DE19831713C2 - 3D adaptive raster monitor (PARM) - Google Patents

Description

The invention relates to a video monitor for autostereoscopic display of 3D Images or scenes in which the viewer's position is adaptive to the stereoscopes View can be included. The 3D raster monitor described here is one Further development of the position adaptive autostereoscopic monitor (PAM), at which, in particular, makes it easy to install for PCs. Ver Different specific grid discs can be displayed on a TFT display (in the laptop or Monitor). By incorporating a special graphics card, through Installation of suitable 3D software and using a trade usual head trackers, a standard PC can be upgraded to a 3D system become. You can get the best 3D view for a desired distance position setting from the monitor. If the position is changed, the 3D view is without readjusted mechanical tracking and switches the stereoscopic view at Do not leave a definition area in front of the monitor, but reduce it then the resolution.

STATE OF THE ART

Stereoscopic film and projection processes have been in use for years Commitment. Polarized light (horizontal / vertical, circular) is mostly used to achieve this separate right and left image. With the advancement of LCD technology, it became possible to electronically control the light transmission of crystals. This did the development of the shutter technique possible, in synchronism with the field frequency, the right and left glasses become opaque and synchronously right and left images appear sequentially on the screen.

Autostereoscopic projections are made with the help of screens with striped lenses carried out with several projection directions. The corresponding one Direction assigned the correct perspective image [2]. A smooth transition from One perspective to the next can hardly be achieved because of the number of projection directions  cannot be increased arbitrarily. With an autostereoscope Display that is intended for only one person, you only use two perspectives, that require a certain line of sight [3], [5]. A fully stereoscopic image like in Representing a hologram is accomplished using "Head Trac king "sensors that control a powerful real-time computer Calculation of the appropriate stereoscopic image perspectives and on the other hand the autostereoscopic screen control for tracking the optical beam path [3], [6], [9]. The exact head position and movement is detected and the associated images are generated at the same time. Beyond that come up too agile VR systems (virtual reality) using those that take getting used to "Head Mounted Displays" are used.

Magnetic resonance and computed tomography are important in medical technology most fields of application for stereoscopic 3D visualizations. To certain searched Powerful special computers are used to calculate perspectives. Kom binary computer visualizations and real-time transmissions of endoscopes one of the most important neurosurgical tools. Stereoscopes are endoscopes already in use. An electronic movement control via an infrared ba sis working "Head Tracking Sensor" will be easy to combine with a personal autostereoscopic screen system.

PROBLEM

To create a stereoscopic representation that is as natural as possible an autostereoscopic method must be used. The viewer is allowed to should not be tied to a specific position in front of the screen, but should can specifically use the means of head movement to actively look out a 3D scene to be able to observe slightly changed viewing directions. This requires one adaptive system, which of course can only be oriented towards one viewer. In addition, such a system should also be used for a wider use of the screen offer the opportunity to switch to a non-stereoscopic display mode to switch.  

When adapting to the observer position by sub-pixel recoding A TFT display [4] has encountered the problem of uneven gaps between the combed right and left picture on the display Troublesome streak have occurred. This local quantization interference should be eliminated. So far, a fixed definition area for the Viewers required in front of the screen in which the 3-D view was possible. At the Leaving this area, the system switched to a two-dimensional view. The goal is to enlarge the domain and at a greater distance reduce a high resolution that is no longer required from the screen to lose the 3D view.

The aim is to open up the possibility of relatively high manufacturing costs large quantities can be drastically reduced thanks to a wide range of applications. It should also avoid having to use two superimposed displays, as is required, for example, in the shutter system [9].

So far, no such 3-D displays are known that solve this problem.

PRINCIPLE OF THE INVENTION

The position adaptive 3-D raster monitor (PARM) is one Further development of the person-adaptive car stereoscopic monitor PAM [3]. He is based on the same principle, but uses instead of one in front of the screen arranged prism grid a special lens grid, in which each Cylinder lens made from a suitable combination of convex-concave sub-cylin derlinsen is composed. Through the cylindrical lens grid are on the Display different, disjoint color sub-pixel areas for the right and left Eye visible. These visible color sub-pixel areas are controlled so that the right eye only sees the right picture and left eye only the left picture.

The control of the subpixels on the display is continuously adaptive to the adjusted the respective position of the viewer in front of the display. This also works 3D view is not lost when a person moves in front of the display. The position information  is supplied by a head tracker.

The problem is solved according to the invention as follows: the cylindrical lens on the grid disc or the cylinder prism on the grid disc additionally have a convex and concave sub-lens structure, so that the visible row of pixels for an image consists of 3 controlled pixels RGB, GBR or BGR by a concave structure is contracted, cf. Beam paths 5 in FIG. 3 and FIG. 4. If there is a display that has sufficient brightness, the convex part of the sub-cylinder lens with the beam path 6 has the focal point on the image plane ( 4 ). If no brightness is to be lost, the convex sub-cylinder lens has its focal point between the lens and the image plane, so that the pixel arrangement appears mirrored over the convex area, cf. Fig. 12. While in the beam path of the concave part 5 the visible area covers about 4 pixels in the order SP 1 , SP 2 , SP 3 , SP 4 , of which only 3 are controlled on the display, appear in the subsequent beam path the convex part ( 6 ) in the mirrored order SP 4 , SP 3 , SP 2 , SP 1 . This arrangement can be continued if necessary. In a subsequent concave part of a sub-cylindrical lens, the same 4 sub-pixels would then appear on the display in the original order, while in a subsequent convex part the mirrored sequence of the same sub-pixels would appear again. This ensures that a pixel from 3 subpixels appears 2 or 4 times in an entire lens. This causes a uniform distribution of the light across the entire lens width and subpixels that are not activated are no longer visible.

The alternating order of the intermeshed pixels on the display on the right, left, right, left, etc. can also be replaced by an order in which two successive right and then two consecutive left image pixels on the display are intermeshed , Fig. 5 shows an arrangement in which there are even 3 right and then 3 left picture pixels in a row. The entire area for the right image (see FIG. 4, FIG. 5) is approximately half as wide as the total lens 2 , so that the area adjacent to it for the left image also makes up about half the total lens width. The individual beam paths through the concave sub-cylindrical lenses ( 5 ) then again cover about 4 sub-pixels, 3 of which are controlled. The beam paths ( 6 ) through the convex sub-cylindrical lenses can, as in FIG. 5, have the focal point on the image plane, on a pixel area that is not controlled or - as already explained in FIG. 12 - the focal point in the middle between the lens and the image plane so that the same area appears in mirror image.

Fig. 2 shows an arrangement in which two right and two left image pixels are controlled on the display, which are then divided into the viewing directions for the right and left eye via the concave portions. In order to always be able to separate the quantizations between the individual pixels, an uncontrolled sub-pixel (na) is provided on average between every two right-hand image pixels, as well as between two left-hand ones, while between the right and left-hand pixel areas there is a larger area of approx controlled (na) is marked so that a clear separation between right and left is achieved.

The uncontrolled sub-pixel area between right and left on the display can adaptively decrease if the distance of the viewer from the display increases. From a certain distance limit, a right subpixel can then mix with a left subpixel. In this case, from this point on, the entire pixel in which an overlap takes place is no longer used, that is, it is left out in control, so that the image remains visible only through the other pixels. This is associated with a loss in the resolution quality of the image, but the stereoscopic view has not yet been lost. This is only possible if, as in Fig. 2 or 5, several right and several left pixels are next to each other.

In Figs. 6 to 11 μmetergenau calculated profile of convex and concave substructures are shown via a lens pitch width and explained in the captions. To clarify the profile, the lens depth is shown enlarged compared to the lens width.

The sub-pixel recoding is - similar to that in PAM. System [3], [4] - made by a digital adaptation encoder. In order to separate the three sub-pixels belonging to a pixel stripe from neighboring stripes in the same viewing direction, a sub-pixel in between is preferably set dark, cf. Fig. 2, (marked with na as not driven). In order to be able to adapt the pixel stripes on the display, which separate the right and left viewing directions, intermeshed more easily, an average of two to three subpixels are preferably not activated at this point (see FIG. 2).

position detection

Infrared head tracking systems that are already used today are suitable as position detectors have the required precision, but are also inexpensive ultrasound systems usable. Would you also like the vertical position and the viewing distance detect, it is advisable to use at least two sensor fields. It can but also CCD cameras with fast image evaluation can be used. ultra Sound measurements are particularly easy if the transmitter is at the head of the Viewer may be attached.

DESCRIPTION OF THE AREAS OF APPLICATION

An autostereoscopic screen has opposite other methods such as the shutter principle or polarization principle with glasses Advantage that the user does not need glasses. To move the stereo effect out To obtain the viewer's gaze is known in previously known systems the lens or prism lens in front of the image surface is mechanically adaptive Corresponding viewer position, updated. Next to it is an autostereoscopic shutter Known screen, but for which the required fast display technology is not yet fully developed at the moment.

The autostereoscopic use remains essentially on one person per screen limits, but what about workstations or in the doctor's operating room anyway  Case is. In addition, of course, additional glasses-bound procedures for a group use can be used.

The most important short-term applications can be seen in medical technology. Here, the reference to a person is not a disadvantage; only judges one anyway Person a 3D CT image: the doctor, the annoying glasses and restricted fields of vision must avoid. If several doctors are working at the same time, several screens can be used to be used. At the same time, an auditorium wants a microsurgical operation Follow stereoscope, this can be done using a projection method with e.g. B. polar light.

In a future digital television system, stereoscopic television will also be used have its place as it is simply closest to natural viewing habits comes and can be used optionally if the digital coding at the Über the additional information for the third dimension.

The TFT displays available today on PCs, laptops and workstations allow one quick changeover to a 3D capability. 3D dimensions available on computers Software can then be used to enable a 3D view if the above-mentioned. Additional equipment is integrated.  

literature

[1] S. Hentschke: Stereoscopic screen. Patent application DE 41 14 023 A1 (

1991

).
[2] R. Börner: Autostereoscopic 3-D imaging by front and rear projection and on flat panel displays. Displays, Vol. 14, No. 1 (

1993

), pp. 39-46.
[3] S. Hentschke: Position-adaptive autostereoscopic monitor (PAM). European patent application EP 0 836 332 A2, (

1997

).
[4] M. Andiel, S. Hentschke, A. Herrfeld, N. Reifschneider: 3D panorama monitor. CeBIT 98 brochure: Cooperation partner in research and innovation. Ministry of Education and Art of Hes (

1998

).
[5] R. Börner: Autostereoskope rear projection and flat screens. Television and Cinema Technology, Vol. 48, No. 11 (

1994

). Pp. 594-600.
[6] S. Hentschke: Person-adaptive autostereoscopic monitor - an option for the television? TV and Cinema Journal No. 5/1996, pp. 242-248
[9] R. Börner: playback device for three-dimensional perception of images. Autostereoscopic Viewing Device for Creating Three Dimensional Perception of Images. German Patent No. DE 39 21 061 A1 (note 1989).
[9] S. Hentschke, person-adaptive autostereoscopic shutter screen (PAAS). Patent specification DE 195 00 315 C1, (

1995

).

Claims (9)

1. Position-adaptive 3D raster monitor (PARM), consisting of a lenticular screen ( 1 ), an RGB flat display ( 3 ), in which the three color subpixels R, G and B are arranged horizontally next to each other, and a subpixel encoder in the digital display control, characterized in that each cylindrical lens ( 2 ) or each cylindrical prism on the grid disc ( 1 ) has a fine structure of two or more consecutive convex and concave sub-cylindrical lenses. 2. PARM monitor according to claim 1, characterized in that the substructure on a cylindrical lens consists of successive sequences of a convex, concave and convex cylindrical lens and the beam path ( 5 ) through the central concave lens on the image plane ( 4 ) covers a visible area, which is about half as wide as the horizontal pitch distance p _ZL on the grid disc (see FIG. 3), while the convex sub-cylindrical lenses have their focal point on the image plane ( 3 ). 3. PARM monitor according to claim 1, characterized in that a sequence of a convex, concave, convex sub-cylindrical lens is arranged on each of the two outer sides of a cylinder prism, so that the beam paths perpendicular to the image surface ( 5 ) through the two concave lenses on the Cover the same visible horizontal area, which is about half the width of the pitch of the cylinder prisms on the screen ( 1 ), and that this area of coverage ( 4 ) includes at least 4 adjacent color subpixels (RGBR) or (GBRG) or (BRGB) ( see Fig. 4), while the convex sub-cylindrical lenses have their focal point on the image plane ( 3 ). 4. PARM monitor according to claim 1, characterized in that each cylindrical lens on the grid disc 2 or 3 juxtaposed combinations of three convex, concave, convex sub-cylinder lenses, in which the vertical to the grid beam striations ( 5 ) through the concave Sub-cylindrical lenses 2 or 3 cover disjoint areas lying next to one another, which together are approximately as wide as half the pitch distance of the large cylindrical lenses on the grid disc and that in each beam path ( 5 ) by a concave sub-cylindrical lens at least three color sub-pixels (RGB) or (GBR) or (BRG) lie on the image plane ( 3 ) (cf. FIG. 5). 5. PARM monitor according to claim 1, characterized in that 2 or 3 or more sub-cylinder lens combinations, as they are assembled in claim 3 in a cylindrical prism, are arranged horizontally one after the other and that the beam paths perpendicular to the grid disc ( 5 ) on the The image plane covers a sub-pixel area that is composed next to one another, but which is disjoint from one cylindrical prism to the other, and that at least 3 color sub-pixels lie in each of these beam paths ( 5 ) through the concave sub-cylindrical lenses (cf. FIG. 2). 6. PARM monitor according to claim 1, characterized in that each cylindrical lens is divided into 4 equal-sized convex or concave partial lenses, such that viewed from right to left on the lens surface for one eye in the first partial lens, the region of 3 or more successive subpixels appears, in the second partial lens the same area appears in the same order for the same eye, e.g. B. B1re, R1re, G1re; and in the next two partial lenses the area of the subpixels that follows on the display, e.g. B. R2re, G2re, B2re, appears repeatedly in the same way for the same eye, while for the other eye the following contiguous disjoint sub-pixel area appears in the same way, e.g. B. R1li, G1li, B1li, etc. cf. Fig. 2. Instead of 4 sub-cylindrical lenses, one cylindrical lens can also contain 6, such that a coherent sub-pixel area is visible in the same way for one eye, while a subsequent coherent disjunct area is visible in the same way for the other left eye. 7. PARM monitor according to claim 1, characterized in that a beam path perpendicular to the cylindrical lenticular disk ( 1 ) through each concave sub-cylindrical lens (which in extreme cases can have a distant focal point) has an area of at least 4 consecutive color sub-pixels SP 1 , SP 2 , SP 3 , SP 4 evenly covered from left to right, while the beam path through each subsequent convex sub-cylindrical lens covers the same area mirrored from right to left (SP 4 , SP 3 , SP 2 , SP 1 ), cf. , Fig. 12. 8. PARM monitor, according to claim 7, characterized in that the convex-concave fine structure consists of several groups of convex and concave sub-cylindrical lenses, in which the beam path through a group the area of at least 4 adjacent color subpixels SP 1- SP 4 covered from right to left and left to right and the next group covered a subsequent area of SP 5-SP 8 from right to left and left to right. 9. PARM monitor according to one of claims 4, 5, 6 or 8, characterized in that at not optimal viewing distance those over the concave sub-cylindrical lenses visible pixel areas that are simultaneously for the right and left eye on the screen are visible, are not controlled and that, if applicable, the pixel values that are no longer applicable to the neighboring right or left.