TW201428684A - Compensation method for image emulation - Google Patents

Abstract

The present invention provides a compensation method for image emulation, which comprises comparing a captured real-time image with a planar image to determine the portion of the real-time image different from the planar image as a main body image; then, selecting a plurality of reference points with x-axis and y-axis positions according to the light source reflection information of the main body image; calculating the ideal z-axis position and the actual projection length for each reference point to derive the compensation value relative to the reference points; extending the main body image to a corresponding position according to the z-axis position of each reference point after compensation, so as to simulate and generate the main body image with curved surfaces; and finally, interweaving the surface-curved main body image and the planar image to obtain an emulated image.

Description

Image simulation correction method

The invention relates to a 3D image processing method, and in particular to a method for image simulation correction.

With the advancement of technology in recent years, the image display technology has also made a significant breakthrough. Therefore, the display image of the 2D plane can not meet the requirements of people's visual senses, thus creating a 3-dimensional stereoscopic display. Born.

The imaging principle of 3D images mainly uses the left and right eyes to see images of different parallaxes respectively, so as to create a simulated visual experience with a sense of distance. At present, the production of 3D images mainly uses two lens modules to obtain two images with different viewing angles, and then synthesized to form a 3D image with stereoscopic perception. However, the use of two lens modules to obtain images of different parallaxes into 3D images not only requires a lot of time for front-end operations, but also uses multiple lens modules, which means that the cost is high, and in order to avoid the synthesized 3D images. In the case of color distortion, it is necessary to use a lens module with high pixels, which leads to a significant increase in cost.

Furthermore, the 3D image produced by the above method has a large file size of 3D images, so there is still a problem of occupying memory space.

In view of this, how to provide a method for correcting image simulation only at low cost is the primary subject of the present invention, so the inventor of the present invention finally has the invention after continually thinking and trialing. Produced.

The main object of the present invention is to provide a method for image simulation correction, which has the effects of low-cost simulation of a planar image and a small image file capacity after synthesis.

To achieve the foregoing objective, the present invention provides a method for image simulation correction, comprising the following steps:
(a) storing a flat image;
(b) capturing a real-time image having the planar image by using a photographing device, comparing different portions of the real-time image according to the planar image, and determining the different portion as a main image;
(c) selecting a plurality of reference points having x and y axial positions according to the light source reflection information of the main image, and calculating an ideal z-axis position of the reference points, and an actual projection length of the reference points, according to The actual projection length of each reference point is obtained as a correction value relative to the reference point, and the ideal z-axis position of each reference point is corrected by the correction value, so that the z-axis position of the reference points after the correction is extended. To the corresponding position, to simulate the generation of the curved body image;
(d) After the surface image of the curved surface is combined with the planar image (Interweaving), a simulated image can be obtained.

Preferably, in step (b), the comparison mode is based on the grayscale difference between the planar image and the real-time image pixel, and when the grayscale difference of the pixel is greater than a predetermined value, the drawing is judged. It is the pixel of the subject image, and is used to obtain the subject image.

Preferably, in step (c), based on the light source reflection information of the subject image, it is determined that those portions of the subject image should belong to a concave surface, and those portions should belong to a convex surface, thereby determining that the subject image is in a real environment. The stereoscopic state, and thereby selecting a plurality of reference points having x, y axial positions.

Further, in the step (c), the corrected z-axis position is further used to calculate a correction value including a positive-angle correction value and a rotation angle correction value by a trigonometric function formula. And the contemporary correction value including the correction value of the orientation angle and the correction value of the rotation angle is used, and then the limit is obtained by the Riemann element imminent law to obtain a correction value which is closer to the true, and the correction value can be used for the reference points. The light source and chromatic aberration are corrected.

The above and other objects and advantages of the present invention will be readily understood from

(no)

Figure 1 is a block diagram of the steps of the present invention.

Figure 2 is a schematic diagram of calculating the actual projection length of the present invention.

Referring to FIG. 1 , a method for image simulation correction provided by the present invention includes the following steps:

(a) First, a planar image is created in a memory, which can be previously captured by a photographic device and stored in the memory.

(b) using the photographing device to capture an instant image having the planar image, and comparing different portions of the instant image according to the planar image, and then determining the different portion as a main image and then taking out the image. The comparison method is based on the grayscale difference between the planar image and the real-time image pixel. When the grayscale difference of the pixel is greater than a predetermined value, the pixel is determined to be the pixel of the subject image, and thereby To get the subject image.

(c) determining, according to the light source reflection information of the subject image, that the portions of the subject image should belong to a concave surface, and those portions should belong to a convex surface, that is, the difference between the reflected information of the light source to determine the subject image in the real environment a stereoscopic state in which a plurality of reference points having x and y axial positions are selected;

Then calculating the ideal z-axis position of each reference point, and the actual projection length of each reference point, and obtaining a correction value relative to the reference point according to the actual projection length of each reference point, and then using the correction value for each reference The ideal z-axis position of the point is corrected, and the z-axis position of the reference points after the correction is extended to the corresponding position to simulate the generation of the curved subject image.

The steps of calculating the ideal z-axis position and the actual projection length of each reference point are as follows:

For example, the Cartesian coordinate equation for a sphere is x ² + y ² = z ² .

Where γ is the radius of the sphere, so x = γcos θ; y = γsin θ; z = C1 (constant)

Bring it into the formula x ² +y ² =z ² to get:

(γcosθ) ² +(γsinθ) ² =z ²

↓

γ ² cos ² θ+γ ² sin ² θ=z ²

↓

γ ² (cos ² θ+sin ² θ)=z ²

Also, cos ² θ+sin ² θ=1

Therefore, γ ² (1)=z ² → γ ² =z ² → γ=z, and thereby the ideal z-axis bit length can be calculated from the known x and y axial positions of each reference point, and z The length of the axial position is equal to the radius γ of the sphere.

When the length of the z-axis position is known, the length of the z-axis bit can be further used to generate a correction value including a positive angle correction value and a rotation angle correction value.

As shown in Fig. 2, according to the trigonometric function formula, z = ρsin 可, and the above formula has obtained γ = z, so γ = ρsin ψ, when sin ψ = 90 ̊, γ = ρ, and ρ is the actual projection length. .

And further consider the difference in the inclination of the main image, so add a variable value ι, so that the right angle coordinate equation of the ball is:

x ² +ιy ² =C1

From the above calculation formula, we can see that x=γcosθ, y=γsinθ, and substituting γ=ρsinψ into the Cartesian coordinate equation of x ² +ιy ² =C1, that is:

(ρsinψcosθ) ² +ι(ρsinψsinθ) ² =C1

↓

ρ ² sin ² ψcos ² θ+ιρ ² sin ² ψsin ² θ=C1

Where ι is the complex variable value, ψ is the correction value of the orientation angle, θ is the rotation angle correction value, and ρsinψ is the actual projection length. And by this, the simulation corresponding to the simulation can be performed according to the degree of freedom of the main action trajectory.

Then take the Riemann element approaching the law to take the limit and get closer to the true correction value:

Where C2, C3, C4, and C5 are each a constant. Thereby, the correction value of the relative reference point can be obtained according to the actual projection length of each reference point, and the ideal z-axis position of each reference point is corrected by the obtained correction value. In detail, the subject image of the present invention extends to the corresponding position according to the corrected z-axis position of each reference point to simulate generating a curved subject image, and then the light source can be used for the curved subject image. And the corrective action of the chromatic aberration, and by the point-to-point simulation and the face-to-face correction, the softness of the curved surface of the subject image is greatly improved, and the subject image after the surface is more natural.

(d) Finally, the surface image of the curved surface and the planar image are synthesized (Interweaving) by a processing unit, and then the simulated image of the planar 2D+ simulated surface 3D is obtained.

The actual operation of the present invention is as shown in Annex 1. The hand shown in the attached photo is in a moving state, so that the information can be reflected according to the light source to determine the concave/convex surface, and thereby select a plurality of x. , the reference point of the y-axis position, for the subsequent simulation surface modification operation, and the correction action of the light source and the chromatic aberration; and the second part is used to display the correction light source and the chromatic aberration correction of the present invention after performing the simulation surface modification operation The front and back images are compared to the photos, and in this actual test, the corrective action of the light source and the chromatic aberration is performed by means of enhanced image contrast, and the corrected photo of the attached image can be clearly seen, the corrected image The red color in the image is more vivid, and the details of the hand are more obvious and real, and the image quality of the high image quality can be achieved by this.

The image simulation correction method provided by the invention can simulate the 2D planar image to generate the simulated 3D image by using only one photographic device, which can effectively reduce the equipment cost and reduce the human load of the user. The invention uses the calculation of the correction values of the reference points, and can perform simulation correction on the light source, the chromatic aberration and the curved surface of the curved subject image, and the point-to-point simulation and the face-to-face correction can greatly enhance the subject image. The softness and realism of the curved surface, and thus the invention, can achieve high image quality only by using a generally low cost lens module. Furthermore, since the present invention is only for the simulation correction operation of the surface image different from the planar image in the real-time image, in addition to effectively reducing the processing time, the effect of greatly reducing the file capacity of the synthesized image can be achieved.

The disclosure of the above embodiments is intended to be illustrative of the invention and is not intended to limit the invention.

In summary, it will be apparent to those skilled in the art that the present invention can achieve the foregoing objectives and is in accordance with the provisions of the Patent Law. Therefore, the inventor of the present invention filed an application according to law.

Claims (6)

A method for image simulation correction includes the following steps:
(a) storing a flat image;
(b) capturing a real-time image having the planar image by using a photographing device, comparing different portions of the real-time image according to the planar image, and determining the different portion as a main image;
(c) selecting a plurality of reference points having x and y axial positions according to the light source reflection information of the main image, and calculating an ideal z-axis position of the reference points, and an actual projection length of the reference points, according to The actual projection length of each reference point is obtained as a correction value relative to the reference point, and the ideal z-axis position of each reference point is corrected by the correction value, so that the z-axis position of the reference points after the correction is extended. To the corresponding position, to simulate the generation of the curved body image;
(d) After the surface image of the curved surface is combined with the planar image (Interweaving), a simulated image can be obtained. According to the method of image simulation correction according to item 1 of the patent application scope, in step (b), the comparison mode is based on the gray scale difference between the plane image and the real-time image pixel, and the gray scale difference of the pixel is When the value is greater than a predetermined value, the pixel is determined to be a pixel of the subject image, and thereby the subject image is obtained. According to the image simulation correction method described in claim 1 of the patent application, in step (c), based on the light source reflection information of the subject image, it is determined that those portions of the subject image should belong to a concave surface, and those portions should belong to The convex surface is used to determine the stereoscopic state of the subject image in the real environment, and thereby select a plurality of reference points having x and y axial positions. According to the method of image simulation correction according to item 1 of the patent application scope, in step (c), the calculated z-axis position is further substituted by the trigonometric function formula to include a positive angle correction value and a rotation angle. Correction value of the correction value. According to the method of image simulation correction according to item 4 of the patent application scope, wherein the correction value of the correction value of the directional angle and the correction value of the rotation angle are included in the contemporary, and then the limit is obtained by the Riemann element imminent law. Approaching the true correction value. The image simulation correction method according to Item 5 of the patent application scope, wherein the correction value is used to correct the chromatic aberration of the light source of each reference point.