CN106991664A - A kind of method that graphics field in image is normalized - Google Patents

Abstract

The present invention relates to a kind of method that graphics field in image is normalized.The normalized algorithm of elliptical image of the present invention can will be processed into the circular image of normal size because of the reasons such as shooting angle and shooting distance ellipse not of uniform size as formed by distorting circular image；For circle, equilateral triangle, square normalized, available for the traffic sign recognition system based on video.

Description

A kind of method that graphics field in image is normalized

Technical field

The present invention relates to a kind of method that graphics field in image is normalized, belong to the skill of image procossing Art field.

Background technology

There is ellipse, scalene triangle, parallel four side close to square in image procossing, in many images Shape or rectangle (abbreviation class square).And these figures in image are often due to camera shooting angle reason, by circular, equilateral Triangle, square occur geometric distortion and formed.Shooting angle is varied, and the distortion to be formed is shot to same image object Shape afterwards also varies.But can become to become after oval, equilateral triangle distortion after circle distortion scalene triangle, It can become rectangular or parallelogram after square distortion.Such as need to circle, triangle, square carry out automatic identification or Match somebody with somebody, then need to be translated into the circle of normal size, the equilateral triangle of normal size, the square of normal size.

Bilinear interpolation algorithm, Hough transform detection algorithm are that image procossing and pattern-recognition are conventional in the prior art Algorithm.

Hough transform detection algorithm is generalized to detection curve, referred to as generalised Hough transform (GHT).Generalised Hough transform It is the effective ways of detection circle, but because round radius, coordinate have three free parameters, using generalised Hough transform, amount of calculation Huge, the substantial amounts of internal memory of needs.Probability Hough transformation (Progressive Probabilistic Hough Transform, contracting It is written as PPHT) it can effectively overcome drawbacks described above.Document：[1] Yuan Li, Ye Lu, merchant build ellipses detections of the Lu based on Hough transform Algorithm [J] Chinese Opticals and Application Optics, in August, 2010,3 (4):379-384. discloses the oval original of Hough transform detection Reason.Document：[2] Li Haibing, easily defends the Chinese image graphics of algorithm [J] for whether having triangle in a kind of efficient detection images of eastern Journal, 2008,13 (3):456-460. disclose Hough transform triangle detection algorithm.Document：[3] Li Qiang soldier, Liu Wen gives Fast Rectangle detection algorithm [J] microcomputer informations based on Hough transform, 2007,23 (31):248-250. disclose Hough transform hough transform algorithm.Document：[4] Zhang Xinghui, Liu Ling, Du Sheng it, wait license plates position and sloped correcting method Study [J] system engineerings and electronic technology, 2004,26 (2):237-239. disclose a kind of parallelogram detection algorithm.But Normalized algorithm is not proposed for special shapes such as ellipse, triangle, class squares.

Perspective transform is also a kind of method commonly used in image procossing.Document [8] good of Dai Qin, Wang Yanjie, Han Guang is based on changing Enter fluoroscopy images correction [J] liquid crystal and the display, 2012,27 (4) of Hough transform and perspective transform:552-556. is disclosed Depending on the detailed process of conversion.

In Traffic Sign Images, circular, equilateral triangle and square-shaped image proportion are very big.For circular, equilateral The normalized of triangle, square, available for the traffic sign recognition system based on video.

The content of the invention

In view of the shortcomings of the prior art, the present invention provides the side that a kind of graphics field in image is normalized Method.

The technical scheme is that：

The profile of graphics field is in a kind of method that graphics field in image is normalized, described image Ellipse, including step are as follows：

According to the principle of perspective transform, new coordinate of the two dimensional image after perspective transform is：

Wherein (x, y) is the pixel coordinate of original image, and (u, v) is the pixel coordinate of image after perspective transform, a, b, c, d, E, f, m, l are perspective transform parameters；

The matrix form of formula (2) is：

Four pixel point coordinates in original image are designated as (x₁, y₁)(x₂, y₂)(x₃, y₃)(x₄, y₄), corresponding perspective becomes The coordinate for changing corresponding points in rear image is designated as (u₁, v₁)(u₂, v₂)(u₃, v₃)(u₄, v₄), it can obtain：

Formula (4) is designated as：B=AM (5)

Then：M=A^-1B (6)

By (2), formula is obtained：

It is represented by with matrix：

The reconstructed formula of perspective transform：

E1) oval general equation is：

Ax²+Bxy+Cy²+ Dx+Ey+1=0 (10)

Long axial rake is θ：

Oval central coordinate of circle is：

The length of oval long and short semiaxis is respectively a and b, satisfaction：

E2) oval minimum enclosed rectangle A_eB_eC_eD_eFour apex coordinates be respectively：

E3 the radius of standard circular region contour after perspective transform) is set as r, and standard circular is minimum external after perspective transform Square E_eF_eG_eH_eFour apex coordinates be respectively：E_e(0,0)；F_e(2r, 0)；G_e(2r, 2r)；H_e(0,2r)；

E4 four points before and after conversion) are obtained into perspective parameter matrix M to substituting into formula (6)；

E5) using reconstructed formula (9) is had an X-rayed, whole integer pixel point coordinates correspondences in the image after perspective transform are obtained Floating number coordinate position in original image；

E6 bilinear interpolation algorithm) is utilized, each pixel in standard picture, asks for each pixel after traversal perspective transform The gray value of point, the standard circular area image after being normalized；

According to currently preferred, the profile of graphics field is triangle in image；

T1) in profile triangle A_tB_tC_tAn external parallelogram A is built on summit_tB_tD_tE_t, profile triangle Base A_tB_tIt is used as parallelogram A_tB_tD_tE_tA side, profile triangular apex C_tIt is used as parallelogram A_tB_tD_tE_tIt is another Individual side D_tE_tMidpoint；

T2) by Hough transform triangle detection algorithm, three apex coordinates of detection profile triangle are respectively：A_t (x₁,y₁)；B_t(x₂,y₂)；C_t(x₃,y₃)；Then, parallelogram A_tB_tD_tE_tThe coordinate on four summits be respectively：A_t(x₁,y₁)； B_t(x₂, y₁)；

T3 the length of side of standard equilateral triangle profile after perspective transform) is set as w, standard equilateral triangle wheel after perspective transform Four apex coordinates of wide boundary rectangle are respectively：F_t(0,0)；G_t(w, 0)；

T4 four points before and after perspective transform) are substituted into formula (6), perspective parameter matrix M is obtained；

T5) using reconstructed formula (9) is had an X-rayed, whole rounded coordinate points in the image after perspective transform are obtained corresponding in original Floating number coordinate in beginning image；

T6 bilinear interpolation algorithm) is utilized, each pixel in standard picture, asks for each pixel after traversal perspective transform The gray value of point, the standard equilateral triangle area image after being normalized.

According to currently preferred, figure is class square in described image：

S1) profile class square A_sB_sC_sD_sThe apex coordinate of correspondence profile is respectively：A_s(x₁, y₁)；B_s(x₂, y₂)；C_s (x₃, y₃)；D_s(x₄, y₄)；If the length of side of standard square region contour is w, standard square after perspective transform after perspective transform E_sF_sH_sG_sFour apex coordinates be respectively：E_s(0,0)；F_s(w, 0)；H_s(w, w)；G_s(0, w)；By four points before and after conversion After all obtaining, formula (6) is substituted into, perspective parameter matrix M is obtained；

S2) using reconstructed formula (9) is had an X-rayed, whole integer pixel points correspondence in new images is obtained in original image Floating number coordinate；

S3 bilinear interpolation algorithm) is utilized, each pixel in standard picture region, is asked for each after traversal perspective transform The gray value of point, the standard square area image after being normalized.

Beneficial effects of the present invention are：

1. the normalized algorithm of elliptical image of the present invention can by because the reasons such as shooting angle and shooting distance by Ellipse not of uniform size is processed into the circular image of normal size formed by circular image distortion；

2. the normalized algorithm of triangular image of the present invention can be by because of reasons such as shooting angle and shooting distances Common triangular image not of uniform size is processed into the equilateral figure of normal size as formed by equilateral triangle pattern distortion Picture；

3. the normalized algorithm of square-shaped image of the present invention can be by because of reasons such as shooting angle and shooting distances Class square-shaped image not of uniform size is processed into the square-shaped image of normal size as formed by distorting square-shaped image.

Brief description of the drawings

Fig. 1 is the preceding oval schematic diagram of conversion in embodiment；

Fig. 2 is schematic diagram round after being converted in embodiment；

Fig. 3 be embodiment in convert first three angular schematic diagram；

Fig. 4 is the schematic diagram of conversion rear triangle in embodiment；

Fig. 5 is the schematic diagram of the preceding class square of conversion in embodiment；

Fig. 6 is schematic diagram square after being converted in embodiment；

Fig. 7 is bilinear interpolation algorithm principle schematic.

Embodiment

With reference to embodiment and Figure of description, the present invention will be further described, but not limited to this.

Embodiment 1

As shown in Figure 1-2.

The profile of graphics field is in a kind of method that graphics field in image is normalized, described image Ellipse, including step are as follows：

According to the principle of perspective transform, new coordinate of the two dimensional image after perspective transform is：

Wherein (x, y) is the pixel coordinate of original image, and (u, v) is the pixel coordinate of image after perspective transform, a, b, c, d, E, f, m, l are perspective transform parameters；

The matrix form of formula (2) is：

Four pixel point coordinates in original image are designated as (x₁, y₁)(x₂, y₂)(x₃, y₃)(x₄, y₄), corresponding perspective becomes The coordinate for changing corresponding points in rear image is designated as (u₁, v₁)(u₂, v₂)(u₃, v₃)(u₄, v₄), it can obtain：

Formula (4) is designated as：B=AM (5)

Then：M=A^-1B (6)

By (2), formula is obtained：

It is represented by with matrix：

The reconstructed formula of perspective transform：

E1) oval general equation is：

Ax²+Bxy+Cy²+ Dx+Ey+1=0 (10)

Long axial rake is θ：

Oval central coordinate of circle is：

The length of oval long and short semiaxis is respectively a and b, satisfaction：

E2) oval minimum enclosed rectangle A_eB_eC_eD_eFour apex coordinates be respectively：

E3 the radius of standard circular region contour after perspective transform) is set as r, and standard circular is minimum external after perspective transform Square E_eF_eG_eH_eFour apex coordinates be respectively：E_e(0,0)；F_e(2r, 0)；G_e(2r, 2r)；H_e(0,2r)；

E4 four points before and after conversion) are obtained into perspective parameter matrix M to substituting into formula (6)；

E5) using reconstructed formula (9) is had an X-rayed, whole integer pixel point coordinates correspondences in the image after perspective transform are obtained Floating number coordinate position in original image；

E6 bilinear interpolation algorithm) is utilized, each pixel in standard picture, asks for each pixel after traversal perspective transform The gray value of point, the standard circular area image after being normalized；

Zoom in and out that (core concept of bilinear interpolation algorithm is in both direction to image using bilinear interpolation algorithm Carry out once linear interpolation respectively) the step of it is as follows：(as shown in Figure 7)

Known function f is in Q₁₁=(x₁,y₁), Q₁₂=(x₁,y₂), Q₂₁=(x₂,y₁) and Q₂₂=(x₂,y₂) four points Value,

Values of the unknown function f to be asked in point P=(x, y).

The first step：Linear interpolation is carried out in x directions, is obtained：

Wherein, R₁=(x, y₁)

Wherein, R₁=(x, y₂)

Second step：Linear difference is carried out in y directions, is obtained：

It is final to can obtain desired result f (x, y).

Embodiment 2

As shown in Figure 3-4.

The method that the graphics field in image is normalized as described in Example 1, except that, image The profile of middle graphics field is triangle；

T1) in profile triangle A_tB_tC_tAn external parallelogram A is built on summit_tB_tD_tE_t, profile triangle Base A_tB_tIt is used as parallelogram A_tB_tD_tE_tA side, profile triangular apex C_tIt is used as parallelogram A_tB_tD_tE_tIt is another Individual side D_tE_tMidpoint；

T2) by Hough transform triangle detection algorithm, three apex coordinates of detection profile triangle are respectively：A_t (x₁,y₁)；B_t(x₂,y₂)；C_t(x₃,y₃)；Then, parallelogram A_tB_tD_tE_tThe coordinate on four summits be respectively：A_t(x₁,y₁)； B_t(x₂,y₂)；

T3 the length of side of standard equilateral triangle profile after perspective transform) is set as w, standard equilateral triangle wheel after perspective transform Four apex coordinates of wide boundary rectangle are respectively：F_t(0,0)；G_t(w, 0)；

T4 four points before and after perspective transform) are substituted into formula (6), perspective parameter matrix M is obtained；

T5) using reconstructed formula (9) is had an X-rayed, whole rounded coordinate points in the image after perspective transform are obtained corresponding in original Floating number coordinate in beginning image；

T6 bilinear interpolation algorithm) is utilized, each pixel in standard picture, asks for each pixel after traversal perspective transform The gray value of point, the standard equilateral triangle area image after being normalized.

Embodiment 3

As seen in figs. 5-6.

The method that the graphics field in image is normalized as described in Example 1, except that, it is described Figure is class square in image：

S1) profile class square A_sB_sC_sD_sThe apex coordinate of correspondence profile is respectively：A_s(x₁, y₁)；B_s(x₂, y₂)；C_s (x₃, y₃)；D_s(x₄, y₄)；If the length of side of standard square region contour is w, standard square after perspective transform after perspective transform E_sF_sH_sG_sFour apex coordinates be respectively：E_s(0,0)；F_s(w, 0)；H_s(w, w)；G_s(0, w)；By four points before and after conversion After all obtaining, formula (6) is substituted into, perspective parameter matrix M is obtained；

S2) using reconstructed formula (9) is had an X-rayed, whole integer pixel points correspondence in new images is obtained in original image Floating number coordinate；

S3 bilinear interpolation algorithm) is utilized, each pixel in standard picture region, is asked for each after traversal perspective transform The gray value of point, the standard square area image after being normalized.

Claims (3)

1. a kind of method that graphics field in image is normalized, it is characterised in that graphics field in image Profile is as follows for ellipse, including step：

According to the principle of perspective transform, new coordinate of the two dimensional image after perspective transform is：

$u=\frac{ax+by+c}{mx+ly+1},v=\frac{dx+ey+f}{mx+ly+1}---\left(2\right)$

Wherein (x, y) is the pixel coordinate of original image, and (u, v) is the pixel coordinate of image after perspective transform, a, b, c, d, e, f, M, l are perspective transform parameters；

The matrix form of formula (2) is：

$\left[\begin{array}{c}u\\ v\end{array}\right]=\left[\begin{array}{cccccccc}x& y& 1& 0& 0& 0& -ux& -uy\\ 0& 0& 0& x& y& 1& -vx& -vy\end{array}\right]\left[\begin{array}{c}a\\ b\\ c\\ d\\ e\\ f\\ m\\ l\end{array}\right]---\left(3\right)$

Four pixel point coordinates in original image are designated as (x₁, y₁)(x₂, y₂)(x₃, y₃)(x₄, y₄), after corresponding perspective transform The coordinate of corresponding points is designated as (u in image₁, v₁)(u₂, v₂)(u₃, v₃)(u₄, v₄), it can obtain：

$\left[\begin{array}{c}{u}_1\\ v_1\\ u_2\\ v_2\\ u_3\\ v_3\\ u_4\\ v_4\end{array}\right]=\left[\begin{array}{cccccccc}{x}_1& y_1& 1& 0& 0& 0& -u_1{x}_1& -u_1{y}_1\\ 0& 0& 0& x_1& y_1& 1& -v_1{x}_1& -v_1{y}_1\\ x_2& y_2& 1& 0& 0& 0& -u_2{x}_2& -u_2{y}_2\\ 0& 0& 0& x_2& y_2& 1& -v_2{x}_2& -v_2{y}_2\\ x_3& y_3& 1& 0& 0& 0& -u_3{x}_3& -u_3{y}_3\\ 0& 0& 0& x_3& y_3& 1& -v_3{x}_3& -v_3{y}_3\\ x_4& y_4& 1& 0& 0& 0& -u_4{x}_4& -u_4{y}_4\\ 0& 0& 0& x_4& y_4& 1& -v_4{x}_4& -v_4{y}_4\end{array}\right]\left[\begin{array}{c}a\\ b\\ c\\ d\\ e\\ f\\ m\\ l\end{array}\right]---\left(4\right)$

Formula (4) is designated as：B=AM (5)

Then：M=A^-1B (6)

By (2), formula is obtained：

$\begin{array}{c}(\mathrm{mu}-a)x+(\mathrm{lu}-b)y=c-u\\ (\mathrm{mv}-d)x+(\mathrm{lv}-e)y=f-v\end{array}---\left(7\right)$

It is represented by with matrix：

$\left[\begin{array}{cc}c-u& f-v\end{array}\right]=\left[\begin{array}{cc}x& y\end{array}\right]\left[\begin{array}{cc}mu-a& mv-d\\ lu-b& lv-e\end{array}\right]---\left(8\right)$

The reconstructed formula of perspective transform：

$\left[\begin{array}{cc}x& y\end{array}\right]=\left[\begin{array}{cc}c-u& f-v\end{array}\right]{\left[\begin{array}{cc}mu-a& mv-d\\ lu-b& lv-e\end{array}\right]}^{-1}---\left(9\right)$

E1) oval general equation is：

Ax²+Bxy+Cy²+ Dx+Ey+1=0 (10)

Long axial rake is θ：

$\mathrm{\θ}=\frac{1}{2}\arctan \frac{B}{A-C}---\left(11\right)$

Oval central coordinate of circle is：

$\begin{array}{cc}{X}_c=\frac{BE-2CD}{4AC-B^2}& Y_c=\frac{BD-2AE}{4AC-B^2}\end{array}---\left(12\right)$

The length of oval long and short semiaxis is respectively a and b, satisfaction：

$a^2=\frac{2\left({\mathrm{AX}}_c^2+{\mathrm{CY}}_c^2+{\mathrm{BX}}_c{Y}_c-1\right)}{A+C+\stackrel{\‾}{{\left(A-C\right)}^2+B^2}}$

$b^2=\frac{2\left({\mathrm{AX}}_c^2+{\mathrm{CY}}_c^2+{\mathrm{BX}}_c{Y}_c-1\right)}{A+C-\stackrel{\‾}{{\left(A-C\right)}^2+B^2}}---\left(13\right)$

E2) oval minimum enclosed rectangle A_eB_eC_eD_eFour apex coordinates be respectively：

$A_e(X_c+asin(\mathrm{\θ}-\frac{\mathrm{\π}}{2})-bcos(\mathrm{\θ}-\frac{\mathrm{\π}}{2}),Y_c-acos(\mathrm{\θ}-\frac{\mathrm{\π}}{2})-bsin(\mathrm{\θ}-\frac{\mathrm{\π}}{2}\left)\right);$

$B_e(X_c+asin(\mathrm{\θ}-\frac{\mathrm{\π}}{2})+bcos(\mathrm{\θ}-\frac{\mathrm{\π}}{2}),Y_c-acos(\mathrm{\θ}-\frac{\mathrm{\π}}{2})+bsin(\mathrm{\θ}-\frac{\mathrm{\π}}{2}\left)\right);$

$C_e(X_c-asin(\mathrm{\θ}-\frac{\mathrm{\π}}{2})+bcos(\mathrm{\θ}-\frac{\mathrm{\π}}{2}),Y_c+acos(\mathrm{\θ}-\frac{\mathrm{\π}}{2})+bsin(\mathrm{\θ}-\frac{\mathrm{\π}}{2}\left)\right);$

$D_e(X_c-asin(\mathrm{\θ}-\frac{\mathrm{\π}}{2})-bcos(\mathrm{\θ}-\frac{\mathrm{\π}}{2}),Y_c+acos(\mathrm{\θ}-\frac{\mathrm{\π}}{2})-bsin(\mathrm{\θ}-\frac{\mathrm{\π}}{2}\left)\right);$

E3 the radius of standard circular region contour after perspective transform) is set as r, the minimum external pros of standard circular after perspective transform Shape E_eF_eG_eH_eFour apex coordinates be respectively：E_e(0,0)；F_e(2r, 0)；G_e(2r, 2r)；H_e(0,2r)；

E4 four points before and after conversion) are obtained into perspective parameter matrix M to substituting into formula (6)；

E5) using reconstructed formula (9) is had an X-rayed, whole integer pixel point coordinates in the image after perspective transform are obtained corresponding in original Floating number coordinate position in beginning image；

E6 bilinear interpolation algorithm) is utilized, each pixel in standard picture, asks for each pixel after traversal perspective transform Gray value, the standard circular area image after being normalized.

2. the method that the graphics field according to claim 1 in image is normalized, it is characterised in that figure The profile of graphics field is triangle as in；

T1) in profile triangle A_tB_tC_tAn external parallelogram A is built on summit_tB_tD_tE_t, the base of profile triangle A_tB_tIt is used as parallelogram A_tB_tD_tE_tA side, profile triangular apex C_tIt is used as parallelogram A_tB_tD_tE_tAnother side D_tE_tMidpoint；

T2) by Hough transform triangle detection algorithm, three apex coordinates of detection profile triangle are respectively：A_t(x₁, y₁)；B_t(x₂,y₂)；C_t(x₃,y₃)；Then, parallelogram A_tB_tD_tE_tThe coordinate on four summits be respectively：A_t(x₁,y₁)；B_t (x₂,y₂)；

T3 the length of side of standard equilateral triangle profile after perspective transform) is set as w, after perspective transform outside standard equilateral triangle profile Four apex coordinates for connecing rectangle are respectively：F_t(0,0)；G_t(w, 0)；

T4 four points before and after perspective transform) are substituted into formula (6), perspective parameter matrix M is obtained；

T5) using reconstructed formula (9) is had an X-rayed, whole rounded coordinate points in the image after perspective transform are obtained corresponding in original graph Floating number coordinate as in；

T6 bilinear interpolation algorithm) is utilized, each pixel in standard picture, asks for each pixel after traversal perspective transform Gray value, the standard equilateral triangle area image after being normalized.

3. the method that the graphics field according to claim 1 in image is normalized, it is characterised in that institute It is class square to state the profile of graphics field in image：

S1) profile class square A_sB_sC_sD_sThe apex coordinate of correspondence profile is respectively：A_s(x₁, y₁)；B_s(x₂, y₂)；C_s(x₃, y₃)； D_s(x₄, y₄)；If the length of side of standard square region contour is w, standard square E after perspective transform after perspective transform_sF_sH_sG_s Four apex coordinates be respectively：E_s(0,0)；F_s(w, 0)；H_s(w, w)；G_s(0, w)；By four points before and after conversion to whole After obtaining, formula (6) is substituted into, perspective parameter matrix M is obtained；

S2) using reconstructed formula (9) is had an X-rayed, floating-point of the whole integer pixel points correspondence in new images in original image is obtained Number coordinate；

S3 bilinear interpolation algorithm) is utilized, each pixel in standard picture region, asks for each point after traversal perspective transform Gray value, the standard square area image after being normalized.