JP2007279643A - Color correction method of dlp projector - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To correct the difference in color characteristic among respective DLP (Digital Light Processing) projectors by photographing images projected on screens, by a camera to obtain binary images and obtaining a correction area on the basis of their upper, lower, right, and left sides. <P>SOLUTION: In The color correction method, respective screens are lit to pick up images by a camera 1, and the correction area is determined from their binary images on the basis of lateral and vertical ends. Next, an image of a test pattern is projected and is picked up by the camera, and difference between average RGB and reference RGB of the correction area is used as an evaluation function, and a setting parameter of a color having the largest absolute value of difference of RGB is changed in a reverse direction so that a sum of squares of the difference is minimized. The color setting parameter meeting a convergence condition is set to DLP projectors. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a color correction method in a DLP (Digital Light Processing: a registered trademark of TI, USA) projector.

  The DLP projector uses digital light modulation technology using a reflective optical switch panel such as DMD (Digital Micro-mirror Device). The DMD is a semiconductor device having a structure in which extremely small size mirrors (micromirrors) are spread in a matrix, and the pixels in the image frame correspond to the micromirrors. The tilt angle of each micromirror is ON / OFF controlled according to drive data (image code). As a result, the light source lamp light is reflected by the micromirror and projected onto the screen through the projection lens or absorbed by the absorber. In order to project a color image, light source lamp light is irradiated through color switching by a color wheel separated into RGB three colors. In this way, an image based on the image code is displayed on the screen (see, for example, Patent Document 1).

  By the way, when installing a DLP projector, predetermined adjustment work is performed to reduce the difference in projection image quality due to equipment installation / use environment (screen characteristics, influence of room lighting, etc.) and aging of the projector, and to perform appropriate image projection. It is necessary to do. However, if there is human intervention in this work, it cannot be made objective and man-hours increase.

  Also, it tends to be inferior in comparison with conventional CRT television monitors in terms of color reproduction of natural moving images. In addition, a large display panel of a multi-screen device combining a plurality of PLD projector devices has different color characteristics of each DLP, and the displayed color and brightness differ even when the color setting parameter setting values are the same. Therefore, it becomes unnatural when the entire screen is viewed with human eyes.

  Therefore, DLP has a sensor that picks up the negative image when the micromirror is OFF, considering that the upper limit of the light quantity of the light source and color reproduction based on a certain range from there to black with zero light quantity. Thus, the color reproduction of a natural moving image is improved by performing white color correction (white balance) at the maximum light amount, which is extremely important for color reproduction, based on image information obtained by the sensor (for example, Patent Document 2). reference).

  In addition, images that continuously change from full-screen white to full-screen black are projected onto the screen, captured by the photographic lens, and the spectrum-resolved light is captured by the image sensor, and the intensity of each wavelength is detected from the output. By analyzing, the output of the light bubble is corrected to reduce the difference in projection image quality due to the secular change of the DLP projector (for example, see Patent Document 3).

Japanese Patent Laid-Open No. 2004-163876 (pages 4-5, FIG. 11) JP 2005-123841 (pages 6-8, FIG. 1) Japanese Patent Laying-Open No. 2005-099478 (pages 4-5, FIG. 1)

  However, the technique described in Patent Document 2 described above is a method for automatically correcting white, and has a first problem that colors other than white cannot be corrected. In addition, in the case of a multi-screen device, it is necessary to mount an image sensor in units of screens, and there is a second problem that it is necessary to correct this because variations in sensors are assumed.

  Furthermore, there is a third problem that the correction area in the screen depends on the mounting position of the sensor, and it is not easy to change the correction area according to the external environment such as external light or the installation position of the DLP projector.

  In the technique described in Patent Document 3 described above, the image projected on the screen is captured by the photographing lens and used to correct the output of the light bubble. This is the color characteristic of each DLP projector in the multi-screen device. It does not correct the difference.

  An object of the present invention is to solve all the above-mentioned problems and to provide a method for correcting a difference in color characteristics of each DLP projector in a multi-screen device.

  The color correction method for a DLP projector according to the present invention is a color correction method for a DLP projector having a multi-screen configuration, a first stage for obtaining at least one correction area for each screen, and a second stage for projecting a test pattern on the multi-screen. The third step of reading the test pattern image in the correction area with the camera, the fourth step of calculating the RGB average values of the image read by the camera in correspondence with the screen, and the difference between the average value and the RGB reference values Repeat step 2 to step 4 for each screen so that the sum of the squares of the two is minimized, and set each RGB color setting parameter for each DLP projector when a predetermined convergence condition is met. And a sixth stage.

  The fifth step includes a first step of adding the difference between the average value and the reference value in correspondence with the screen for the correction region, and a second step of changing the setting parameter for the color having the largest added value in a direction in which the difference is reduced. Have.

  In the fifth step, the first step of storing the sum of squares of the difference between the average value and the RGB reference values in the memory and the first step while changing the R color setting parameter within a predetermined range are repeated. 2 steps, a third stage in which the first step is repeated while changing the G color setting parameter in a predetermined range, a fourth step in which the first step is repeated while changing the B color setting parameter in a predetermined range, and a second step You may make it have the 5th step which repeats the said step-the said 4th step.

  The first stage includes a first step in which each screen is blacked on the outer periphery and sequentially displayed in white, a second step in which an image is read with a camera, and binarization of the read image by binarization. A third step of extracting a predetermined area as a detection area on the basis of up, down, left, and right adjacent to another screen in the image, and a fourth step of setting a predetermined area at the center of the detection area as a correction area.

  The first effect of the present invention is to turn on each screen constituting a multi-screen, take an image projected on the screen with a camera to obtain a binarized image, and obtain a correction area based on the upper, lower, left, and right sides. Therefore, the difference in the color characteristics of each DLP projector in the multi-screen device can be corrected without depending on the external environment or the installation position of the DLP projector.

  The second effect of the present invention is that the correction area can be easily changed by designating, so that various multi-screen (large display panel) configurations can be supported.

  The third effect of the present invention is that it is possible to optimize colors that do not depend on nonlinear characteristics of color setting parameters and variations of each DLP projector. Since the color setting parameters of DLP projectors are non-linear, accurate optimization is not possible with ordinary least squares methods (sequential optimization methods such as the gradient method and Newton method), and annealing methods can also be used. The calculation cost is huge and impractical.

  However, in the present invention, the color setting parameter of the DLP projector having the largest absolute value of the difference as an evaluation function using the difference between the average RGB value obtained by taking an image of the correction area test pattern with the camera and the reference RGB value. Since the color setting parameter when the evaluation function is minimized is adopted, the color can be accurately corrected even if the color output from the DLP projector is not linear with respect to the set value. is there.

  Next, embodiments of the present invention will be described in detail with reference to the drawings. When unevenness is large in one screen screen, multi-screen color correction is performed. FIG. 1 shows an apparatus for executing a color correction method in a DLP projector according to the present invention. This apparatus is used to correct a difference in color characteristics of n DLP projectors # 1 to #n in a multi-screen apparatus composed of n DLP projectors.

  Since the DLP projector and the screen have a one-to-one correspondence, images are projected onto the n screens by the n DLP projectors. A camera 1 is installed in front of the screen so that images on the entire screen can be taken. The DLP projectors # 1 to #n and the camera 1 are connected to a personal computer (PC) 2 and controlled. That is, the PC 2 controls the operations of the DLP projectors # 1 to #n and the camera 1 to project a correction image on the screen, obtains an image from the camera 1, performs a calculation process, and based on the result, the DLP projector # The correction parameters are set to 1 to #n.

  FIG. 2 is a flowchart showing the procedure of this color correction method. First, the camera 1 is set at a predetermined position, and the PC 2 is connected to the DLP projectors # 1 to #n to prepare for the color correction operation (step S1 in FIG. 2). Here, in order to perform color correction, there is a big problem of which part of the screen is to be corrected. In this regard, it is known that when there is a color / brightness difference at the boundary of the screen, human vision is easy to feel. Therefore, an area near the boundary of each screen (referred to as a detection area) is detected, and the center of the detection area is set as a correction area.

(1) Extraction of detection area and setting of correction area For example, in the case of a 2 × 2 screen configuration, as shown in FIG. 3, hatched portions A, B, C, and D are used as detection areas. This area is detected (step S2 in FIG. 2), which is executed by a program in the PC2. The procedure will be described in detail below.

  First, the variable x = 1 corresponding to the screen numbers # 1 to #n is set, the display pattern of the DLP projector x is white (RGB numerical values are all 255), and the display pattern of the other DLP projectors is black (RGB numerical values). Are all set to 0). Since the DLP projector projects such a display pattern on the screen, the DLP projector takes a picture with the camera 1 and loads it into the PC 2. In this case, in order to reduce the influence of noise and diplomacy, five images are taken and the time average is taken.

  Color correction is performed on an image with a white display pattern, but it is desirable to use a high-quality portion of the image. Therefore, in order to eliminate the peripheral portion of the image, 5 pixels are made black on each of the upper, lower, left and right sides of the image. . Then, the RGB numerical value is converted into a luminance value Y by the following equation for each pixel.

Y = 0.3R + 0.59G + 0.11B
Next, the luminance value Y is binarized (0 and 255), and areas other than the largest connected area of pixels in the image are erased. This eliminates noise.

  In the binarized image obtained as described above, the centroids of the detection areas A, B, C, and D are obtained. First, in order to detect the maximum value and the minimum value of the u direction (direction from the left to the right of the image), the u direction of the binarized image is scanned from the left (the smaller number), and the u coordinate Count the number of “255” in the value. Then, the difference Dui (i = 0 to width) from the previous scan is calculated. Width represents the width of the camera 1 image. As a result, the u coordinate value at which Dui is the maximum value is the minimum value (min_u) of the correction region, and the u coordinate value at which Dui is the minimum value is the maximum value (max_u) of the detection region. This means that in the horizontal direction, the position where the image goes from black to white is min_u, and the position where the image goes from white to black is max_u.

  Further, in order to detect the maximum value and the minimum value of the v direction (from the top to the bottom of the image) coordinates of the binarized image obtained as described above, the u direction of the binarized image is moved upward (numerical). And the number of “255” in the v coordinate value is counted. Then, a difference Dvi (i = 0 to height) from the previous scanning is calculated. height represents the height of the camera 1 image. As a result, the v coordinate value at which Dvi is the maximum value is the minimum value (min_v) of the correction region, and the v coordinate value at which Dui is the minimum value is the maximum value (max_v) of the correction region. This means that in the vertical direction, the position where the image goes from black to white is min_v, and the position where the image goes from white to black is max_v.

  Now, assuming that the short side of the rectangular detection region in FIG. 3 is DT1, the long side is DT2, and the offset value is Mg, the coordinates of the center of gravity of the detection regions A, B, C, and D can be expressed as follows. Note that the values of DT1, DT2, and Mg are set to appropriate values by visual evaluation.

A: (uc1, vc1) = (max_u-Mg-DT1 / 2, vc)
B: (uc2, vc2) = (uc, min_v + Mg + DT1 / 2)
C: (uc3, vc3) = (min_u + Mg + DT1 / 2, vc)
D: (uc4, vc4) = (vc, max_v-Mg-DT1 / 2)
As a result, since the coordinates of the center of gravity of the detection areas A, B, C, and D are obtained for the screen # 1, x is incremented by one and the above processing is repeated until x = n. Also for # 4, the coordinates of the center of gravity of the detection areas A, B, C, D are obtained.

  FIG. 3 shows a case where detection areas A, B, C, and D are detected on the screen # 1. ID numbers are assigned to the detection areas A, B, C, and D, and the detection areas A, B, C, and D are in the counterclockwise order with the right end as a reference. This is also performed for screens # 2 to # 4. In the case of the 2 × 2 screen configuration, the detection areas B and C on the screen # 1 are not in contact with other screens, and as described above, human vision is not affected by the color / brightness difference at the boundary of the screen. Because it is sensitive and deviates from the idea of making the boundary part a detection area, the detection areas B and C are excluded. FIG. 4 shows the final detection area of all the screens thus obtained.

  The detection area is visually confirmed on the screen (step S3 in FIG. 2). As a result, for example, when the extraction state is abnormal, such as when a person crosses in front of the camera 1, the process returns to step S1 and the correction area detection process in step S2 is performed again. If the extraction state is normal, the center of each detection area is set as a correction area.

  If the unevenness in one screen is small, as shown in FIG. 5, the center of gravity coordinates of the screen are obtained by following the procedure of step S2, and predetermined square areas a, b, c centered on the center of gravity coordinates are obtained. , D are set as detection areas, and the center of the detection areas a, b, c, d is used as a correction area to perform multi-screen color correction. The detection areas A, B, C, and D in FIG. 3 and the detection areas a, b, c, and d in FIG. 5 are switched by, for example, key operation of the PC 2 according to the state of the screen.

(2) Setting the standard RGB Now what is the standard for correction? Reference RGB is defined and set (step S4 in FIG. 2), and each DLP projector is adjusted. A DLP projector has, as a default, setting parameters for changing the color, brightness, contrast, and the like of an image. Here, in the case where the screen is bright and dark, the darkest color setting parameter among the defaults of each DLP projector for each RGB is set as a reference value. This is because, when adjusting the color, it is easy to darken, but it is difficult to brighten due to the characteristics of DLP. For example, in the case of a 2 × 2 DLP projector, the default RGB of each DLP projector is as follows.

DLP1 (RGB) = (100,85,75)
DLP2 (RGB) = (120,95,85)
DLP3 (RGB) = (150,45,150)
DLP4 (RGB) = (110,23,184)
In this case, the reference Rref, Gref, and Bref are as follows.

Rref, Gref, Bref = 100,23,75
The reference RGB determined in this way is input by a key operation of the PC 2, for example, and can be used by a program.

(3) Execution of color correction Next, color correction is performed using a test pattern (steps S5 to S12 in FIG. 2). The program executed on the PC 2 displays a test pattern on the multi-screen (Step S5 in FIG. 2). The test pattern is a unique pattern output from the DLP projector. For example, the gray 80 test pattern displays a gray image close to white, and the gray 50 test pattern displays a substantially gray image. Here, gray80 is used to optimize the parameters when the screen is bright, and gray50 is used to optimize the parameters when the screen is dark.

  The camera 1 reads the image of the test pattern gray 80 displayed on the entire multi-screen and transmits it to the PC 2. In PC2, a variable N = 0 indicating the number of times of color correction is set (step S6 in FIG. 2), and color correction is performed so that the color displayed on each screen matches the reference color (step S7 in FIG. 2). This color correction is performed in parallel for each DLP projector.

The color correction program executed on the PC 2 defines an evaluation function J as shown in the following equation in order to optimize the color correction.

  In this equation, n is the number of correction regions, which is 2 in the example of FIG. 4 and 1 in the example of FIG. Further, i is the number of the correction area, Ri, Gi, and Bi are the average values of R, G, and B of the correction area i. The correction area i is extracted from the gray 80 image transmitted from the camera 1. It is obtained by calculation.

Color correction is optimized by minimizing the value of the evaluation function J. For this purpose, evaluation functions Jr, Jg, and Jb are defined for each of R, G, and B in the program.

  The absolute values of the evaluation functions Jr, Jg, Jb are calculated by the above formula, and the largest setting parameter among them is changed. For example, when the absolute value of the evaluation function Jr is the largest, the R parameter is changed. The direction of change is opposite to the sign of the evaluation function Jr. That is, if Jr <0, the R parameter is incremented by 1 for example, and if Jr> 0, the R parameter is incremented by −1. 1 of ± 1 does not need to be 1, and may be 1 after starting from a slightly large value and becoming close to convergence. Next, when the absolute value of Jg becomes the largest, the same process as the process for the R parameter is performed for the G parameter.

  When the above color correction is performed for the variable N = 0, N + 1 → N (step S8 in FIG. 2), the convergence condition is satisfied (step S9 in FIG. 2), or the number limit is reached (step in FIG. 2). The color correction (step S7 in FIG. 2) is repeated until S10). That is, an image of the test pattern gray 80 displayed on the entire multi-screen is read according to the changed color setting parameter and transmitted to the PC 2. The PC 2 calculates absolute values of the evaluation functions Jr, Jg, and Jb so that the color displayed on each screen matches the reference color, and changes the largest setting parameter among them. In this process, the value of the evaluation function J is calculated from Ri, Gi, Bi at that time. The convergence condition is that the value of the evaluation function J is not more than a predetermined value.

  The above is the optimization of the color correction by the test pattern gray80, but the same color correction is performed for the test pattern gray50 (steps S4 to S10 in FIG. 2). Then, when color correction is performed for all of the specified test patterns (in this case, gray80 and gray50) (step S11 in FIG. 2), it is finally visually confirmed whether color correction has been completed (step S12 in FIG. 2). If not, start again from the reference color setting (step S4 in FIG. 2).

(4) Execution of further color correction The color correction described in (3) is performed while paying attention to any one of R, G, and B for any test pattern. Is performed for a plurality of combinations of R, G, and B in order to combine both test patterns and perform color correction with higher quality.

  First, the color correction program executed on the PC 2 loads the test pattern gray80 to all DLP projectors and changes the R parameter within a predetermined range. Each time, the image of the test pattern gray80 displayed on the entire multi-screen is read by the camera 1 and transmitted to the PC 2, and the value of the evaluation function J is calculated. All the values of the evaluation function J are stored in the memory.

  Next, the test pattern gray50 is loaded, and the value of the evaluation function J is calculated by changing the R parameter within a predetermined range in the same manner as described above. The sum of the value of the evaluation function J calculated here and the value of the evaluation function J stored in the memory is taken and stored in the memory.

  Next, the test pattern gray80 is loaded again, the G parameter is subjected to the same processing as that performed for the R parameter, and is overwritten and saved in the memory. Also, the test pattern gray50 is loaded again, and the same processing as that performed for the R parameter is performed for the G parameter, and is overwritten and saved in the memory.

  Further, the test pattern gray80 is loaded again, and the same processing as that performed for the R parameter and the G parameter is performed on the B parameter, and is overwritten and saved in the memory. Further, the test pattern gray50 is loaded again, and the same processing as that performed for the R parameter and the G parameter is performed on the B parameter, and is overwritten and saved in the memory.

  In the process described above, the value of the evaluation function J is calculated, and the process is repeated until a predetermined satisfaction condition is satisfied or the limit number of times is reached.

  In any of the methods (3) and (4), the R parameter, the G parameter, and the B parameter that give the value of the evaluation function J when the processing is completed are the optimum color parameters. The optimized color parameters are stored in each DLP projector, and the code data at the time of practical use is displayed in color on the screen by the color parameters. As described above, each DLP projector is color-corrected with the RGB color collected by the camera 1 from each correction area under the same reference RGB, so that even if each DLP projector has a difference in color characteristics, it is appropriate. So that it is not as unnatural as possible with the human eye.

The figure which shows the apparatus for performing the color correction method in the DLP projector of this invention The flowchart which shows the procedure of the color correction method of this invention Diagram illustrating the detection area on the screen The figure which illustrates the detection area in the multi-screen Another diagram illustrating a detection area in a multi-screen

Explanation of symbols

1 Camera 2 PC
A to D detection area

Claims (4)

In the color correction method in a multi-screen DLP projector,
A first step for determining at least one correction area for each screen;
A second stage of projecting a test pattern on a multi-screen;
A third stage of reading an image of the test pattern in the correction area with a camera;
A fourth step of calculating an average value of each RGB of the image read by the camera in correspondence with the screen;
A fifth stage in which the procedures of the second stage to the fourth stage are repeated in correspondence with the screen so that the sum of the squares of the difference between the average value and the RGB reference values is minimized;
A color correction method for a DLP projector, comprising: a sixth step of setting each RGB color setting parameter in each DLP projector when a predetermined convergence condition is satisfied. The fifth stage includes
A first step of adding a difference between the average value and the reference value for the correction area in correspondence with a screen;
2. The color correction method for a DLP projector according to claim 1, further comprising a second step of changing a setting parameter for the color having the maximum added value in a direction in which the difference is reduced. The fifth stage includes
A first step of storing in a memory a sum of squares of differences between the average value and the RGB reference values;
A second step of repeating the first step while changing the R color setting parameter within a predetermined range;
A third stage in which the first step is repeated while changing the G color setting parameter within a predetermined range;
A fourth step of repeating the first step while changing the color setting parameter of B within a predetermined range;
The color correction method for a DLP projector according to claim 1, further comprising a fifth step of repeating the second step to the fourth step. The first stage includes
A first step in which each screen is blacked on the outer periphery and sequentially displayed in white;
A second step of reading the image with a camera;
A third step of binarizing the read image and extracting a predetermined area as a detection area on the basis of upper, lower, left and right adjacent to another screen in the binarized image;
4. The color correction method for a DLP projector according to claim 1, wherein the fourth step is to set a predetermined area at the center of the detection area as a correction area.

Images