CN105072433A - Depth perception mapping method applied to head track virtual reality system - Google Patents

Abstract

The invention discloses a depth perception mapping method applied to a head track virtual reality system. The depth perception mapping method comprises the steps of: 1, obtaining a three-dimensional picture of scene objects, judging that whether the head state of the three-dimensional picture of the scene objects is changed, and if no, shooting the three-dimensional picture of the scene objects by using a dual-camera with a camera axle spread which is same with the distance of two eyes of a user; and if yes, entering the next step; 2, adjusting the distance from the zero parallax projection plane of the dual-camera to the center position of the dual-camera to allow the distance from the center of the dual camera to the projection plane to be equal to the distance from a physical display screen to the center of the two eyes of the user; 3, determining a suitable depth perception region according to the screen size and a viewpoint position, performing scale transformation on the scene objects on two sides of the projection plane along the depth direction and extruding to the projection plane, so as to allow the scene depth range to be within the determined suitable depth perception region; and 4, performing skew transformation on the scene objects according to perspective projection geometry, and shooting the three-dimensional picture of the scene objects by the dual-camera.

Description

Be applied to the perceived depth mapping method that head follows the tracks of virtual reality system

Technical field

The present invention relates to field of computer multimedia, particularly relate to a kind of perceived depth mapping method being applied to head tracking virtual reality system.

Background technology

Stereo display is the image parallactic by simulating human binocular vision, makes beholder form the visually-perceptible of depth when watching image.In virtual reality system, stereo display effectively can strengthen the sense of reality and the property immersed.But in stereo display, due to the conflict that human eye is converge like the spokes of a wheel at the hub and focus on, also can cause visual fatigue.The best method that software solves visual fatigue reduces conflict that is converge like the spokes of a wheel at the hub and that focus on by the parallax of reduction binocular images.Parallax minimizing means that perceived depth also can reduce.Therefore when beholder is when watching stereo-picture, the perceived depth scope that suitable is had.The method be compressed to by perceived depth within suitable perceived depth scope is called depth map.

In virtual reality system, traditional depth-compression method is that the wheelbase by reducing eyes camera realizes depth map.But because depth map can cause perceptual object to produce distortion in both direction: one is that to produce crimp along view direction be that object seems to be tending towards flat; Two is that perceptual object is cut sth. askew using the center line of eyes camera as reference position distortion.In head tracking system, because the eyes camera taking stereo-picture moves along with the motion of head, the datum line of the distortion of cutting sth. askew of perceptual object also can move along with the motion of head, thus causing perceptual image can move along with the motion of head, the movement of this image is referred to as " image drift ".

Summary of the invention

In order to solve the image drift problem that traditional depth map method produces in head tracking system, the invention provides a kind of perceived depth mapping method being applied to head tracking virtual reality system.The method utilizes carrys out photographic images with actual eyes apart from identical camera axis distance, utilizes fixing datum of deformation line of cutting sth. askew to carry out depth-compression conversion to virtual scene.The method not only can eliminate the image drift problem in head tracking virtual reality system, can also realize linear processes depth map flexibly.

To achieve these goals, the present invention is by the following technical solutions:

Be applied to the perceived depth mapping method that head follows the tracks of virtual reality system, comprise:

Step one: the stereo-picture obtaining object scene, and judge whether the head state of the stereo-picture of object scene changes, if do not change, then adopt and the stereo-picture of user's eyes apart from the double camera photographed scene object of identical camera axis distance; If change, then enter next step;

Step 2: 0 parallax perspective plane of adjustment double camera, to the distance of double camera point midway, makes double camera mid point two oculocentric apart from identical to user with physical display screen to the distance on perspective plane;

Step 3: determine perceived depth Suitable Area according to screen size and viewpoint position, carries out scale transformation by both sides, perspective plane object scene along depth direction and extrudes to perspective plane, perceived depth Suitable Area scene depth scope being positioned at determine;

Step 4: according to perspective projection geometry, object scene is cut sth. askew conversion, utilize double camera to carry out the stereo-picture of photographed scene object.

In described step 4, also passing the projection on the axis at center, perspective plane perpendicular to perspective plane as the perspective reference point calculating conversion of cutting sth. askew using double camera mid point.

In described step 3, the method that both sides, perspective plane object scene carries out scale transformation along depth direction is comprised linear scale transformed mappings and non-linear zoom transformed mappings.

To have an X-rayed reference point for the origin of coordinates, with the perpendicular bisector on perspective plane for Z axis, take horizontal direction as X-axis, vertical direction is that Y-axis sets up coordinate system, and model vertices coordinate is from (x, y, z) be converted to after convergent-divergent and conversion of cutting sth. askew (x ', y ', z ');

Carry out mapping along the enterprising line linearity scale transformation of depth direction to model in this coordinate space, the scale transformation on depth direction is:

$z^{\′}=\frac{F-f^{\′}}{F-f}(z+F)-F$

Wherein, f be camera cutting identity distance far away from, f' be the distal border of perceived depth Suitable Area to camera distance, F is double camera mid point distance projection plane.

Object scene cutting sth. askew in the horizontal and vertical directions in both sides, perspective plane is transformed to:

$\begin{array}{cc}{x}^{\′}=\frac{{\mathrm{xz}}^{\′}}{z},& y^{\′}=\frac{{\mathrm{yz}}^{\′}}{z}\end{array}.$

When both sides, perspective plane object scene carries out non-linear zoom transformed mappings along depth direction, adopt different compression ratios in the different degree of depth.

When both sides, perspective plane object scene carries out non-linear zoom transformed mappings along depth direction, compression ratio reduces along with both sides, perspective plane object scene and reducing of perspective plane distance.

To have an X-rayed reference point for the origin of coordinates, with the perpendicular bisector on perspective plane for Z axis, take horizontal direction as X-axis, vertical direction is that Y-axis sets up coordinate system, and model vertices coordinate is from (x, y, z) be converted to after convergent-divergent and conversion of cutting sth. askew (x ', y ', z ');

Map along the enterprising line nonlinearity scale transformation of depth direction model in this coordinate space, the scale transformation on depth direction is:

$z^{\′}=\frac{(F-f^{\′})(F-n^{\′})(2F-f-n)(z+F)}{\left(\right(F-f^{\′}\left)\right(F-n)-(F-n^{\′}\left)\right(F-f\left)\right)\left|z+F\right|+(2F-f^{\′}-n^{\′})(F-f)(F-n)}-F$

Wherein, f be camera cutting identity distance far away from, n be nearly cutting identity distance from, f', n' are respectively the far and near border of perceived depth Suitable Area to camera distance, and F is the distance of double camera mid point to projection plane.

Object scene cutting sth. askew in the horizontal and vertical directions in both sides, perspective plane is transformed to:

$\begin{array}{cc}{x}^{\′}=\frac{{\mathrm{xz}}^{\′}}{z},& y^{\′}=\frac{{\mathrm{yz}}^{\′}}{z}\end{array}.$

Beneficial effect of the present invention:

(1) when the camera axis of actual beholder's eyes Distance geometry shooting stereo-picture is apart from time equal, perceptual object does not produce distortion, therefore, in head tracking system, employing of the present invention can removal of images drift phenomenon apart from taking stereo-picture apart from identical camera axis with the eyes of actual user.

(2) in order to realize depth-compression, virtual scene is carried out convergent-divergent and cuts sth. askew conversion to reduce perceived depth by the present invention, adopt simultaneously and take stereo-picture with user's eyes apart from identical camera axis distance, perceived depth can be reduced when removal of images drifts about like this, alleviate visual fatigue.

Accompanying drawing explanation

Fig. 1 is the perceived depth mapping method schematic flow sheet being applied to head tracking virtual reality system of the present invention;

Fig. 2 is the conversion schematic diagram of cutting sth. askew in model convergent-divergent in the depth direction and horizontal direction.

Embodiment

Below in conjunction with accompanying drawing and embodiment, the present invention will be further described:

As shown in Figure 1, the perceived depth mapping method being applied to head tracking virtual reality system of the present invention, comprising:

Step one: the stereo-picture obtaining object scene, and judge whether the head state of the stereo-picture of object scene changes, if do not change, then adopt and the stereo-picture of user's eyes apart from the double camera photographed scene object of identical camera axis distance; If change, then enter next step;

Step 2: 0 parallax perspective plane of adjustment double camera, to the distance of double camera point midway, makes double camera mid point two oculocentric apart from identical to user with physical display screen to the distance on perspective plane;

Step 3: determine perceived depth Suitable Area according to screen size and viewpoint position, carries out scale transformation by both sides, perspective plane object scene along depth direction and extrudes to perspective plane, perceived depth Suitable Area scene depth scope being positioned at determine;

Step 4: according to perspective projection geometry, object scene is cut sth. askew conversion, utilize double camera to carry out the stereo-picture of photographed scene object.

Further, in step 3, the method that both sides, perspective plane object scene carries out scale transformation along depth direction is comprised linear scale transformed mappings and non-linear zoom transformed mappings.

For the object be taken for the animal model shown in Fig. 2:

User eyes are adopted to carry out the image of this animal of stereo display apart from the right and left eyes image that the double camera of identical camera axis distance takes the animal shown in Fig. 2 respectively, the object that such perceptual object is namely taken does not produce distortion for the animal model shown in Fig. 2, can reach the object of removal of images drift phenomenon like this.Wherein, in the present embodiment by the detailed process of animal model linear scale transformed mappings and non-linear zoom transformed mappings be:

First, to have an X-rayed reference point for the origin of coordinates, with the perpendicular bisector on perspective plane for Z axis, take horizontal direction as X-axis, vertical direction is that Y-axis sets up coordinate system;

Then, the apex coordinate of this animal model is converted to after convergent-divergent and conversion of cutting sth. askew from (x, y, z) (x ', y ', z '); Also pass the subpoint on the axis at center, perspective plane perpendicular to perspective plane for perspective reference point with double camera mid point, and calculate the conversion of cutting sth. askew of object according to this reference point;

Carry out mapping along the enterprising line linearity scale transformation of depth direction to the animal model shown in Fig. 2 in this coordinate space, the scale transformation on depth direction is:

$z^{\′}=\frac{F-f^{\′}}{F-f}(z+F)-F$

Object scene cutting sth. askew in the horizontal and vertical directions in both sides, perspective plane is transformed to:

$\begin{array}{cc}{x}^{\′}=\frac{{\mathrm{xz}}^{\′}}{z},& y^{\′}=\frac{{\mathrm{yz}}^{\′}}{z}\end{array}$

Wherein, f be camera cutting identity distance far away from, f' be the distal border of perceived depth Suitable Area to camera distance, F is double camera mid point distance projection plane;

When both sides, perspective plane object scene carries out non-linear zoom transformed mappings along depth direction, different compression ratios is adopted to carry out non-linear zoom transformed mappings at different depths; When both sides, perspective plane object scene carries out non-linear zoom transformed mappings along depth direction, compression ratio reduces along with both sides, perspective plane object scene and reducing of perspective plane distance.

To have an X-rayed reference point for the origin of coordinates, with the perpendicular bisector on perspective plane for Z axis, take horizontal direction as X-axis, vertical direction is that Y-axis sets up coordinate system, and model vertices coordinate is from (x, y, z) be converted to after convergent-divergent and conversion of cutting sth. askew (x ', y ', z ');

Carry out mapping along the enterprising line nonlinearity scale transformation of depth direction to model in this coordinate space, the scale transformation on depth direction is:

$z^{\′}=\frac{(F-f^{\′})(F-n^{\′})(2F-f-n)(z+F)}{\left(\right(F-f^{\′}\left)\right(F-n)-(F-n^{\′}\left)\right(F-f\left)\right)\left|z+F\right|+(2F-f^{\′}-n^{\′})(F-f)(F-n)}-F$

Object scene cutting sth. askew in the horizontal and vertical directions in both sides, perspective plane is transformed to:

$\begin{array}{cc}{x}^{\′}=\frac{{\mathrm{xz}}^{\′}}{z},& y^{\′}=\frac{{\mathrm{yz}}^{\′}}{z}\end{array}$

Wherein, f be camera cutting identity distance far away from, n be nearly cutting identity distance from, f', n' be the far and near border of perceived depth Suitable Area to camera distance, F is the distance of double camera mid point to projection plane.

Realize depth-compression in the present embodiment, virtual scene carried out convergent-divergent and cut sth. askew conversion to reduce perceived depth, when removal of images drifts about, finally reaching minimizing perceived depth, alleviate the object of visual fatigue.

By reference to the accompanying drawings the specific embodiment of the present invention is described although above-mentioned; but not limiting the scope of the invention; one of ordinary skill in the art should be understood that; on the basis of technical scheme of the present invention, those skilled in the art do not need to pay various amendment or distortion that creative work can make still within protection scope of the present invention.

Claims (9)

1. be applied to the perceived depth mapping method that head follows the tracks of virtual reality system, it is characterized in that, comprising:

Step one: the stereo-picture obtaining object scene, and judge whether the head state of the stereo-picture of object scene changes, if do not change, then adopt and the stereo-picture of user's eyes apart from the double camera photographed scene object of identical camera axis distance; If change, then enter next step;

Step 2: 0 parallax perspective plane of adjustment double camera, to the distance of double camera point midway, makes double camera mid point two oculocentric apart from identical to user with physical display screen to the distance on perspective plane;

Step 3: determine perceived depth Suitable Area according to screen size and viewpoint position, carries out scale transformation by both sides, perspective plane object scene along depth direction and extrudes to perspective plane, perceived depth Suitable Area scene depth scope being positioned at determine;

Step 4: according to perspective projection geometry, object scene is cut sth. askew conversion, utilize double camera to carry out the stereo-picture of photographed scene object.

2. a kind of head that is applied to follows the tracks of the perceived depth mapping method of virtual reality system as claimed in claim 1, it is characterized in that, in described step 4, also passing the projection on the axis at center, perspective plane perpendicular to perspective plane as the perspective reference point calculating conversion of cutting sth. askew using double camera mid point.

3. a kind of head that is applied to follows the tracks of the perceived depth mapping method of virtual reality system as claimed in claim 1, it is characterized in that, in described step 3, the method that both sides, perspective plane object scene carries out scale transformation along depth direction is comprised linear scale transformed mappings and non-linear zoom transformed mappings.

4. a kind of head that is applied to follows the tracks of the perceived depth mapping method of virtual reality system as claimed in claim 3, it is characterized in that, to have an X-rayed reference point for the origin of coordinates, with the perpendicular bisector on perspective plane for Z axis, take horizontal direction as X-axis, vertical direction is that Y-axis sets up coordinate system, model vertices coordinate is converted to after convergent-divergent and conversion of cutting sth. askew from (x, y, z) (x ', y ', z ');

Carry out mapping along the enterprising line linearity scale transformation of depth direction to model in this coordinate space, the scale transformation on depth direction is:

$z^{\′}=\frac{F-f^{\′}}{F-f}(z+F)-F$

Wherein, f be camera cutting identity distance far away from, f' be the distal border of perceived depth Suitable Area to camera distance, F is double camera mid point distance projection plane.

5. a kind of head that is applied to follows the tracks of the perceived depth mapping method of virtual reality system as claimed in claim 4, it is characterized in that, object scene cutting sth. askew in the horizontal and vertical directions in both sides, perspective plane is transformed to:

$\begin{array}{cc}{x}^{\′}=\frac{{\mathrm{xz}}^{\′}}{z},& y^{\′}=\frac{{\mathrm{yz}}^{\′}}{z}\end{array}.$

6. a kind of head that is applied to follows the tracks of the perceived depth mapping method of virtual reality system as claimed in claim 3, it is characterized in that, when both sides, perspective plane object scene carries out non-linear zoom transformed mappings along depth direction, adopt different compression ratios in the different degree of depth.

7. a kind of head that is applied to follows the tracks of the perceived depth mapping method of virtual reality system as claimed in claim 6, it is characterized in that, when both sides, perspective plane object scene carries out non-linear zoom transformed mappings along depth direction, compression ratio reduces along with both sides, perspective plane object scene and reducing of perspective plane distance.

8. a kind of head that is applied to follows the tracks of the perceived depth mapping method of virtual reality system as claimed in claim 6, it is characterized in that, to have an X-rayed reference point for the origin of coordinates, with the perpendicular bisector on perspective plane for Z axis, take horizontal direction as X-axis, vertical direction is that Y-axis sets up coordinate system, model vertices coordinate is converted to after convergent-divergent and conversion of cutting sth. askew from (x, y, z) (x ', y ', z ');

Map along the enterprising line nonlinearity scale transformation of depth direction model in this coordinate space, the scale transformation on depth direction is:

$z^{\′}=\frac{(F-f^{\′})(F-n^{\′})(2F-f-n)(z+F)}{\left(\right(F-f^{\′}\left)\right(F-n)-(F-n^{\′}\left)\right(F-f\left)\right)\left|z+F\right|+(2F-f^{\′}-n^{\′})(F-f)(F-n)}-F$

Wherein, f be camera cutting identity distance far away from, n be nearly cutting identity distance from, f', n' are respectively the far and near border of perceived depth Suitable Area to camera distance, and F is the distance of double camera mid point to projection plane.

9. a kind of head that is applied to follows the tracks of the perceived depth mapping method of virtual reality system as claimed in claim 8, it is characterized in that, object scene cutting sth. askew in the horizontal and vertical directions in both sides, perspective plane is transformed to:

$\begin{array}{cc}{x}^{\′}=\frac{{\mathrm{xz}}^{\′}}{z},& y^{\′}=\frac{{\mathrm{yz}}^{\′}}{z}\end{array}.$