CN106780339B

CN106780339B - A method of it is inconsistent to solve interchannel brightness

Info

Publication number: CN106780339B
Application number: CN201611224484.3A
Authority: CN
Inventors: 李赟晟; 王勇; 王凯; 叶红磊
Original assignee: Shanghai Integrated Circuit Research and Development Center Co Ltd; Chengdu Image Design Technology Co Ltd
Current assignee: Shanghai IC R&D Center Co Ltd; Chengdu Image Design Technology Co Ltd
Priority date: 2016-12-27
Filing date: 2016-12-27
Publication date: 2019-02-05
Anticipated expiration: 2036-12-27
Also published as: WO2018121221A1; CN106780339A

Abstract

The present invention provides a kind of methods that solution interchannel brightness is inconsistent, comprising the following steps: provides two images, in the edge joint position of two images, sets a width range；Piece is surrounded by set width；Calculate the brightness of every row pixel in set width range；Using the coordinate of every row pixel value in set width range as abscissa and ordinate, to obtain point set；According to the method for seeking optimal solution in Ransac algorithm, point set is fitted into straight line, obtains the optimal solution model of point set；The match value that each pixel in piece side is found out in optimal solution model adjusts the brightness of the pixel of piece side corresponding position according to the match value.Method of the invention can be automatically performed making up for interchannel difference in brightness, optimize image effect.

Description

A method of it is inconsistent to solve interchannel brightness

Technical field

The present invention relates to image sensor technologies fields, and in particular to a method of it is inconsistent that brightness is built in solution channel.

Background technique

During image imaging joint, there is the inconsistent problem of brightness of image, and have in the centre of image Apparent gap can make image seem apparent splicing gap, and conventional application method is: a channel is added one Value obtains a channel multiplied by a value identical so that two interchannel brightness are consistent or two interchannel adjust gains Brightness, or the emergence processing at splicing gap, so that seeming smooth excessive, brightness change seems unobvious.

But first two method changes unanimous circumstances primarily directed to two channel luminances, not in face of interchannel brightness change Unanimously, that is to say, that effect is unobvious in the case that possible top half and lower half portion variation are different, and a kind of last emergence The case where will affect the clarity of image, facing fine definition is infeasible.

Summary of the invention

In order to overcome the above problems, the present invention is directed to a kind of method that solution interchannel brightness is inconsistent, can know automatically The optimal solution of other interchannel numerical value change improves image effect.

In order to achieve the above object, the present invention provides a kind of method that solution interchannel brightness is inconsistent, packets Include following steps:

Step 01: two images being provided, in the edge joint position of two images, set a width range；Piece is by set Width surrounds；

Step 02: calculating the brightness of every row pixel in set width range；

Step 03: using the coordinate of every row pixel value in set width range as abscissa and ordinate, thus Obtain point set；

Step 04: according to the method for seeking optimal solution in Ransac algorithm, point set being fitted into straight line, is obtained a little The optimal solution model of set；

Step 05: finding out the match value of each pixel in piece side in optimal solution model, adjusted according to the match value The brightness of the pixel of whole piece side corresponding position.

Preferably, in the step 02, the brightness of every row pixel is each pixel in corresponding line

The summation of value is divided by set width.

Preferably, in the optimal solution model, the brightness of every row pixel on the piece left side, which is equal to, is spelled

Stitch the brightness of the pixel of the corresponding line on the right.

Preferably, the step 04 specifically includes:

Step 041: the minimum sampling of setting integrates as the model of n and a sample set (P), and sample set (P) is in every row Point set of the pixel value as transverse and longitudinal coordinate, number of samples is greater than n in sample set (P)；

Step 042: the subset (S) of the sample set P comprising n sample is randomly selected from sample sets (P), simulation obtains just Beginningization model (M)；

Step 043: the sample of given threshold t will be less than in the complementary set of sample sets (P) with the error of initialization model (M) Collection and subset (S) constitute new set (S*), are interior point sets in set (S*)；

Step 044: if number >=N of set (S*), then it is assumed that initialization model (M) is correct；If gathering the number of (S*) =N extracts new subset (S) again, repeats the above steps, and obtains the interior point set of maximum number, and corresponding model is exactly optimal Solve model.

Preferably, in the step 042, initialization model (M) is obtained using least square method.

Of the invention solves the inconsistent method of interchannel brightness, can be automatically performed making up for interchannel difference in brightness, Optimize image effect.

Detailed description of the invention

Fig. 1 is the inconsistent image schematic diagram in the channel of a preferred embodiment of the invention

Fig. 2 is the optimal solution model schematic of a preferred embodiment of the invention

Fig. 3 is the flow diagram for solving the inconsistent method of interchannel brightness of a preferred embodiment of the invention

Specific embodiment

To keep the contents of the present invention more clear and easy to understand, below in conjunction with Figure of description, the contents of the present invention are made into one Walk explanation.Certainly the invention is not limited to the specific embodiment, general replacement known to those skilled in the art It is included within the scope of protection of the present invention.

Below in conjunction with attached drawing 1~3 and specific embodiment, invention is further described in detail.It should be noted that attached drawing It is all made of very simplified form, using non-accurate ratio, and only to facilitate, clearly reach aid illustration the present embodiment Purpose.

Referring to Fig. 3, in the present embodiment, a method of it is inconsistent to solve interchannel brightness, comprising the following steps:

Specifically, as shown in Figure 1, the image inconsistent there are channel chooses certain width in the edge joint position of two images Degree, as shown in Figure 1, there are pieces between gray image and white image.Here the width set is i_weight.

Step 02: calculating the brightness of every row pixel in set width range；

Specifically, the brightness of every row pixel is the summation of each pixel value in corresponding line divided by set width.Please after It is continuous refering to fig. 1, the brightness on the jth row piece left side be Mean_left (1, j)=(I (i, j)+I (i-1, j)+...+I (i-i_ Weight+1, j))/i_weight, it can similarly calculate the brightness Mean_right (1, j) on the right of jth row piece.

Specifically, please continue to refer to Fig. 1, such as the abscissa of a pixel I (i, j) is i, ordinate j, to constitute One coordinate, and then the point set in the width range set.

Specifically, referring to Fig. 2, the brightness of every row pixel on the piece left side is equal on the right of piece in optimal solution model The brightness of the pixel of corresponding line.

This step 04 specifically includes:

Step 042: the subset (S) of the sample set P comprising n sample is randomly selected from sample sets (P), simulation obtains just Beginningization model (M)；Here it is possible to obtain initialization model (M) using least square method.

Step 044: if number >=N of set (S*), then it is assumed that initialization model (M) is correct；If gathering the number of (S*) =N extracts new subset (S) again, repeats the above steps, and obtains the interior point set of maximum number, and corresponding model is exactly optimal Solve model

Specifically, since the pixel intensity on the left of piece is equal to the pixel intensity on the right side of piece, Mean_left=f (Mean_right), all pixels of piece left or right side can be mapped with the respective pixel in optimal solution model, To complete brightness adjustment.

Although the present invention is disclosed as above with preferred embodiment, right embodiment is illustrated only for the purposes of explanation, and It is non-to limit the present invention, those skilled in the art can make without departing from the spirit and scope of the present invention it is several more Dynamic and retouching, the protection scope that the present invention is advocated should be subject to claims.

Claims

1. a kind of solve the inconsistent method of interchannel brightness, which comprises the following steps:

Step 01: two images being provided, in the edge joint position of two images, set a width range；Piece is by set width It surrounds；

Step 02: calculating the brightness of every row pixel in set width range；

Step 03: using the coordinate of every row pixel value in set width range as abscissa and ordinate, to obtain Point set；

Step 04: according to the method for seeking optimal solution in Ransac algorithm, point set being fitted into straight line, obtains point set Optimal solution model；It specifically includes:

Step 041: the minimum sampling of setting integrates as the model of n and a sample set (P), and sample set (P) is with the pixel in every row It is worth the point set as transverse and longitudinal coordinate, number of samples is greater than n in sample set (P)；

Step 042: the subset (S) of the sample set P comprising n sample is randomly selected from sample sets (P), simulation is initialized Model (M)；

Step 043: the sample set of given threshold t will be less than in the complementary set of sample sets (P) with the error of initialization model (M), and Subset (S) constitutes new set (S*), is interior point set in set (S*)；

Step 044: if number >=N of set (S*), then it is assumed that initialization model (M) is correct；If gathering the number=N of (S*), Again new subset (S) is extracted, is repeated the above steps, the interior point set of maximum number is obtained, corresponding model is exactly optimal solution mould Type；In the optimal solution model, the brightness of every row pixel on the piece left side is equal to the brightness of the pixel of the corresponding line on the right of piece；

Step 05: finding out the match value of each pixel in piece side in optimal solution model, spelling is adjusted according to the match value Stitch the brightness of the pixel of side corresponding position.

2. the method according to claim 1, wherein the brightness of every row pixel is corresponding line in the step 02 In each pixel value summation divided by set width.

3. the method according to claim 1, wherein being obtained initially in the step 042 using least square method Change model (M).