WO2013178213A2

WO2013178213A2 - Method for automatic imaging correction of a projector

Info

Publication number: WO2013178213A2
Application number: PCT/DE2013/000291
Authority: WO
Inventors: Heiko CORNELIUS
Original assignee: Cornelius Heiko
Priority date: 2012-06-01
Filing date: 2013-05-28
Publication date: 2013-12-05
Also published as: WO2013178213A3; DE102012010814A1

Abstract

The invention relates to a method for automatic imaging correction of a projector (100)for digitized images, wherein the projector (100) comprises a camera (102) for capturing the projected image and wherein the captured projected image is compared with at least one image projected under normal conditions and on the basis of the comparison the projection optics (101) of the projector (100) are modified and/or the projected image is distorted in the projector (100) until the captured projected image is identical with the at least one image captured under normal conditions in the range of a preselected tolerance interval. According to the invention, the method and the corresponding projector permit an automatic imaging correction.

Description

Method for automatic image correction of a projector

[01] The invention relates to a method for automatic imaging correction of a projector for digitized images, wherein the projector has a camera for recording the projected image and wherein the recorded projected image is compared with at least one image projected under normal conditions and depending on the comparison the projection optics of the projector are changed and / or the projected image in the projector is distorted until the recorded, projected image is identical to the at least one image taken under normal conditions in the range of a preselected tolerance interval and a corresponding projector for the projection of digitized Images to a projection screen.

[02] Projectors for the projection of digitized images on any projection surface, usually a screen, receive via a dedicated interface a signal of a digitized image and project the image on the projection surface via a corresponding projection optics. It is necessary that the geometry of the projection surface and the projection optics are matched to each other in order to obtain a sharp and distortion-free image. In addition to the coordination of the distance of projection optics and screen is also to take into account that the projector often has to be positioned in relation to the projection surface so that a central projection is not possible.

CONFIRMATION COPY The projector projects from the direction of the projector on an oblique screen or the other way round, the projector is not centered and perpendicular in relation to the screen. The placement problem goes hand in hand with the need for an image correction. Typical aberrations, if the projection optics are not set correctly, are the chromatic aberration that causes a slightly different image for different colors. This error is due to the dispersion of the glass used and not optimally adjusted projection optics makes this error noticeable. Another mistake is the so-called distortion, in which a picture in the form of a horizontally lying quadrilateral is distorted to a bulbous, bulging ton or - inversely thereto - to a pillow with extended corners. The two-dimensional image to be projected appears to be three-dimensionally collapsed or expanded for a viewer. The slanted arrangement of the image plane of the projector and the projection plane creates trapeze effects in which a projected quadrilateral is distorted to a trapezoid when projected onto the projection plane. Finally, if a curved or non-planar projection plane is used, the projected image experiences a corresponding distortion.

[03] In order to optimally set up a projector for its use, a considerable amount of time is required to use the projector, depending on the spatial conditions. Especially for mobile use of projectors, for example for quick setups in sales presentations or sudden events that projects should be, the time required for setting is not acceptable. Also, for optimal setting a certain exercise is necessary that not everyone has.

[04] The object of the invention is therefore to automate the imaging correction. The invention is solved by the method according to claims 1 to 5 and by the projector according to claims 6 to 10.

[05] According to the invention, it is proposed to equip the projector with a camera which records the projected image and to use a control device which internally compares the taken, projected image with an image taken under normal conditions and the projection optics for so long Depending on the comparison within the projector, the digital image itself is distorted and / or distorted until the recorded, projected image is identical to an image that was taken under normal conditions, whereby a predetermined degree of tolerance is accepted.

[06] In order to determine the degree of tolerance within which the recorded projected image should be considered identical to the image taken under normal conditions, the RMS error of individual selected pixels is determined. If the RMS error of the measured deviation as a tolerance value is less than 5% of the image diagonal, preferably less than 1% of the image diagonal, then the images are considered identical and the control loop comes to a standstill. [07] The RMS error, in English "root-mean-square" or German the "quadratic mean" is determined by the mathematical calculation of the root mean square of all the distances from the nominal position of each selected pixel to the measured position of this pixel , where each individually determined distance is included as a fraction of the image diagonal in the calculation.

[08] In order to be able to perform the image correction during the projection of a user-desired image to be projected, it is preferred if the projector inserts a test image into the desired image to be projected only briefly, the time duration preferably being less than 1 / 18 s lasts. This time is just no longer perceived by the human eye when the test image is temporally bound by two identical images, ie when an identical or nearly identical image is displayed before and after the test image. Depending on the type of image to be projected, however, another time period is advantageous. For moving digitized images, it is suitable to display a test image within two individual images, which are shown in typical frequency, ie 1/24 s. For moving pictures to European standards, a time of less than 1/25 s or 1/50 s is suitable. Times of less than 1/30 s or 1/60 s are suitable for US standard images. The image correction in this case is carried out exclusively by these briefly recorded, projected test images [09] The test image is designed so that image analysis algorithms can easily and quickly identify selected points, whether the captured, projected image is distorted or not. This recorded, projected test image is then compared with a same test image, which has been recorded under normal conditions, ie ideal projection conditions with optimally adjusted projection optics, and a control algorithm determined on the basis of the comparison result, the control strategy according to predetermined rules to equalize the projected image. In order to be able to identify selected pixels with a simple algorithm that is real-time-capable for a control, it has proved to be advantageous if the recorded, projected image is masked so that only a few pixels identified by the algorithm are present in the Picture are available.

[10] The ideal control strategy is to correct a detected chromatic aberration by varying the corresponding lens distances or lens positions within the projection optics. Also, such distortion errors as pillow errors can be compensated. Keystone errors are corrected by varying the position of the image plane within the projector. Additional distortions that are irregular, for example due to a curved or off-screen projection surface, are achieved by digital inverse distortion of the image. It has proven to be advantageous if the projector in each image dimension at least one higher resolution by a factor of 2 than the image to be projected. At the time of this application, although high-resolution projectors are known with an image score of well over 2,000 in the horizontal and well over 1,000 in the vertical. Nevertheless, image resolutions for moving pictures and also for simple presentations on the order of 640 x 480 pixels are still quite common. The high-resolution projector would thus be able to equalize a distorted image due to the higher resolution.

[1 1] The invention will be explained in more detail with reference to the following figures. It shows:

1 Sketching of a barrel-shaped distortion, FIG. 2 Sketching of a cushion-shaped distortion, FIG.

FIG. 3 Sketching of a keystone error,

FIG. 4 a sketch of a chromatic aberration,

5a, 5b, 5c sketch of the masking of picture elements,

FIG. 6 a sketch of an alternative test pattern,

Figure 7 Sketching of an irregular distortion due to a non-planar projection surface and

FIG. 8 shows a projector according to the invention.

FIG. 1 shows the result of a distortion due to a barrel-shaped distortion. A rectangle 1 to be projected, which assumes the dashed form when ideally projected, is distorted by an improper adjustment of the projection optics to a barrel-shaped rectangle 2 according to the solid line, the sketch chosen here exaggerating the error for better visualization. Projection optics are like a telephoto lens to it able to independently vary the aperture angle and sharpness. If the opening angle is too large or too small, this distortion effect is produced so that the barrel-shaped distortion effect can be corrected by varying the opening angle. As the lenses within the projection optics gradually align, the curved lines of the barrel-shaped rectangle move inward according to the arrows drawn, leaving the imaged corners of the ideally projected and barrel-shaped rectangles contiguous. In this image correction, the aperture angle, the sharpness and the magnification of the digital image must be set simultaneously so that the change in the aperture angle does not change the image size, so that then the vertices of the two drawn figures, ideally projected Rechteckl and barrel-shaped rectangle. 2 stay virtually unchanged. FIG. 2 shows the aberration that is inverse to the aberration in FIG. 1, the pincushion distortion. This could be understood as an over-correction of barrel distortion or vice versa. In this distortion, an ideally projected rectangle 1 is distorted into a rectangle 3 which is in the form of a pillow. Even with this distortion error exaggerated here, the pincushion distortion effect is corrected by a variation of the opening angle. As the lenses within the projection optics gradually align, the curved lines of the barrel-shaped rectangle shift outward according to the arrows drawn, with the imaged corners of the ideally projected and the square pillow-shaped rectangles remain on each other. In this imaging correction, the aperture angle, the sharpness and the magnification of the digital image must be set simultaneously so that the change in the aperture angle does not change the image size, so that then the vertices of the two drawn figures, ideally projected rectangle 1 and rectangular shaped cushions 3 remain virtually unchanged.

[14] The trapezoidal distortion is an aberration due to a relative skew of the image plane and the projection plane. This keystone error is outlined in FIG. In this case, a rectangle 1 projected with an ideal projection is distorted to a rectangle 4 trapezoidal. Due to the misalignment of the image plane and the projection plane, the central projection enlarges a point further away from the projection optics or displays it further from the center of the projection. This projection error can be compensated by an inclination of the image plane within the projector, in the ideal state, the image plane is aligned parallel to the projection plane.

Finally, FIG. 4 shows a sketch of the error of a chromatic aberration. In this case, a white, ideally projected rectangle 1 - here in a keystone error - widened or narrowed by the chromatic aberration to different figures, which is represented by the various trapezoids, trapezoidal 5, trapezium 6 and trapezium 7. When using a continuum spectrum, ie with a light bulb as a projection lamp, The three trapezoids 5, 6 and 7 melt into a rainbow spectrum. If, on the other hand, different monochromatic light sources, such as lasers, are used as the projection light, then the three colors of the trapezoids 5, 6 and 7 separate into different trapezoids. Depending on the projection optics, this error can also be corrected to a predetermined extent by a relative shift of different lens groups within the projection optics. An ideal projection optics does not show these errors, but due to the unavoidable dispersion, the different refractive indices of the lens lenses, this error is almost permanent, but - depending on the type of lens - still within certain limits adjustable.

FIG. 5a, FIG. 5b and FIG. 5c show a masking of an imaged projection error, by means of which the automatic method for automatic imaging correction is explained. In the projection of a rectangle 1, which would ideally assume the dashed contour of the rectangle 1 on the projection surface, a barrel-shaped distortion distorts the tonally distorted rectangle 2, shown in lines. A camera immediately adjacent to the projector, preferably with very small lens dimensions, records the projected image of the projector. The fact that the camera has very small dimensions compared to the projection optics, the focus on "infinite" can provide sharp images even at 2m distance to a distant projection screen. Due to its small dimensions, the camera will also show only a very small parallax error due to its off-center position with respect to the camera Central projection comes about. This camera delivers the image of the barrel-shaped distorted rectangle, whereby the rectangle is masked as a simple test image within a computing unit, within the camera or already during the projection. By masking with a cross-shaped mask 10, which is placed over the recorded image of the barrel-shaped distorted rectangle 2 according to Figure 5b, the image is formed in accordance with the solid line in Figure 5c. The positions of the four sections of the barrel-shaped distorted rectangle 2 are compared within the control device by a computing unit with a stored image of an ideally projected rectangle, according to the dotted lines remaining after masking. An image recognition algorithm can quickly capture these very few image contents, determine the focus of the identified image content, and determine the distance to the ideal center of gravity that lies within the area of the dotted lines. This mask results in distance vectors distance vector 20, distance vector 21, distance vector 22 and distance vector 23, which each have a direction and a length, for example as a fraction of a picture diagonal. By comparing the distance vectors with a table, the ideal control strategy for correcting barrel distortion, pincushion distortion and trapezoidal distortion can be read. In the present case, this table has at least 3 states for each vector, namely "up", "down", "right" for the two top and bottom image contents and "right", "left" and "right" for the two at the Pages lying image content. This results in 3 high 4, so 81 possibilities of the identified Deviation combinations. By a superimposition of different distortion errors, for example, two different trapezoidal distortions, which arise through a corresponding, distorted by two axes, projection plane, and by a barrel-shaped or pincushion distortion so very different vector patterns can be created by appropriate variation of the position parameters of the projection optics and the image plane can be corrected. However, this type of correction, namely the assignment of vector parameters to a table from which individual control strategies are retrieved, is only a simple example of a control strategy.

[17] A completely different type of control strategy is known, for example, in the image equalization for the reading of two-dimensional matrix codes. In this case, as shown in FIG. 6, an easily recognizable pattern 8, which is identified by an image recognition algorithm, is inserted into the corners of a rectangle. When identifying the individual picture elements, the position of the photographed image with respect to the camera position can be precisely determined and thus the control strategy can be called up.

[18] A special case exists when there is a non-planar projection surface, which is shown in FIG. In this case, the mapping of each pixel or the image of a group of pixels, for example a quadrilateral of 4, 9, 16 or more points, must be determined with the projected position. This results in the distance vectors shown in FIGS. 5a, 5b and 5c, of which two rows, vector row 30 and vector row 31, are shown in FIG. This vector matrix, once detected, is then used for a nonuniform distortion corrector of the image due to a nonplanar projection surface, with an image processor in the projector artificially inversely distorting the image content to the detected aberration resulting in near aberration free projection. In order to determine this error, it is therefore necessary to determine the aberration with a plurality of test images, which can also be displayed as individual images in a sequence of desired projected images. Finally, FIG. 8 shows a projector 100 according to the invention which has projection optics 101, a camera 102 for recording the projected image, and also a control unit 103 as a computing unit which is capable of analyzing image contents and accordingly the projection optics set to motorized. Furthermore, the control device can adjust the image plane and also adjust individual lens groups to each other in order to achieve an ideal image during projection by the variation of the image plane and lens distances. The exact control strategy depends on the selected projection optics, which can have a large number of setting parameters depending on the design. The control strategy is then dependent on a lens model implemented in the control device.

[20] In addition to the possibility of visually correcting aberrations, the control device may also have an image processor that changes the digital signal of an image to be projected so that individual Pixels or individual pixel groups within the image shifted, one-dimensional or two-dimensional compressed or expanded, possibly even distorted.

[21] A special type of image correction is the correction of detected brightness gradients and color gradients. The brightness can be adjusted averaged. This type of correction can be adjusted using the average measured brightness of the captured projected image. When projecting onto projection surfaces inclined relative to the projector, brightness differences inevitably arise, which can be corrected by reversing the brightness distribution of the individual image contents. Also, the color gradient of a uniform white image can be corrected by a corresponding color mask. Shading can not be corrected by adjusting the projector parameters. However, a detected shading can be used as a control signal to give, for example, the migration of shading as a signal to the projector or to the projecting computer in the manner of a computer mouse signal.

LIST OF REFERENCE NUMBERS

1 rectangle, deal projected

2 rectangles, barrel-shaped

3 rectangles, pillow-shaped

4 rectangles, trapezoidal

5 rectangles, trapezoidal, first color

6 rectangles, trapezoidal, second color

7 rectangles, trapezoidal, third color

8 patterns

10 mask

20 distance vector

21 distance vector

22 distance vector

23 distance vector Vector series of vector series Projector Projection optics Camera

control device

Claims

claims:

1. A method for automatic image correction of a projector (100) for digitized images,

- wherein the projector (100) has a camera (102) for receiving the projected image and

wherein the recorded, projected image is compared with at least one image projected under normal conditions, and

depending on the comparison, the projection optics (101) of the projector (100) are changed and / or the projected image in the projector (100) is distorted until the recorded, projected image with the at least one image taken under normal conditions is in the range of a preselected Tolerance interval is identical.

2. The method of claim 1, wherein the identity of the captured projected image with the image taken under normal conditions is determined via the RMS error of individual selected pixels, and wherein the RMS error of the measured deviation as a tolerance value is less than 5 % of the picture diagonal.

3. The method according to any one of claims 1 or 2, wherein in the image to be projected time at least one test image superimposed and the Image defect is corrected on the basis of the imaginary test image, the duration of the insertion being less than 1/18 s, preferably less than 1/24 s, more preferably less than 1/25 s, particularly preferably less than 1/30 s, very particularly preferably about 1 / 50 s or 1/60 s and wherein the imaging correction coincides solely with the comparison of the recorded, projected test image with the at least one test image recorded under normal conditions.

The method of any one of claims 1 to 3, wherein the captured, projected image is masked prior to comparison with a mask (10).

Method according to one of claims 1 to 4,

- wherein the chromatic aberration and the distortion error is corrected by varying the lens position in the projection optics (101), and

- wherein other image distortions are made by a corresponding inverse digital distortion of the image to be projected.

A projector (100) for projecting digitized images onto a projection screen comprising a camera (102) for taking the projected image and a control device (103) which varies the projected image based on the comparison of the captured projected image with an image projected under normal conditions until the captured projected image is identical to the image projected under normal conditions in the range of a preselected tolerance interval.

A projector according to claim 6, comprising a digital signal processor that distorts the image to be projected through a mapping function that is variable by the control device (103), in which individual pixels in the digital image are spatially displaced.

A projector according to claim 6, comprising electrically adjustable projection optics (101) for correcting the distortion error, the image sharpness and the chromatic aberration.

Projector according to one of Claims 6 to 8, characterized by a resolution which is higher by at least a factor of 2 than corresponds to the image to be projected.