CN105898274B - 2D deepening 3D image longitudinal storage method based on RGB compression - Google Patents

Abstract

The invention discloses a longitudinal storage method of 2D plus depth 3D images based on RGB compression. Divide the image with a resolution of A Within, the top‑half Color is placed in the A x 1/3B area at the top, and the bottom‑half Color is placed in the A x 1/3B area in the middle; the depth map is compressed vertically to 2/3 and then compressed and placed through the RGB channel In the lower A Within the 1/2A x 1/3B area; the 2D depth image is converted into a multi-view view through a conversion circuit or software. As a result, the 3D display quality can be greatly increased without any other overhead.

Description

A longitudinal storage method of 2D plus depth 3D image based on RGB compression

Technical field:

The invention belongs to the technical field of image processing, and in particular relates to a longitudinal storage method for 2D-enhanced 3D images based on RGB compression.

Background technique:

At present, the format of 3D display source files based on 2D plus depth is mostly 2D and depth image formats with half on the left and half on the left. As shown in Figure 1, if the resolution of the target display screen is AxB, and the image to be displayed is a 2D AxB and an AxB depth map, the display area is first divided into two symmetrical parts, 0.5AxB on the left and right sides. Compress the 2D color image by half horizontally and place it in a 0.5AxB area on the left, and then compress the depth image by half horizontally and place it in a 0.5AxB area on the right. The 2D depth-enhanced image is converted into a multi-view view through a proprietary circuit or software.

The disadvantage of this format is that since the 2D image and the depth image each account for half, the information of the 2D image is compressed by half in the horizontal direction, while the depth image on the right is only a grayscale image, occupying three channels of RGB, which appears unbalanced. Also wastes space.

The information disclosed in this Background section is only for enhancing the understanding of the general background of the present invention and should not be taken as an acknowledgment or any form of suggestion that the information constitutes the prior art that is already known to those skilled in the art.

Invention content:

The object of the present invention is to provide a longitudinal storage method for 2D plus depth 3D images based on RGB compression, so as to overcome the above-mentioned defects in the prior art.

To achieve the above object, the present invention provides

A 2D plus depth 3D image longitudinal storage method based on RGB compression, the steps of which are:

(1) Divide the picture with a resolution of AxB into three equal parts vertically, and then divide the lower part into two equal parts horizontally;

(2) Compress the 2D color image vertically to 2/3, and then place it in the Ax2/3B area, where the top-half Color is placed in the top Ax1/3B area, and the bottom-half Color is placed in the middle Ax1/3B area Inside;

(3) Compress the depth map to 2/3 vertically and place it in the lower Ax1/3B area through RGB channel compression, where the depth map corresponding to the top-half Color is placed in the 1/2Ax1/3B area on the lower side, The depth map corresponding to bottom-half Color is placed in the 1/2Ax1/3B area on the other side of the lower part;

(4) Convert the 2D depth-enhanced image into a multi-view view through a conversion circuit or software.

Preferably, in the above technical solution, the RGB channel compression method converts the RGB format into the YCbCr format, and adopts the method of placing two depth points for each natural pixel point; each natural pixel point has three channels, and the two depth points take up Two channels, the remaining one channel is used for CbCr correction.

Preferably, in the above technical solution, the YCbCr format is 420 format, for each 2x2 natural pixel in space The brightness of each point is extracted and compressed separately, and CbCr is shared by the four 2x2 points.

Preferably, in the above technical solution, R1 and B1 place the depth information of the positions D(0,0) and D(1,0), and R2 and G2 place the depth information of the positions D(2,0) and D(3,0) Information; G1 places the same content as R1, that is, D(0, 0) position depth information; B2 places the same content as R2, that is, D(2, 0) position depth information; place other positions according to this rule depth information.

Preferably, in the above technical solution, when the YCbCr format is restored to an RGB image, CbCr is corrected, and Dcb and Dcr are the correction amounts of Cb and Cr, then:

R1＝1.164*(Y1-16)+1.596*(Cr-128)

G1＝1.164*(Y1-16)-0.813*(Cr-128)-0.392*(Cb-128)

B1＝1.164*(Y1-16)+2.017*(Cb-128)

R2＝1.164*(Y2-16)+1.596*(Cr-128)

G2＝1.164*(Y2-16)-0.813*(Cr-128)-0.392*(Cb-128)

B2＝1.164*(Y2-16)+2.017*(Cb-128)

Since R1=G1, the formula for adding the correction amount is:

1.596*(Cr+Dcr-128)＝-0.813*(Cr+Dcr-128)-0.392*(Cb+Dcb-128) (1)

Since R2=B2, the formula for adding the correction amount is:

1.596*(Cr+Dcr-128)＝2.017*(Cb+Dcb-128) (2)

Solve (1) and (2) to get the values of Dcr and Dcb; then put the values of Dcr and Dcb into the YCbCr to RGB formula to obtain the corrected R', G', B' depth values:

R'＝1.164*(Y-16)+1.596*(Cr+Dcr-128)

G'＝1.164*(Y-16)-0.813*(Cr+Dcr-128)-0.392*(Cb+Dcb-128)

B'=1.164*(Y-16)+2.017*(Cb+Dcb-128).

Compared with the prior art, the present invention has the following beneficial effects:

In the case that the size of the final display screen remains unchanged, the amount of information in the original 2D color image part is increased by 50% compared with the original method. The human eye is more sensitive to the horizontal resolution and less sensitive to the reduction of the vertical resolution. This method does not reduce the horizontal resolution. Resolution, so the definition will be higher than the horizontal arrangement. Due to the multiplexing of RGB channels, the amount of information in the original depth map is also increased by 50% compared with the original storage method. Thus, the 3D display quality can be greatly increased without other overhead.

Description of drawings:

Fig. 1 is a schematic diagram of the prior art;

Figure 2 is a schematic diagram of YCbCr in 420 format;

Fig. 3 is a schematic diagram of the longitudinal storage method of 2D plus depth 3D images based on RGB compression in the present invention;

Fig. 4 is a detailed schematic diagram of the longitudinal storage method of 2D plus depth 3D images based on RGB compression in the present invention;

FIG. 5 is a schematic diagram of a pixel storage area where a depth map area is placed.

detailed description:

Specific embodiments of the present invention are described in detail below, but it should be understood that the protection scope of the present invention is not limited by the specific embodiments.

Unless expressly stated otherwise, throughout the specification and claims, the term "comprise" or variations thereof such as "includes" or "includes" and the like will be understood to include the stated elements or constituents, and not Other elements or other components are not excluded.

As shown in Figure 3-4, a 2D plus depth 3D image longitudinal storage method based on RGB compression, the steps are:

(1) Divide the screen with a resolution of 1920╳1080 into three equal parts vertically, and then divide the lower part into two equal parts horizontally;

(2) Compress the 2D color image vertically to 2/3, and then place it in the 1920╳720 area, where the top-half Color is placed in the top 1920╳360 area, and the bottom-half Color is placed in the middle 1920╳360 area Inside;

(3) The depth map is compressed vertically to 2/3 and placed in the lower 1920╳360 area after RGB channel compression. The depth map corresponding to the top-half color is placed in the 960╳360 area on the lower side, and the bottom- The depth map corresponding to halfColor is placed in the 960╳360 area on the other side of the lower part;

(4) Convert the 2D depth-enhanced image into a multi-view view through a conversion circuit or software.

As shown in Figure 2, the RGB format is converted to the YCbCr format, and two depth points are placed for each natural pixel point; each natural pixel point has three channels, and two depth points occupy two channels, leaving one The channel is used for CbCr correction.

YCbCr format is 420 format, for each 2x2 natural pixel in space The brightness of each point is extracted and compressed separately, and CbCr is shared by the four 2x2 points.

As shown in Figure 5, R1 and B1 place the depth information of positions D(0,0) and D(1,0), and R2 and G2 place the depth information of positions D(2,0) and D(3,0); G1 places the same content as R1, that is, D(0, 0) position depth information; B2 places the same content as R2, that is, D(2, 0) position depth information; place the depth of other positions according to this rule information.

Preferably, in the above technical solution, when the YCbCr format is restored to an RGB image, CbCr is corrected, and Dcb and Dcr are the correction amounts of Cb and Cr, then:

R1＝1.164*(Y1-16)+1.596*(Cr-128)

G1＝1.164*(Y1-16)-0.813*(Cr-128)-0.392*(Cb-128)

B1＝1.164*(Y1-16)+2.017*(Cb-128)

R2＝1.164*(Y2-16)+1.596*(Cr-128)

G2＝1.164*(Y2-16)-0.813*(Cr-128)-0.392*(Cb-128)

B2＝1.164*(Y2-16)+2.017*(Cb-128)

Since R1=G1, the formula for adding the correction amount is:

1.596*(Cr+Dcr-128)＝-0.813*(Cr+Dcr-128)-0.392*(Cb+Dcb-128) (1)

Since R2=B2, the formula for adding the correction amount is:

1.596*(Cr+Dcr-128)＝2.017*(Cb+Dcb-128) (2)

Solve (1) and (2) to get the values of Dcr and Dcb; then put the values of Dcr and Dcb into the YCbCr to RGB formula to obtain the corrected R', G', B' depth values:

R'＝1.164*(Y-16)+1.596*(Cr+Dcr-128)

G'＝1.164*(Y-16)-0.813*(Cr+Dcr-128)-0.392*(Cb+Dcb-128)

B'=1.164*(Y-16)+2.017*(Cb+Dcb-128).

The foregoing descriptions of specific exemplary embodiments of the present invention have been presented for purposes of illustration and description. These descriptions are not intended to limit the invention to the precise form disclosed, and obviously many modifications and variations are possible in light of the above teaching. The exemplary embodiments were chosen and described in order to explain the specific principles of the invention and its practical application, thereby enabling others skilled in the art to make and use various exemplary embodiments of the invention, as well as various Choose and change. It is intended that the scope of the invention be defined by the claims and their equivalents.

Claims (4)

1. A 2D plus depth 3D image longitudinal storage method based on RGB compression, its steps are: (1) Divide the picture with a resolution of A×B into three equal parts vertically, and then divide the lower part into two equal parts horizontally; (2) Compress the 2D color map vertically to 2/3, and then put it into the A×2/3B area, where the upper half of the color map is placed in the top A×1/3B area, and the lower half of the color map is placed in In the A×1/3B area in the middle; (3) Compress the depth map to 2/3 vertically and place it in the lower A×1/3B area through RGB channel compression, where the depth map corresponding to the upper half of the color map is placed in 1/2A×1 on the lower side In the /3B area, the depth map corresponding to the lower half of the color map is placed in the 1/2A×1/3B area on the other side of the lower part; the RGB channel compression method converts the RGB format into the YCbCr format, using each natural The method of placing two depth points on a pixel point; each natural pixel point has three channels, two depth points occupy two channels, and the remaining channel is used for CbCr correction; (4) Convert the 2D depth-enhanced image into multiple views through a conversion circuit or software. 2. The method according to claim 1, characterized in that: the YCbCr format is a 420 format, for each 2×2 natural pixel in space The brightness of each point is extracted and compressed separately, and CbCr is shared by the four points of 2×2. 3. The method according to claim 2, characterized in that: R1 and B1 place the depth information of D(0,0) and D(1,0), R2 and G2 place D(2,0) and D( 3,0) position depth information; G1 places the same content as R1, that is, D(0,0) position depth information; B2 places the same content as R2, that is, D(2,0) position depth information; This rule places the depth information of the remaining positions. 4. The method according to claim 3, characterized in that: when the YCbCr format is restored to an RGB image, CbCr is corrected, and Dcb and Dcr are the correction amounts of Cb and Cr, then: R1＝1.164*(Y1-16)+1.596*(Cr-128) G1＝1.164*(Y1-16)-0.813*(Cr-128)-0.392*(Cb-128) B1＝1.164*(Y1-16)+2.017*(Cb-128) R2＝1.164*(Y2-16)+1.596*(Cr-128) G2＝1.164*(Y1-16)-0.813*(Cr-128)-0.392*(Cb-128) B2＝1.164*(Y1-16)+2.017*(Cb-128) Since R1=G1, the formula for adding correction amount is: 1.596*(Cr+Dcr-128)＝-0.813*(Cr+Dcr-128)-0.392*(Cb+Dcb-128) (1) Since R2=B2, the formula for adding the correction amount is: 1.596*(Cr+Dcr-128)＝2.017*(Cb+Dcb-128) (2) Solve (1) and (2) to get the values of Dcr and Dcb; then substitute the values of Dcr and Dcb into the YCbCr to RGB formula to obtain the corrected R’, G’, B’ depth values: R'＝1.164*(Y1-16)+1.596*(Cr+Dcr-128) G'＝1.164*(Y1-16)-0.813*(Cr+Dcb-128)-0.392*(Cb+Dcb-128) B'=1.164*(Y1-16)+1.596*(Cr+Dcb-128).