TW588289B - 3-D digital image processor and method for visibility processing for use in the same - Google Patents

Abstract

A three-dimensional (3-D) digital image processor and a method for processing a visibility for use in a displaying procedure of a 3-D digital image are disclosed. The 3-D digital image processor includes a depth map generator, a memory device and a rendering engine. The method includes steps of presetting a depth map according to a plurality of pixels received, the depth map storing the pixels and reference depths corresponding thereto, and receiving a pixel data and proceeding a visibility test with reference to the depth map, thereby determining whether to proceed a rendering operation on the 3-D digital image by the pixel data.

Description

588289 _Case No. 91123147 V. Description of the Invention (1) Field of the Invention The present invention relates to a three-dimensional digital image processor and the visibility processing method applied thereto, especially the three-dimensional digital image processor and personal computer system. Apply the visibility processing method above. BACKGROUND OF THE INVENTION In a three-dimensional space image application, an object system in a scene is represented by a three-dimensional space image model. For example, using a polygon mesh, the surface of an object can be modeled by several interconnected polygons. The surface rendering process (: rendering process) usually provides the model data required for the rasterizing process by transforming the endpoints on the geometric primitives. The so-called rasterization generally refers to the process of calculating the pixel value of a pixel in visual space based on the geometric basic elements projected onto or covering the pixel. Please refer to the first figure, which is a functional block diagram of a 3D graphics engine. The 3D graphics engine mainly includes a transform-lighting engine for geometric calculations; a setting engine 12 (setup engine) is used to initialize the basic element; a scan converter 13 is used to derive pixel coordinates; a color calculator 14 is used to generate smooth colors; a texture unit 15 is used To deal with textures; a transparency

Page 6 588289 / 〇3 3. ίδ _ Case No. 91123147, the base month day is __ V. Description of the invention (2) The alpha blending unit 16 is used to produce transparent and translucent effects, a depth test The unit 17 (depthtestunit) is used to remove the hidden surface composed of pixels; a display controller 18 (disp 1 ay control ler) is used to accurately display images on the display 21 and so on. The 3D graphics engine receives and executes the instructions stored in the instruction queue 10, and the memory controller 19 accesses the graphics memory 20 through the memory bus. The instruction queue 1 0 is a F 丨 F 0 unit for storing instruction data, which receives the instruction data from the controller 1 through a system bus. In a given three-dimensional image scene, there may be multiple polygons projected on the same area on the projection plane at the same time. 'This way', there will be some basic elements that cannot be seen in this scene 'The depth test unit 17 described above is used to remove the hidden surface composed of pixels. Therefore, many hidden surface removal algorithms have been developed. One of the most well-known algorithms is a Z-buffer algorithm that uses a Z-buffer to store the depth value of each drawing point. Method (Z-buffer alg0rithm). The core of this algorithm is a sphericity comparison mechanism for the depth value of each point received and the depth value that was originally stored in the B buffer. For a point (χ 'y) on a plane (facet), its depth value can be interpolated from the depth value of the vertex of the plane to obtain the corresponding depth value of the' coordinate (X, y) by the z_buffer Obtain and perform a depth test to determine which one is closer to the viewer by comparing the two depth values. The Z-buffer is then updated with a closer depth value. So ’the Z-buffer reflects every point in the projection plane so far

Page 7 588289 ^ ^ 19 __ Case No. 91123147 'Month :: Day 佟 i_ V. The closest depth value state of the invention description (3). For example, suppose the viewer is located at the origin 0 of the Z coordinate, and the viewing direction is toward the positive Z-axis. Therefore, the z-buffer can retain the smallest z value of each drawing point so far. The Z-buffer algorithm is also the easiest way to remove hidden surfaces in modern computer graphics systems. The pseudocode of the Z-buffer algorithm is shown below.

For (each polygon) {

For (each pixel in polygon ’s projection) {Calculate pixel ’s z value (source-z) at coordinates (χ, y);

Read destination-z from Z-buffer (x, y); If (source-z is closer to the viewer) Write source-z to Z-buffer (x, y); It is known that one of the main problems of modern 3D applications is Overlay drawing (〇verdraw). Most graphics processors have no way of knowing which parts of the scene are visible and which ones need to be covered before the shading process begins. Because the core of the Z-buffer algorithm lies in the depth comparison mechanism between the depth value of each input pixel and the depth value stored in the Z-buffer. During the depth comparison process, many pixels are written to the daytime buffer (f r a m e b u f f r r) and then overwritten by new pixels closer to the viewer. Repeated drawing means

Page 8 588289 JI: The page number 91123147 is being replaced 3 Amendment 11 I " ··· 1 ·· LassaaiaaccB — ^ ―—- rf ·· 1 _ 5. Description of the invention (4) This kind of screen buffer Repeated writing of pixels in the device. The number of repeated depictions in a scene is called depth complexity and represents the ratio of total shaded pixels to visible pixels. For example, if a scene has a depth complexity of 4, it means that the shaded pixels are 4 times as many pixels as are actually visible in the scene. In a complex 3D scene, a large number of objects overlap each other. From the perspective of the depth comparison mechanism, polygons (basic elements) in the order from the back are preferred. Pixels with larger depth values (away from the viewer) will be discarded after depth comparison processing, because overlapping pixels with smaller depth values (closer to the viewer) have already been drawn. Otherwise, the new pixels will be colored, and the current depth and color values will be re-written in the corresponding pixel positions of the depth buffer and day buffer. In visible pixels, if not discarded in the early stages of drawing lines, this shading process obviously consumes a lot of processing and memory resources. The second figure is an example of a top-view drawing scene. The viewer's field of view is indicated by a dotted line. Visible objects in the scene are represented by black dotted lines. As shown in the second figure, most of the objects in this example are hidden, which greatly reduces the efficiency of the graphics shading system due to repeated drawing. Traditional drawing hardware attempts to solve this problem with Z-ordering to eliminate some redundant information. Although the foregoing method can avoid the memory bandwidth required for the pixel visibility test one by one, it cannot overcome the problem of repeated rendering, and still retains a considerable amount of unnecessary calculations and memory requirements. For example, if the basic elements of the geometry are depicted in order from back to front (from far to near), a group of pixels will pass the visibility test, and undesired repetitive rendering will occur. Therefore, although the Z-buffer algorithm is easily implemented in software or hardware, it does not require pre-sequencing. Z-buffer reflects projection so far

588289 Case No. 91123147 $ Zhijin Narrow: 铒 Amendment V. Description of the Invention (5) The closest depth value of each point on the plane. However, as mentioned earlier, if the coloring order of objects is from back to front, the traditional Z-buffer algorithm cannot solve the problem of repeated drawing. Therefore, how to resolve the lack of conventional means is the main purpose of developing this case. Summary of the Invention A plurality of functions of the image of the present case are stored in the image energy space number according to a two-dimensionality mapping function value of the visibility measurement and the depth data to be referenced. According to the two-dimensional system of the point is first established for the display pixel storage. The visibility of the prime points depends on the number of coordinates of the above-mentioned experimental system and the sounding value is more than facing the above-mentioned depth coordinates. An intermediate point of the process can be used to predict the poor image of the image and test. The concept of the image includes the following visibility. Refer to It ’s conception of approaching the color-viewing action in the pixels corresponding to the depth value corresponding to the depth to be measured. The mapping function is used to enter the depth processing method of the number of cases. The application method includes the following steps: Corresponding parameters have been established to determine whether to perform the color shading with the image point as described in the visibility step: take out the pixel degree value; according to the second test depth value; and the depth value closer to the observer, it is convenient Third degree step: According to the number, the depth of depth of the deep examination of the depth map prime point data action. The method uses the dimensional coordinate input in the point data to compare the parameter depth value, not to map with the degree of acceptance of the image space number; and a function to the three of which the inclusion is included in the depth of examination and when the prime point asset is described. Visibility processing method, wherein the pre-method includes the following steps: mapping these pixel functions to corresponding original parameters

Page 10 588289 Change:] month to year correction la. Description of the invention _______________ —; fruit value .................. ........... 1 value = deep mapping function of preset reference depth values such as f and f and the pixels themselves = j and update action to determine whether to update the deep root t The original reference depth values. According to the above conception, the A comparison and update action described in the present case includes the following step df processing method, wherein :: the depth of the prime point itself: 2 the original i: far reference is maintained when the depth value of the original reference depth value ': The depth value is closer to the first one: According to the above concept, the value of 4ί in this case remains unchanged. g. The method of establishing the depth mapping function is more processing method 'which is pre-made': $ requires another-visibility test only; step: confirm that the image is not updated, the comparison and update are based on the above conception 'as described in this case The "j degree test" is a transparency color mixing test. Another method, wherein another aspect of the other aspect: the reading device includes:-a depth mapping function processor to establish a plurality of pixels in advance-; f is based on the mapping function stored in the mapping function of the pixel, and The depth should be related; a note 愔 w your two •,-the opposite of the dimensional coordinate and the depth value, which is provided for: Hai :; set 'belief f is widely connected to the depth mapping function to generate f, which stores the mapping function ; And-surface shading introduction, the five-degree degree ratio and corresponding pixel points perform a surface shading action in the image, and the surface shading engine

$ 11 pages of the invention description (γ) Try to recall the depth mapping function in the device-see if the i tone = 中 in the visibility measurement is to use the pixel data to perform surface shading on the three-dimensional digital image pixel point . Device, I = t. The visibility test of the black row of three-dimensional digital image processing pixels described in this case includes the following steps: taking out the two-dimensional coordinates at one of the two-dimensional coordinates contained in t and one to be measured Depth value; According to and; Degree mapping function corresponding to a reference depth value; Depth of the person: Which of the measured depth values is closer to the observation of color movement: "The surface color engine does not use the pixel data to make a surface basis According to the above conception, the third generation of the life gates described in this case, where the deep-producing map of Lexu Mountain and the digital image processing coordinates is turned on and off. The generating device is to create a complex image of these images. Then, the original reference depth values of these preset reference depths are separately compared and updated; the original reference depth sounds in the depth value projection function of the edge pixels themselves are determined to determine whether to update The depth mapping = according to the conception of the description, the depth mapping function described in this case, and the spatial digital image processing: the operation system includes the following steps: the comparison and update performed by the farm home Contact value and the pixel The ice value is closer to the observer's ice value; when the depth value of the pixel j replaces the depth mapping function, the new reference depth value of the pixel itself; and when the parameter = f original reference depth Value becomes a test depth value closer to the depth of the observer ⑽ ⑽ 疋 Replace Λ page

Degree values Japan ------ ',' Keep the original reference depth value unchanged. According to the conception described above, the three-dimensional digital image processing described in this case is to perform the following steps: the depth mapping function generating device further performs the following steps: confirmation = speech point does not need to perform another visibility test before performing the comparison and Update crying, according to the above conception, the third-degree spatial digital image processing master described in this case is one of the visibility testing is the transparency color mixing test. „F According to the above conception, the three-dimensional digital image processing described in this case = ,,, and more includes a picture buffer 'signal connected to the surface coloring index = 丄 It is used by the surface coloring engine to perform surface coloring operations. For a description of the preferred embodiment of the pixel, please refer to the third figure, which is a functional block diagram of a preferred embodiment developed for the 3D graphics engine in this case, in which a transform-lighting engine 31 is used for Geometry calculation; setup engine 32 (setup engine) to initialize the basic element; scan converter 33 (scan converter) to derive pixel coordinates; color calculator 34 (color calculator) to produce smooth colors; texture unit 35 (texture unit) used to process textures; transparency blending unit 36

(alpha blending unit) is used to produce transparent and translucent effects; depth test unit 37 (d e p t h t e s t u n i t) is used to remove hidden surfaces composed of pixels; and display controller 38 (display contfQl lerO

Page 13 588289% Ύ Η: MU9 j \ __ ii No. 1 911HU7 V. Description of the invention (9) Used to accurately display the image on the display 41. The surface shading engine 44 is composed of a color calculator 34, a texture unit 35, a texture unit 35, an alpha bl ending unit umt, and a depth test unit 37. composition. The 3D graphics engine receives and executes the instructions stored in the instruction queue 40. The memory controller 39 stores the graphics memory 3 through the memory bus. As for the instruction queue 3, it is a F for storing instruction data? 〇It receives the command data from the controller 3 through the system bus. The main feature of this case is that a depth map function generator 42 (depth map generator) is added between the transform- ^ ghtlng engine 3l and the setup engine 32 (setup engine). After the conversion-lighting engine (transf〇rm-Hghting englne) 3 ^ processed pixel data, take out the three-dimensional space = set one-dimensional coordinates (X, Y) and a depth value Z to establish a depth Mapping function J, and the depth mapping function is mainly used to store the correspondence of each image in the display screen-the correspondence between the dimension depth private (χ, Υ) and its corresponding reference depth value Zr-but because the 3 bowel scenes are mostly A plurality of objects are overlapped back and forth to make up &-what the picture does not,) 'Therefore, in order to obtain the positive HΓ cloth of the entire 3D video scene, & t ϊ ί (X, Y) When the data of pixels with different depths is used, a ratio r: Cang / clothing value Zr and the depth value Z of the pixel itself are performed: pound action, # to determine whether to update (A; ^ ... in the depth mapping function Baicalensis test respiratory droplets value itself of the pixel depth value, where

Page 14 588289

Case No. 91123147, description of the invention (10) _ This value is closer to the observer's depth value; (b) When the pixel point is close to the observer's depth value, the pixel itself ^ the original value of the depth value depth mapping function The inflammation person's deafness / woodiness value is replaced— (Ot. ^ Λ / # ^ ^ Keep the original reference depth value unchanged. ', When the ice degree value, in this way, the depth map is passed at all pixels τ ★ 4 After 2, we can get a set up-M, mapping function generating device map). The depth mapping function 22 and the mapping function (depth memory device 43, or memory ;: independent setting memory (tempOrary memry) shown in the figure. Because of I-shaped; self-memory 30 of the first For temporary color rendering (render), it can be painful by referring to the deep trajectory function of the subsequent subsequent surface = unnecessary overdraw motion: received when performing the surface coloring motion; 2 ::, my ：: Use the depth mapping function that has been established! The prime point data, Vlsiblllty test, to determine whether or not 仃 —Visibility test The image in the degree space digital image ^ δ Hai pixel point data to ^^, (viSlbllity ° The following steps are included in the pixel data: 3) U) Take out (b) Enter the radiance ^ and a depth value to be measured according to the two-dimensional coordinates; the depth value; Reference: It is closer to the depth of the observer, and "When the depth of the heart, the depth value is not the same as the pixel value = the value is closer: Observe. Compared with the above depth mapping function, the surface is more colored. You. '"In the two-dimensional coordinates and the depth value to be measured, = it for example 588289 ----- Case No. 91123147 Year P: 1: 4 Amendment V. Description of Invention (11) Texture (te X ture), color (c 〇 〇 〇) and other data are ignored, so the system can be greatly reduced Consumption of computing power and memory bandwidth, but because the test to determine whether a pixel data is up or discarded is not only a single visibility test, but there are other such as transparency Various color visibility tests (a 1 phab 1 ending test) or operation of transparency (vi s ibi1i ty test). The transparency color mixing test (aipha b 1 ending ΐ est) is the pixels that will be received. The alpha value in the point data is compared with a reference transparency value (reference aipha va 1 ue). If the pixel fails the transparency blending test, the pixel data will also be discarded without going Update the data in the frame buffer and Z buffer defined in the graphic body 30. The problem is, The lightness value is generated by operations such as texture mapping and alpha blending test. The texture mapping requires access to a large amount of texture data from the texture register, and the transparency blending requires the target frame Buffer data, used to mix the source color and the target color. For the transparency of 3D drawing scenes, the color is mixed. The foreground object must be mixed with the drawn background object. Because the coloring action of each pixel is not only determined by its depth value, the preliminary depth mapping function mentioned above cannot meet the actual needs. To improve this problem, the above depth mapping function is better than the update action. The embodiment is a flowchart of the steps shown in the fourth figure. For a point (X, y) on a polyhedron, the depth value to be measured can be within the depth of the vertex of the polyhedron.

Page 16 588289 Case No. 91123147 Revised

&. ¾ .; 3 < Μ. Amendment 5. Interpolation of the invention description, which is inferred from the number which is more than the mapping function depth test. Lightness and color mixing depth value. The coloring stage will be followed by a step to establish a perfect embodiment. Transparency and visibility. Create updates due to bandwidth. In summary, the action of the color in the memory is to move and enter the scene to set up any protector. (12) and coordinates. Use Close View. However, it also needs to be tested. It will not be decided. See the depth step of the first step. After calculation, the reference depth that can still be (X, y) can be compared from the depth map function's two depth values. Perform a depth test to determine the depth, and then update the depth with a closer depth value. If a pixel is day out or discarded depends not only on timing based on another visibility test, such as the above-mentioned penetration, the depth mapping function The reference coordinate of (X, y) is modified. In this way, the visibility test will be left on the surface. This figure is a comparison of the preferred embodiment of the initial mapping function in the case of surface coloring. It is mainly aimed at pixels that need to perform other visibility calculations, such as, and so on, and perform comparison and update operations after performing another depth test. Most of the data in the depth mapping function does not need to be described by conventional methods to save a lot of system resources and memory. This case is created in advance with a small amount of information and temporarily stored in the depth mapping function, which can provide the surface shading engine in Performing reference, this will save a lot of unnecessary over-drawing, save a lot of system resources and memory bandwidth, and make the speed increase. Therefore, the invention of this case can be modified by people who are familiar with this technique, but it is not as good as the scope of the patent application.

Page 17 588289 jX: 3. 19 _Case No. 91123147_I ^ Month Revision _ Simple Illustration Simple Illustration This case has a deeper understanding through the following drawings and detailed description: First picture: its It is a functional block diagram of the 3D graphics engine. The second picture: it is a schematic diagram of a top view example of a 3D scene. The third figure is a functional block diagram of a preferred embodiment developed for the 3D graphics engine in this case. The fourth figure is a flowchart of the steps in the preferred embodiment of the present embodiment in which a comparison operation of the depth mapping function is performed. Fifth figure: It is a flow chart of the steps of a preferred embodiment of a depth mapping function developed during the surface coloring phase developed in this case. The components included in the drawings in this case are listed below: Controller 1 Conversion-Lighting Engine 11 Setting Engine 12 Scan Converter 13 Color Calculator 1 4 Mapping Unit 1 5 Transparency Mixing Unit 1 6

Page 18 588289 only. Case No. 91123147 U-day_correction diagram brief description of depth test unit 1 7 display controller 1 8 memory controller 1 9 graphics memory 2 0 display 2 1 instruction queue 1 0 depth mapping function generation Device 4 2 Memory device 4 3 Surface shading engine 44 Controller 3 Conversion-lighting engine 3 1 Setting engine 32 Scan converter 3 3 Color calculator 3 4 Mapping unit 35 Transparency mixing unit 3 6 Depth testing unit 3 7 Display controller 3 8 Memory controller 3 9 Graphics memory 4 0 Display 4 1 Instruction queue 3 0 Depth mapping function generator 4 2

Page 19

Claims (1)

588289 Ά 9Π2.ΊΗΤ] Visualization of digital images in degree space VI. Patent application i. Visibility processing methods are applied to a process, the method includes the following steps: According to the received plurality of pixels to establish in advance-depth number, the depth mapping function system These pixels are stored with their reference depth values, as well as the depth number that has been established and referenced: the number of gates i ί1 to determine whether the surface coloration of the pixel in the three-dimensional work shirt image of the pixel is dynamic. In the scope of the patent application, the first tip and item 1 of this test are the methods of processing the visibility returned by the sentencers. In the case of visibility measurement, you include the following steps: ^ Take out the degree value contained in the pixel data of '丄. T, Depth the depth value according to the two-dimensional coordinates; and] ^ ^ depth mapping function to compare the depth value of the inviter corresponding to the reference depth value, and when reading which one of the measured depth values is compared, the pixel is not used. ^ 1 The taste value is closer to the observer's ~ 4 lean material for surface coloring action. Corresponding parameter mapping function of degree mapping function. Among them, the dimensional coordinate of the month * and a reference depth to be measured, a reference value close to the observation depth value 3. If the scope of the patent application is first established, the depth mapping function " The second method includes the following steps: Enter the relative original reference, location, and quasi-coordinates and enter the depth mapping function to correspond to the knee outer ridge 7, the hydration value; preset the 4 reference ^ / wood & value to the pixels The depth value of the point itself 588289 Case No. 91123147 Man / jx Xianbanbei Year, month, month, month, day, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, year, month, month, month, year, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, month, year, month, month, or month 4. The visibility processing method as described in item 3 of the scope of patent application, wherein the comparison and update action includes the following steps: compare the original reference depth value with the depth value of the pixel itself, which is closer to the observer's depth value ; When the depth value of the pixel point is closer to the depth value of the observer, replacing the original reference depth value of the depth mapping function with the depth value of the pixel point itself to become a new reference depth value; and when the reference depth value is When the value is closer to the depth value of the observer, the original reference depth value is maintained unchanged. 5. The visibility processing method described in item 3 of the scope of the patent application, wherein the method of establishing the depth mapping function in advance further includes the following steps: the pixel point is not required to perform another visibility test before the comparison and update operation is performed. 6. The visibility treatment method described in item 5 of the scope of patent application, wherein the other visibility test is a transparency blending test. 7. A three-dimensional spatial digital image processor, comprising: a depth mapping function generating device, which is based on a plurality of pixels received to establish a depth mapping function in advance, and the depth mapping function is j number 9 also 3147 VI. The scope of the patent application contains 588289 1Φ of these pixels, which means that the relationship between the coordinate and depth of the storage area and the depth value is adjusted, and the signal is connected to the M relationship; For the depth mapping function to be stored; and, a function generating device, a surface coloring bow, which will receive ~ i degrees ::?:? The corresponding pixel points in the image are two? : It is expected that the function, the visibility test mapping maps the surface of the pixel in the digital image between m:: Data 8. The three-dimensional device as described in item 7 of the scope of patent application, the visibility test performed It contains 1 sequence of numbers / image processing to extract one of the 乂 degree values contained in the pixel data; I depth ΪΠΓ coordinates are entered into the depth mapping function to correspond-reference to compare the reference depth value with the measured depth value Which is closer to the depth value of the person, and when the reference depth value is closer to the depth value of the observer, then the surface shading engine is controlled not to perform a 2 | color action with the pixel data. Furthermore 9 · The three-dimensional digital image processing as described in item 7 of the scope of patent application 1§ 'wherein the depth mapping function generating device inputs the -dimensional | coordinates of these pixels into the depth mapping function and corresponds to the relative After the original reference depth value, the preset reference depth values and the depth dimension of the pixels themselves and the dimensional coordinate and — depth to be measured Page 22 588289 An Ikt case number 91123U7, amended by six points to apply for patent scope __ Do not perform / compare and update actions to determine the original reference depth values in the ~ θ function. ', Whether to update the depth mapping 1 10. As described in item 9 of the scope of the patent application, wherein the depth mapping function generates a second image processing operation including the following steps: The comparison performed by Jin And update = compare the original reference depth value with the pixel point, the depth value closer to the observer; if not, the clothing value, why the depth value of the pixel itself is closer to the depth value of the observer, the depth And the depth value of the body replaces the original reference list of the depth mapping function as a new reference depth value; and when the reference depth is closer to the observer's depth value, the original value is maintained. The degree value does not change. , Its application. ^ Three-dimensional space digital image processing described in item 9 of the patent scope. The image ^ f The depth mapping function generating device further performs the following steps: confirm the action. The point and point do not need to perform another visibility test before performing the comparison and update. The third-degree spatial digital image processing described in item 11 of the patent scope of the patent application. The other visibility test is a transparency blending test. Device, as described in item 7 of the scope of the patent application for digital image processing in three degrees. It also includes a daylight buffer, and the signal is connected to the surface coloring. Page 23 588289 Case No. 91123147羌 V% · Revision 6. The scope of the patent application engine is used to write the pixel data when the surface shading engine performs the surface shading action. Page 24