EP1699036A1 - LED display with high resolution - Google Patents

Abstract

The method involves defining one of masks (7) at each of light sources (4) that include a set of pixels (11) in a surrounding area of the sources. Color values of the pixels contained in the masks are determined for colors of the sources, and brightness values are calculated for the sources from the color values. Color transformation of image data of images is carried out in a color space before calculating the brightness values. An independent claim is also included for a display device for displaying an image.

Description

The invention relates to a method for displaying an image on a display unit, in particular an LED display unit, according to the preamble of patent claim 1, and to a corresponding display apparatus according to the preamble of claim 12.

Such display devices are particularly useful for displaying images at major events, e.g. Concerts or football matches. However, they are also used as road signs, e.g. in traffic control systems, or used as display signs. The physical display elements (LEDs) of the display devices known from the prior art are arranged at least in triplets, one of these groups consisting of at least three color values which are linearly independent in the color space (usually red, green and blue). As a result, the representation of a pixel, represented by a color value within the color space spanned by these colors, is possible.

Fig. 1 shows a well-known from the prior art RGB display device, such as is used as a traffic road sign for traffic. The display device comprises a display panel 1 with a multiplicity of LEDs 4, as well as a data processing device 2 connected thereto. LEDs of three different colors R, G, B are in each case combined in triangular groups 3, of which only a few are shown here by way of example. The DV device 2 processes the image data of the image to be displayed and controls the LEDs 4 so that the image appears as sharp as possible on the display board. The image data can either be stored locally or supplied via an input (Data). The DV device 2 usually comprises a microprocessor with image processing software and a memory device.

FIG. 2 shows the pixels 11 of a high resolution image 10 in a rectangular grid, as well as a plurality of display elements 4 (LEDs) of a physically low resolution display device on which the image 10 is to be displayed. The LEDs 4 are combined into triangular LED groups 3. Each LED group 3 generates a pixel 8 which is at the center of the group. The color value of the generated pixel 10 is dependent on the brightness of the individual LEDs 4 of a group 3. The spacing of the LED groups 3 defines the physical resolution of the LED array.

• a) "Average Downsampling": dabei wird um jeden Bildpunkt 8 bzw. virtuellen Bildpunkt 8' eine Maske 9 gelegt, die einen Betrachtungsbereich definiert und mehrere Pixel 11 des Originalbildes 10 umfasst. Aus den Farbwerten dieser Pixel wird dann ein Mittelwert gebildet, der den Farbwert für den Bildpunkt 8 bzw. virtuellen Bildpunkt 8' darstellt.
• b) "Bicubic Downsampling": hierbei wird ebenfalls um jeden Bildpunkt 8 bzw. virtuellen Bildpunkt 8' eine Maske 9 gelegt, die einen Betrachtungsbereich definiert. Die Farbwerte der im Betrachtungsbereich liegenden Pixel 11 werden dann gewichtet, wobei die Gewichtung vom Abstand der Pixel 11 zum Bildpunkt 8 bzw. virtuellen Bildpunkt 8' abhängt. Die gewichteten Farbwerte werden danach gemittelt. Diese Methode ist relativ rechenintensiv, erzielt jedoch die beste Qualität der Darstellung.
• c) "Subsampling": hierbei werden die Farbwerte des Bildpunkts 8 bzw. virtuellen Bildpunkts 8' lediglich durch die Farbwerte des nähesten Pixels 11 ersetzt.

In order to increase the resolution of the display panel 1 with respect to the physical resolution, it is known to consider in addition to the LED groups 3 and additional LED groups 3 ', each composed of a subset of LEDs 4 adjacent groups 3. Each of these additional groups 3 'comprises at least one LED 4 of each color (eg red, green and blue). The pixel 8 'resulting in the center of these additional LED groups 3', which is composed of the sum of the color values of the associated LEDs 4, is commonly referred to as "virtual pixel" 8 '. The color value which the pixel 8 or the virtual pixel 8 'is to display is usually calculated from the color value of the original image 10 at the coordinate of the pixel 8 or the virtual pixel 8' and the color values of the pixels 11 in the immediate vicinity. In the case of a virtual pixel, the color values of the neighboring pixels 8 are also taken into account. Various standard methods are known in computer graphics for the calculation of the color value to be displayed:

• a) "Average downsampling": in this case, a mask 9 is laid around each pixel 8 or virtual pixel 8 ', which defines a viewing area and comprises a plurality of pixels 11 of the original image 10. From the color values of these pixels, an average value is then formed, which represents the color value for the pixel 8 or virtual pixel 8 '.
• b) "bicubic downsampling": in this case, a mask 9 is likewise placed around each pixel 8 or virtual pixel 8 ', which defines a viewing region. The color values of the pixels 11 lying in the viewing area are then weighted, the weighting depending on the distance of the pixels 11 to the pixel 8 or virtual pixel 8 '. The weighted color values are then averaged. This method is relatively computationally intensive, but achieves the best quality of presentation.
• c) "subsampling": in this case, the color values of the pixel 8 or virtual pixel 8 'are merely replaced by the color values of the closest pixel 11.

Known display devices have the disadvantage that the resolution is essentially determined by the spacing of the LED groups and is therefore relatively low.

It is therefore the object of the present invention to substantially improve the effective resolution and contour sharpness of a display device.

This object is achieved according to the invention by the features specified in the patent claim 1 and in claim 12 features. Further embodiments of the invention are the subject of dependent claims.

An essential idea of the invention is to define each of the light sources a separate mask, the plurality of pixels in the vicinity of the light source comprises determining the color values of the pixels contained in the mask in each case for the color of the light source (eg in the case of a green LED the color values of the individual pixels for the color green) and from the individual color values the brightness value for to calculate the considered light source. In contrast to the prior art mentioned above, not the pixels 8 or "virtual pixels" 8 'form here, but the individual light sources are the center of gravity for the downsampling. Thus, the brightness value for a light source is calculated from the color values (for the color of the light source) of those pixels that lie in the vicinity of the coordinate of the light source. This has the advantage that the resolution of the display device can be significantly increased in comparison to the prior art described above and the sharpness of the contours can be substantially improved by the use of different transformation masks.

According to the invention, before the downsampling, ie before the calculation of the brightness values for the display elements, a color transformation is carried out in which the color data of the original image are transformed into the color space of the physical display elements (eg LEDs). According to a preferred embodiment, this is the high-resolution. Original image converted by color transforms in-part images with the same resolution, wherein the target color space of this transformation is characterized by the colors of the physical display elements. This means that in the case of an RGB display, the original image is converted into partial images for the colors red (R), green (G) and blue (B) of the display elements. Thus, there are always so many partial images as there are different colored display elements (LEDs). The masks are then placed in the sub-images at the respective positions of the corresponding color light sources (in the case of a partial image of, for example, the color green at the positions of the green light sources) and the brightness values are calculated. Optionally, the masks could also be on the Pixel raster of an image (eg the original image), which still contains all the color information, are placed and the individual color values (in the color space of the display elements) are then determined from the total color information. In both cases, however, a color transformation occurs before downsampling.

The individual light sources of a display device according to the invention are preferably distributed uniformly over the display area. By this arrangement and the use of the method described above, it becomes possible to take advantage of the effect that the brightness resolution of the human eye is greater than the resolution of color perception. The perceived image quality is thus maximized.

According to a preferred embodiment of the invention, the color values of the pixels contained in a mask are weighted by means of weighting factors and the brightness value for the considered light source is calculated from the weighted color values. The weighted color values are preferably averaged for this purpose.

The weighting factors of the pixels of a single mask are preferably selected such that each pixel after a superposition is weighted equally well with adjacent masks for the same color, preferably with exactly the same weight. This has the significant advantage that over the entire surface of the display unit a homogeneous color impression can be generated.

If one considers the masks defined around the light sources of a color, they are preferably dimensioned so large that they overlap spatially. In the overlapping areas, it is therefore possible to selectively generate a desired color impression by means of a plurality of adjacent light sources (of the same color). The light sources can then be controlled so that creates the desired color impression in the space between the light sources. The resolution of the display device can thereby be significantly increased.

The masks defined around the light sources are preferably dimensioned so that they do not contain another light source of the same color.

The masks defined around the light sources are preferably dimensioned to be close to the boundary of the pixels that at least partially contain a light source of the same color. The overlap area is thereby maximized.

According to a special embodiment of the invention, the masks are also defined such that the associated light sources lie approximately in the center, preferably exactly in the center of the masks. The masks are also preferably formed symmetrically.

• Gemäß einer ersten Ausführungsform der Erfindung wird ein Verfahren gewählt, bei dem die in einer Maske enthaltenen Farbwerte gewichtet werden, wobei die Gewichtungsfaktoren eine Funktion des Abstandes der in der Maske enthaltenen Pixel von der Lichtquelle sind. Die so gewichteten Farbwerte werden dann gemittelt.
• Gemäß einer zweiten Ausführungsform der Erfindung wird ein Downsampling-Verfahren gewählt, bei dem die in einer Maske enthaltenen Farbwerte gewichtet werden, wobei der Gewichtungsfaktor eines Pixels von der Anzahl der superponierten Masken (derselben Farbe) abhängt, in denen das Pixel enthalten ist. D.h. Pixel, die in weniger Masken enthalten sind erhalten einen höheren, und Pixel, die in mehr Masken enthalten sind, einen kleineren Gewichtungsfaktor. Die Gewichtungsfaktoren werden dabei so gewählt dass alle Pixel (betrachtet man die Superposition) möglichst gleich stark bewertet werden. Die so gewichteten Farbwerte werden dann gemittelt. Dieses Verfahren hat insbesondere den Vorteil, dass keine komplizierten Abstandsberechnungen durchgeführt werden müssen.

In principle, any desired downsampling method can be used to calculate the brightness value from the color values contained in the mask. The following are some preferred procedures:

• According to a first embodiment of the invention, a method is selected in which the color values contained in a mask are weighted, wherein the weighting factors are a function of the distance of the pixels contained in the mask from the light source. The weighted color values are then averaged.
• According to a second embodiment of the invention, a downsampling method is selected in which the color values contained in a mask are weighted, wherein the weighting factor of a pixel depends on the number of colorants superponed masks (the same color) in which the pixel is contained. That is, pixels that are contained in fewer masks receive a higher, and pixels that are included in more masks, a smaller weighting factor. The weighting factors are chosen so that all pixels (considering the superposition) are rated as equally strong as possible. The weighted color values are then averaged. This method has the particular advantage that no complicated distance calculations must be performed.

wobei d_ij der Abstand des Pixels von der zugehörigen Lichtquelle und c eine Konstante ist, die vorzugsweise derart gewählt wird, dass sämtliche Pixel einer Maske nach einer Superposition mit benachbarten Masken für dieselbe Farbe etwa gleich stark gewichtet werden.In the downsampling method according to the first embodiment of the invention, as mentioned, the weighting factors are calculated as a function of the distance of the pixel to the associated light source. The following relationship can be used for this: $${\mathrm{H}}_{\mathrm{i}\mathrm{j}}=\frac{1}{{\left(d_{ij}+c\right)}^2}.$$

where d _{ij is} the distance of the pixel from the associated light source and c is a constant which is preferably chosen such that all pixels of a mask after a superposition are weighted equally equally with adjacent masks for the same color.

The weighting factors h _{ij of} a mask are preferably mathematically represented so that they form the sum of a power of two. To calculate a weighted average, the weighted values are known to be summed and divided by the sum of all weighting factors. If the sum of all weighting factors is a power of 2, the division can be easily performed by a floating point operation.

A display device according to the invention essentially comprises a display unit, such as a display panel, for example. and a data processing device with software that processes the image data of the image to be displayed and drives the light sources. The IT device works for this purpose according to one of the above methods.

According to a preferred embodiment of the invention, the display unit comprises light sources with a total of four colors, preferably R, G, B, Y. The light sources (R, G, B, Y) are preferably uniformly distributed and arranged in a square arrangement. As a result, a particularly dense and high-resolution representation can be achieved.

The light sources of a color are preferably evenly distributed over the display unit. This has the advantage that the resolution in both dimensions of the display unit is the same size.

According to a preferred embodiment of the invention, not only the light sources of a color, but all the light sources are evenly distributed over the display unit.

• Fig. 1 eine aus dem Stand der Technik bekannte LED-Anzeigevorrichtung;
• Fig. 2 die Zuordnung eines anzuzeigenden Bildes zu den LEDs der Anzeigevorrichtung;
• Fig. 3 die wesentlichen Verfahrensschritte eines Verfahrens zur Berechnung von Helligkeitswerten für die LEDs einer Anzeigevorrichtung;
• Fig. 4 eine LED-Anzeigeeinheit gemäß einer ersten Ausführungsform der Erfindung;
• Fig. 5 benachbarte Masken für die Farbe Grün;
• Fig. 6 benachbarte Masken für die Farbe Rot;
• Fig. 7 eine LED-Anzeigeeinheit gemäß einer zweiten Ausführungsform der Erfindung;
• Fig. 8 benachbarte Masken für die Farbe Rot bei der Anzeigeeinheit von Fig. 7;
• Fig. 9 die Berechnung von Gewichtungsfaktoren für die Farbe Rot bei der Anzeigeeinheit von Fig. 7;
• Fig. 10 die Überlappung von Masken für die Farbe Blau bei der Anzeigeeinheit von Fig. 7;
• Fig. 11 die Berechnung von Gewichtungsfaktoren für die Farbe Blau bei der Anzeigeeinheit von Fig. 7;
• Fig. 12 benachbarte Masken für die Farbe Grün bei der Anzeigeeinheit von Fig. 7;
• Fig. 13 die Berechnung von Gewichtungsfaktoren für die Farbe Grün bei der Anzeigeeinheit von Fig. 7;
• Fig. 14 eine LED-Anzeigeeinheit gemäß einer dritten Ausführungsform der Erfindung;
• Fig. 15a bis 15c die Berechnung von Gewichtungsfaktoren für die Farben rot, grün und blau bei der Anzeigeeinheit von Fig. 14;
• Fig. 16 benachbarte Masken für die Farbe rot bei der Anzeigeeinheit von Fig. 14;
• Fig. 17 die LED-Anzeigeeinheit gemäß der zweiten Ausführungsform der Erfindung bei höherer Auflösung des anzuzeigenden Bildes;
• Fig. 18 benachbarte Masken für die Anordnung von Fig. 17 gemäß einer ersten Ausführungsform; und
• Fig. 19 benachbarte Masken für die Anordnung von Fig. 17 gemäß einer zweiten Ausführungsform.

The invention will now be described by way of example with reference to the accompanying drawings. Show it:

• Fig. 1 is a known from the prior art LED display device;
• 2 shows the assignment of an image to be displayed to the LEDs of the display device;
• 3 shows the essential method steps of a method for calculating brightness values for the LEDs of a display device;
• 4 shows an LED display unit according to a first embodiment of the invention;
• Fig. 5 adjacent masks for the color green;
• FIG. 6 shows adjacent masks for the color red; FIG.
• 7 shows an LED display unit according to a second embodiment of the invention;
• Fig. 8 shows adjacent masks for the color red in the display unit of Fig. 7;
• Fig. 9 shows the calculation of weighting factors for the color red in the display unit of Fig. 7;
• Fig. 10 shows the overlap of masks for the color blue in the display unit of Fig. 7;
• Fig. 11 shows the calculation of weighting factors for the color blue in the display unit of Fig. 7;
• Fig. 12 shows adjacent masks for the color green in the display unit of Fig. 7;
• Fig. 13 shows the calculation of weighting factors for the color green in the display unit of Fig. 7;
• 14 shows an LED display unit according to a third embodiment of the invention;
• Figs. 15a to 15c illustrate the calculation of weighting factors for the colors red, green and blue in the display unit of Fig. 14;
• Fig. 16 shows adjacent masks for the color red in the display unit of Fig. 14;
• 17 shows the LED display unit according to the second embodiment of the invention at a higher resolution of the image to be displayed;
• FIG. 18 shows adjacent masks for the arrangement of FIG. 17 according to a first embodiment; FIG. and
• Fig. 19 adjacent masks for the arrangement of Fig. 17 according to a second embodiment.

With regard to the explanation of Figs. 1 and 2, reference is made to the introduction to the description.

3 shows the fundamental principle of the transformation of image data (R, G, B) of the image 10 into brightness values (R ", G", B ") for the control of the individual LEDs 4. The image data R, G, B of the original image 10 are first converted into the color space of the LED display unit 1 by means of a transformation 5, color values R ', G', B 'being transformed for each pixel 11 (see Fig. 2) of the original image 10. In block 6, a geometrical Assignment of the image pixels 11 to the LEDs 4 of the display panel to determine which of the pixels 11 are located in the area of which LEDs 4. Finally, in block 7, the downsampling is performed, whereby the brightness values R ", G", B "for the LEDs 4 are generated.

4 shows the LEDs 4 of a display panel 1, as well as the pixel grid of an image 10 to be displayed, which is superimposed behind the LED arrangement in order to illustrate the geometric assignment of the pixels 11 to the LEDs 4. The display panel 1 includes a plurality of LEDs 4 in the colors red (R), green (G), blue (B) and yellow (Y) evenly distributed over the surface of the display panel 1. In each case four of the LEDs 4 of different colors R, G, B, Y are arranged square. The colors blue B and green G, or red R and yellow Y lie diagonally opposite each other.

This results in a uniform distribution over the surface of the display unit 1 arrangement of all colors, which is for the quality of the presentation of advantage.

The arrangement of the LEDs 4 is chosen here such that each LED 4 is associated with exactly one pixel 11 of the image 10 to be displayed. The picture 10 has z. B. a resolution of 1024x768 pixels.

In order to display this image 10, a separate brightness value L is calculated for each of the LEDs 4 and the following procedure is carried out: First, a mask 7 is placed around each of the LEDs 4 of a specific color (eg, color green G) comprising a plurality of pixels 11 in the Environment of the associated LED 4 includes. 4 shows an example of such a mask 7 for the color green, wherein for reasons of clarity, the masks 7 of the other green LEDs are not shown.

Thereafter, the color values k of the pixels 11 contained in the mask 7 for the color of the light source 4 (for example, green) are determined, and from these color values, a brightness value L for the LED 4 under consideration is calculated. In principle, any desired downsampling method can be used to calculate the color or brightness value. A preferred method is explained in more detail below with reference to FIGS. 5 and 6 by way of example.

Fig. 5 shows the masks 7 of several adjacent LEDs 4 of the same color (here for the color green G). The individual masks 7 are formed in such a way that the associated LED 4 lies in the center of the mask and the circumference of the mask 7 reaches as far as the boundaries of the pixels 11 which contain a light source 4 with the same color (here green G). In the case of the LED arrangement of FIG. 4, this results in masks 7 with 3 × 3 = 9 pixels 11.

The masks 7 of adjacent LEDs 4 (of the same color) each overlap one pixel row or column. The in the numbers 1,2,4 contained in the middle mask 7 indicate in how many masks 7 the relevant pixel 11 is contained.

In the case of the downsampling method proposed here, the smallest common multiple is now calculated from the individual numbers 1, _{2, 4} (KGV = ₄ ) and this value is divided by the corresponding number 1, _{2, 4} for each pixel 11. This results in the following weighting or transformation mask H for the LED 4 of the color green: $$\mathrm{H}=\left[\begin{array}{ccc}1& 2& 1\\ 2& 4& 2\\ 1& 2& 1\end{array}\right]$$

For the color values k of the pixels 11 contained in the mask 7 for the color green, a color value mask K results here, where: $$\mathrm{K}=\left[\begin{array}{ccc}{K}_{11}& K_{12}& K_{13}\\ K_{21}& K_{22}& K_{23}\\ K_{31}& K_{32}& K_{33}\end{array}\right]$$

The brightness value for this green LED 4 is thus calculated as: $$\mathrm{L}=\mathrm{H}\cdot \mathrm{K}/\mathrm{\Sigma }{\mathrm{H}}_{\mathrm{i}\mathrm{j}}=\frac{{\displaystyle \Sigma H_{ij}\cdot K_{ij}}}{{\displaystyle \Sigma H_{ij}}}$$

This corresponds to the calculation of a weighted average for the color values contained in the mask 7.

The value of the mask multiplication must be divided by the sum of the weighting factors h _ij . The weighting mask H is here preferably represented in a form in which the sum of the entries h _ij one 2 power, so that the otherwise necessary division can be easily replaced by a shift operation.

In the case of the weighting of the individual color values k _ij used here, it is also ensured in particular that all pixels 11 are rated equally. That is to say, with a superposition of all adjacent masks 7 (of the same color), a superposed transformation mask H _sup is obtained for the middle, fat mask 7, in which all entries h _{ij are} identical: $${\mathrm{H}}_{\sup }=\left[\begin{array}{ccc}1+1+1+1& 2+2& 1+1+1+1\\ 2+2& 4& 2+2\\ 1+1+1+1& 2+2& 1+1+1+1\end{array}\right]=\left[\begin{array}{ccc}4& 4& 4\\ 4& 4& 4\\ 4& 4& 4\end{array}\right]$$

The color value k _{ij of} each pixel is thus equally taken into account as a whole. As a result, the image 10 to be displayed can be displayed particularly clearly and evenly.

Fig. 6 shows the masks 7 for the LEDs 4 of the color red. Because of the uniform distribution of all the LEDs 4 over the area of the display unit 1, the masks 7 for the color red have the same size as for the color green. This results in an identical weighting mask H as for the color green. The brightness values L for the red LEDs 4 are calculated in the same way as described above with respect to the green LEDs.

The same method is now also applied to the LEDs 4 of the colors blue B and yellow Y and calculated for each of the LEDs 4, a brightness value L. The image 10 can thus be displayed completely and with high resolution.

Fig. 7 shows the LED arrangement of another display unit 1 and an image 10 to be displayed, which is superimposed behind the arrangement. The display unit 1 in this case comprises LEDs 4 in the three colors R, B, G, each in the form of equilateral triangles are arranged. The LEDs 4 are again distributed uniformly over the surface of the display unit 1.

To calculate the brightness values L for the individual LEDs 4, a mask 7 is again placed around each of the LEDs 4 and the color values k _{ij of} the pixels contained in the mask 7 are weighted for the color of the respective LED 4. The size of the masks 7 is again chosen here so that they reach as far as the boundaries of the adjacent pixels 11, in which an LED 4 with the same color R, G, B is at least partially contained. Therefore, different sizes result for the masks 7 of the colors R, G, B.

FIG. 8 shows the masks 7 for adjacent LEDs 4 of the color red R. Each of the masks 7 comprises 3 × 3 = 9 pixels 11. In the middle, bold mask 7 again values are indicated which indicate in how many masks 7 the relevant one Pixel 11 is included.

In contrast to the LED arrangement of FIG. 4, the individual LEDs 4 are not exactly in the middle of a pixel 11. The red LEDs 4 are, for example, slightly to the right outside the center. In this case, a downsampling method is used in which the color values k _ij contained in the masks 7 are each weighted with a weighting factor h _ij which depends on the distance d _{ij of} the LED 4 under consideration to the center of the respective pixel (see FIG. 9) ). For this a quadratic function is used similar to the bicubic downsampling, where: $${\mathrm{H}}_{\mathrm{i}\mathrm{j}}=\frac{1}{\left(d_{ij}+c^2\right)}$$

Here d _{ij is} the distance of the LED 4 to the center of the respective pixel and c is a constant. With c = 1.4, a transformation mask H for the color red results: $$\mathrm{H}=\left[\begin{array}{ccc}0.124& 0.187& 0.148\\ 0.164& 0.465& 0.220\\ 0.124& 0.187& 0.148\end{array}\right]$$

The constant c is selected such that the individual pixels 11 of a mask 7 - if one also considers the adjacent masks 7 for the same color - are weighted approximately equally. That is, the superposed weighting mask H has about the same size entries. If the middle, bolded mask 7 is superimposed on the adjacent masks 7 in the manner illustrated in FIG. 8, then the standard deviation relative to an average of the weighting factors h _{ij is} minimal for c = 1.4. For the superpositioned transformation mask H: $${\mathrm{H}}_{\sup }=\left[\begin{array}{ccc}0.124+0.220+0.124& 0.187+0.187& 0.148+0.164+0.148\\ 0.164+0.148+0.148& 0.465& 0.220+0.124+0.124\\ 0.124+0.220+0.124& 0.187+0.187& 0.148+0.164+0.148\end{array}\right]=\left[\begin{array}{ccc}0.468& 0.374& 0.460\\ 0.460& 0.465& 0.468\\ 0.468& 0.374& 0.460\end{array}\right]$$

Fig. 10 shows the masks 7 for LEDs of the color blue. The masks 7 again comprise 3x3 = 9 pixels 11 and are formed square. In contrast to the representation of FIG. 8, however, the individual blue LEDs 4 are located slightly to the left outside the center of the masks 7.

FIG. 11 again shows the distances d _ij , this time a blue LED 4, to the centers of the individual pixels 11. The relative arrangement is exactly mirrored to the arrangement of FIG. 9. Therefore, a mirrored weighting mask H results for which applies: $$\mathrm{H}=\left[\begin{array}{ccc}0.148& 0.187& 0.121\\ 0.220& 0.465& 0.164\\ 0.148& 0.187& 0.124\end{array}\right]$$

FIG. 12 shows the masks 7 for the LEDs 4 of the color green G. The masks 7 in this case comprise only 2x3 = 6 pixels 11, since in the lateral adjacent pixels 11 (see FIG. 7) already another LED 4 of the color green G is included.

FIG. 13 shows the distances d _{ij of} a green LED 4 to the centers of the pixels 11 contained in the mask 7.

From the calculation of the weighting factors h _ij according to the above-mentioned formula, the following transformation mask results with c = 1: $$\mathrm{H}=\left[\begin{array}{cc}0.22& 0.22\\ 0.44& 0.44\\ 0.22& 0.22\end{array}\right]$$

The brightness values L for each of the LEDs 4 in the colors R, G, B are in turn given by: $$\mathrm{L}=\mathrm{H}\cdot \mathrm{K}/\mathrm{\Sigma }{\mathrm{H}}_{\mathrm{i}\mathrm{j}}$$

Where H is the associated weighting and K is the associated color mask.

Fig. 14 shows an H-shaped LED array according to another embodiment of a display unit 1 with LEDs in three colors R, G, B. Although the LEDs 4 of a single color R, G, B are evenly distributed over the surface of the display unit 1, the distribution of all LEDs 4 is non-uniform. FIG. 14 also shows, by way of example, the masks 7 for the LEDs 4 of the colors red, green and blue.

15a, 15b and 15c respectively show the distances d _{ij of} the LEDs 4 to the centers of the pixels 11 contained in the mask 7 for the colors red, green and blue.

Fig. 16 shows the masks 7 of a plurality of adjacent LEDs 4 for the color red. With a constant c = 1.9, the following weighting mask results after calculation with the above formula: $$\mathrm{H}=\left[\begin{array}{ccc}0.091& 0.119& 0.091\\ 0.119& 0.277& 0.119\\ 0.091& 0.119& 0.091\end{array}\right]$$

Due to the symmetrical arrangement of the LEDs 4, symmetrical weighting masks H with identical entries also result for the other colors G, B. Because of the unequal distribution of the LEDs 4 of all three primary colors R, G, B over the total area of the display unit 1, however, a coarser image resolution is shown.

The aforementioned weighting mask can in turn be normalized so that the sum of all entries gives 256. In this case: $$\mathrm{H}=\left[\begin{array}{ccc}21& 27& 21\\ 27& 64& 27\\ 21& 27& 21\end{array}\right]$$

In this arrangement, 8x8 pixels 11 of the image require 10 48 LEDs 4.

FIG. 17 shows a further LED arrangement in which the individual LEDs 4 are distributed uniformly over the surface of the display unit 1. Three LEDs 4 of different colors R, G, B are here each arranged in triangles. This arrangement substantially corresponds to the arrangement of FIG. 7, wherein the assignment of the individual pixels 11 of the image 10 to the LEDs 4 has been selected differently. In the present Case lie the individual LEDs 4 respectively in the center of the associated pixels 11. Between two adjacent LEDs 4 is at least one pixel 11, in which no LED 4 is arranged. The horizontal distance between two LEDs 4 is also much greater than the vertical distance. The masks 7 of the individual LEDs 4 are therefore substantially larger and may, for example, comprise 5 × 5 = 25 pixels 11. The size of the masks 7 for the three primary colors R, G, B is chosen identical here.

FIG. 18 shows the overlapping of a plurality of square masks 7 using the example of adjacent LEDs 4 of the color red. Optionally, it would also be possible to use masks 7 which have a smaller extent in the vertical direction than in the horizontal direction.

The calculation of the brightness values L for the individual LEDs is preferably carried out again as a function of the distance according to the above formula h _ij = f (d _ij ).

Fig. 19 shows another alternative mask shape that lends itself well to the downsampling method. As the distance from the center increases, the number of overlaps increases. This is favorable for an accurate and sharp presentation of the image 10.

LIST OF REFERENCE NUMBERS

1

display unit

2

Data processing device

3

LED group

4

LED

5

color transformation

6

mapping function

7

mask

8th

virtual pixel

9

mask

10

image to be displayed

11

Pixels of the image 10

12

Downsampling process

R

red

G

green

B

blue

Y

yellow

Claims (16)

A method of displaying an image (10) consisting of a plurality of pixels (11) on a display unit (1) having a plurality of light sources (4) of one or more colors (R, G, B, Y) whose brightness values (L) for displaying the image (10) are set, characterized in that the following steps for calculating an associated brightness value (L) are performed for each of the light sources (4): Defining a separate mask (7) around each of the light sources (4) comprising a plurality of pixels (11) in the vicinity of the light source (4), Determining the color values (k) of the pixels (11) contained in the mask (7) for the color of the light source (4), and - Calculating the brightness value (L) for the light source (4) from the color values (k). A method according to claim 1, characterized in that prior to calculating the brightness values (L), a color transformation of the image data of the image (10) in the color space determined by the color (R, G, B, Y) of the light sources (4) is performed. A method according to claim 2, characterized in that the image (10) to be displayed is converted into a plurality of partial images in the colors (R, G, B, Y) of the light sources (4) with the same resolution as the image (10). Method according to claim 1, 2 or 3, characterized in that the masks (7) are defined as large the masks (7) of adjacent light sources (4) of the same color (R, G, B, Y) spatially overlap. Method according to one of the preceding claims, characterized in that the color values (k) contained in a mask (7) are weighted with weighting factors (h _ij ) and a weighted mean value is calculated. A method according to claim 5, characterized in that the weighting factors (h _ij ) of a mask (7) are determined such that each pixel (11) of the mask (7) - one also takes into account the weighting factors (h _ij ) of adjacent masks (7) the same color (R, G, B, Y) - is weighted as equally as possible. Method according to one of claims 5 or 6, characterized in that the weighting factors (h _ij ) are a function of the distance (d _ij ) of a pixel (11) from the light source (4). Method according to one of claims 5 or 6, characterized in that the weighting factor (h _ij ) of a color value is a function of the number in how many adjacent masks (7) for the same color (R, G, B, Y) that pixel ( 11), and is not a function of the distance (d _ij ) of a pixel (11) from the light source (4). Method according to one of Claims 6 to 8, characterized in that the sum of all weighting factors (h _ij ) of a mask (7) corresponds to a value which can be represented as a power of 2. Method according to one of the preceding claims 6 to 9, characterized in that a weighting factor (h _ij ) is calculated according to the following function: ${\mathrm{H}}_{\mathrm{i}\mathrm{j}}=\frac{1}{{\left(d_{ij}+c\right)}^2}.$

where d _{ij is} the distance of the pixel (11) from the associated one Light source (4) and c is a constant which is chosen so that all pixels (11) are weighted as equally strong. Method according to one of the preceding claims, characterized in that the size of the masks (7) is chosen so that they contain exactly one physical light source (4) of a particular color whose brightness value (L) is calculated. A display device for displaying an image (10) having a plurality of pixels (11), comprising a display unit (1) having a plurality of light sources (4) of one or more colors (R, B, G, Y) and a data processing device ( 2) processes the image data and calculates brightness values (L) for the light sources (4), characterized in that the data processing means performs (2) following the calculation of the brightness values (L) steps of: Defining a separate mask (7) around each of the light sources (4) comprising a plurality of pixels (11) in the vicinity of the light source (4), Determining the color values (k) of the pixels (11) contained in the mask (7) for the color of the light source (4), and - Calculating the brightness value (L) for the associated light source (4) from the color values (k). Display device according to Claim 12, characterized in that the size of the masks (7) is chosen to be large enough to reach the boundary of the pixels (11) having a light source (4) of the same color (R, B, G, Y) at least partially. Display device according to claim 12 or 13, characterized in that all the light sources (4) are evenly distributed. Display device according to Claims 12 to 14, characterized in that the display unit (1) has light sources (4) in four different colors (R, B, G, Y). Display device according to claim 15, characterized in that the display device has uniformly distributed light sources (4) in four colors (R, G, B, Y), wherein four light sources (4) of different colors are arranged in a square manner.

Images