate,Represent and round downwards.

The third aspect, The embodiment provides a kind of device obtaining Image Edge-Detection operator, this device bag Include:

Region divides module, in the target image, marks off four neighborhoods including first object pixel；Institute Stating first object pixel is line number index and the pixel of columns index all minimums in described four neighborhoods；

Interpolation processing module, for described target image carries out bilinear interpolation process, optional one is inserted described four Pixel in neighborhood, as the second target pixel points, is set up described first object pixel and is existed with described second target pixel points Position relationship in described four neighborhoods；

First determines module, for according to described position relationship and bilinear interpolation formula, determines Image Edge-Detection Functional relation, described Image Edge-Detection functional relation is for calculating the gray value of described first object pixel with described The difference of the gray value of the second target pixel points；

Second determines module, for determining Image Edge-Detection operator according to described Image Edge-Detection functional relation.

In conjunction with the third aspect, in the first possible embodiment of the third aspect, described interpolation processing module includes:

Interpolation process unit, for default interpolation multiplying power, carries out bilinear interpolation process to described target image, it is thus achieved that The interpolation image of described target image；

Pixel chooses unit, in described interpolation image, selects to insert the second target picture in described four neighborhoods Vegetarian refreshments, described second target pixel points is positioned at row pixel second from the bottom and row pixel second from the bottom in described four neighborhoods and intersects Position.

In conjunction with the first possible embodiment of the third aspect, the embodiment that the second in the third aspect is possible In, described first determine module specifically for:

According to described position relationship, and bilinear interpolation formula, determine following about described first object pixel Gray value and the gray value of described second target pixel points between the first functional relation；

D (u, v)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+ 1)；

When determining each pixel in described target image as first object pixel, right in described interpolation image The second target pixel points answered；

Extract all described second target pixel points in described interpolation image, according to the first mesh described in described target image All described second target pixel points are rearranged by the queueing discipline of mark pixel, form reference picture；

According to described second target pixel points position coordinates in described reference picture and described first functional relationship Formula, determines between the gray value of the following gray value about described first object pixel and described second target pixel points Second functional relation；

D (i, j)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+ 1)；

According to described second functional relation, determine following Image Edge-Detection functional relation；

$\begin{array}{c}G(i,j)-D(i,j)=(v+u-uv)*G(i,j)+u(v-1)*G(i,j+1)+(u-1)v*G(i+1,j)-uv*\\ G(i+1,j+1)\end{array};$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents with reference to figure D The gray value of the pixel of the i-th row jth row in Xiang；G (i, j+1) represents the ash of the pixel of the i-th row jth+1 row in target image Angle value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target figure The gray value of the pixel of i+1 row jth+1 row in Xiang；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to the first corresponding mesh The fore-and-aft distance of mark pixel.

In conjunction with the embodiment that the second of the third aspect is possible, at the third possible embodiment of the third aspect In, described second determine module specifically for:

Described Image Edge-Detection functional relation is transformed to the Image Edge-Detection functional relationship of following convolution form Formula；

$G(i,j)-D(i,j)=\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right]\left[\begin{array}{cc}G(i,j)& G(i,j+1)\\ G(i+1,j)& G(i+1,j+1)\end{array}\right];$

Image Edge-Detection functional relation according to described convolution form determines following Image Edge-Detection operator；

$\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right];$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents with reference to figure D The gray value of the pixel of the i-th row jth row in Xiang；G (i, j+1) represents the ash of the pixel of the i-th row jth+1 row in target image Angle value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target figure The gray value of the pixel of i+1 row jth+1 row in Xiang；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to the first corresponding mesh The fore-and-aft distance of mark pixel.

Fourth aspect, embodiments provides the above-mentioned Image Edge-Detection operator of a kind of employing and enters image to be detected The device of row Image Edge-Detection, this device includes:

3rd determines module, for according to for the default interpolation multiplying power of image to be detected and described Image Edge-Detection Operator, determines the concrete Image Edge-Detection operator for described image to be detected；

Image processing module, increases one-row pixels for the lower section of last column pixel at the image to be detected obtained Point, increases string pixel, obtains actually detected image on the right side of its last string pixel；The one-row pixels point increased leads to Crossing the row pixel second from the bottom acquisition replicating image to be detected, the string pixel of increase is by replicating falling of image to be detected Number secondary series pixel obtains；

Image Edge-Detection module, is used for using described concrete Image Edge-Detection operator to enter described actually detected image Row Image Edge-Detection, thus obtain the gray value information of the edge-detected image of described image to be detected.

In conjunction with fourth aspect, in the first possible embodiment of fourth aspect, the described 3rd determines that module is concrete For:

Obtain default interpolation multiplying power M for image to be detected；

According to described default interpolation multiplying power M, determine the value of u and v according to the following equation；

$u=v=\frac{M}{1+M};$

Value according to described u and v and described Image Edge-Detection operator, determine that described concrete Image Edge-Detection is calculated Son.

In conjunction with the first possible embodiment of fourth aspect, the embodiment that the second in fourth aspect is possible In, described Image Edge-Detection module includes:

Gray value data acquiring unit, for obtaining the gray value data of described image to be detected, and according to described to be checked The gray value data of altimetric image determines the gray value data of described actually detected image；

Image Edge-Detection unit, for according to the following equation, uses described concrete Image Edge-Detection operator with described The gray value data of actually detected image carries out convolution, thus obtains the gray value of the edge-detected image of described image to be detected Information；

Wherein, D represents the gray value of pixel in the edge-detected image of image to be detected, and I ' represents described actually detected The gray value data of 2 × 2 data windows in image, A represents described concrete Image Edge-Detection operator, | | represent absolute value fortune Calculate,Represent and round downwards.

The technical scheme that embodiments of the invention provide can include following beneficial effect: by side disclosed by the invention Method, it is thus achieved that Image Edge-Detection operator, when being applied in the Image Edge-Detection of digital picture, owing to having merged window data Weighted average (the bilinear interpolation processing procedure of image) and object pixel minor shifts (the down-sampled processing procedure of image) Effect, be greatly reduced noise, the image edge information obtained is abundanter, and the profile of image becomes apparent from.Meanwhile, inspection Survey the data window needed smaller, that is operand is smaller, so, Image Edge-Detection operator disclosed by the invention exists In the technique of image edge detection of digital picture, the suitability is more preferable.

The embodiment of the present invention is it should be appreciated that it is only exemplary reconciliation that above general description and details hereinafter describe The property released, the disclosure can not be limited.

Accompanying drawing explanation

Accompanying drawing herein is merged in description and constitutes the part of this specification, it is shown that meet the enforcement of the present invention Example, and for explaining the principle of the present invention together with description.

In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing In having technology to describe, the required accompanying drawing used is briefly described, it should be apparent that, for those of ordinary skill in the art Speech, on the premise of not paying creative work, it is also possible to obtain other accompanying drawing according to these accompanying drawings.

The gray value data of a kind of target image that Fig. 1 provides for the embodiment of the present invention；

Fig. 2 is the gray value data of the interpolating pixel point of each pixel of target image in Fig. 1；

Fig. 3 is the gray value information of the edge-detected image of target image in Fig. 1；

The schematic flow sheet of a kind of method obtaining Image Edge-Detection operator that Fig. 4 provides for the embodiment of the present invention；

The schematic flow sheet of the another kind of method obtaining Image Edge-Detection operator that Fig. 5 provides for the embodiment of the present invention；

Fig. 6 uses above-mentioned Image Edge-Detection operator that image to be detected is carried out figure for the one that the embodiment of the present invention provides Schematic flow sheet as the method for rim detection；

Fig. 7 uses above-mentioned Image Edge-Detection operator to carry out image to be detected for the another kind that the embodiment of the present invention provides The schematic flow sheet of the method for Image Edge-Detection；

The structured flowchart of a kind of device obtaining Image Edge-Detection operator that Fig. 8 provides for the embodiment of the present invention；

The structured flowchart of the another kind of device obtaining Image Edge-Detection operator that Fig. 9 provides for the embodiment of the present invention；

Figure 10 uses above-mentioned Image Edge-Detection operator to carry out image to be detected for the one that the embodiment of the present invention provides The structured flowchart of the device of Image Edge-Detection；

Figure 11 uses above-mentioned Image Edge-Detection operator to enter image to be detected for the another kind that the embodiment of the present invention provides The structured flowchart of the device of row Image Edge-Detection.

Detailed description of the invention

For the technical scheme making those skilled in the art be more fully understood that in the present invention, real below in conjunction with the present invention Execute the accompanying drawing in example, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described enforcement Example is only a part of embodiment of the present invention rather than whole embodiments.Based on the embodiment in the present invention, this area is common The every other embodiment that technical staff is obtained under not making creative work premise, all should belong to present invention protection Scope.

Before specifically introducing embodiments of the invention, first introduce some skills relevant to technical scheme disclosed by the invention Art principle: a width target image is processed by bilinear interpolation after amplifying, obtain the interpolation image of this target image, lead to afterwards Cross down-sampled process, then the size by the size reduction of described interpolation image to target image, obtain the reference of this target image Image.The information (including gray value information) of reference picture is essentially identical with the information of target image, only exists one between the two The other minor shifts of individual sub-pixel, the gray value of each pixel (line number rope corresponding with position in reference picture in target image Draw identical with columns index) the difference of gray value of pixel, after taking absolute value, it is possible to obtain the rim detection of target image The gray value information of image, i.e. obtains the Image Edge-Detection result of the target image characterized with gray value information, uses these Gray value information can draw out edge-detected image.Herein, by target image after Image Edge-Detection processes, according to It is edge-detected image, in edge-detected image, Ren Menke that the gray value information that detection obtains draws the image definition obtained To be clearly seen the profile of target image, that is, in edge-detected image, the marginal position of corresponding target image is highlighted aobvious Showing, other position of corresponding target image is black, and therefore, people, in edge-detected image, can only see target image Profile, this profile is the edge of target image.

From above-mentioned know-why, by target image is carried out bilinear interpolation process, for each in target image After individual pixel determines an interpolating pixel point, calculate the gray value of each pixel in target image and insert with corresponding The difference of gray value of value pixel, and after difference is taken absolute value, it is possible to obtain the edge-detected image of target image Gray value information, such as, shown in Fig. 1 is the gray value data of a certain width target image, illustrated therein is in target image The gray value of each pixel；Fig. 2 is illustrated that the gray scale of the interpolating pixel point that each pixel of target image is corresponding Value, wherein, in target image, certain pixel position in FIG and the interpolating pixel corresponding with this pixel point are in fig. 2 Position corresponding (line number index and columns index identical)；In Fig. 1, the gray value of each pixel is corresponding with position in Fig. 2 respectively After the gray value of each pixel subtracts each other, difference takes absolute value, and just obtains the gray value information of the edge-detected image of target image, Fig. 3 is illustrated that the gray value information of the edge-detected image of target image, according to the gray value information shown in Fig. 3, it is possible to paint Go out the edge-detected image of target image, edge-detected image is highlighted the marginal position that position is target image.

When specifically calculating, can in the target image, with each pixel as target pixel points, build one and include this Target pixel points interior four neighborhoods (in digital picture, the region of four pixel adjacent one another are compositions, be defined as one Four neighborhoods), after target image is carried out bilinear interpolation process, using certain pixel of inserting in four neighborhoods as this target picture The interpolating pixel point of vegetarian refreshments, carries out identical process to each pixel in target image, i.e. can get in target image The interpolating pixel point that each pixel is corresponding, afterwards to the gray value of pixel each in target image respectively with corresponding interpolation The gray value of pixel carries out difference operation, and takes absolute value operation result, it becomes possible to obtain the rim detection of target image The gray value information of image.

Based on this, the invention provides a kind of method obtaining Image Edge-Detection operator, obtained by the method for the present invention The Image Edge-Detection operator arrived, as long as data window 2 × 2 data window of detection, operand is the least.Further, This Image Edge-Detection operator is owing to having merged window data weighted average (the bilinear interpolation processing procedure of image) and mesh The effect of mark pixel minor shifts (the down-sampled processing procedure of image), greatly improves noise robustness, and the suitability is more preferable.

Below in conjunction with the accompanying drawings, the specific embodiment of the present invention is discussed in detail.

As shown in Figure 4, Fig. 4 is illustrated that the flow process of a kind of method obtaining Image Edge-Detection operator disclosed by the invention Figure, the method includes:

Step 11, in the target image, marks off four neighborhoods including first object pixel；Described first object Pixel is line number index and the pixel of columns index all minimums in described four neighborhoods.

In the present invention, arbitrarily choose a width digital picture as target image, after obtaining target image, at target image In, arbitrarily choose a pixel as first object pixel, and in this target image, mark off four neighborhoods, neighbours Territory includes first object pixel, and, in this four neighborhood, first object pixel is positioned at the upper left corner of four neighborhoods, its Line number index and columns index are the most minimum in four neighborhoods, herein, use line number index and columns to index, represent that pixel exists Line number residing in digital picture and columns.

Step 12, described target image is carried out bilinear interpolation process, an optional picture inserted in described four neighborhoods Vegetarian refreshments, as the second target pixel points, sets up described first object pixel with described second target pixel points at described four neighborhoods In position relationship.

With default interpolation multiplying power, after target image is carried out bilinear interpolation process, obtain the interpolation image of target image, In interpolation image, arbitrarily choose a pixel inserted in four neighborhoods as the second target pixel points, it is preferred that for it After the effect of Image Edge-Detection more preferable, select to be positioned in four neighborhoods, row pixel second from the bottom and row pixel second from the bottom The pixel of some intersection location is as the second target pixel points.Wherein, interpolation multiplying power refers to and image is carried out interpolation processing Time, the multiplying power that image is zoomed in and out, herein, the multiplying power zooming in and out image is defined as interpolation multiplying power, presets interpolation The value of multiplying power can arbitrarily set.

Step 13, according to described position relationship and bilinear interpolation formula, determine Image Edge-Detection functional relation, Described Image Edge-Detection functional relation is for calculating the gray value of described first object pixel and described second target picture The difference of the gray value of vegetarian refreshments.

After obtaining target image, it is possible to obtain the gray value data of target image, herein, the gray value data of image Including the gray value that each pixel in image is corresponding.

In interpolation image, determine first object pixel and second target pixel points position relationship in four neighborhoods it After, according to bilinear interpolation formula, it is possible to other pixel in the gray value of employing first object pixel and four neighborhoods Gray value calculates the gray value of the second target pixel points, that is is capable of determining that the gray value and second of first object pixel Gray value function relational expression between the gray value of target pixel points, after deforming this gray value function relational expression, it is possible to Obtain difference functions relational expression, be used for calculating the difference of the gray value of the gray value of first object pixel and the second target pixel points Value.

Using each pixel in target image as first object pixel, determine each first object picture The second target pixel points that vegetarian refreshments is corresponding, uses above-mentioned difference functions relational expression, calculates every a pair first object pixel After the difference of the gray value of gray value and the second target pixel points, difference is carried out signed magnitude arithmetic(al) and can be obtained by target image The gray value information of edge-detected image, thus, difference functions relational expression is defined as Image Edge-Detection letter by the present invention Number relational expression, that is use this Image Edge-Detection functional relation gray value and second to every a pair first object pixel After the gray value of target pixel points carries out computing, operation result is taken absolute value and is obtained with the edge detection graph of target image The gray value information of picture, i.e. obtains the Image Edge-Detection result of the target image characterized with gray value information, according to these ashes Angle value information, it is possible to draw out the edge-detected image of target image, the position that is highlighted in edge-detected image is target The marginal position of image.

Step 14, determine Image Edge-Detection operator according to described Image Edge-Detection functional relation.

Using any one width digital picture as target image, step 11 can be used to the method for step 13, it is thus achieved that should The gray value information of the edge-detected image of target image, but this process operand is relatively big, for simplified operation step, reduces Operand, can be deformed into convolution form by Image Edge-Detection functional relation, afterwards can be from the image limit of convolution form Extracting general Image Edge-Detection operator in edge detection function relational expression, this Image Edge-Detection operator can apply to appoint In the Image Edge-Detection of what digital picture.

A kind of method provided by the present embodiment, it is possible to obtain general Image Edge-Detection operator, by this image limit When edge detective operators is applied in the technique of image edge detection of image to be detected, it is only necessary to the data window of 2 × 2, can be right Image to be detected carries out Image Edge-Detection, and operand is the least, simultaneously because use this Image Edge-Detection operator to be checked Altimetric image carries out the process of Image Edge-Detection, has merged window data weighted average (the bilinear interpolation processing procedure of image) And the effect (the down-sampled processing procedure of image) of object pixel minor shifts so that testing result is affected by noise less, After detection, in the edge-detected image that the gray value information of edge-detected image obtained according to detection is drawn, image to be detected Edge become apparent from.

As it is shown in figure 5, Fig. 5 is illustrated that the stream of the another kind of method obtaining Image Edge-Detection operator disclosed by the invention Cheng Tu, the method includes:

Step 21, in the target image, marks off four neighborhoods including first object pixel；Described first object Pixel is line number index and the pixel of columns index all minimums in described four neighborhoods.

When being embodied as, the gray value of described first object pixel is designated as G (i, j), by described four neighborhoods other The gray value of pixel be designated as respectively G (i, j+1), G (i+1, j), G (i+1, j+1).

Step 22, described target image is carried out bilinear interpolation process, an optional picture inserted in described four neighborhoods Vegetarian refreshments, as the second target pixel points, sets up described first object pixel with described second target pixel points at described four neighborhoods In position relationship.

Step 23, according to described position relationship, and bilinear interpolation formula, determine about described first object pixel The first gray value function relational expression between gray value and the gray value of described second target pixel points of point.

In the present embodiment, the first gray value function relational expression is as follows:

D (u, v)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+ 1)。

After target image is carried out bilinear interpolation process, obtain the interpolation image of target image, the first gray value function Relational expression shows, in interpolation image, specifies first object pixel and the second target pixel points position in four neighborhoods and closes After system, it is possible to utilize other pixel in bilinear interpolation formula and the gray value of first object pixel and four neighborhoods Gray value, calculates the gray value of the second target pixel points, that is exists according to first object pixel and the second target pixel points Position relationship in interpolation image, it is thus achieved that the gray value of first object pixel and the first of the gray value of the second target pixel points Gray value function relational expression.

When step 24, each pixel determined in described target image are as first object pixel, described interpolation The second target pixel points corresponding in image.

For each pixel in target image, during as first object pixel, in corresponding interpolation graphs In Xiang, all there is second target pixel points corresponding.In interpolation image, determine each first object pixel pair The second target pixel points answered.

Step 25, extract all described second target pixel points in described interpolation image, according to institute in described target image State the queueing discipline of first object pixel, all described second target pixel points are rearranged, form reference picture.

All second target pixel points in interpolation image are individually proposed, according to corresponding first object pixel at mesh Queueing discipline in logo image, rearranges the second target pixel points proposed, and forms the image that a width is new, herein, incites somebody to action This new image definition is reference picture.

Step 26, according to described second target pixel points position coordinates in described reference picture and described first ash Angle value functional relation, determines the gray scale of the gray value about described first object pixel and described second target pixel points The second gray value function relational expression between value.

In the present embodiment, the second gray value function relational expression is as follows:

D (i, j)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+ 1)。

For a pair first object pixel the most corresponding and the second target pixel points, the second target pixel points is in reference Position coordinates in image, corresponding with first object pixel position coordinates in the target image, both are at respective image In line number index and columns index the most identical.Based on this, in conjunction with first object pixel and the second target pixel points in interpolation Position corresponding relation in image, the first gray value function relational expression can become the second gray value function relational expression.

Step 27, according to described second gray value function relational expression, determine Image Edge-Detection functional relation.

In the present embodiment, Image Edge-Detection functional relation is as follows:

$\begin{array}{c}G(i,j)-D(i,j)=(v+u-uv)*G(i,j)+u(v-1)*G(i,j+1)+(u-1)v*G(i+1,j)-uv*\\ G(i+1,j+1)\end{array}.$

For target image and reference picture, the number of pixel and the number of pixel in target image in reference picture Identical, and with target image in reference picture, the queueing discipline of pixel is the most identical, therefore, the size of reference picture and mesh The size of logo image is the most identical, and reference picture is essentially identical with the information of target image, only exists a sub-pix between the two The skew of rank.Therefore, by the gray value of the first object pixel in corresponding target image with in reference picture The gray value of the second target pixel points carries out difference operation, and after the result of difference operation is carried out signed magnitude arithmetic(al) again, i.e. The gray value information of the edge-detected image of target image can be obtained.

Based on this, after the second gray value function relational expression deformation, it is thus achieved that difference functions relational expression, for calculating the first mesh The gray value of mark pixel and the difference of the gray value of the second target pixel points.Herein, this difference functions relational expression is defined For Image Edge-Detection functional relation, by this Image Edge-Detection functional relation, to each picture in target image After vegetarian refreshments all carries out corresponding computing, it is possible to obtain the gray value information of the edge-detected image of target image, i.e. obtain with ash The Image Edge-Detection result of the target image of angle value information representation, according to these gray value informations, it is possible to draw out target figure The edge-detected image of picture, is highlighted, in edge-detected image, the marginal position that position is target image.

Step 28, described Image Edge-Detection functional relation is transformed to the Image Edge-Detection functional relationships of convolution form It it is formula.

In the present embodiment, the Image Edge-Detection functional relation of convolution form is as follows:

$G(i,j)-D(i,j)=\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right]\left[\begin{array}{cc}G(i,j)& G(i,j+1)\\ G(i+1,j)& G(i+1,j+1)\end{array}\right].$

Step 29, Image Edge-Detection functional relation according to described convolution form determine that Image Edge-Detection is calculated Son.

In the present embodiment, Image Edge-Detection operator is as follows:

$\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right].$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents with reference to figure D The gray value of the pixel of the i-th row jth row in Xiang；G (i, j+1) represents the ash of the pixel of the i-th row jth+1 row in target image Angle value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target figure The gray value of the pixel of i+1 row jth+1 row in Xiang；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to the first corresponding mesh The fore-and-aft distance of mark pixel.

When specifically a certain digital picture to be detected being carried out Image Edge-Detection, to be detected for this according to obtain Default interpolation multiplying power M of digital picture, according to formulaThe span of N is 1～the value of M, M and N can be with Meaning is arranged, after calculating the value of u and v, it is possible to obtains corresponding concrete Image Edge-Detection operator, uses concrete image limit Edge detective operators can carry out Image Edge-Detection to digital picture to be detected, it is not necessary to 21 to the step of computing above-mentioned steps again 28, The gray value information of the edge-detected image of described digital picture to be detected can be obtained, can be by this Image Edge-Detection operator It is applied in the Image Edge-Detection of arbitrary digital picture.

Make u=0 in Image Edge-Detection operator, landscape images edge detection operator can be obtainedFor logarithm The transverse edge of word image detects；In like manner, make v=0 in Image Edge-Detection operator, the inspection of longitudinal image border can be obtained Measuring and calculatingFor the longitudinal edge of digital picture is detected.

The method that the present embodiment provides, processes by target image carries out bilinear interpolation and after down-sampled process, adopts Difference is done, it is thus achieved that the rim detection of target image with the gray value of the gray value of target image with the image after Duplex treatment The gray value information of image, and then obtain general Image Edge-Detection operator, use this Image Edge-Detection operator to be checked When altimetric image carries out Image Edge-Detection, owing to having merged window data weighted average (the bilinear interpolation processing procedure of image) And the effect of object pixel minor shifts (the down-sampled processing procedure of image) so that testing result is affected by noise less, In the edge-detected image that after detection, the gray value information of the edge-detected image that basis obtains is drawn, the profile of image to be detected Become apparent from.Further, this Image Edge-Detection operator landscape images edge detection operator can be obtained, it is possible to figure to be detected As carrying out transverse edge detection, in like manner, this Image Edge-Detection operator longitudinal Image Edge-Detection operator, energy can also be obtained Image to be detected enough carries out longitudinal edge detection, and application is more flexible, and the suitability is more preferable.

As shown in Figure 6, Fig. 6 is illustrated that the above-mentioned Image Edge-Detection operator of a kind of employing disclosed by the invention is to be detected Image carries out the flow chart of the method for Image Edge-Detection, and the method includes:

Step 31, according to for the default interpolation multiplying power of image to be detected and described Image Edge-Detection operator, determine Concrete Image Edge-Detection operator for described image to be detected.

After determining image to be detected, obtain default interpolation multiplying power M that image to be detected is relevant, according to default interpolation times Rate M calculates the value of u and v in Image Edge-Detection operator, wherein,The span of N is 1～M, M and N Value can arbitrarily be arranged, and then determines concrete Image Edge-Detection operator, in order to for the image limit of image to be detected afterwards In edge detection.

Step 32, increase one-row pixels point in the lower section of last column pixel of the image to be detected obtained, at it The right side of rear string pixel increases string pixel, obtains actually detected image；The one-row pixels point increased is treated by duplication The pixel of the row second from the bottom of detection image obtains, and the string pixel of increase is by replicating the second from the bottom of image to be detected The pixel of row obtains.

Concrete Image Edge-Detection operator is the operator of 2 × 2, when using this operator to carry out Image Edge-Detection, needs Want the data window of 2 × 2, for last column pixel and the gray value letter of last string pixel of image to be detected Breath, it is impossible to use this operator to detect, accordingly, it would be desirable to increase a line in the lower section of last column pixel of image to be detected Pixel, so as to use this operator that the gray value information of last column pixel of image to be detected is detected, with Reason, needs the right side of the last string pixel at image to be detected to increase string pixel, so as to use this operator pair The gray value information of the last string pixel of image to be detected detects.

In order to ensure the string pixel of one-row pixels point and the increase increased, the gray value of image to be detected will not be believed Breath causes too much influence, and the one-row pixels point increased in image to be detected and the string pixel of increase are in the following manner Obtain: the one-row pixels point of increase obtains by replicating the row pixel second from the bottom of image to be detected, the string pixel of increase Point obtains by replicating the row pixel second from the bottom of image to be detected.

Step 33, use described concrete Image Edge-Detection operator that described actually detected image is carried out image border inspection Survey, thus obtain the gray value information of the edge-detected image of described image to be detected.

In image to be detected, increase one-row pixels point and string pixel, it is thus achieved that after actually detected image, examine from reality In altimetric image, line number index and columns index are the pixel of first to start, and chooses 2 × 2 successively in actually detected image The gray value data of data window, carries out convolution with concrete Image Edge-Detection operator, travel through complete actually detected image it After, just can obtain the gray value information of the edge-detected image of image to be detected, i.e. obtain with gray value information characterize to be checked The Image Edge-Detection result of altimetric image, according to these gray value informations, it is possible to draw out the edge detection graph of image to be detected Picture, is highlighted, in edge-detected image, the marginal position that position is image to be detected.

The method using the present embodiment to provide, as long as slightly dealing with image to be detected, can use the present invention to provide Image Edge-Detection operator carries out Image Edge-Detection to the image to be detected after processing, and then obtains the edge of image to be detected The gray value information of detection image, whole calculating process, operand is the least, and due to the image limit using the present invention to provide Edge detective operators, when image to be detected is carried out Image Edge-Detection, testing result is affected by noise less, after detection, according to In the edge-detected image that the gray value information obtained is drawn, the edge of image to be detected becomes apparent from.

As it is shown in fig. 7, Fig. 7 is illustrated that the above-mentioned Image Edge-Detection operator of another kind of employing disclosed by the invention is to be checked Altimetric image carries out the flow chart of the method for Image Edge-Detection, and the method includes:

Step 41, obtain for the default interpolation multiplying power of image to be detected.

When being embodied as, preset difference value multiplying power is designated as M.

Step 42, according to described default interpolation multiplying power, determine the value of u and v.

In the present embodiment, according to formulaDetermine the value of u and v.

During above-mentioned acquisition Image Edge-Detection operator, if each second target pixel points is respectively positioned on accordingly The position that in four neighborhoods, row pixel second from the bottom and row pixel second from the bottom intersect, the reference picture obtained and target image Skew maximum, the Image Edge-Detection operator obtained in this manner, after being applied to, image to be detected is carried out figure As, in rim detection, in the edge-detected image obtained, the edge of image to be detected becomes apparent from.In this manner, corresponding The value of u and v according to formulaDetermine.

Step 43, according to the value of described u and v and described Image Edge-Detection operator, determine described concrete image border Detective operators.

In the present embodiment, concrete Image Edge-Detection operator is designated as A.The value of u and v is substituted into Image Edge-Detection calculate Son calculates, just can obtain concrete Image Edge-Detection operator A.

Step 44, increase one-row pixels point in the lower section of last column pixel of the image to be detected obtained, at it The right side of rear string pixel increases string pixel, obtains actually detected image；The one-row pixels point increased is treated by duplication The row pixel second from the bottom of detection image obtains, and the string pixel of increase is by replicating the row second from the bottom of image to be detected Pixel obtains.

Step 45, obtain the gray value data of described image to be detected, and according to the gray value number of described image to be detected According to the gray value data determining described actually detected image.

The gray value data of the row pixel second from the bottom of image to be detected is increased to last column of image to be detected The lower section of the gray value data of pixel, increases to be checked by the gray value data of the row pixel second from the bottom of image to be detected The right side of the gray value data of the last string pixel of altimetric image, just can get the gray value data of actually detected image.

Step 46, described concrete Image Edge-Detection operator is used to carry out with the gray value data of described actually detected image Convolution, thus obtain the gray value information of the edge-detected image of described image to be detected.

When being embodied as, according to formulaUse described concrete Image Edge-Detection operator and described reality The gray value data of detection image carries out convolution, thus obtains the gray value letter of the edge-detected image of described image to be detected Breath.

Wherein, D represents the gray value of pixel in the edge-detected image of described image to be detected, and I ' represents described reality The gray value data of 2 × 2 data windows in detection image, A represents described concrete Image Edge-Detection operator, | | represent definitely Value computing,Represent and round downwards.

After the gray value data using concrete Image Edge-Detection operator and actually detected image carries out convolution, the volume obtained Long-pending result there will be floating number, it is contemplated that the gray value of digital picture the most in round figures, carries out rounding operation to convolution results, Rounding operation refers to herein: takes and rounds result less than the maximum integer of floating number as corresponding floating number, rounds downwards.

After the gray value information of the edge-detected image being calculated image to be detected according to above-mentioned formula, i.e. obtain with ash After the edge detection results of the image to be detected of angle value information representation, the gray value information according to obtaining is drawn, To the edge-detected image of image to be detected, in edge-detected image, the marginal position of image to be detected is for being highlighted.

Use the method that the present embodiment provides, due to during determining concrete Image Edge-Detection operator, u's and v Value is according to formulaDetermining, image to be detected is entered by the concrete Image Edge-Detection operator using this mode to determine After row Image Edge-Detection, according to the gray value information of the edge-detected image obtained, in the edge-detected image of drafting, to be checked The edge of altimetric image is the most clear.

With the method for above-mentioned acquisition Image Edge-Detection operator and use above-mentioned Image Edge-Detection operator to be detected The method that image carries out Image Edge-Detection is corresponding, the embodiment of the invention also discloses a kind of acquisition Image Edge-Detection operator Device and use above-mentioned Image Edge-Detection operator that image to be detected carries out the device of Image Edge-Detection.

As shown in Figure 8, Fig. 8 is illustrated that the structure of a kind of device obtaining Image Edge-Detection operator disclosed by the invention Block diagram, this device includes:

Region divides module 51, in the target image, marks off four neighborhoods including first object pixel； Described first object pixel is line number index and the pixel of columns index all minimums in described four neighborhoods；

Interpolation processing module 52, for described target image carries out bilinear interpolation process, optional one is inserted described Pixel in four neighborhoods, as the second target pixel points, sets up described first object pixel and described second target pixel points Position relationship in described four neighborhoods；

First determines module 53, for according to described position relationship and bilinear interpolation formula, determines that image border is examined Surveying functional relation, described Image Edge-Detection functional relation is for calculating gray value and the institute of described first object pixel State the difference of the gray value of the second target pixel points；

Second determines module 54, for determining that Image Edge-Detection is calculated according to described Image Edge-Detection functional relation Son.

Use the device that the present embodiment provides, it is possible to obtain a kind of general Image Edge-Detection operator, by this image limit Edge detective operators is applied in the technique of image edge detection of image to be detected, while reducing operand, improves greatly Noise resisting ability so that after detection, the gray value information obtained according to detection is drawn in the edge-detected image obtained, to be checked The profile of altimetric image becomes apparent from.

As it is shown in figure 9, Fig. 9 is illustrated that the knot of the another kind of device obtaining Image Edge-Detection operator disclosed by the invention Structure block diagram, this device includes: region divides module 51, and interpolation processing module 52, first determines that module 53 and second determines mould Block 54；

Wherein, interpolation processing module 52 includes:

Interpolation process unit 521, for default interpolation multiplying power, carries out bilinear interpolation process to described target image, Obtain the interpolation image of described target image；

Pixel chooses unit 522, in described interpolation image, selects to insert the second target in described four neighborhoods Pixel, described second target pixel points is positioned at row pixel second from the bottom and row pixel phase second from the bottom in described four neighborhoods The position handed over；

First determine module 53 specifically for:

According to described position relationship, and bilinear interpolation formula, determine following about described first object pixel Gray value and the gray value of described second target pixel points between the first functional relation；

D (u, v)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+ 1)；

When determining each pixel in described target image as first object pixel, right in described interpolation image The second target pixel points answered；

Extract all described second target pixel points in described interpolation image, according to the first mesh described in described target image All described second target pixel points are rearranged by the queueing discipline of mark pixel, form reference picture；

According to described second target pixel points position coordinates in described reference picture and described first functional relationship Formula, determines between the gray value of the following gray value about described first object pixel and described second target pixel points Second functional relation；

D (i, j)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+ 1)；

According to described second functional relation, determine following Image Edge-Detection functional relation；

$\begin{array}{c}G(i,j)-D(i,j)=(v+u-uv)*G(i,j)+u(v-1)*G(i,j+1)+(u-1)v*G(i+1,j)-uv*\\ G(i+1,j+1)\end{array};$

Second determine module 54 specifically for:

Described Image Edge-Detection functional relation is transformed to the Image Edge-Detection functional relationship of following convolution form Formula；

$G(i,j)-D(i,j)=\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right]\left[\begin{array}{cc}G(i,j)& G(i,j+1)\\ G(i+1,j)& G(i+1,j+1)\end{array}\right];$

Image Edge-Detection functional relation according to described convolution form determines following Image Edge-Detection operator；

$\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right];$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents with reference to figure D The gray value of the pixel of the i-th row jth row in Xiang；G (i, j+1) represents the ash of the pixel of the i-th row jth+1 row in target image Angle value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target figure The gray value of the pixel of i+1 row jth+1 row in Xiang；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to the first corresponding mesh The fore-and-aft distance of mark pixel.

The Image Edge-Detection operator that the device provided by the present embodiment is obtained is flat owing to having merged window data weighting All (the bilinear interpolation processing procedure of image) and effects of object pixel minor shifts (the down-sampled processing procedure of image), Being applied in the Image Edge-Detection of image to be detected, testing result is affected by noise less, according to obtain after detection Gray value information, draws in the edge-detected image obtained, and the profile of image to be detected becomes apparent from.

As shown in Figure 10, Figure 10 is illustrated that the above-mentioned Image Edge-Detection operator of a kind of employing disclosed by the invention is to be checked Altimetric image carries out the device of Image Edge-Detection, and this device includes:

3rd determines module 61, for examining according to default interpolation multiplying power and the described image border for image to be detected Measuring and calculating, determines the concrete Image Edge-Detection operator for described image to be detected；

Image processing module 62, for increasing a line picture in the lower section of last column pixel of the image to be detected obtained Vegetarian refreshments, increases string pixel, obtains actually detected image on the right side of its last string pixel；The one-row pixels point increased Obtaining by replicating the row pixel second from the bottom of image to be detected, the string pixel of increase is by replicating image to be detected Row pixel second from the bottom obtains；

Image Edge-Detection module 63, is used for using described concrete Image Edge-Detection operator to described actually detected image Carry out Image Edge-Detection, thus obtain the gray value information of the edge-detected image of described image to be detected.

Using the device that the present embodiment provides, when image to be detected is carried out Image Edge-Detection, testing result is by noise Affecting less, after detection, according in the edge-detected image that the gray value information obtained is drawn, the edge of image to be detected is more Clearly.

As shown in figure 11, Figure 11 is illustrated that the above-mentioned Image Edge-Detection operator of another kind of employing disclosed by the invention is treated Detection image carries out the device of Image Edge-Detection, and this device includes: the 3rd determines module 61, image processing module 62, image Edge detection module 63；

Wherein, the 3rd determine module 61 specifically for:

Obtain default interpolation multiplying power M for image to be detected；

According to described default interpolation multiplying power M, determine the value of u and v according to the following equation；

$u=v=\frac{M}{1+M};$

Value according to described u and v and described Image Edge-Detection operator, determine that described concrete Image Edge-Detection is calculated Son.

Image Edge-Detection module 63 includes:

Gray value data acquiring unit 631, for obtaining the gray value data of described image to be detected, and treats according to described The gray value data of detection image determines the gray value data of described actually detected image；

Image Edge-Detection unit 632, for according to the following equation, uses described concrete Image Edge-Detection operator and institute The gray value data stating actually detected image carries out convolution, thus obtains the gray scale of the edge-detected image of described image to be detected Value information；

Wherein, D represents the gray value of pixel in the edge-detected image of image to be detected, and I ' represents described actually detected The gray value data of 2 × 2 data windows in image, A represents described concrete Image Edge-Detection operator, | | represent absolute value fortune Calculate,Represent and round downwards.

Use the device that the present embodiment provides, when image to be detected is carried out Image Edge-Detection, owing to determining specifically During Image Edge-Detection operator, the value of u and v is according to formulaDetermine so that after detection, according to obtaining The gray value information of edge-detected image, in the edge-detected image of drafting, the edge of image to be detected is the most clear.

Each embodiment in this specification all uses the mode gone forward one by one to describe, identical similar portion between each embodiment Dividing and see mutually, what each embodiment stressed is the difference with other embodiments.Especially for device or For system embodiment, owing to it is substantially similar to embodiment of the method, so describing fairly simple, relevant part sees method The part of embodiment illustrates.Apparatus and system embodiment described above is only schematically, wherein as separating The unit of part description can be or may not be physically separate, and the parts shown as unit can be or also Can not be physical location, i.e. may be located at a place, or can also be distributed on multiple NE.Can be according to reality The needing of border selects some or all of module therein to realize the purpose of the present embodiment scheme.Those of ordinary skill in the art In the case of not paying creative work, i.e. it is appreciated that and implements.

It should be noted that in this article, such as the relational terms of " first " and " second " or the like is used merely to one Individual entity or operation separate with another entity or operating space, and not necessarily require or imply these entities or operate it Between exist any this reality relation or order.And, term " includes ", " comprising " or its any other variant are intended to Contain comprising of nonexcludability, so that include that the process of a series of key element, method, article or equipment not only include those Key element, but also include other key elements being not expressly set out, or also include for this process, method, article or set Standby intrinsic key element.In the case of there is no more restriction, statement " including ... " key element limited, it is not excluded that Other identical element is there is also in including the process of key element, method, article or equipment.

Below it is only the detailed description of the invention of the present invention, it is noted that those skilled in the art are come Saying, under the premise without departing from the principles of the invention, it is also possible to make some improvements and modifications, these improvements and modifications also should be regarded as Protection scope of the present invention.

Claims (14)

1. the method obtaining Image Edge-Detection operator, it is characterised in that including:

In the target image, four neighborhoods including first object pixel are marked off；Described first object pixel is institute State line number index and the pixel of columns index all minimums in four neighborhoods；

Described target image carries out bilinear interpolation process, and an optional pixel inserted in described four neighborhoods is as second Target pixel points, sets up described first object pixel and closes with described second target pixel points position in described four neighborhoods System；

According to described position relationship and bilinear interpolation formula, determine Image Edge-Detection functional relation, described image limit Edge detection function relational expression is for calculating the gray value of described first object pixel and the gray scale of described second target pixel points The difference of value；

Image Edge-Detection operator is determined according to described Image Edge-Detection functional relation.

Method the most according to claim 1, it is characterised in that described described target image is carried out at bilinear interpolation Reason, an optional pixel inserted in described four neighborhoods, as the process of the second target pixel points, specifically includes:

With default interpolation multiplying power, described target image is carried out bilinear interpolation process, it is thus achieved that the interpolation graphs of described target image Picture；

In described interpolation image, select to insert the second target pixel points in described four neighborhoods, described second target pixel points It is positioned at the position that in described four neighborhoods, row pixel second from the bottom and row pixel second from the bottom intersect.

Method the most according to claim 2, it is characterised in that described public according to described position relationship and bilinear interpolation Formula, determines the process of Image Edge-Detection functional relation, specifically includes:

According to described position relationship, and bilinear interpolation formula, determine the following ash about described first object pixel The first functional relation between the gray value of angle value and described second target pixel points；

D (u, v)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+1)；

When determining each pixel in described target image as first object pixel, correspondence in described interpolation image Second target pixel points；

Extract all described second target pixel points in described interpolation image, according to first object picture described in described target image All described second target pixel points are rearranged by the queueing discipline of vegetarian refreshments, form reference picture；

According to described second target pixel points position coordinates in described reference picture and described first functional relation, really Make second between the gray value of the following gray value about described first object pixel and described second target pixel points Functional relation；

D (i, j)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+1)；

According to described second functional relation, determine following Image Edge-Detection functional relation；

$\begin{array}{c}G(i,j)-D(i,j)=(v+u-uv)*G(i,j)+u(v-1)*G(i,j+1)+(u-1)v*G(i+1,j)-uv*\\ G(i+1,j+1)\end{array};$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents in reference picture D The gray value of the pixel of the i-th row jth row；G (i, j+1) represents the gray scale of the pixel of the i-th row jth+1 row in target image Value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target image The gray value of the pixel of middle i+1 row jth+1 row；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to corresponding first object The fore-and-aft distance of pixel.

Method the most according to claim 3, it is characterised in that described true according to described Image Edge-Detection functional relation Determine the process of Image Edge-Detection operator, specifically include:

Described Image Edge-Detection functional relation is transformed to the Image Edge-Detection functional relation of following convolution form；

$G(i,j)-D(i,j)=\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right]\left[\begin{array}{cc}G(i,j)& G(i,j+1)\\ G(i+1,j)& G(i+1,j+1)\end{array}\right];$

Image Edge-Detection functional relation according to described convolution form determines following Image Edge-Detection operator；

$\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right];$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents in reference picture D The gray value of the pixel of the i-th row jth row；G (i, j+1) represents the gray scale of the pixel of the i-th row jth+1 row in target image Value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target image The gray value of the pixel of middle i+1 row jth+1 row；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to corresponding first object The fore-and-aft distance of pixel.

5. the Image Edge-Detection operator that a kind uses described in Claims 1-4 any one carries out image to image to be detected The method of rim detection, it is characterised in that including:

According to for the default interpolation multiplying power of image to be detected and described Image Edge-Detection operator, determine for described to be checked The concrete Image Edge-Detection operator of altimetric image；

One-row pixels point is increased, at its last string pixel in the lower section of last column pixel of the image to be detected obtained Right side increase string pixel, obtain actually detected image；The one-row pixels point increased is by replicating falling of image to be detected Several second row pixels obtain, and the string pixel of increase obtains by replicating the row pixel second from the bottom of image to be detected；

Use described concrete Image Edge-Detection operator that described actually detected image is carried out Image Edge-Detection, thus obtain institute State the gray value information of the edge-detected image of image to be detected.

Method the most according to claim 5, it is characterised in that described basis is for the default interpolation multiplying power of image to be detected And described Image Edge-Detection operator, determine the process of the concrete Image Edge-Detection operator for described image to be detected, Specifically include:

Obtain default interpolation multiplying power M for image to be detected；

According to described default interpolation multiplying power M, determine the value of u and v according to the following equation；

$u=v=\frac{M}{1+M};$

Value according to described u and v and described Image Edge-Detection operator, determine described concrete Image Edge-Detection operator.

Method the most according to claim 6, it is characterised in that the described concrete Image Edge-Detection operator of described employing is to institute State actually detected image and carry out Image Edge-Detection, thus obtain the gray value letter of the edge-detected image of described image to be detected The process of breath, specifically includes:

Obtain the gray value data of described image to be detected, and determine described reality according to the gray value data of described image to be detected The gray value data of border detection image；

According to the following equation, described concrete Image Edge-Detection operator is used to enter with the gray value data of described actually detected image Row convolution, thus obtain the gray value information of the edge-detected image of described image to be detected；

Wherein, D represents the gray value of pixel in the edge-detected image of image to be detected, and I ' represents described actually detected image In the gray value data of 2 × 2 data windows, A represents described concrete Image Edge-Detection operator, | | represent signed magnitude arithmetic(al),Represent and round downwards.

8. the device obtaining Image Edge-Detection operator, it is characterised in that including:

Region divides module, in the target image, marks off four neighborhoods including first object pixel；Described One target pixel points is line number index and the pixel of columns index all minimums in described four neighborhoods；

Interpolation processing module, for described target image carries out bilinear interpolation process, optional one is inserted described four neighborhoods In pixel as the second target pixel points, set up described first object pixel with described second target pixel points described Position relationship in four neighborhoods；

First determines module, for according to described position relationship and bilinear interpolation formula, determines Image Edge-Detection function Relational expression, described Image Edge-Detection functional relation is for calculating the gray value and described second of described first object pixel The difference of the gray value of target pixel points；

Second determines module, for determining Image Edge-Detection operator according to described Image Edge-Detection functional relation.

Device the most according to claim 8, it is characterised in that described interpolation processing module includes:

Interpolation process unit, for default interpolation multiplying power, carries out bilinear interpolation process to described target image, it is thus achieved that described The interpolation image of target image；

Pixel chooses unit, in described interpolation image, selects to insert the second target pixel points in described four neighborhoods, Described second target pixel points is positioned at the position that in described four neighborhoods, row pixel second from the bottom and row pixel second from the bottom intersect Put.

Device the most according to claim 9, it is characterised in that described first determine module specifically for:

According to described position relationship, and bilinear interpolation formula, determine the following ash about described first object pixel The first functional relation between the gray value of angle value and described second target pixel points；

D (u, v)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+1)；

When determining each pixel in described target image as first object pixel, correspondence in described interpolation image Second target pixel points；

Extract all described second target pixel points in described interpolation image, according to first object picture described in described target image All described second target pixel points are rearranged by the queueing discipline of vegetarian refreshments, form reference picture；

According to described second target pixel points position coordinates in described reference picture and described first functional relation, really Make second between the gray value of the following gray value about described first object pixel and described second target pixel points Functional relation；

D (i, j)=(1-u) (1-v) * G (i, j)+u (1-v) * G (i, j+1)+(1-u) v*G (i+1, j)+uv*G (i+1, j+1)；

According to described second functional relation, determine following Image Edge-Detection functional relation；

$\begin{array}{c}G(i,j)-D(i,j)=(v+u-uv)*G(i,j)+u(v-1)*G(i,j+1)+(u-1)v*G(i+1,j)-uv*\\ G(i+1,j+1)\end{array};$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents in reference picture D The gray value of the pixel of the i-th row jth row；G (i, j+1) represents the gray scale of the pixel of the i-th row jth+1 row in target image Value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target image The gray value of the pixel of middle i+1 row jth+1 row；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to corresponding first object The fore-and-aft distance of pixel.

11. devices according to claim 10, it is characterised in that described second determine module specifically for:

Described Image Edge-Detection functional relation is transformed to the Image Edge-Detection functional relation of following convolution form；

$G(i,j)-D(i,j)=\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right]\left[\begin{array}{cc}G(i,j)& G(i,j+1)\\ G(i+1,j)& G(i+1,j+1)\end{array}\right];$

Image Edge-Detection functional relation according to described convolution form determines following Image Edge-Detection operator；

$\left[\begin{array}{cc}v+u-uv& v(u-1)\\ u(v-1)& -uv\end{array}\right];$

Wherein, (i j) represents the gray value of the pixel of the i-th row jth row in target image to G；(i j) represents in reference picture D The gray value of the pixel of the i-th row jth row；G (i, j+1) represents the gray scale of the pixel of the i-th row jth+1 row in target image Value；(i+1 j) represents the gray value of the pixel of i+1 row jth row in target image to G；G (i+1, j+1) represents target image The gray value of the pixel of middle i+1 row jth+1 row；U represents in interpolation image, and the second target pixel points is to corresponding The lateral separation of first object pixel；V represents in interpolation image, and the second target pixel points is to corresponding first object The fore-and-aft distance of pixel.

The Image Edge-Detection operator that 12. 1 kinds use described in Claims 1-4 any one carries out image to image to be detected The device of rim detection, it is characterised in that including:

3rd determines module, for calculating according to default interpolation multiplying power and the described Image Edge-Detection for image to be detected Son, determines the concrete Image Edge-Detection operator for described image to be detected；

Image processing module, increases one-row pixels point for the lower section of last column pixel at the image to be detected obtained, Increase string pixel on the right side of its last string pixel, obtain actually detected image；The one-row pixels point increased passes through The row pixel second from the bottom replicating image to be detected obtains, and the string pixel of increase is by replicating the inverse of image to be detected Secondary series pixel obtains；

Image Edge-Detection module, is used for using described concrete Image Edge-Detection operator that described actually detected image is carried out figure As rim detection, thus obtain the gray value information of the edge-detected image of described image to be detected.

13. devices according to claim 12, it is characterised in that the described 3rd determine module specifically for:

Obtain default interpolation multiplying power M for image to be detected；

According to described default interpolation multiplying power M, determine the value of u and v according to the following equation；

$u=v=\frac{M}{1+M};$

Value according to described u and v and described Image Edge-Detection operator, determine described concrete Image Edge-Detection operator.

14. devices according to claim 13, it is characterised in that described Image Edge-Detection module includes:

Gray value data acquiring unit, for obtaining the gray value data of described image to be detected, and according to described mapping to be checked The gray value data of picture determines the gray value data of described actually detected image；

Image Edge-Detection unit, for according to the following equation, uses described concrete Image Edge-Detection operator and described reality The gray value data of detection image carries out convolution, thus obtains the gray value letter of the edge-detected image of described image to be detected Breath；

Wherein, D represents the gray value of pixel in the edge-detected image of image to be detected, and I ' represents described actually detected image In the gray value data of 2 × 2 data windows, A represents described concrete Image Edge-Detection operator, | | represent signed magnitude arithmetic(al),Represent and round downwards.