CN110288548B - Integer pixel value image floating point method - Google Patents

Abstract

The invention discloses a floating point method for an integer pixel value image, which comprises the following steps: step 1: high spot coordinates (r) in the input image₀,c₀) And an image matrix mat, the size of the mat is width × height, the coordinates of the highlight points are determined manually, and r represents the line number of the light source points in the matC represents the column number of the light source point in mat; defining the light source point as the position of the strongest light source in the image or the position of the point nearest to the light source point in the image; step 2: respectively calculating the distances from the highlight point to four boundaries of the upper boundary, the lower boundary, the left boundary and the right boundary of the image, finding the maximum distance and recording the maximum distance as maxR; and step 3: creating a zero floating point matrix dstMat with the size of width multiplied by height; and 4, step 4: changing the value of the c column r row of the dstMat to the value of the highlight point; and 5: from a high light point (r)₀,c₀) Expanding to realize floating point formation; the invention has the advantages that: the invention can realize floating point based integer pixel value in a picture with real world contents.

Description

Integer pixel value image floating point method

Technical Field

The invention relates to the technical field of image processing, in particular to an integer pixel value image floating point method.

Background

In real life, human eyes perceive optical signals as analog signals, and a camera or a camera performs discretization processing on the optical signals in real life and stores the optical signals in real life as digital signals, so that signal capture of the human eyes and the camera has certain difference. Due to the storage mode of the image, much information is lost after the picture of the real world is mapped in the image.

Disclosure of Invention

In order to overcome the defect that a lot of information is lost after a real-world picture is mapped in an image, the invention provides an integer type pixel value image floating point method, which enables pixels to influence other pixels according to the position of a light source point in the image and floats the pixel value of the integer type image.

The technical scheme of the invention is as follows:

an integer pixel value image floating point method is characterized by comprising the following steps:

step 1: high spot coordinates (r) in the input image₀,c₀) And an image matrix mat, the size of the mat is width × height, the coordinates of highlight points are determined manually, r represents the row number of the light source points in the mat, and c represents the column number of the light source points in the mat; defining the light source point as the position of the strongest light source in the image or the position of the point nearest to the light source point in the image;

step 2: respectively calculating the distances from the highlight point to four boundaries of the upper boundary, the lower boundary, the left boundary and the right boundary of the image, finding the maximum distance and recording the maximum distance as maxR;

and step 3: creating a zero floating point matrix dstMat with the size of width multiplied by height;

and 4, step 4: changing the value of the c column r row of the dstMat to the value of the highlight point;

and 5: from a high light point (r)₀,c₀) Expanding to realize floating point, and the specific steps are as follows:

step 5.1: defining the calculation formula of the dynamic moving average as shown in (1):

dma(pre,now)＝(1-α)*now+α*pre (1)

wherein, now and pre are floating point decimal values, alpha is the influence rate of diffuse reflection, and alpha is more than or equal to 0 and less than or equal to 1;

step 5.2: starting from the light source point, the rectangle expands, the expansion radius d gradually expands from 1 to maxR by step size of 1, and coordinate points (rs, cs) and (re, ce) of the upper left corner and the lower right corner of the rectangle after each expansion are calculated according to the formulas (2) - (5):

step 5.3: during each expansion, the value of the boundary of the row-column expansion rectangle is calculated, and the specific formula is as follows:

dstMat_rs，cs＝dma(dstMat_rs+1，cs+1，mat_rs，cs) (6)

dstMat_rs,ce＝dma(dstMat_rs+1,ce-1,mat_rs,ce) (7)

dstMat_re,cs＝dma(dstMat_re-1,cs+1,mat_re,cs) (8)

dstMat_re,ce＝dma(dstMat_re-1,ce-1,mat_re,ce) (9)

dstMat_rr,cs＝dma(dstMat_rr,cs+1,mat_re,cs),rr∈(rs,re) (10)

dstMat_rr,ce＝dma(dstMat_rr,ce-1,mat_re,ce),rr∈(rs,re) (11)

dstMat_rs,cc＝dma(dstMat_rs+1,cc,mat_rs,cc),cc∈(cs,ce) (12)

dstMat_re,cc＝dma(dstMat_re-1,cc,mat_re,cc),cc∈(cs,ce) (13)

wherein, mat_r,cAnd dstMat_r,cAnd respectively representing the values of the r row and the c column of the matrix mat and the dstMat, wherein after the operation is finished, the dstMat is the finally obtained floating point matrix.

The invention has the advantages that: the invention can float the integer pixel value in a picture with real world contents according to the diffuse reflection principle of light, can integrate surrounding pixel information into the current pixel point, is beneficial to helping other algorithms to obtain the surrounding information of the current pixel point, and in addition, prolongs the value range of the pixel point.

Drawings

FIG. 1 is a highlight point illustration;

in the figure: the left image point A is a highlight point, the right image point A is a highlight point, the point B is an image center point, and the point C is a light source point (such as the sun) outside the image.

Detailed Description

The following examples are given to illustrate specific embodiments of the present invention.

The integer pixel value image floating point method includes the following steps:

step 1: high spot coordinates (r) in the input image₀,c₀) And an image matrix mat, the size of the mat is width × height, the coordinates of highlight points are determined manually, r represents the row number of the light source points in the mat, and c represents the column number of the light source points in the mat; defining the light source point as the position of the strongest light source in the image or the position of the point nearest to the light source point in the image;

step 2: respectively calculating the distances from the highlight point to four boundaries of the upper boundary, the lower boundary, the left boundary and the right boundary of the image, finding the maximum distance and recording the maximum distance as maxR;

and step 3: creating a zero floating point matrix dstMat with the size of width multiplied by height;

and 4, step 4: changing the value of the c column r row of the dstMat to the value of the highlight point;

and 5: from a high light point (r)₀,c₀) Expanding to realize floating point, and the specific steps are as follows:

step 5.1: defining the calculation formula of the dynamic moving average as shown in (1):

dma(pre,now)＝(1-α)*now+α*pre (1)

wherein now and pre are floating point decimal values, now is a current floating point pixel value, pre is a floating point pixel value of a certain pixel around the current point (which specific surrounding is determined by a place for calling the dma formula), α is a diffuse reflection influence rate, α is greater than or equal to 0 and less than or equal to 1, and in this example, α is greater than or equal to 0.1;

step 5.2: starting from the light source point, the rectangle expands, the expansion radius d gradually expands from 1 to maxR by step size of 1, and coordinate points (rs, cs) and (re, ce) of the upper left corner and the lower right corner of the rectangle after each expansion are calculated according to the formulas (2) - (5):

step 5.3: during each expansion, the value of the boundary of the row-column expansion rectangle is calculated, and the specific formula is as follows:

dstMat_rs，cs＝dma(dstMat_rs+1，cs+1，mat_rs，cs) (6)

dstMat_rs,ce＝dma(dstMat_rs+1,ce-1,mat_rs,ce) (7)

dstMat_re,cs＝dma(dstMat_re-1,cs+1,mat_re,cs) (8)

dstMat_re,ce＝dma(dstMat_re-1,ce-1,mat_re,ce) (9)

dstMat_rr,cs＝dma(dstMat_rr,cs+1,mat_re,cs),rr∈(rs,re) (10)

dstMat_rr,ce＝dma(dstMat_rr,ce-1,mat_re,ce),rr∈(rs,re) (11)

dstMat_rs,cc＝dma(dstMat_rs+1,cc,mat_rs,cc),cc∈(cs,ce) (12)

dstMat_re,cc＝dma(dstMat_re-1,cc,mat_re,cc),cc∈(cs,ce) (13)

wherein, mat_r,cAnd dstMat_r,cAnd respectively representing the values of the r row and the c column of the matrix mat and the dstMat, wherein after the operation is finished, the dstMat is the finally obtained floating point matrix.

The embodiments described in this specification are merely illustrative of implementations of the inventive concept and the scope of the present invention should not be considered limited to the specific forms set forth in the embodiments but rather by the equivalents thereof as may occur to those skilled in the art upon consideration of the present inventive concept.

Claims (1)

1. An integer pixel value image floating point method is characterized by comprising the following steps:

step 1: high spot coordinates (r) in the input image₀,c₀) And an image matrix mat, the size of the mat is width × height, the coordinates of the highlight points are determined manually, r represents the row number of the highlight points in the mat, and c represents the column number of the highlight points in the mat; the highlight point is the position of the strongest light source in the image or the position of the point closest to the light source point in the image;

step 2: respectively calculating the distances from the highlight point to four boundaries of the upper boundary, the lower boundary, the left boundary and the right boundary of the image, finding the maximum distance and recording the maximum distance as maxR;

and step 3: creating a zero floating point matrix dstMat with the size of width multiplied by height;

and 4, step 4: changing the value of the c column r row of the dstMat to the value of the highlight point;

and 5: from a high light point (r)₀,c₀) Expanding to realize floating point formation;

the specific steps of the step 5) are as follows:

step 5.1: defining the calculation formula of the dynamic moving average as shown in (1):

dma(pre,now)＝(1-α)*now+α*pre (1)

wherein, now and pre are floating point decimal values, now is the current floating point pixel value, pre is the floating point pixel value of a certain pixel around the current point, alpha is the diffuse reflection influence rate, and alpha is more than or equal to 0 and less than or equal to 1;

step 5.2: starting from the highlight point, the rectangle expands, the expansion radius d gradually expands from 1 to maxR by step size of 1, and coordinate points (rs, cs) and (re, ce) of the upper left corner and the lower right corner of the rectangle after each expansion are calculated according to the formulas (2) - (5):

step 5.3: during each expansion, the value of the boundary of the row-column expansion rectangle is calculated, and the specific formula is as follows:

dstMat_rs，cs＝dma(dstMat_rs+1，cs+1，mat_rs，cs) (6)

dstMat_rs,ce＝dma(dstMat_rs+1,ce-1,mat_rs,ce) (7)

dstMat_re,cs＝dma(dstMat_re-1,cs+1,mat_re,cs) (8)

dstMat_re,ce＝dma(dstMat_re-1,ce-1,mat_re,ce) (9)

dstMat_rr,cs＝dma(dstMat_rr,cs+1,mat_re,cs),rr∈(rs,re) (10)

dstMat_rr,ce＝dma(dstMat_rr,ce-1,mat_re,ce),rr∈(rs,re) (11)

dstMat_rs,cc＝dma(dstMat_rs+1,cc,mat_rs,cc),cc∈(cs,ce) (12)

dstMat_re,cc＝dma(dstMat_re-1,cc,mat_re,cc),cc∈(cs,ce) (13)

wherein, mat_r,cAnd dstMat_r,cAnd respectively representing the values of the r row and the c column of the matrix mat and the dstMat, wherein after the operation is finished, the dstMat is the finally obtained floating point matrix.

Images