DE3035213A1 - Flight simulator with picture development device - incorporates digital picture data storage and transformation system, enabling stereo picture formation, with zoom facility - Google Patents

Abstract

The procedure relates to the capture and replay of land pictures for flight simulators, especially for training in aircraft pursuit and shooting. The information is obtained in the form of a projected picture of the scene, using digital data transformation using interpolated data to fill in between two adjacent pictures. Brightness values are coded and stored, and reproduces to form a stereo picture. A picture contrast facility is included, to allow for the effect of darkness, fog and low lying clouds. A two level chroma key operation method is used. The plane of the simulated picture is normal to the sight line and there is a device to determine the magnification required, but it is kept constant during the pursuit procedure.

Description

Process and device for the recovery and recovery

giving of terrain images for vision simulators The invention relates to a method and a device for the acquisition and reproduction of terrain images for vision simulators, preferably for a flight simulator for flight and shooting training of aircraft pilots, with real aerial photographs as a database and storage of the Image signals on video disks with individual addressing of the images.

Visual simulators of the aforementioned type and methods of extraction the images required for them are known from the prior art. From the US PS 2 385 291 is known a flight simulator with which for the first time a realistic Simulation of an overflown area from the pilot's point of view became possible. For this purpose, an image previously recorded in flight was projected, which responded moved along the flight path.

Another well-known method of flight simulators works with one Terrain model with TV camera moving over it. The picture from the TV camera will be depicted on a viewing device and should correspond to the view wiliikel as it is offers a pilot belm flight over the actual terrain.

This method, which is very common at the moment, has a number of decisive factors Disadvantages. So is z. B. the maneuvering space for an overflying aircraft to about 5 x 10 km limited. With these model sizes, however, there are already problems with the Tracking speed the recording camera. Moving targets are very difficult to implement and have not yet been used. The model has a large space requirement with high energy consumption, which is for the intensive Illumination of the model is necessary. The recording camera can only be used by means of elaborate Mechanics are tracked.

Flight simulators with a synthetic image structure are also known become, e.g. B. US-PS 3 911 597 and DE-AS 25 58 501. With such computer-controlled Systems an in principle unlimited maneuvering range is possible, however there are real limits due to the effort required to create models and the storage capacity of the computer used. In order to still get the model image in real time with existing systems To be able to create from individual surfaces, the level of detail of the created Limited scene. This is a disadvantage that is even more unfavorable affects as the insufficient degree of detailing in a model site over the a television camera is driving. On the simulators, this appears artificial in the picture, for example like in a cartoon, since the computer-generated image has a different visual structure has like a real picture. The creation of real barbeque and lighting effects is very computationally intensive and so far only possible with inferior quality.

If you still wanted to achieve a satisfactory simulation, so a very high computing capacity would have to be provided, which is currently only with special hard-wired computers is possible, but only to a limited extent Numbers are required and are therefore extremely costly and high maintenance problems raise.

In addition to these flight simulators that have been used in practice so far Systems are also known which use real aerial photographs as a database. These Recordings already contain the correct perspective and concealment or visibility the viewed terrain. One such method is in AGARD PG-268 20, p 1, last paragraph.

The advantage of this method is that no imaging is done needs to be, but only an image transformation must be carried out must, which requires significantly less computational effort, it is computing capacity through Storage capacity replaced. The system has so far been implemented in two versions. To the one will be a vision system with film as mass storage, the other will be like known from the AGÄRD report, a vision system with video disk as mass storage device used. The main problem with both versions is that a very large Mass image storage with sufficient access speed must be available, in addition, the image transformation unit must operate at a very high speed have, since z. B. the bandwidth of a single video channel for 625 television lines is already 5 MHz.

The object of the invention is to provide a method with which terrain images can be obtained and used for visual simulation, which is represented by a high resolution and detail of the terrain image over a wide viewing angle so that even low-level flight situations with overlay of ground and air targets as well as realistic pilot training for the combined use of sensors possible will.

In addition, a device is to be specified with which the highly detailed Terrain imagery visual simulation can be realized.

The process of obtaining and displaying terrain images for Vision simulators with real aerial photographs as a database and storage of the image signals on video disks with individual addressing of the images is characterized in that: a) in addition to each terrain image, a second synthetically generated image is stored is that for each image point of the terrain image the associated distance from the recording location in the radial direction, this information in the form of a distance mask image is saved; b) for a continuous transition of the projected image Movement from one location to the next is a digital image transformation in the form of an interpolation of the reproduced terrain views between the two is carried out at the closest reception sites; c) any playback projector a display unit has its own image transformation and target fade-in unit is connected upstream, with targets to be mapped on a separate image plate from all directions in question are mapped and in an image processing unit the target image is rotated to the correct position and zoomed in to the correct size, and that the target is also in the form of a video image and a distance mask image is saved; d) the format of the recorded terrain images is matched to the viewing angle of the display projectors that an image transformation unit of a projector can only see one image of the terrain in the database and thus only has to perform one transformation.

Further inventive method training s and a device for carrying out the method are specified in the subclaims.

The advantages of the method and the device used to implement it are based on the known use of real aerial photographs with which a the highest possible realism of the simulated ground view can be achieved, in that, that in principle arbitrarily large image angles can be reproduced, that a high one Resolution of the terrain background can be achieved by changing the terrain model some optical disks that can be completely exchanged at any point Targets with high resolution can be faded in, and in that low-level flights are carried out can be, without the restrictions that for mechanical reasons in the previously known simulators are available.

The terrain model can be easily reproduced by duplicating the image plates can be reproduced, so that it is possible to also use other simulators in a simple manner Way to equip with the terrain models. It is also of great advantage that to a large extent common devices from commercial computer and television technology are used the can, so that the financial outlay can be kept low. By the possibility the higher level of detail of the generated image over a large field of view areas of application for simulators that were incomplete with previous systems could be covered. So are z. B. Air floor simulations possible where are perceived and identified by the low-flying aircraft from ground targets, and with which Low-altitude helicopter flights, e.g. B. anti-tank helicopters can be practiced. In addition, dangerous conditions in the landing approach and low-level flight can be simulated, since this is the case a detailed view of the ground is required for an accurate elevation estimate. as well it is easy to train operators for remote-controlled aircraft at low altitude. The reconnaissance from the aircraft can also be carried out using several sensors in a network, e.g.

Eye, IR vision device, millimeter wave radar, as well as the realistic one Training in the use of eye and IR vision devices against ground targets.

The observer can move freely within the maneuvering area possible, with the highest speed and rate of turn, the aircraft in this Can reach altitude range.

The invention is explained in more detail in the figures; They show: FIG. 1 the construction of the terrain model; 2 shows the course of the transformation from the terrain model level into the projector level; 3 shows the computational sequence of the interpolation between neighboring viewpoints; 4 shows the block diagram of a device for visual simulation.

FIG. 5 partial steps of the affine distortion; FIG. According to Fig. 1, the terrain model is built from two five images that are doubled cover. The field of view of each display projector is adjusted so that a projector only sees an image of the terrain model. This creates a piecing together multiple image parts to one projector image avoided.

The arrangement of the database is chosen so that a horizontal Angle of view of 720 (angle of view due to oblique projection 80 ° x 64 °) and a vertical one of 64 °, with the image reaching 120 above and 520 below the horizon. It follows that two by five images are required for one vantage point. As can be seen from the figure, the first set is 1 of five stereo images against offset the second set 2 of images by half an image width. The diagonal of the projector image field is equal to or smaller than half the image width. Through this a projector only sees one image from the database at a time. The advantage the use of plane images as a database is that for the transformation of the image from the terrain image level in the database into the projector image level a straight-line projection from one plane onto a second plane at an angle to it Level can be used. A straight projection is computationally easier to be carried out as a non-straightforward, which is important for real-time computation.

Instead of the method described above, a single continuous All-round recording with vertical line direction can be used.

In this process, four in dots arranged in a square with a side length of 10 meters from continuous all-round images in like a cylinder projection with a CCD line scanner. These four All-round images with a number of lines of z. B. 9 x 575 lines are recorded on video tape stored in 9 subsequent images each.

From these four all-round images, a stereo evaluation can be made for each pixel to determine its distance from the recording location. In the Choice of nine standard television pictures with 4: 3 aspect ratio for storage for each one all-round image, each of the nine partial shots is 400 wide x 51.80 high. With this arrangement, a comparable amount of memory is required and comparable Height of the all-round image thus the resolution of the floor image is approx. 38% higher than when using a flat image base.

The projector image can simply be read out from neighboring image base recordings can be put together, with only one direction of transformation for a projector is to be considered.

The disadvantage of this method is that the transformation from the cylindrical terrain image plane no longer true to the straight line in the plane projector image plane is and is therefore computationally much more complex, so that real-time calculations the application of this procedure is made more difficult.

According to FIG. 4, there is a device for visual simulation a central control computer 10 of the vision system, the flight and vision data from a simulation computer takes over and the control of the individual assemblies and Sub-computer performs. This is connected to an interface system 11, which all Tact, Includes control and data lines. There are disk players for each of the 10 recording directions 01 to 020 available, each with a video disc player containing the video image and the second bili-disk player contains the associated distance mask image. the Bili disk players O1 to 020 are with a microprocessor image readout control Mistake.

There is also an optical disc player 12 with images of targets all of the possible recording directions and another bili-turntable 13 with the associated distance mask images of the target. Stand with it in parallel, all image and distance information for a recording location in analog Shape available.

The information reaches an Ereuzschienenverteiler 14, the the analog video signals on the associated projector chain 22 with bilitransformation unit 15 with 22 inputs and 22 outputs with 11 playback projectors.

There is therefore a total of 11 projector chains with 11 playback projectors 22 is present, one being shown in FIG. The image transformation unit 15 consists of analog-to-digital converters 16, a unit 17 for supplementing the projector image section in the sky and earth areas that are not covered by the image base, one Contrast manipulation computer 18 for simulating haze, fog or low hanging Clouds, a computer 19 for inserting the target range mask into the video image, a computer 20 which transforms the terrain image into the projector image plane and to correct the image on a new interpolated observer was standing as well as for splitting off the target range mask signal from the transformed video image, including digital to analog converter for video and target key mask signal, contour enhancer 21, which slightly differentiates the video output signal, 2-level chroma key picture mixer 23 for overlaying targets in the terrain image and saving images: .in the target entrance of the image mixer 24, which enables the target image to be shifted in parallel.

A target image transformation unit 25 is also present from analog-to-digital converter-for.

Video image and distance mask image signal, a calculator to insert the target removal mask into the video image, image distortion calculator for perspective important processing of the target image, a computer for splitting off the distance line image from the transformed target image, digital-to-analog converter for video image and target key signal image, Two-level chroma key image mixer 26 for masking out the visible part of the Target and a crossbar distributor 27, the target image signal to the target image input through the correct projector chain. The fade-in of the target in the background of the terrain is carried out as described in FIG.

The image transformation in FIG. 4 is essentially as follows in front of you. For explanation, the function of the contrast manipulation computer is first shown in FIG. 6 18 described in more detail. The contrast manipulation computer 18 takes a point by point Processing of the image contrast in the terrain image, with the location or address of the pixels are not changed.

The computer reads in the etching mask belonging to the image and changes the color of the individual pixels to blue and reduces the contrast corresponding to the Eatfern1nng; of the distance mask point belonging to the image point and the amount of haze to be simulated.

Furthermore, the contrast manipulation computer 18 changes the color within a distance mask uniformly into white, around fog or deep-lying To simulate cloud bases.

During the image transformation from the terrain image to the projection plane and the interpolation between the shooting viewpoints can when using The transformation of flat images as the image basis into several simple sub-steps be disassembled. This transformation essentially corresponds to a projection from one level to a second inclined level, as shown in FIG and 3 is shown. The individual steps are: a) Parallel displacement of the projector axis in the middle of the new picture; b) Rotating the grid direction from the terrain row direction in the projector line direction so that the center line of the image is parallel to the projector line lies; c) Projecting the image from the terrain image plane into the projector image plane by affine distortion; d) Correction of the image to the new (interpolated observer position) by affine distortion); e) Zoom the image into the correct one Size when moving along the projector axis.

Sub-steps c) and d) can become a resulting sub-step be summarized.

For the computational treatment of the transformation, it is useful to split this resulting step into sub-steps, one line at a time allow parallel processing of the image. So this with the available hardware can be carried out, a constant magnification factor must be used within an image line in line direction and perpendicular to it.

In order to carry out an affine, straight-line distortion of the image, must first be a picture distortion in.

Line direction can be carried out so that the right and left edges of the image are parallel and vertical. Then the image is rotated 90 °. Now the picture will again so distorted that the left and right edges of the picture are now sloping also run parallel and vertically, making the project image section rectangular assumes, see Fig. 5. The image is then rotated back by 90 °. By Zooming the image is then enlarged to a format that fills the projector image, see Fig. 2. This transformation can also be carried out by the sequence of operations according to FIG. 7, consisting of Parallel shift and rotation of the plane through the terrain image axis and the projector image axis perpendicular to the line direction, distortion of the image so that the projector image section becomes rectangular, parallel rotation of the grid direction and enlargement to the projector image size can be achieved with correction for keystone projection.

In the case of the image transfer, there is only one change in the position of the individual Pinpoints made, the content (brightness, color value) remains untouched.

The removal mask image can be processed in one operation are transformed with the video image, since in the image transformation as described above, the content of the pixels is no longer changed and the available hardware a higher one.

Number of bits per pixel available than is absolutely necessary is.

To process the distance mask image, the digitized Range mask image reads out the range level that corresponds to the range of the target is closest. Along the boundary line of this distance area is now the image into a foreground area that obscures the target and a background area, which is covered by the target, split up and this information n the lowest bit level of the video image, so that this now also includes the image formation in addition to the Contains foreground-background division with regard to the target to be faded in.

This combined image is subjected to the transformation, as described above, subject. After the transformation has been carried out, the foreground background information becomes split off again from the video image and parallel in the form of a black / white key image output to the video image.

Blank page

Claims (13)

Method and device for the acquisition and reproduction of terrain images for vision simulators claims 1. Method for obtaining and reproducing Terrain images for vision simulators, preferably for a flight simulator for the Flight and shooting training for pilots, with real aerial photographs as a database and storage of the image signals on video disks with individual addressing of the images, in this way that a) in addition to å each terrain image, a second synthetically generated image is stored, which is associated with each pixel of the Terrain image the associated distance from the recording location in the radial direction indicates, this information being stored in the form of a distance mask image will, b) for a continuous transition of the projected image when moving one Location to the next a digital image transformation in the form of an interpolation of the reproduced terrain views between the two closest recording locations is carried out, c) each display projector of a display unit has its own Image transformation and target fade-in unit is connected upstream, whereby to be mapped Targets shown on a separate image plate from all relevant directions and rotated the target image into the correct position in an image processing unit and is zoomed to the correct size, and that the target is also in the form of a Video image and a distance mask image is stored, d) the format the recorded terrain images to the viewing angle of the display projectors is coordinated that an image transformation unit of a projector only can see an image of the terrain in the database and can therefore only carry out one transformation got to. 2. The method according to claim 1, characterized in that g e k e n n -z e i c h n e t, that the associated distances from the recording location are coded as brightness levels be stored in the form of the distance mask image, and that for this purpose the aerial photographs of the database recorded as stereo images with a large stereo base will. 3. The method according to claim 1 or 2, characterized g e k e n nz e i c h n e t that a distance-dependent deterioration in visibility conditions through haze, fog or low-hanging clouds through manipulation of the bili-contrast is simulated depending on the distance level. 4. The method according to claim 1, characterized in that g e k e n n -z e i c h n e t, that using the distance information for the pixels and the direction the line of sight vector a missile is moving and its range by the observer, the point of impact of a missile in the field is determined and correct is pictured. 5. The method according to one or more of the preceding claims, thereby g e k e n n n z e i c fr n e t that on the basis of the known distances for å moving targets to each terrain pixel in the correct obscuration in the terrain any point and that the insertion of the targets on analog Paths are done in a two-level chroma key process. 6. The method according to any one of the preceding claims, characterized g e it does not indicate that the image base is composed of two stereo panorama recordings which in turn consist of a large number of flat stereo images that come together the all-round view of 3600 overlap twice, with the first set of stereo images is offset from the second set of stereo images by half an image width, and the diagonal of the projector image field is equal to or less than half that Image width. 7. The method according to claim 1, characterized in that g e k e n n -z e i c h n e t, that for each recording location from four points arranged in a square Continuous all-round images were recorded in the manner of a cylinder projection with suns and that for each of these four all-round images for each pixel there is a stereo evaluation is carried out to determine its distance from the recording location, and that the projector image is composed by simply reading out neighboring image base recordings will. 8. The method according to claim 5, characterized in that g e k e n n -z e i c h n e t, that by reading out the dividing line between covered and non-covered terrain image parts from the distance mask image and using this signal as a key signal in one 2-level chroma key video image mixer applying moving target into one Terrain image is made with the target being the background in the auto-tey process is superimposed and the foreground is the target according to the target range key signal is superimposed. 9. The method according to claim 5, characterized in that g e k e n n -z e i c h n e t, that by dividing the target image into distance zones and controlling the changes in visibility of the target with radial movement through a distance mask image as a key signal a sudden appearance in a picture mixer using the chroma key method of the target is avoided in radial movements. 10. The method according to one or more of the preceding claims, in that a transformation and interpolation of saved air recordings for projection in the flight simulator through image transformation is carried out, which is divided into the individual steps parallel shift, Rotation, affine straight line distortion, and magnification of the image. 11. The method according to claim 10, characterized in that g e k e n n -z e i c h n e t that the affine straight-line distortion is divided into a distortion with line-by-line constant enlargement factor perpendicular and parallel to the line direction, so that the left and right edges of the projector image section in the terrain image are vertical run to the direction of the line, then rotate by 90 ° again distortion, as described above and reverse rotation by 90d. 12. The method according to any one of the preceding claims, characterized g e does not indicate that the processing of the video image and the distance mask image takes place in one operation by putting both pieces of information into one pixel data word inscribed in superimposed layers and the combined image is transformed according to the preceding claims. 13. The method according to one or more of the preceding claims, in that a transformation and interpolation of stored aerial photographs for projection in the flight simulator through image transformation is carried out, which is divided into the individual steps parallel shift, Rotation of the connecting plane from the aerial image axis to the projector image axis perpendicular to the Line direction, distortion of the image with line-by-line constant enlargement factors in and perpendicular to the line direction, parallel rotation of the grid and enlargement of the picture.