TW200818114A - System and method for automated calibration and correction of display geometry and color - Google Patents

Abstract

Various embodiments are described herein for a system and method for calibrating a display device to eliminate distortions due to various components such as one or more of lenses, mirrors, projection geometry, lateral chromatic aberration and color misalignment, and color and brightness non-uniformity. Calibration for distortions that vary over time is also addressed. Sensing devices coupled to processors can be used to sense display characteristics, which are then used to compute distortion data, and generate pre-compensating maps to correct for display distortions.

Description

200818114 IX. Description of the invention: I: Technical field to which the invention belongs 3 Priority claim This application claims the application of the prior US Provisional Patent Application No. 60/836,940 and May 11, 2007 on August 11, 2006. Priority is claimed in U.S. Provisional Patent Application Serial No. 60/836,940. FIELD OF THE INVENTION Various embodiments are discussed with reference to calibration of display devices. L·Tltr 10 BACKGROUND OF THE INVENTION Most image display devices exhibit some form of geometric or color distortion. Such distortions can have a variety of factors, such as geometric background, non-ideal properties of various optical components in the system, under-alignment of various components, complex display surfaces and optical paths that cause geometric distortion, and flaws within the 15 panels, etc. . Depending on the system, the amount of distortion can vary greatly from undetectable to very annoying. The effects of the aforementioned distortions may also vary, and may result in changes in the color of the image, or changes in the shape or geometry of the image. SUMMARY OF THE INVENTION 3 In one feature, at least one embodiment described in this specification provides a display calibration system for use with a display device having a viewing surface. Such a display calibration system includes: at least one sensing device adapted to sense information relating to at least one of a shape, a scale, a 5200818114 boundary, and an orientation of the viewing surface; and at least one processor And coupled to the at least one sensing device and adapted to calculate characteristics of the display device based on information sensed by the at least one sensing device. In another feature, at least one embodiment described in this specification provides a display calibration system for use with a display device having a viewing surface. Such a display calibration system includes: at least one sensing device adapted to sense information of a test image displayed on the viewing surface; and at least one processing coupled to the at least one sensing device The at least one processor is adapted to calculate display distortion based on the sensed information. These precompensation maps can be implemented by surface functions. When the pre-compensation map is applied to the input image data before display, a display image formed on the viewing surface is substantially free of distortion. In another feature, at least one of the implementations illustrated in this specification provides a display calibration system for use with a display device having a viewing surface. The display calibration system includes: at least one image sensing device adapted to sense information from a test image displayed on the viewing surface; and at least one coupled to the at least one image sensing device The processor, the at least one processor, is adapted to cause 20 to calculate display distortion based on the sensed information, so that the viewing surface is divided into small pieces according to the strictness of display distortion in each small piece, and A precompensation map relating to display distortion within each patch is generated to provide substantially no distortion of the displayed image formed on the viewing surface when the precompensation map is applied to the input image material prior to display. 6 200818114 In another feature, at least one embodiment described in this specification provides a display calibration system for use with a display device having a viewing surface. Such a display calibration system includes: at least one image sensing device adapted to independently sense color information relating to at least one color component of a test image displayed on the viewing surface 5; and at least one coupling a processor to the at least one image sensing device, the at least one processor adapted to calculate color non-uniformity based on the sensed information, and to generate at least one color correction map associated with the at least one color component Therefore, when the at least one color correction map, 10 is applied to the input image material before the display, a display image formed on the viewing surface is substantially free of at least one color unevenness. In another feature, at least one embodiment described in this specification provides a display calibration system for use with a display device having a viewing surface. The display calibration system includes: at least 15 image sensing devices adapted to sense information of individual color component test images displayed on the viewing surface; and at least one coupled to the at least one image a sensing device and a processor of the display device, the at least one processor adapted to independently calculate geometric display distortion based on the sensed information, and to independently generate at least one color component having at least 20 levels A pre-compensation map, such that when the at least one color correction map is applied to the input image material before display, a display image formed on the viewing surface is substantially free of at least one color dependency geometric distortion. In another feature, at least one embodiment described in this specification provides a display calibration method that can be used in a projection system 7 200818114 having a curved viewing surface, the method comprising the steps of: using multiple a projector that projects a different portion of an image onto a corresponding portion of the curved viewing surface; and causes each portion of the image to be substantially focused on a corresponding portion of the curved viewing surface 5 to maximize the image The focus of the optimisation is integrally formed on the curved viewing surface. In another feature, at least one embodiment described in this specification provides a display calibration method that can be used in a projection system having a curved viewing surface, the method comprising the steps of: 10 measuring from the curved viewing a plurality of distances from the surface to the focal plane of the projected image; and offsetting the focus plane until the functions of the plurality of distances are minimized to obtain an optimized focus. Brief Description of the Drawings 15 For a better understanding of the embodiments and/or related implementations of the present specification, and for the purpose of clarity, it will be understood that At least one exemplary embodiment and/or related implementation is shown: Figure 1 is a simplified diagram of an exemplary embodiment 20 of an automated calibration and calibration system; and Figures 2a and 2b are diagrams of a curved surface geometry; 3 is an illustration of an example of overflow, underflow, and mismatch in geometric distortion, and Figure 4 is an illustration of an example of a calibration image test pattern; 8 200818114 Figure 5 is a calibration geometry and involved An illustration of various coordinate spaces; Figure 6 is an illustration of an exemplary embodiment of a calibration data generator, 5 Figure 7 is an illustration of an optimization of the scale and origin; Figure 8 is a multiple An illustration of an exemplary embodiment of a color calibration data generator; Figure 9 is an illustration of a setting related to color unevenness calibration; Figure 10 is a color unevenness correction FIG. 11 is an illustration of an exemplary embodiment of a warpage data generator; FIG. 12 is an illustration of a display correction related tile segmentation; Figure 13 is an illustration of an exemplary embodiment of a digital warping unit; Figure 14 is a schematic view of the arrangement of the shape and relative orientation of the viewing surface; Figure 15 is an illustration of a defocusing test pattern. Figure 16 is an illustration of a focus test pattern; 20 Figure 17 is a partial illustration of an exemplary embodiment of a calibration system consisting of a multi-projector and a curved screen; Figure 18 is a Partial illustration of an exemplary embodiment of a calibration system comprising a multi-projector and a curved screen of Figure 17 showing the focus plane of different projectors; 9 200818114 Figure 19 is a minimization of a distance function An illustration of an example of a focus technique; Figure 20 is another image of a multi-projector and a curved screen that adjusts the position of the projector to make the image Partial illustration of an exemplary embodiment of a focus optimized calibration system; FIG. 21 is a partial illustration of an exemplary embodiment of a calibration system using multiple cameras; Partial illustration of an exemplary embodiment of a rear projection television (RPTv) of a self-calibrating display that allows self-calibration and dynamic distortion correction; Figure 23 is a calibration system with multiple projectors and multiple sensing devices Partial illustration of an exemplary embodiment; Figure 24 is a partial illustration of an exemplary embodiment of a calibration system using physical edges and boundaries of the viewing surface; 15 Figure 25 is a use of a focusing technique Partial illustration of an exemplary embodiment of a calibration system that determines the shape of a curved display screen; and the use of a focusing technique to determine the shape of a wavy eight m screen A partial diagram of an exemplary embodiment of a barebones system is shown. I: EMBODIMENT OF THE PREFERRED EMBODIMENT The detailed description of the preferred embodiment is to be understood that the simplicity and clarity of the illustrations are considered to be appropriate. The number of test numbers may be repeated between the figures to indicate 200818114. Corresponding or similar components. In addition, numerous specific details are set forth to provide a thorough understanding of the embodiments and/or embodiments disclosed herein. However, it will be understood by one of ordinary skill in the art that the embodiments and/or implementations described herein may not require such specific details. In other instances, some of the methods, procedures, and components of the present invention are not described in detail, and are not intended to confuse the embodiments and/or implementations described herein. In addition, the description should not be taken as limiting the scope of the embodiments described in the specification, but rather the structure and/or operation of the embodiments illustrated in the specification. 10 Some of the important distortions associated with some display devices include distortion caused by the lens assembly, distortion from the mirror (curved or planar) reflection group, distortion caused by projection geometry, such as off angle And rotary projection (trapezoidal 'rotation) and projection on the curved screen; lateral chromatic aberration and distortion for each color, such as under-alignment and 15 misconvergence in multiple microdisplay devices; color and luminance ( Distortion due to brightness) unevenness, and optical focusing problems (spherical aberration, astigmatism, etc.). The first group sees the geometric distortion in the final image, that is, the shape of the wheeled image is not preserved. Chromatic aberration is also a geometric distortion; however, the distortion varies with each color component. Such distortions are common in projection (front or rear) 20-type display devices, and collectively referred to as geometric distortion. Chromaticity and brightness non-uniformity affects all display devices, so a signal that means a fixed brightness or chromaticity is seen to vary across the surface of a display device, or is different from its expectations. feel. This type of distortion may be caused by a variety of varying brightness, optical path lengths across display 11 200818114, and panel (eg, LCD, UX) S, eMule) Caused by the response. Focusing on the associated distortion blurs the image and causes the different points on the object plane to be focused on different image planes. This specification refers to the implementation of the financial, (4) is a discussion related to the focus and depth of focus 10 15 20 The embodiments presented in this specification illustrate the device to eliminate or reduce at least some of the above-mentioned distortions: = law. This embodiment automates both the calibration data and the resulting corrections and the application of the calibration. Also aimed at the long-term changes in their short-term changes related to the filming ^ delete paragraph (production _ system (four): characterization of the display is not fast, shooting the display device above the special test style observed, for example, "through a similar The sensing device of the high-resolution camera; and the data required for the removal of the image (four) from this (four). The calibration phase involves the electronic correction of the image, which causes the image to be pre-deficient and promotes the top of the image. The distortion-free image is also proposed as a mechanism for achieving the best test pattern for focusing and shooting. Figure 1 is a simplified diagram showing an exemplary embodiment of an automated calibration and calibration system. It can correct images of (6) on the viewing surface 16 of a display device. Such an automated calibration and calibration system includes: a test image generator 14, a sensing device 11, a calibration data generator 12, a warp generator b And a digital warping unit 15. The display device may be a television (rear projection television, LCD, plasma, (four)... a front projection system (also I3', with a projection II of the screen), or any other The current image system 12 200818114: There is a = the film to see the surface, see the surface =-= the boundary or frame that is different from the background will be a surrounding crane (four) (four) (viewing the surface) the boundary is not necessarily a frame or a certain physical form. Usually, a boundary = any connection with the entity viewing table (4) can be linked by a certain type of background 10 15 ==. For example, a display is projected onto the display by the display. And the rectangular outline displayed inside the physical screen frame can be confirmed as the boundary. In the exemplary embodiment of the present invention, the viewing surface 16 is regarded as a calibration and correction point. The actual display device is located in an area within the identified boundary of the frame itself, at least in some cases. The boundary is also referred to as the viewing surface bezel shown as the n-view surface 16 in the first towel. For a curved surface with varying depths, the display can take two main observation points. The viewing plane can be thought of as a focus plane to make the image the correct form 'which may differ from the entity above Looking at the surface 16, or only a portion of the body of the viewing surface 16. All points on the focal plane have the same depth of focus. In this case, the entity of the sensing device (i.e., the observer) The mark or field of view will determine the focal plane boundary (see Figure 2a). When the viewing surface bezel is available, the system 20 is used to determine the orientation of the camera relative to the viewing surface 16. Alternatively, the entire screen may be Viewing, so that the physical screen frame forms the above-mentioned curved boundary (see Figure 2b). Here, the different points on the screen have different water coke/juju. The calibration and correction operation is focused on In order to match the final image with the curved boundary. 13 200818114 Five observation points such as 5H can be combined to identify the different display areas of the calibration and the correctness. For example, the boundary can be a combination of upper and lower physical screen frames, except for the image rims captured above, which are outside the particular focus plane. A curved boundary may also be forced to cry on a flat display by smashing a curved-shaped veranda. This can be regarded as a special case in which the boundary is curved. Shape, but the screen itself is flat, that is, has an infinite radius of curvature. 10 15 20 For some distortions involving changes in shape or geometry, view the image viewed on surface 16 (in correction) Previously, it may be completely included (overflow). This is shown in Figure 3. In the case (hook, the shadow $ABCD overflows and can completely contain the surface frame 18 of the viewing, and in the case In the case, the image is completely covered (underflow). The situation ((〇 is a middle case (mismatch), wherein the image portion is partially covered by the viewing surface 16 described above. All three cases may be from the front Or caused by the rear projection system, and can be corrected by the system. The test image generator 提供 provides some images containing special patterns designed for the calibration program; these images are also called calibration test patterns. The calibration test styles that can be used most often are: • Rule (non-joined) 样式 line style, circle, square, horizontal, and; degree = line, concentric style, rectangle and above The color shirt version can be used for lateral chromatic aberration: positive = uniform sentence correction. The shapes in these styles are also called the shape of the shape and shape, and have their well-defined shape characteristics. That is, the position, scale, boundary, color, and any other 14 200818114 defined parameters are known. Several exemplary correction patterns are shown in panels (4) to (4) of Figure 4. They are used The guidelines (center position, radius, etc.) showing these characteristics are not part of the test pattern. The color and shape of these test patterns are also used to exchange black and white, replaced by color. Black and white, different colors related to different shapes in the pattern, different shapes in a pattern, and changes in grayscale and chromaticity. These versions using the style of the primary color are used to correct Lateral chromatic aberration. An example The color style is displayed in the panel (the magic, in which the 10 horizontal lines, vertical lines, and their intersections, all of the different colors. Each style presents some specific characteristics, the most notable of which The center position of the shapes and their boundaries, which are mathematically considered to be points and lines, respectively. The sensing device 11 can record the calibration test seen on the viewing surface 16 To correct the geometric distortion, the sensing device 11 may be a camera. The resolution and shooting format of the camera can be selected according to the accuracy required in the above correction. When the unevenness of the chromaticity and brightness is corrected The sensing device 11 may be a color analyzer (for example, a photometer or a spectrometer). In this exemplary embodiment, to correct geometric errors, the sensing device 20 can be placed relative to Any position of the above display device. Such a degree of freedom in the positioning sensing device 11 may be based on the fact that the camera image can be accommodated by the sensing device 11 to allow for the inclusion of distortion components. Unless the sensing device 11 directly views the viewing surface 16 (i.e., directly above), there will be a trapezoidal component due to the sensing device. This 15 200818114 variant may occur in up to three axes, which are considered to be multi-axis trapezoidal distortion components. Furthermore, since the optical device 'such as a camera of the sensing device u has its own distortion', there is also an optical distortion component that is taken into consideration. Other types of sensing devices 11 have other inherent distortions. The combined distortion introduced by the above camera or sensing device 11 will be referred to as camera distortion. This age machine distortion is compensated by mosquitoes when generating the amount of clothing. To account for the camera distortion described above, in at least one exemplary embodiment, physical reference marks are used, and their undistorted orientation/shape 10 is known. These markers are captured by the camera and the camera distortion described above can be determined by comparing their orientation/shape in the captured image to their undistorted orientation/shape. A natural mark is the border (boundary) itself, which is known to be a fixed orientation and shape 15 (U is a distortion-free rectangle in the real world). The frame is also a reference for the school to operate 70%. In other words, the corrected image should be linear with respect to the frame. Therefore, when correcting geometric distortion, the image captured by the above machine should include such viewing glory borders (ie, frame 18) 〇20 in another exemplary embodiment in which the boundary is undetectable, j The sensor in the machine is used to sense the signal from the above-mentioned transmission = ' to determine (4) the camera distortion of the viewing surface 16. This measurement will yield p maps of the viewing surface 16 as seen by the camera. When Tian Haozheng Maotian and chromatic aberration, the camera will capture κ group images, where 16 200818114 κ is the number of color components, for example, the three primary colors are deleted. At least some of the test patterns in Figure 4 will be repeated for each color component. The brightness and color (brightness and chrominance) corrections are complete, regardless of the geometric correction correlation. In some projection systems, these brightness and color corrections are done after the correction of geometric distortion. In a flat panel display device in which geometric distortion is not present, brightness and color correction are directly accomplished. In an exemplary embodiment, a sensory device, such as a color analyzer, is placed directly at or near the viewing surface 16 for color information. In this case, the above-mentioned correction related to the location of the shock is not necessary. The sensing device 11 may capture the entire image or the assets at a particular point. In the latter case, data from the grid lines above the screen needs to be taken. If the sensing device 11 is in a trapezoidal position relative to the viewing surface 16, it is similar to the camera above, and its correction due to positioning needs to be completed. 15 For some display devices with geometric distortion, brightness and color correction should be done after the geometric correction has been completed. This means that the display device is the first to include the geometric distortion of the color dependent person. The geometrically corrected color-related correction allows for any additional color distortion introduced by the above geometric correction to be taken into account, and to ensure that the area containing the last image (i.e., no background) is corrected. In this exemplary embodiment, the calibration data generator 12 can analyze the images, and can retrieve the data in the format used by the warpage generator 13, which can then provide warpage data. For the digital warping operation, the digital warping operation can usually be described, so that the mathematical transformation between the coordinates of the wheel-compensated image is performed according to equation (1). Don't enter image coordinates and output shadow (1) In equation (1), cover the space coordinates of the input pixels, and give the above-mentioned mapping to the round. Space J: the color of the pixel 'mark' and e; given the space dome of the corresponding pixel wheel excellent / pixel, ί is only an RGB value. The equation is gone. In the form of a three-primary color system, it is in the form of a grid line. The format of the above correction is difficult. The correction is necessary to use a grid line such as the 60 Hz frame rate for video. Therefore, the ▲ method is applied to convert equation (1) into a more hardware-efficient 2-cursor generator, and the generator 12 is composed of three sub-generators: "β海 calibration data condition, lateral color, And color non-uniformity. Finely calibrating the geometry separately. In the following, the above calibration materials will be discussed first. In the examples listed below, the calibration of the are, /, pieces is a type of line style, such as The preliminary sample of the knife, the style of the display. When the styles in the panels (4) to (4) of the 4th time of the line/training can also be =, . - a grid line, they are similar to the test of the grid type. The image provides a set of shapes centered on the known locations in the input space described above. These centers can be designated as (χζ), where i covers the range of the Nightmare T-oil-shaped shape. There is a total here

The number is the shape of Afx#, starting from the upper left corner from the R angle, and proceeding along the 18-20 200818114 column of the above test pattern 'and %' as the resolution of the test pattern. The test pattern resolution does not need to match the native resolution of the display device. When displayed, the center of the shape in the above test pattern will be transformed into some other value indicated by the κ heart due to geometric distortion. These shapes are also distorted, that is, - a circle will be distorted into a number, and so on. The coordinate system defines the display space at the origin relative to the upper left corner of the bezel 18 of the viewing surface 16 described above. Let the heart indicate the resolution of the display device (in the bezel 18) in an arbitrary measurement unit, and the seat bars (") are also in the same measurement unit. The display space is equivalent to 10 in the real world or The viewer space. In other words, the corrected image is necessarily undistorted in the display space. 15 The camera can capture the above-described distorted grid pattern image and transmit the image to the calibration data generator 12. The camera The resolution is indicated as %. In the embodiments recited in this specification, the camera resolution does not have to match the display, and in addition, the camera can be placed anywhere. The coordinates of the center of the camera space are (<〇, defined as the upper left corner of the above-mentioned captured image. /, tooth ~ £ you are shooting the image, which is the observation point from the above camera' and the calibration operation is necessary In the above-mentioned real world observation material, that is, from the correction, _ Cheng light must (4) to the observation point of the above camera, also known as the camera distortion. As in the above two example embodiments, this The completion of the use of the view = border 18 mark. Therefore, the camera image should also capture the viewing surface frame 18. In the real world, the viewing surface edge (four) is defined by 20 200818114 coordinates: upper left corner: (〇 ,〇) Upper right corner: (%, 〇) (2)

Lower left corner: (0, / / D) U Lower right corner · (WD, HD) In this camera image, the coordinates become: Upper left corner: (4x, > 4) Upper right corner: (3⁄43⁄4, meaning 77 ?) .. Bottom left _ : οάιΆι) Bottom right corner · (Xc8R, ycBJi) Figure 5 illustrates various spaces and coordinate systems. Although these images are displayed as black circles on a white background, all test styles 5 can be colored and other shapes or topography used (see, for example, Figure 4). The three conditions displayed in the display and camera space correspond to: case (a) when the image overflows to completely cover the viewing surface bezel 18, condition (b) when the image is fully fit into the viewing surface When the bezel 18 is under or underflow, and condition (c) is an intermediate condition or mismatch, wherein the image is not fully positioned within the viewing surface bezel 18. These conditions are called projection geometry categories. It should be noted that when the input and camera space are bounded by pixels, the display space may be pixels, millimeters, or some other unit. The above-mentioned display distortion indicated by Λ can be described as a map generated by the program (4) in a functional manner. fD: (x:, yn-^(x0di, y°di) (4) This means that the correction (/ί) is the inverse function generated in Equation 4, and its 20 200818114 is listed in Equation 5. :ix〇di^y°di)(xi ^y°i) (5) The digit warping unit 15 will apply the above-mentioned correction # to the input image so that it is warped before display (pre-distortion) . The above two maps are defined in the forward direction: the function domain is the input shadow image, and the range is the output image. As you can see, an electronic correction circuit is more efficient and correct, using an inverse function architecture to generate the image. In an inverse sacred architecture, the output image of the circuit is generated by mapping the pixels in the output to the input via the correction map, and then performing color filtering (ie, assigning) in the input space. Color value). This also means that the correction map is expressed in the form of an inverse function, which will be marked as /r. Due to the correction in the inverse function form, the distortion map itself is displayed (9^(/#疒=9). The map or warpage data required by an inverse function architecture correction unit is only the display distortion map. Therefore, The above-mentioned ruled line data to be generated by the calibration data generator 12 is defined in equation (6). (6) 15 It should be noted that the terms "grid line" and map are often used interchangeably. The images taken from the camera are taken out, and they are located in the camera space. The captured images correspond to the map defined in equation (7). Λ :(<,>0 — («) (7) This will be referred to as the map of the complete image map, which can be regarded as a combination of the display 20 distortion map /d and the camera distortion map /c, which can be decremented to generate the program 21 200818114 Required in (8) /r. fc · (xdi ? y°di) (x°ci 9 y°ci) fF:fcfD=fcfw3fwH (8) /c subtracted from /D, just on The link of the two maps (or function compound). In addition, when the display coordinate system scale and origin may not be commercial, The coordinates (4, %) need to be such that the correct pixel f and the 5 origin are achieved. This will be discussed in more detail below. An exemplary embodiment of the above calibration data generator 12 is shown in the sixth In the figure, the %\圮 camera image of a test pattern is first analyzed to capture the center of the shape (<,,·^); this will produce A. The shape center in the above camera space is input. The shape center in space is at a corresponding position after being mapped by the 10 dedicated display and camera distortion. For some image areas that overflow the viewing surface 16, the shapes will be unavailable. These are external shapes, on the back. In projection televisions, or in the case of a front projection system, they are usually invisible because they will lie in a background on a possibly different plane. Therefore, only the viewing surface 16 described above is defined as EFGH (see 15). The shape of Figure 5) will be analyzed. The center of these shapes can be found using various image processing algorithms. One method involves using a threshold mechanism to transform the captured image into a binary (black) The image, the shape of the binary image, allows the pixels to be identified and marked. The centroid of each group of classified pixels will then approximate the center of the shape by 20. The threshold can be analyzed by analysis. The histogram of the above image is automatically determined. The histogram may be the brightness or specific hue of the image taken. 22 200818114 These shots are also age and border. This step may use different/take the viewing surface. Coordinates, the coordinates of the frame are required. ~ Image. To determine the camera distortion, the camera distortion will be _ a target machine without optical distortion, then the a 'only these four corners, the falling side: distortion, And to decide the border of the border - necessary. The line equation parameterization <4 can be determined by the edges of the 10 15 20 corners of the line or by the wires. The shape of the elements is located within the viewing surface 16. A solid rectangular grid line with cc cc ""^, such as (1⁄2, ~) in the display space, can also be 1/4 上, ϋ, or projected onto the viewing surface 16 to provide additional The mark, 笪+ » ', 攸 in the camera space, will be imaged as ^, / ^). This grid can be regarded as the camera calibration (CC) grid. The coordinates of the border and the decision of the border : Also known as display characterization. <:,, breakpoint view' The camera lens and the surface of a curved screen are indistinguishable. In both cases, the standard σ has a spear border The image is curved. Therefore, a curved screen can also be addressed within the framework of 4 camera distortion and a combined CC grid. Correcting the camera and the piano distortion will also ensure the final image. In contrast to the curved frame 1° in terms of curved surface correction, the CC ruled line can be formed by attaching a mark to the frame 18 at a regular distance (as measured on the screen) to enable them to (10) (4) of the frame I8. Their marks may also be attached to the frame] s, S It should be noted that although the screen is curved, it is a _ & ^ -, 隹 surface, so it can be calibrated via the above-mentioned two-dimensional CC grid 23 200818114 line. Λ special edge (frame 18 or additional The CC grid line or mark can be prepared by using a standard image processing method such as edge detection, for example. Knowing or waiting for the position of the edge, a line equation can be matched to the edge to 5 and the line The intersections provide four corner and CC grid coordinates. These edges and CC grid coordinates can be defined as shown in equation (9), where 'Nec is the number of points in the calibration grid of the camera. (9) ('(7)' recognized)) ~ upper edge dk (7) w right edge I (8) 绫 left 5 - camera calibration grid For some display devices (such as those with curved screens), a CC line from the body mark , may not be immediately available. In this case 10, the edge equations can be used to mathematically establish the cc grid. The degrees of freedom that exist are related to how the points are arranged along the edges. 'and how to interpolate to The interior of the bezel 18. Regardless of the method selected, 'the last image will match the bezel 18, provided that the domain coordinates (see discussion of sorting) are properly selected. One arrangement is along the equilateral 15 The edges are equidistantly arranged such that they can then be linearly interpolated into the interior. If the manufacturer provides the specifications for the optical distortion that the camera is labeled, then these specifications can be combined with the above-described perspective distortion. Replace or generate the camera calibration grid with the spare end, which is shown in equation (1〇). (10) The optical component of the camera distortion can be resolved before the monitor is calibrated 24 200818114 ^Because it is The camera position and orientation are irrelevant. The shells ', ' in equations (7) and (9) are collectively referred to as camera calibration data. - Once the coordinates have been taken, they will need to match correctly with the above =! Mathematically, the sorting will be based on each range coordinate to create the above-mentioned complete image map, which has "" The above #1 procedure does not provide any description of the second =: information. These centers will not necessarily be determined in the order in which they match the shape ordering in the upper ^ style. 10 15 Can be used = (4) Test patterns shown in panels (4) and (4) of Figure 4, Γ = Γ etc. - Some of the shadows captured from these test styles can also be placed in this graded shape. The bar code is lifted (4) in horizontal and vertical, and the I-series is in equation (1)). The domain coordinate ("°), its {rl)N + s ^ (11) , face ==, it is important to decide which bar code and shape are in the background area of the viewing i (the viewing surface frame 18 External) =: An image with high contrast, then an appropriate threshold (in the city of Yicheng City), single (four) is indeed Γ == measurement. If the external shape is two I: wide = Comparing, you can decide which shape and number of barcodes. It is necessary to consider any missing barcode (the outside of the border 18). - The order of the number of ^ can be flashed - one, borrowed (four), etc. The bar codes of (4) 25 20 200818114 may also be used to implicitly number them. The camera calibration data also needs to be sorted, wherein the domain coordinates are within the display space. However, here, The program is simpler because all the topography (by definition) are located within the border 18. In most cases, the coordinate comparison is sufficient to determine the ordering. In the case of the CC grid, the ordering The above-mentioned coordinate network will be assigned, which are the domains related to the CC grid line referred to above as the domain CC grid line. Mark (in the display space). The value of the CC grid line in the field will depend on whether the grid corresponds to the entity marker, or whether it is mathematically established. In the former case, the marker has 10 coordinates. , the CC grid line of the domain can be generated. For the latter, the CC grid line is selected to have some degree of freedom. If the last image matches the above-mentioned frame 18 (ie, geometric category (a)) Then, the CC grid points above the edges are necessarily mapped to the corresponding edges above the rectangular EFGH. This means that the edges need to be mapped as follows: 15 The upper edge θ passes {(〇,〇), (%, 〇)} The right edge of the line "Through the lower edge of the line of {(%,0),(%,凡)}<=> by the line of {(〇, /^), (%, /^)} left In addition to the constraints of the line {(0,0), (0,仏)}, the CC grid points of the fields can be selected in any reasonable way. When the capture and sorting are completed, The map/" can be found using equation (8). The camera calibration data is used to first create the inverse function camera distortion map /c_1. In the most common case of pure perspective camera distortion (ie, /c =//), only four corner points are needed. 26 (12) 200818114 — (ο,0) (xdcTR^ydcTR) (WD^ ) (x^L, y^R) — (XtBR, D, HD) This (reverse) perspective transformation is generated by Equation 13. ^ =fcX'\Xc^c)^ cixc ^byc -Kc gxc+hyc -^\ yd =fcy~\Xc^c)^ dxc-heyc +£ Team kiss c+l (13) Here, (\,h) is the coordinates in the above display space, and (\,foot) is above The coordinates within the camera space. Using equation (12), eight linear equations are obtained. This can be used to find solutions for defining the coefficients 5 {a, b, c, d, e, f, g, h} of the above-described perspective transformation. When the camera distortion includes an optical distortion component / (?, or will be corrected for a curved frame, the edge equation or cc grid, 10 signal system is used to determine the inverse function camera distortion map 'One method is to use the CC grid because it provides the loss of direct information at the internal point, not just the information on the edge. The cc line is in the = program (10). The grid can be matched or Interpolate with the least squares of the "basis function group defined in the square. One choice is:," or by a spline base to obtain a pair of equations (14) two - spline spline fit Or interpolation. 27 (14) 200818114 / wide: (4e, e) - (4e, #), the fit or interpolation of the grid line ^ = / cx l (Xc ^ c) yd ^ fcy \xc ^yc) dO=(/r,/T) The /c-1 and the coordinates (<·,·^) calculated during the step of capturing the camera calibration data above, the map/r is by the chain The result is as follows: («)>«,%) where "·,%) is generated by equation G5) - fc'ffixUyt) = fc'ix^yli) The camera has perspective distortion & =/ ϋ:,β) (15) /di ^fcyV{X〇ci,y〇ci) The camera has perspective distortion + optical distortion, and the link can use the complete image mapping range for its domain to evaluate the camera anti-aliasing map. The middle diagram in Figure 5, and the map (in the form of an inverse function) needed to correct the distortion of the display. As mentioned earlier, the grid only contains those on the viewing surface. a point within the border 18. In terms of the distortion of the overflow (in case::), a plurality of pixels (corresponding to the center of the shape) of the above-mentioned domain space (that is, from the viewpoint of display distortion), in the grid The lines are defined as ^ not empty _, and do not have their coordinates. In the example money column, the digital warping unit 15 of the electronic warfare unit is 兀I 玟Immediately, it will be processed. :: inter-pixel; an inverse function architecture correction unit related to the domain of the empty asset 钭 2 is the output image produced. Therefore, the above-mentioned missing grid line shake system needs to be purely calculated 'money by _ rape and practice steps End 28 15 200818114 As in the above calculation of camera distortion, Where ruled line or may be fit (in

The positive line is more dense. And by evaluating the above inputs, the correction map can now be adopted as a correction grid for the function at any point array above the space, and the package 10 contains the points at which they are missing. In order to maintain the initial grid («)—«, %), the parent-related interpolation form is used to define a grid array of new regular intervals above the input space as shown in Equation 16. = , which contains the array {(«)} (16) The array can be made denser, with the opposite and the > rows. Evaluating the top of the array can be based on Equation 17, which produces the calibration corrected 15 grid lines (heart, *^), which include their missing points and may be denser. (1⁄2,~) = «,%'), if the brother ·) = («) and ", less 1" in the display frame, the combination of their cooperation and interpolation, can also be used by the brother, so that it can be provided Missing data interpolation related to the interpolation related to internal data. The final stage in the generation of the above calibration data is to fix the scale and origin. The correction line is within the display space and defines the upper phase of the upper phase 20 with respect to the viewing surface bezel 18. The above-mentioned display space list 29 200818114 (scale) is arbitrary and may be different from the one used in the input space. Before the data can be used by the warpage generator 13, the origins and scales need to be consistent with the input space. This can be considered as an optimization of the origin and scale. 5 Consider the middle diagram of Figure 5, which should be rectangular with respect to the viewing surface bezel 18 when the correction is applied. Referring to Fig. 7, such a rectangle containing the corrected image will be referred to as an active rectangle A'BfCT)f. This active rectangle is necessarily located within the optical envelope of the above image (ABCD) and within the viewing surface border (EFGH). The origin and the target 1 degree need to be selected, and the upper left corner of the active rectangle corresponds to (〇, 〇)' and the width of the rectangle multiplied by the height is %, which is the pixel resolution of the input image. (See Figure 7). It should be noted that the input space related to the above correction is actually an output image related to the electronic correction in an inverse function architecture, and the input image related to the above correction once the shift and the shift have been completed, actually It is spatially equivalent to the display (ie, the output space associated with the correction). If the upper left corner and the scale of the active rectangle are on the display coordinate space, respectively, (乂, given, all the grid coordinates, then it is calibrated as shown in equation (18) and Shift.

Wd nd v J , the above value which can determine the % of the rectangular coordinate value, can be selected as an integer value, as long as the (four) maintains the above-mentioned viewing surface frame to see the ratio. Using equation (18), the display spatial scale (bottom 30 200818114 map) can be transformed into the rounded image size required for the correction (top graph) in Figure 7. There is a degree of freedom in the determination of the active rectangle; however, adding a certain natural constraint can simplify the selection. In order to maximize the pixel resolution of the above corrected image, the rectangle should be selected to be as large as possible. If the corrected image is to have the same aspect ratio as the input image, the aspect ratio of the rectangle should match the input image. The various restrictions C1 to C4 series are as follows. 10 C1) The active rectangle is contained in the optical envelope ABCD. C2) The active rectangle is included in the viewing surface border EFGI1. C3) The active rectangular area is maximized. C4) The active rectangle aspect ratio is set so as to equal the input image {^dlhd=WTIHT) 〇15 to solve the above-mentioned active rectangle (that is, to determine the relevant constraints, and to become numerically optimized) One problem. All of the above restrictions can be placed in a mathematical form that allows for the use of various optimization methods to solve the top problem. One possible approach is to minimize the use of restrictions. 20 and rewrite the constraints in the form of the equation or inequality, and define a function to be minimized (or maximized). For the edge of the border (see equation (9)) and the outermost grid point (see equation (17) The line equations of )) can be used to formulate their constraints C1 and C2 into the form of inequalities, that is, four rectangular corners are located in (<=) the lines. The constraint C4 has become a kind Equation 31 200818114, and constraint C3 becomes a function of the above maximization, that is, maximizing the region of the active rectangle. In the case of the reality (a) of Figure 5, where the image is Overfill the watch 16. The viewing surface bezel 18 provides a natural rectangle that automatically satisfies the constraints C1 to C3. By adjusting the scale of the display to the test image, the parameters can be based on equation (19). set as.

Wd^WD= WT The above corrected image will exactly match the viewing surface bezel 18, which is ideal for using the entire viewing surface bezel 18. Thus, in the case of the case - the optimization step in Fig. 6 simply means that equation (19) is used, i.e., the points do not need to be scaled or shifted. This optimization step can also be used to achieve a change in the aspect ratio by modifying the constraint 4 as shown in equation (20). Ί—α (20) Continuing with equation (10), the aspect ratio of the above corrected image will become 15 α. The degree of freedom in the choice of such aspect ratio allows the image to be included in an image screen having different aspect ratios in a display device ((4) "(4) or pillar-boxed. By adjusting the scale and shift, The image can also be easily scanned over the viewing surface 16 (i.e., image overflow) and under (four) (i.e., 'image underscan). Therefore, using the surface function facilitates easy over-scanning and owuating Scanning situation 32 200818114 The last calibration data generated by the above calibration data generator 12 is the grid data given by equation (21) /V. / w: (xt di) (21) The above discussion focuses on all Distortion in the case where the corrections for the primary colors are the same. In these cases, the same ruled line data indicates all 5 color-related corrections, which may be referred to as single color correction. However, in terms of lateral color distortion The grid data is different for all primary colors, and the system requires multiple color corrections, which can be referred to as multi-color correction. Any geometric distortion common to all primary colors can be included in the horizontal In the correction; therefore, the previous implementation of the calibration data generator 12 described above can be considered as a special case of the multiple color correction described below. The exemplary data generator 12 described above is an exemplary implementation of lateral color correction, It is shown in Figure 8. As can be seen, this is similar to the implementation of the above-mentioned single color correction (see the previous section) in which K times are performed, where K is the number of primary colors. For Chiyou. For the most common three primary colors RGB, (/l, /2, /3) = (foot. Take the steps and details related to the calibration of each primary color and the above is related to the single color. The descriptions are the same, but with the following modifications: These test styles are now based on the primary color being calibrated as above. For example, when calibrating red color, all test samples = 2〇4, panel (4) To (1)), they will have their appearance (circles, lines, brothers, etc. (closed, material), and the above color samples "have not all image processing steps, such as capturing their centers And side, use color Color image. The threshold value can be adjusted to process the above-mentioned positive and the normal color. The image processing is independent of the color. The system is due to the lateral direction of the camera lens itself. Caused by color distortion: The camera calibration data is different for different primary colors, and needs to be calculated separately for all primary colors. The system can be configured to correct lateral color distortion within the camera itself. Testing from different primary colors The image pattern is similar to that of the display device and can be used to generate the camera registration data. The (multi-color) calibration data of the camera can be completed independently of the display calibration, and only needs to be completed once. In generating the 1 〇 camera calibration data, it should be turned off to be a display device having zero or very small (i.e., much less than the camera) lateral color distortion. If such a display device is not available, some additive markings can be used to provide a solid grid with known coordinates. The final result associated with the multi-color camera calibration described above is an inverse 15 function camera distortion that depends on the primary colors defined by equation (22) above. Fck · (3⁄4 , y^k) (χ^, yc^k), A: = l.. .ΛΓ, the fit and interpolation of the grid (2) After any missing data has been calculated, the κ The obtained ruled line (similar to equation (17)) is defined in equation (23). fm: (xi 5^ (χώ5ykdi)

k = l...K i = l...MkxNk (23) Here, the number of points per grid may vary depending on the test used and any resampling done. 34 200818114 The test patterns associated with these primary colors may be subject to different projection geometry categories (see Figure 5). Some of the test patterns associated with the primary colors may be completely overlaid on the viewing surface frame 丨8 as in panel (a) of Figure 5, while others may be completely located in panel (b) of Figure 5 Inside the border. When the optimal 5 is performed, the above-mentioned active rectangle is necessarily located in the viewing surface frame 16, and the spatial intersection of the image envelopes in each color-related image envelope ABCDk will be used. . What this means is that a single optimization is done, and the constraints are taken into account the envelope of all primary colors ABCDk. This optimization determines the coordinates of the active rectangle shared by all primary colors. These coordinates are then used to calibrate and shift the grid lines according to equation (18). The output of the optimization step is a grid line, which can be used to generate calibration data relating to all primary colors as indicated in equation (24). /W, /) ~> (χ'Κ) k = 1.., Κ xNk (24) These data sets are used by the warpage generator 13. 15 In the exemplary embodiment, the color and brightness, or simply the color non-uniformity calibration data is generated, and the geometric distortion (the type of heart has been corrected) is performed. The color unevenness may be caused by Caused by several origins, such as due to projection geometry (trapezoidal angle) to the viewing surface 16: changes in path length, flaws in the microdisplay panel, etc. 20 Jingyou a geometrically corrected display device In this case, the test pattern image is presented in the frame 18 as a rectangle (ie, a live _) that may be matched to the scale. The origin uses the above activity = 35 200818114 The upper left corner, rather than the upper left corner of the viewing surface bezel 18. The test patterns used are only the additive versions of the user of the single color geometry described above; that is, in terms of correcting the primary color κ, The topography (circles, lines) will be colored. This is the same as the one used to correct the horizontal color. For the 5 shell, the gray value (maximum white, half maximum) can be used. "usually To identify any color component being corrected; it may be a component of luminance, RGB or YCbCr, or a component in any other color space that can be measured by the sensing device 11. The sensing device 11, possibly A camera or a color analyzer (also known as a spectrometer, photometer, etc.). For greater accuracy, a photometer or spectrometer should be used. These color analyzers may Capture the entire image (ie, multiple points) or data at a single point. The sensing device 11 is arranged so as to be as close as possible to the viewing surface 16. Their single point color analyzer will actually be placed The coordinates of the coordinates (also 15 at the center of the shape) are obtained on the screen, and the information about the coordinates is obtained. Although the multi-point color analyzer and the camera can be placed at an arbitrary position, the lifting is performed. Accuracy is obtained by arranging them close to the viewing surface 16 and as close as possible to the center. Figure 9 illustrates a viewing surface, a single point color analyzer 92, and multi-point color analysis. The exemplary configuration of the above-mentioned color unevenness is similar to that of correcting geometric distortion. Fig. 10 is an exemplary example of a calibration data generator 12' related to color non-uniformity. The data captured by the single-point color analyzer 92 includes the primary color value C'I and the corresponding spatial coordinates at all the points to be measured. At 36 200818114, Α = 1··Χ can identify positive The color to be analyzed. The original color value specified by C is also known because its test pattern is clearly defined. This produces equation (25), which is the above-mentioned ruled line for explaining color unevenness distortion. The data is called the color distortion map. /dc · (xi ,C^) -> (xf ,C°ki) (25) It should be the main thinking, the 4 space coordinates, and will not be Color unevenness distortion changes. The original color value, for a given test pattern, will usually be a fixed value; this means that all non-moon rooms and pixels are of the same color. More than one set of measurements s = may be formed, where each set corresponds to a test pattern having a different fixed color value (such as saturation or grayscale at different levels). To simplify its notation, this single indicator will also cover different measurement sets across the equations shown in equation (26). These spatial coordinates are the same for each group. The discussion below applies to each group (i.e., test style). 15 In the case of the multi-point color analyzer 93 which can be a camera as described above, the above-mentioned photographed data corresponds to the entire image. In this case, some shadows: the processing system needs to be completed before the grid is obtained. These shapes: H (the smashed domain coordinates (9), the illusion will be calculated. The completion of the extraction and sorting steps used during the eve geometry correction. ^The center of the above-mentioned shape is also the color value at the center of the shape It will be calculated. Correcting the color value can be obtained by averaging the pixel color values in the identified nearest neighbors of the above-mentioned shot according to equation (27). 37 20 200818114 c,Hraf, A = filter coefficient (27) The neighborhood of Γ = is the color value in the captured image in the neighborhood of the above center. For the four closest averaging points, the filter coefficients are = 1/4" = 1 ... 4 The final result is the ruled line data defined in equation (25). It should be noted that: (1) only the domain coordinates are needed because the color distortion does not change the spatial coordinates; (ii) There is no missing data because the image is not geometrically distorted and is within the viewing surface 16; and (iii) there is no need to calculate the distortion of the sensing device and perform the chaining because it is not completed. Geometric correction. 10 Depending on the sensing used Depending on the type of shot and the format of the shot, it may be necessary to have a color space transformation that leads the color data to the color space of the display. For example, a spectrometer may produce data based on chroma values, and The display device and the electronic correction unit (which is a processor) require RGB values. A color transformation may be performed by a 15 matrix multiplication or by a more complex nonlinear equation. In other words, the grid data for all primary colors will be used. Usually, this transformation is in the form shown in equation (28). (28) If no color distortion occurs, then a fixed color test pattern is used. In other words, all the coordinates (χΓ, 〇, the color value, should be measured as a constant G. 20 The constant of this measurement, may not be equal to the original fixed pixel value. On the big 38 200818114 most of the display, these The measured value is proportional to the original value, wherein the proportional constant λ is fixed in the absence of color distortion and will be present in the presence of color distortion There is a spatial variation. Therefore, the color distortion map of the display can be represented as shown in equation (29).

Cli = A(x%y;)C°ki^X(x%y:) = ^ (29) 5 Normally, the input and measured color values will be determined by some kind of known display color function given by 3. The association is associated with a parameter vector as shown in equation (30). C0ki = MX,c〇ki) (30) If there is color distortion, there will be a spatial change in 3. The parameter at a given coordinate (<, brother °) can be determined by analyzing the data for a different set of s = U as shown in equation (31), where the s indicator is clearly shown. s = (31) What is needed at each coordinate is a sufficient number of values. The analysis may approximate the Λ by doing a match with the data. Similarly, the inverse of the function / / 1 can be calculated by analyzing the same data 15 in the opposite direction as shown in equation (32). (<, fire, C'L)—(<, brother°, 〇=> Q (32) The inverse function also depends on some parameters called color correction parameters, which can be known from明白 明白 明白 明白 明白 明白 明白 , , , , , , 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 明白 特定 特定 特定 特定 特定 39 The one-number map is in the form shown by the linear least squares fit of the shoes.

Loss: Under the generality, it can be assumed to be in the form of a polynomial. The above reference also allows adjustment of the last fixed chromaticity, since in some cases it may be necessary or desirable to reduce the original Q value of the turn. Here, the adjustment to increase or decrease the 4 parameter can increase the value of the above inversion by a simple scale factor. Once the inverse function (at each center coordinate) is known, the above corrected color map that corrects the color non-uniformity distortion is given by equation (34). (34) 15 The spatial variation and correction of the color distortion are fully explained by the parameter J and its inverse function 1, respectively. Therefore, the above (base) calibration data, which is labeled as /^, can be used to fully illustrate the above-described grid data associated with the color correction parameters in accordance with equation (35). Fwck: (35) For the most common case of equation (29), these parameters are generated according to equation 20 (36). 40 200818114 r^\〇 ik (36) The above-mentioned ruled line may be made dense by a new sampling with a suitable fit or interpolation. The above uses a similar note to the geometric calibrator. The new grid is given by equation (37). Year old fwck * (Xi i ) ^ik = ikr } (37) k = \...K / = 1.. .Mck x Nck r = 1 This is the data output of the above calibration data generator 12. The complete data output of the above-mentioned sub-generators (i.e., each of the rows in the first graph) is determined by equation (38). f’ m ·· (xi, fire, —(x’kdi,ykdi) / ·· (Ά,乂,= μ’*)

k = l".K i = \...MkxNk (38) j = \^MckxNck r = 1.../? If no lateral colors exist, then the K grid lines are the same, ie Only one geometric correction grid is calculated and output. This calibration data is input to the warpage generator 13. 1〇 As mentioned in the article, the grid data is not directly used by the electronic calibration. Although a squall line representation is the most common format, it is inefficient in terms of body and body, mainly because it needs to store the outer material (the coordinates related to each pixel), and it is not easy. The ground is manipulated 1 (such as; in terms of changes in degree). Some pre-existing technical systems that use a checklist are not optimal for the same reason. The warpage generator 13, 200818114 can convert the ruled line expression defined in (38) into the warpage data, which is a type of expression of the above correction, in the form of hardware application. The system is efficient. If the electronic correction unit can directly use the ruled line data, the above-described ruled line for re-sampling all pixels can be used, and it is no longer necessary to use the warp generator 13 to generate warpage data. The commemorative data of the mouth is produced according to the data requirements of the above electronic correction unit. An electronic correction unit that applies geometry and color transformations using a variety of architectures. Most cell systems use geometric correction related inverse function maps, and the above-described grid lines are designed for an inverse function architecture. A 10 efficient electronic correction architecture, such as the issued US Patent Application No. 2006-0050074 A1 entitled "System and meth〇df〇r epresenting a general two dimensional transformation" (representing a system of general two-dimensional transformations) The method described in the method is based on the linear function expression of the above-mentioned grid-line material. The warp generator 13 converts the ruled line 15 data into a function representation. Fig. 11 is a view showing an exemplary embodiment of the above-described warp-producing mushroom 13. A two-dimensional grid line (the general function expression of the heart brother, can be written as shown in equation (39). u = PiB^y) (39) The square equation (39) defines a cross-domain (^ force A two-dimensional surface function, which is a linear combination of the above-mentioned basis functions, a combination of the coefficients of the combination, called the surface coefficient, given by %. The coefficients are constants, and the plate does not cross the domain. There is a change. These basis functions do not have to be in line 42 200818114. Only the combination of them should be linear. At least some of the basis functions may be extremely nonlinear; therefore, Equation 4 is not in the form 'system^ It is sufficient to represent the correction line. The number of the number and their number are obtained by the electronic correction unit. The earth bottom letter is realized and evaluated in the hard towel. Wei Qu produced 01' because they are The necessary coefficients in the number function and the corresponding surface Bij(X,y) = X1 yj can determine the exemplary embodiments, the use of the standard, -2, may be written as the equation (4〇)示.一('少)ξΧχ>ν = 0···χ"7 = α ' (40) 10 due to these basis functions Is known, #咅靖7m, 丄 surface expression, two as silk (4)) medium ah, from grid line value to surface coefficient change ut α. (41) 15 20 on the expression of the benefits, health from The line value needs to be pinned for each pixel. In the case of a cell, the surface coefficients, the group, calculate the grid values, and thus the surface coefficients that need to be stored. , in terms of the number is quite small. To determine the number of such factors, the number of the original grid lines can be accurate =. The number of the increase factor, that is, the basis function, will be increased. Accuracy. Or, if the field is distinguished by 43 200818114 as a small piece, a smaller number of basis functions can be used, and each piece uses a different surface function. The piece structure is based on the display distortion in each piece. The rigor of the solution is established. This solution allows for greater flexibility in matching the complexity of the combined surface to the distortion. 5 For example, the more complex a distortion, the more small pieces are used. Small piece /7=1.. related coefficient The system is marked as <. In the following, without loss of generality, a polynomial form of the token will be used, which can be easily adapted to another substrate. The entire surface will be in equation (42). The form specified.

UJ i成·丄x,j = l』y (42)

/7 = 1...P (especially, > small piece; 7 10 A single surface system corresponds to a single small piece, which is equal to the entire output image (domain) described above. An exemplary small piece is distinguished and displayed In Fig. 12. The slice distinguishing can be initialized to a starting structure, such as 16 small pieces arranged symmetrically in 4x4. The arrangement of the small pieces (i.e., the number of small pieces 15 and the boundary of each small piece), The system is called the patch geometry D, which is in the form specified in equation (43). (43) Small piece of corpse = ((4/)14 Given a piece of geometrical condition, the coefficients can be used according to the above equation ( 38) The linear least squares fit of the data is calculated. The fit should be limited to ensure continuity at the boundary of the die. 44 200818114 Once the surface is determined, an error analysis is completed. And, as shown in equation (44), compare the coordinate network values with the calculated values. Error { =| u{ (44) The errors (ΕιτοιΟ values are compared with an allowable level Compare if the maximum error is less than Or equal to the tolerance level, that is, 5 nipc (£m^) < £max, the surface coefficients are retained and output from the warpage generator 13 as the warpage data described above. If the maximum error is large, the small piece geometry will be further subdivided and refined, and the coefficients will be recalculated and the error analysis repeated. The surface representation in equation (38) can be written as As shown in Equation (45).

Uk(x^y) = Y,ai/PxlyJ i,j ^k(^y) =

iJ (u,v) = YjCy PUiVj (45)

iJ

k = l...KP = L"Pk i = Q...Lkx,j = 0...Lky It should be noted that the above-mentioned rule of the grid expression is no longer needed because the function form is the entire space and not only It is only defined at a separate coordinate group. These indicators (〖, / /) can now be given the indices, or more generally the above-mentioned basis functions. The indicator k identifies the original 15 colors, and the indicator P identifies the above-mentioned small pieces. The surface is evaluated for the small piece of the domain coordinates. The number of patch arrangements and basis functions can vary with respect to the primary colors. Additional variations to the above format, for example, 45 200818114 can be obtained by changing the basis function of each tile. The above-mentioned geometric correction related domain space 'has been marked as, and its system corresponds to the output image two (in an inverse function architecture), and the range space has been relabeled as (')' and its system corresponds to This input image space. 5 For this color correction, the domain space has been re-marked as (w, v). The color correction is performed for a geometrically correct image. It is believed that the color correction potential must be applied before the input image having the coordinate space (w, v) has been warped before being corrected for geometry. If the electronic branch is positive, the color correction is applied when the input image has been corrected for the geometry and the warp is applied, then the above coefficients need to be adjusted for the new order corrected by the application, that is, Need a reordering step. In this case, the color parameters are defined in the (X,) space. First, a new grid force A can be obtained, which is defined in the space (x, j〇 from the above surface) as shown in equation (46). 7 (46) (xtk, y,) - again\ 15 The ruled line can then be matched as mentioned above, and the coefficients are calculated, which is now the output image space described above. The color correction surface coefficients use the same mark. The analysis will now use the above-mentioned reordered grid lines. The final output of the warpage generator 13 is the coefficient group 20 (adjusted if necessary) in equation (47), which collectively forms the warp described above. Information. 46 (47) 200818114

k = L..K p = x ... Pk i = Q".Lkx,j = Q.JLky Item: contains all the poor information defining the geometry of the patch associated with the primary color k. The (β,6) data selection Twist, Bessie, the above-mentioned geometric warping data or transformation of the calibratable plastic 1 to 4, and the color warping or transformation of the above-mentioned turbidity type 7 distortion of the correct type 5. 5 recording her music unit 15 System, a kind of processor, and an electronic correction unit functioning as the system. The term "electronic correction unit", this specification is used interchangeably with the term "digital Zhaoqu unit". In actual use, the digital enchantment Single 兀 15, can apply facial music data to the digital input image (as ι〇Λ) 'by pre-distortion or warping the input images. These input images are in the two-dimensional space and color space. The warp is warped. The spatial softness is based on the geometric warpage, and the color warping is completed, according to the color (4). The predistortion can be established to eliminate the display °. Distortion to produce an undistorted image displayed on the viewing surface 16 An exemplary embodiment of the digital light flexing unit of the above-mentioned correctable geometric and color non-uniformity is shown in FIG. 13 (omitted in this figure) 5 Hai digit distortion unit 15, including Two main blocks: one that is applied to the first block of the light (that is, the geometric input of the input shadows can only be used to correct the color unevenness and distort the wheel in the color space Entering the second warped block of the image. Here, the color correction occurs after 4 geometric corrections, however, this discussion can be easily applied to the reverse order. 47 200818114 When -(4) calibration is not required, both Blocks can be bypassed. Each block I has two components: _ a surface evaluation component that evaluates each pixel (w, for each primary color (this indicator is omitted) in equation (10) The surface of the 丨疋, by which the necessary coordinates k'v~} ' and _ pixels are generated, which actually use the necessary coordinates: to calculate the pixel color value 就. In terms of geometric correction, the pixel generates $ wave steps, one of which Some of the neighbors of a pixel having a coefficient of ~X·· "the pre-centered coefficient applied to the current pixel being processed ^). 10 15 In the case of the first and second, the filtering The coefficients are placed outside the system and will be in the shape of the unit. In terms of the color, the "pixel generation will take the pixel values from the image of the above geometric surface, and the application Equation (33), to determine its new color value. The pixel generation step is summarized in the neighborhood of equation (48), /e "~ C'i = Z and WQ), r = l; j (48 r These steps are performed for each primary color. The G system represents the geometrically corrected intermediate color value. The wave and color of the 桉正古# 4, ^ only the details of the private party depends on the structure of the above hardware. - A simple seamer, perhaps just averaging the four nearest neighbors, in which case m - may use - elliptical neighbors, the shape of which depends on the surface of the above 1 (4) acobian (image ya Comparable), and the filter coefficients can be obtained using a rider chopper generating algorithm. In this case, the neighbors 48 20 200818114 coordinates (, r, vyer), and J 匕 need to be used to evaluate Jacobian. For the same reason, the early color correction between the mouth and the mouth of the solidification is related to using only one linear correction as defined in equation (49).

Ci = ^i2 1 (49) A complex color correction, which may be used, uses a cubic polynomial as defined in the equation GO}. ^ = ^(ζ)3+Α·3(ς)2+Λ,2(^) + Λ, (5〇) These color parameters and surface coefficients are calculated, and the digital warping unit 15 is known. Architectural details. The final result of the digital content unit 15 is that the vector symbol used to indicate all the primary color components is overwritten by the equation (1) in Equation (8) 10 below to mathematically correct the correction. Input image 〇 output image (W,, V,, A), ie, e'·) (51) The turtle or pre-compensated output image is input to the display device (not shown), wherein Projected onto the viewing surface 16 without visual distortion, thus completing the automated calibration and correction described above. Once the calibration and calibration procedures are completed, normal (no test style) images and video are transmitted to the display device. This multi-color geometry calibration and correction has been discussed in conjunction with lateral color correction. However, it can be used to calibrate and correct any distortion in which the primary color components have geometric distortion. Other applications include: distortion caused by under-representation; and optical components are arranged relative to one another or a rear projection display relative to the chassis or casing in the device, plus the color points = 49 2 〇〇 8l8li4 Under-convergence due to multiple microdisplay devices of different magnifications. ~ In the projection system, this color calibration and correction is done on a geometrically aligned image. This means that the color correction also takes into account any non-uniformity introduced by the geometric warping itself. An image of the interesting music on the geometry will have some different areas containing the same color or brightness content due to the calibration and the wave program. In fact, the more the area is set, the greater the change in brightness and color. This is automatically compensated for by color correction made after geometric warping. Therefore, the system can automatically complement the color unevenness caused by the above geometric warping program. 10 In another aptamer, the system can be integrated into a single circuit and passed to a digital calibration and warping unit. These calibration data and interesting songs are produced|§12 and 13, which are components that can be executed on any processor. The test image generator 14 can also be replaced by a set of stored images output by the processor. Using a built-in processor in the above-described hardware, a single circuit solution can be given for the entire calibration and calibration procedure described above. In addition, the hardware can be integrated with a camera to be integrated into a display device to provide a self-aligning display device. In this aptamer, only one processor is needed to receive the sensing information from the at least one image sensing device and calculate the distortion of the display, and generate some pre-compensation maps, that is, warpage. Maps and color maps (also known as geometric warpage and color warping), and applying the pre-compensation maps to their input image data to make the displayed image on the viewing surface generally There will be no distortion in the upper. However, in other cases, using more than one processor may be more efficient. Thus, to implement the embodiments described in the teachings of the present specification 50 200818114, at least one processor is required. Various types of sensors can be integrated into the display device (rather than or in conjunction with the cameras) to act as the sensing device 11. In an exemplary embodiment shown in Fig. 14, a sensor 143 is a 5-way sensing device that is used independently or in conjunction with a camera 142 to measure the viewing. The distance of a certain number of points on surface 141. This plane does not need to be flat. The relative angles of the cameras 142 to the viewing surface 141 are calculated from the measured distances and the angles at which the sensing distances are opposite each other. Further, if the shape of the screen is not flat, 10 can be calculated by such a method. In the example shown in Figure 14, the denser lines on the right side of the screen will indicate that the sensor 143 is closer to the normal view of the screen, and the less dense pattern on the left side, Indicates that it is far from the general observation above the left side. Various types of sensors 143 can be used to include infrared sensors, and the like. In this exemplary embodiment, the display screen (i.e., viewing surface 141) is depicted, and does not require a physical structure, and the camera 142 can be arbitrarily arranged. Another exemplary embodiment constitutes a self-calibrating display device with dynamic calibration and calibration so that the calibration and calibration procedures can be performed at any time to correct for distortion without requiring external resources. This allows for correction of distortions that may change over time, such as a projector-related trapezoidal distortion, or some field calibration of a rear projection display device such as an RPTV. The calibration system is disposed within the outer casing or chassis of the RPTV to provide self-calibration in this situation. Other important distortions of long-term changes are the optical components 51 200818114 Internal changes in physical motion, deb, and ^ degrees. For example, in a rear projection display device, the curvature of a one-sided mirror may be due to the weight or the π sound. :: ❿ slightly changed 'This will require dynamic calibration and calibration 5 10 15 20 when the time is turned on' or the change in the distortion is detected Λ again; and the righteous system will be executed. In the field of display devices, such as television systems, important ^, , , _ can be used for 'dynamic calibration and calibration becomes special & 'after the initial calibration and correction, the future is not straight, A small amount of change in the component in the long term. In a controlled conditional background period = such as manufacturing work, the number-curve unit can be used to model the expected distortion in the field for a long time, i = 1...N. These distortions can then be corrected for use in order to use the system described in the exemplary embodiments mentioned above; however, 'two electronic correction units may be used, = pseudo-distortion, and the other is used to test the Automatically generated calibration assets. The correction related warpage data relating to the N test conditions can be stored in the display device. In the field and for a long time, with the development of a small amount of distortion, the correction of the N warps was made, and the one that best corrected the loss was selected. Therefore, the entire system is not necessary, only the digital music is required to be built in the display device, because the calibration system is completed during the manufacturing process, and the viewing correction (4) is stored in the display farm. in. To automate the selection of appropriate calibration data as described above, the sensors within the display screen can be used to test their specific test patterns. The image test pattern associated with the best detection of the distortion described above is thus implanted. This procedure can be performed when the display is turned on for dynamic correction and 52 200818114 calibration. As shown in Figures 15 and 16, in an exemplary embodiment, the calibration system is adapted to find the best projector focus on a viewing surface. This is done by displaying a test pattern on the viewing surface, such as a set of specific numbers of parallel lines. The image is then captured and scanned by the electronic calibration unit to find a comparison of dark areas and 焭 intervals in the test patterns. The projector is focused and then shifted, and the contrast is re-measured. This will continue until the largest comparison is found. This maximum contrast corresponds to the best focus. This is inferior to the viewing surface 151 and is preferably displayed on the side viewing surface 161 with better focus. This same technique can be used to focus the sensing device. Some physical markers with sharp edges, such as the screen frame of the display screen (i.e., the 'viewing surface'), are taken and analyzed for maximum contrast. If necessary, a properly colored test pattern can be displayed to improve the contrast between the 15 marks and the background. The sensing device is focused and then shifted, and the 5H contrast is re-measured. This maximum contrast setting provides the best focus associated with the sensing device. The sensing device focuses the focus prior to focusing the display device. In another exemplary embodiment, some of them are shown in Figures 17 and 18 of Figure 20. The calibration system is used in a display device having curved screens 171 and W1 and multiple projectors 1 through 3, respectively. The projectors span the entire area of the curved screens 171 and 181 and are controlled by the same electronic component. The geometric calibration is done for each of the projectors 1 through 3, and is mapped to corresponding regions of the screens 171 and 181. In addition, the 53 200818114 geometry calibration rotates and translates each projector image to align it with a field-matched projector image. In particular, in the overlapping regions, the pixels corresponding to the pixels are overlaid on top of each other. It should be noted that the mappings from the different projectors 3 to 3 on the screens 171 and 181 have different angles of incidence and are varied according to the curves of the screens 171 and 181. The electronic components having or having the maps represented by the curved screens 171 and 181, such as warp data, correct for angular variations across the screens 171 and 181. In addition to the geometric calibration, the color calibration of each of the projectors 1 through 3 is done to ensure visually identical color characteristics across all of the projector areas. The electronic component is also adapted to distinguish between pixel color and brightness in or between projectors 1 through 3 to achieve a uniform redundancy and color mapping across the curved screens 171 and 181. It should be noted that any number of individual shots can be used, and such overlapping areas can be shared among many projectors while applying the same calibration technique. For a projection on a curved screen, this focus problem is always critical. This stems from the fact that a projector has a flat focusing plane and the screen is curved, and as such, different portions of the screen have different distances from any of the focal planes. Looking at a portion of the screen 20, the image may appear to be more focused than another portion of the screen. To overcome this problem, a technique can be used to minimize the out-of-focus when projecting with a single projector, an exemplary embodiment of which is shown in FIG. In this case, the calibration system minimizes the sum of the squared distances of a series of normals from the curved screen 191 to the focus plane 193 54 200818114 in the manner in which the above-described projected focus plane is disposed. If the screen wishes its center to be more focused than the side, the segments that connect the central portion of the screen to the focal plane give a greater weight. In this case, the optimum focus plane can be calculated in advance based on the known shape of the screen 5 described above. The intersection of the best focus plane and the screen described above produces a point on the screen with the best focus image and, in turn, a maximum contrast. Under the calculated and known best plane and maximum contrast point, an image test pattern similar to that used in Figure 16 will be projected onto the screen, and the image will then be captured, with 10 and contrast analysis. If the maximum contrast position of the captured image is consistent with the previously determined maximum contrast point, and within a certain tolerance, the projected image is located on the optimal focus plane. If the maximum contrast point does not match the previously determined maximum contrast point, the projector focus will be adjusted and the program will repeat until and after a match is obtained. 15 It should be noted that this technique can be applied to some one-dimensional bending (eg, cylindrical, zero-space curvature, etc.) or two-dimensional bending (eg, spherical, non-zero spatial curvature, etc.) screen. In another exemplary embodiment shown in Fig. 20, in addition to the calibration already explained, the focus problem is addressed by multiple projectors projecting images at not the same angle. As shown in this figure, the defocus problem can be substantially eliminated by illuminating a particular area of the curved screen 201 by the projector at a particular angle. The angles are such that each projection axis is substantially normal to its corresponding portion of the screen portion, and each of the focusing planes is nearly tangent to the center of the portion of the curved screen 201. To optimize the same technology as shown in Figure 18 200818114 苐19. Or you can keep it tangential to the screen. The focus of each segment here is the user, the center of each focus segment. In an exemplary embodiment, the calibration system can match the focus of the overlapping regions of the multiple projectors plus the conditions and brightness of the image. And color, by which a smooth and seamless and focused image is produced on the screen 201. As a result of this technology, the rounds of the "curtains" and the reduction of the angle between the tangent planes have become much less serious. A system for multi-color geometric calibration - a sensing device has been discussed. The system can be used to calibrate the color (non-geometric) distortion in the above sensing device. Using a calibrated and calibrated display device, some fixed color patterns are displayed on the glory and recorded by the sensing device; the same-style used to calibrate the color distortion of the display can be made S. Knowing the value of the I color, you can get a camera color map similar to the program (25). From the color map, it is possible to determine the color correction parameters associated with the above camera, which will have spatial variations in the presence of color distortion. The model associated with this correction, for example, can be a linear least squares fit. These calibration parameters fully characterize the calibration data associated with the color distortion of the camera. This color correction has been rendered in terms of primary colors and brightness. The system can be adapted to handle correction and adjustment of an arbitrary color. These test patterns or various colors (not only primary colors or gray scales) can be used to obtain a color map of the above display in a manner similar to equation (31), which is shown in equation (52). (52) 56 200818114 - Here each 'e' can produce a color vector in a primary color knives with all components and not just one feature. The color used by the set can be selected as some resampling of the vectors in the entire color space described above. This inverse function map is now represented by equation (53). ^=7;>(λ-,〇Τ) = Σα\^(〇7),, = 1...λ λΊ x'uj (53) Here, each color parameter is a length, especially a vector ( The number of primary colors). According to the previous mark··

Jf» __ ^ ilr ir — · (54) u,” However, this is not only the re-women's volleyball in which the color parameters become a single equation, because these basis functions are now attached to the entire color space and not only in one dimension. The color space (ie, a primary color) is defined above. 10 In terms of the form of a polynomial, the basis functions are defined in equation (55).矣m = ) ί2···(ς^ (55) These parameters can be further generalized by importing the two-dimensional Euvy plate structure shown in equation (56) in the color space. Ig = {c\c°gk<ck<cik9k^Lt.K} (56) This can be added to the equal-color 15 color parameter as shown in equation (57). 57 (57) 200818114 Also " irq A, A, i\rq i2rq This produces a general k: change (the center of the shapes) at each spatial grid point in the color space. This calibration color data is now defined by the program (58). ...(58) . In the absence of any distortion, the grid will be a single 4-element warp generator at each coordinate, which can be converted into a surface function of the form specified in equation (59). Ο) = Σ0νν ij

k = L..K, Γ = Q (59) P = '-Pk 1 2 Λ) After taking the digital light unit, the polynomial will be evaluated to use the equation 3) to consume pure. ^ 10 ... at the parent's coordinates, with a - general color map. Valley final correction of any sitting; I: ® not / /, W敕, colors. This system includes execution 1 and port. Positive 'such as white point adjustment, contrast adjustment, and hue are irrelevant to the display's not 疋' and A's no (four) field. All such adjustments, the specific function of the shirt space, 'in the color, and the mouth can be approximated by a function y to achieve the above-mentioned as specified in equation (53), so that the small features of the additional features under the " * Just within the space can force the color patch to be positive. The school's early 7L line was corrected to limit the number of 58 15 200818114 to some specific colors, while leaving the others unchanged. This includes selective tone calibration where specific tones are corrected and other tonals are not touched. With the general color calibration and calibration of the system, high color accuracy will be achieved in the display device. The system can also be used for color customization by providing some custom color parameters that can be calculated external to the system and can be input to the warp generator 13. Similarly, some custom geometric effects (specific effects) can be achieved by providing the custom geometry line «, yi) to the warp generator 13. In another exemplary embodiment shown in Fig. 21, two cameras Cml and Cm2 are mounted on a projector 213. An input image is provided to the projector 213, which in turn produces a corresponding projected image pattern on a viewing surface 211 15 20 . The two cameras Cml*cm2 are used to capture the projected image pattern on the viewing surface 211. The system further includes a processor (not shown but previously described). The relative positions of the two cameras Cml and Cm2 are known to the processor. The two cameras Cml and Cm2 can be staggered horizontally, vertically, or both horizontally and vertically with respect to the projector 213. The processor can determine distortion parameters including the angle of the projector 213 relative to the viewing surface 211 based on a comparison of two captured images from the two cameras Cml and Cm2. The above-mentioned electronic correction unit (which has not been calibrated and then applies a warp transformation to the input image, thereby correcting the projected image of the above achievement, is substantially free of distortion. The system and method can be used in In a rear projection television (RPTV), for example, 59 200818114 has one or more cameras mounted to the RPTV in a fixed position and orientation as seen in the exemplary embodiment shown in FIG. These cameras can also be installed in other ways. These cameras can capture some of the patterns projected onto the rPTV screen. From the perspective of the camera, the RpTV screen view 5 may have some associated keystone distortion. Where the calibration system is part of the display device described above, the display can be self-calibrated as discussed above. In another portion of the exemplary embodiment shown in Figure 23, there are several projectors. P1 to P3 are used to project an image onto a 10 curved screen 231. At the same time, there are several cameras cml to Cm3, which are used to capture each projection. The images projected by the devices P1 to P3. The number of the cameras cml to Cm3' and the number of the projectors ?1 to ?3 are arbitrary in this embodiment. In one case, each camera Cini Up to Cm3, can be used to capture images from all projectors 1 to 3. These cameras are: 1111 to 15 Cm3, which can be staggered horizontally and vertically with each other. Each projector P1 to P3, It is adapted to project a known pattern or test image onto the curved screen 231 for calibration. Based on the cameras (: 1111 to (::1113), a processor (not shown but not shown) As previously explained, the distortion parameters including the shape and relative orientation of the curved surface 231 described above can be calculated. The parameters can then be used by the processor to generate a warp transformation that can be used during normal use. Applied to the input images provided to each of the projectors ρι to P3. The warp transformation of each of the projectors is such that it pre-compensates for display distortion suffered by the particular projector. In addition, each projector P1 to P3 related The degree can be analyzed such that the total brightness of the projected image on the viewing surface 231 60 200818114 is consistent. In addition, the processor can align the pixels in the overlapping regions and the sorrow of the overlapping pixels. The seamless image quality is distributed between the different projectors. In one embodiment of the system of Fig. 23, the brightness and color 5 data can also be taken by the cameras Cml to Cm3. This data will then be used by the processor to adjust the intensity of each pixel and match the side edges of the different adjacent images. The total brightness and color of all projectors 1 to P3. , can also be normalized by the processor. 10 15 2 〇 丨仞 丨仞,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, In the case of a camera, it is used to capture a projected image that has been projected with or without a pattern. At this point, the camera is also used to detect the shape, scale, relative orientation, and boundary of a viewing surface 241. The boundary side edges may be the side edges of a drop-down viewing surface (i.e., a telescopic projector screen), or the corners of a room, and the like. A processor not only previously described), but also the direction of the side edges of the image and the pattern of the test image can be analyzed to calculate characteristics of the viewing surfaces, such as shapes, dimensions, boundaries, and relative orientations. In this calculation, the second distortion can be determined based on the projection and then the image of the image = the complexity of the 'electronic correction unit' (ie, processing the two distortion parameters. In the case of the early phase 4, the electronic correction unit is the target of the viewing angle of the viewing surface. In the more complicated electronic correction unit, the shape of the viewing surface can be determined = 'the curved or irregular viewing surface. The electronic correction unit, in other words, the distortion parameters related to the curved mirror, such as the pin cushion shape or the barrel distortion, ., = through, the 呑 呑 61 61 200818114 and other distortion parameters are collected, the appropriate pre-compensation of the imaginary map Applied to the input image data, the image is visually undistorted. 5 10 15 20 In a other embodiment, the system of Fig. 24 is attached to the body mark or side edge. In the presence of it, it is adapted to correct the projection onto a flat surface. From the above projection Φ mu —. Straight — “... both trapezoidal and lens distortion. In this system, there is a camera. Attached in a fixed orientation To the projector, the calibration and calibration are performed in a two-step procedure. In the first step, a complete calibration procedure uses a number of test image patterns that can be used to store the camera in the Including the known trapezoidal distortion angle of the zoom level and the image of the style captured by the household wheel under the lens distortion parameter. In addition, any additional capital required for the correction, such as (4) data, can also be stored. This step can be Executed in a factory that assembles the projector, and can be considered factory calibration. This second step

The system occurs in the field towel where the projection H is made. The projection (4) can project the same pattern used in the first step, which can then be taken by the camera. The pattern of such live shots is such that, in conjunction with the styles captured by the factory, the factory is compared to the stored distortion parameters to determine the distortion parameters of the projector in the field. Knowing the distortion parameters in the scene, the calibration surface can be retrieved if it has already been stored, or it can be created in real time to read the trapezoidal distortion and lens distortion of the projector. Since the comparison is done with the previous storage of the Zisaki image, the real Zhejiang or mark (such as the screen frame) is not needed. The data stored in the above factory does not need to be a complete image, but it may be grid data, or other parameters that characterize the different distortions of the 62 818114 level. The heart is only her order, the camera will be corrected by a simple four-point line style image to correct the trapezoidal distortion. In this case 5 10 (only need 4:: 2 in the coffee chart, by a 2x2 grid below him). Turn the county true words, knot any lens lost two: to determine the distortion. These four points can be placed at any position to know their position (before and after the projection), which is sufficient, -^ 抆. This method can also be combined with any of the adjustments of the f-position, and the heart of the #健^" is a lens shift or a translucent αΓ simple translation. It is possible to have 15 gentlemen-shaped ^ straight and turn The average school song, the sister line (that is, no ladder ~ ® is executed. Then, the correction warp (the appropriate zoom level; the mirror distortion is applied in the city, and the calibration can make the point only: shape The right is repeated over and over again. The keystone correction can be combined with the telescopic lens correction phase chain or functionally combined to obtain a final map that corrects all projections = distortion. The lens correction only needs to be calibrated in a factory. The sequence period is calculated and stored once. The keystone correction is then performed using the camera in the field, and the lens correction is performed. Another exemplary embodiment is shown in Figure 25, and the 20 series and - For the case of the projection on the curved screen 251, in order to determine the map of the shape and distance contained in the curved screen 251, a two-dimensional image style, for example, a plaid image pattern is projected onto the viewing surface. /camera It is used to capture the above projected image. The electronic correction unit (that is, the processor not shown but previously described) is then adapted to calculate the contrast introduced in each of the lines of the 2008 200818114 = flower pattern. By constantly changing the optimal track of 5 10 15 20 for each nowhere above, it is found to be a function of over, , and distance. In this way, the surface surface map of the above surface will be determined. The accuracy of the map and (iv) depends on the complexity of the projection style and the number of focal lengths attempted. Also deer, the main, the technology will also produce the angle of the camera, and thus _ point = 杈影 relative to the above Observe the angle of the normal of the surface. The x _ ~ side of the electronic school is early, and the distortion parameter associated with the shape of the viewing surface and the angle of the sound at each point is calculated, and then an interesting Change or use - a suitable person already stored. She, +, > from the scorpion in the application to the top; L input image data, will achieve a non-visually distorted image matching the viewing surface Special 1± another The exemplary embodiment, which is shown in Fig. 26, is related to the case of a wavy screen 261. The above said 'and Η 9 ττΑΔ , 2 5 |^1 The technique 'can also be used to determine the shape and relative orientation of the wavy viewing surface at each point. This is a material = regular viewing surface that can be used in the display device. A map of the surface, the electronic correction is a _, q flat 7 ° (not shown but not explained), you can use this map to configure the warp transformation applied to the wheel of the wheat image. Once this A warp transformation is applied to the input image, the projected image is visually distorted and matches the characteristics of the viewing surface. While the above description provides various exemplary embodiments, it should be understood that Some of the features and/or functions of the illustrated embodiments may be modified in the spirit and principles of the operation of the embodiments of the description of the invention. Accordingly, the above description is intended to be illustrative and not restrictive, and it is understood by those skilled in the art that other variations and modifications can be practiced without departing from the practice. The scope defined in the scope of patent application 5 is attached. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a simplified diagram of an exemplary embodiment of an automated calibration and calibration system; FIGS. 2a and 2b are diagrams of a curved surface geometry; 10 Figure 3 is an overflow of geometric distortion An illustration of an example of a calibration image test pattern; Figure 5 is an illustration of an example of a calibration geometry test; and Figure 5 is an illustration of a calibration geometry and various coordinate spaces involved; An illustration of an exemplary embodiment of a calibration data generator, Figure 7 is an illustration of an optimization of the scale and origin; Figure 8 is an exemplary embodiment of a multiple color calibration data generator FIG. 9 is an illustration of a setting related to color unevenness calibration; FIG. 10 is an illustration of an exemplary embodiment of a calibration data generator related to color unevenness correction; FIG. An illustration of an exemplary embodiment of a warpage data generator, 65 200818114 Figure 12 is an illustration of a slice segmentation associated with display correction; Figure 13 is an example of a digital warp unit An illustration of an embodiment; Figure 14 shows a schematic diagram of the shape and relative orientation of the tooth surface viewing surface; Figure 15 is an illustration of a out-of-focus test pattern; Figure 16 is a focus test pattern Example diagram; Figure 17 is a partial illustration of an exemplary embodiment of a calibration system consisting of a multi-projector and a curved surface; 10 胄18-system for displaying different projections Partial illustration of an exemplary embodiment of a calibration system consisting of a multi-projector and a curved glory of Figure 17; Figure 19 is an example of a focusing technique that minimizes a distance function An exemplary view of an exemplary embodiment of a calibration system consisting of a multi-projector and a curved screen to adjust the position of the projector to optimize image focus; The figure is a partial illustration of an exemplary embodiment of a calibration system using multiple cameras; I calibrated and allowed a visual motion (RPTV) example 2 〇 22 系 呈 呈 古 具有 可使 可使 可使 可使 可使 可使 可使 可使State loss Partial illustration of a backprojection electrical embodiment of a calibrated integrated calibration system; multiplex multiplicity followed by _ 66 200818114 Figure 24 is an exemplary embodiment of a calibration system using physical edges and boundaries of the viewing surface described above Partial illustration of Figure 25; Figure 25 is a partial illustration of an exemplary embodiment of a calibration system using a focusing technique to determine the shape of a curved display screen; and Figure 26 is a focusing technique A partial illustration of an exemplary embodiment of a calibration system that determines the shape of a wavy display screen. [Main component symbol description] 1-3...projector 143...sensor 11...sensing device 151...viewing surface 12...calibration data generator 161... viewing surface 12'.. Calibration data generator 171, 181..... Curve screen 13... Warpage generator 191... Curve screen 14... Test image generator 193... Focus plane 15... Digital warpage unit 211... View surface 16.. Viewing surface 213...Projector 91...Viewing surface 231...Surface screen 92...Single point color analyzer 241...Viewing surface 93...Multi-point color analyzer 251···Surface screen 141 ...view surface 261...wavy screen 142...camera 67

Claims (1)

200818114 X. Patent Application Range: 1. A display calibration system for use with a display device having a viewing surface, the display calibration system comprising: at least one sensing device adapted to sense the viewing At least one processor related to at least one of a shape, a scale, a boundary, and an orientation; and The information is measured to calculate the characteristics of the display device. The display calibration system of claim 1, wherein the at least one sensing device is further adapted to sense a test image displayed on the viewing surface, and wherein the at least one processor The system is further adapted to calculate the distortion of the display based on the sensed test images and the characteristics of the display devices. 15 3. The display calibration system of claim 2, wherein the one to less than one processor is further adapted to generate a pre-compensation map based on the display loss-true, so that The pre-compensation map is applied to the input image data before the display, so that the display image formed on the viewing surface is substantially free of distortion. 20. The display calibration system of claim 3, wherein the display distortion varies over time, and wherein the display calibration system is adapted to dynamically calibrate the display device to precompensate for changes Distortion. 5. The display calibration system of claim 3, wherein the processor of 68 200818114 v is adapted to correct at least one of the following: ^ an overflow condition, wherein an image of the display, The system is larger than the viewing surface, an underflow condition, wherein, a displayed image, the system 1, the watch table φ; and a mismatch situation, the towel, a part of the shirt, overflowing the viewing The surface, as well as other parts of the display image, underflow the viewing surface. 6. The display calibration system of claim 3, wherein the display device is a rear projection display device having a housing, and the display calibration system is disposed in the housing . 7. The display calibration system of claim 3, wherein the sensing device is further adapted to sense at least one brightness information and color information, and wherein the at least one processor, Intro: Steps are adapted to compensate for at least one brightness uniformity and color unevenness. 8. If applying for a full-time m item 3 display calibration system, wherein the display is not a system, it also includes some optical components with additional distortion, and wherein the 'at least one processor is stepped into The adaptation is such that the additional distortion is linked to the display distortion to pre-compensate for both the additional distortion and display distortion. 9_ The display calibration system of claim 2, wherein the display is distorted by at least one of geometric distortion, optical distortion, convergence, under-alignment, and lateral chromatic aberration. 1〇·^ Apply for the patent scope! Display calibration system, wherein the at least one sensing device is adapted to sense a distance from the viewing surface to a majority of 69 200818114 points, and wherein the at least one processor is adapted to Equidistual distances are used to calculate the relative position and relative orientation of the viewing surface. 11. The display calibration system of claim 2, wherein the one to five less sensing devices are adapted to sense different portions of a test image on the viewing surface at various focal lengths, and wherein The at least one processor is adapted to determine a highest contrast in different portions of the test image and to calculate a distance to a different portion of the viewing surface based on the highest contrast of the determination to calculate the viewing surface 10 shape and relative orientation. 12. The display calibration system of claim 2, wherein the at least one sensing device has some sensor distortion, and wherein the at least one processor is further adapted to calculate the sensing Distortion, and the distortion of the sensors 15 is taken into account when calculating the distortion of the displays. 13. The display calibration system of claim 12, wherein the sensor distortion is caused by at least one sensing device having an axis that is not parallel to a normal direction of the viewing surface. 14. The display calibration system of claim 2, wherein the one or more image sensing devices are composed of a plurality of image sensing devices, and the plurality of image sensing devices are disposed differently from the at least one processor. And wherein the at least one processor is adapted to compare different sensed test images from different sensing devices and to locate the different sensed images and the different sensing devices based on the different sensed devices , to calculate the display of the display 70 200818114 distortion. 15. The display calibration system of claim 2, wherein the at least one image sensing device is adapted to sense information about a test image having four marks on the viewing surface, and wherein 5 The at least one processor is adapted to calculate keystone distortion based on the sensed information. 16. The display calibration system of claim 3, wherein the at least one sensing device is further adapted to sense at least one of the brightness information and color information, and wherein the at least one processing The apparatus is further adapted to correct at least one of brightness non-uniformity and color unevenness caused by the pre-compensation maps. 17. A display calibration system for use with a display device having a viewing surface, the display calibration system comprising: at least one sensing device adapted to sense display on a surface of said viewing table 15 Information of the test image; and at least one processor coupled to the at least one sensing device, the at least one processor adapted to calculate display distortion based on the sensed information, and to generate some pre-compensation maps In order to compensate for the display distortion, the pre-compensation maps can be implemented by a surface function, so that when the pre-compensation maps are applied to the input image data before the display, a display image formed on the viewing surface, There is basically no distortion. 18. The display calibration system of claim 17, wherein the at least one processor is further adapted to link various distortions and to generate surface functions that pre-compensate for distortion of the links. 71 200818114 19. The display calibration system of claim 17, wherein the surface functions are polynomials. A display calibration system of 17 items, wherein the processor is further adapted to adjust the surface functions to further compensate for at least one of an overscan condition and an underscan condition. 21 • A display calibration system for use with a display device having a viewing surface, the display calibration system comprising: at least one image sensing device adapted to sense a display from the viewing surface Information of the test image; and at least one processing state coupled to the at least one image sensing device, the at least one processor is adapted to calculate display distortion based on the sensed information, such that each piece is The severity of the display distortion within the 'dividing the viewing surface into small patches, and the pre-compensation maps associated with the display distortion in each patch, so that when such pre-compensation maps are applied to the input image material before display, A display image formed on the viewing surface is substantially free of distortion. 22. A display calibration system for use with a display device having a viewing surface, the display calibration system comprising: at least one image sensing device adapted to independently sense the viewing surface Displaying at least one color component related to the color component of the test image; and processing the processor to the at least one image sensing device, the at least one processor adapted to sense the 72 200818114, to calculate color non-uniformity, and to generate at least one color correction map related to at least one color component, so that when the at least one color correction map is applied to the input image material before display, one on the viewing surface In the display image formed above, there is generally no at least one color that is not uniform. 23. A display calibration system for use with a display device having a viewing surface, the display calibration system comprising: at least one image sensing device adapted to sense a display on the viewing surface Information of an individual color component test image; and 10 at least one processor coupled to the at least one image sensing device and the display device, the at least one processor adapted to independently calculate based on the sensed information Geometrically displaying distortion and causing at least one pre-compensation pattern associated with at least one color component to be independently generated such that when the at least one color correction map is applied to the image data before display, one is formed on the viewing surface The display image generally has no geometric distortion of at least one color dependency. 24. A display calibration method that can be used in a projection system having a curved viewing surface, the method comprising the steps of: projecting different portions of an image to the curved viewing surface using a plurality of projectors And correspondingly, each portion of the image is substantially focused on a corresponding portion of the curved viewing surface such that the image is optimally focused and integrally formed on the curved viewing surface. 25. The method of claim 24, wherein the method further 73 200818114 comprises the steps of independently positioning and orienting each of the plurality of projectors such that a projection axis of each of the projectors is substantially A portion corresponding to the curved viewing surface is oriented to optimize focus and minimize geometric distortion. 5 26. A display calibration method that can be used in a projection system having a curved viewing surface, the method comprising the steps of: measuring a majority distance from the curved viewing surface to a focal plane of the projected image; and offsetting The focus plane is until the function of the majority of the distances is minimized to obtain an optimized focus. 74