KR20060126344A - Fast anti-aliasing method - Google Patents

Abstract

A fast anti-aliasing method is provided to remove multiplication and division operation processes for area calculation of figures by judging whether sub-pixels are overlapped with segments. X-coordinates and Y-coordinates of sub-pixels extracted while scanning a figure along segments are stored in a sub-pixel information storing unit to extract sub-pixels overlapped along the segments of the figure by using figure information stored in a figure information storing unit(S11). It is judged whether residual segments of the figure stored in the figure information storing unit exist(S12). If the residual segments do not exist, the sub-pixels stored in the sub-pixel information storing unit are aligned on X and Y coordinates(S20). Intensity of each pixel is calculated through the sub-pixels aligned in the sub-pixel information storing unit and a result of the calculation is stored in a display buffer to reflect the pixel intensity in physical display(S30).

Description

Fast anti-aliasing method

1 to 4 are views showing a general antialiasing treatment scheme,

5 to 7 is a view showing a problem of the conventional antialiasing processing method,

8 is a functional block diagram illustrating a configuration of a display apparatus to which a fast antialiasing method is applied according to an embodiment of the present invention;

9 is an operation flowchart showing a fast antialiasing processing method according to an embodiment of the present invention;

10 to 12 illustrate a tenth step S10 in the fast antialiasing processing method of FIG. 9;

13 to 15 illustrate a process of drawing a triangle with 16 pixels;

16 to 35 are views illustrating a process of recording coordinates, intensities, and coverages of pixels in the fast antialiasing processing method of FIG. 9;

36 is a view showing a result value of the twentieth step S20 in the fast antialiasing processing method of FIG. 9;

37 is a view illustrating a result value of a thirtieth step S30 in the fast antialiasing processing method of FIG. 9;

<Explanation of symbols for the main parts of the drawings>

10: calculator 11: subpixel information extraction module

12: subpixel information alignment module 13: pixel color value calculation module

20: memory 21: figure information storage unit

22: subpixel information storage unit 23: display buffer

30: display unit

The present invention relates to a high speed anti-aliasing processing method in vector graphics, and more particularly, in the process of extracting a pixel overlapping each line segment with a figure forming a vector graphic, Rather than calculating the area of overlapping pixel parts, divide the pixel into several subpixels, perform a simple calculation to see if a particular subpixel overlaps the shape, and how many of the subpixels that make up a pixel The present invention relates to a high-speed antialiasing method that can eliminate the multiplication and division operations due to area calculations by determining the color intensity of pixels by counting overlaps, so that mathematically described figures can be shown more quickly.

As is well known, due to the rapid development of the computer, telecommunications, telecommunications industry and related technologies, modern society is increasing the use value of display devices. Among these, a case where a mathematically described figure is actually to be shown on a display device is frequently increasing, and related technologies are being actively developed.

However, when the mathematical figure is represented by the display device, image distortion is caused by the characteristics of the device. For example, in the case of representing a triangle, it should be drawn as shown in FIG. 1, but in the actual display device, each coordinate is inevitably specified in pixel units. This phenomenon is called a staircase phenomenon. In order to eliminate this, computer graphics performs an antialiasing process. Antialiasing refers to a process of reducing stepping by adjusting the brightness intensity of pixels at the boundary where stepping occurs. When the antialiasing process is performed on the figure of FIG. 2, it is shown as shown in FIG. 3.

In this case, the pixel intensity of the boundary portion is determined by calculating what percentage of a pixel the actual triangle represents, and an algorithm for calculating this is called area-sampling. When there is a pixel present at one boundary portion of the triangle as shown in FIG. 4, the brightness intensity of the pixel is calculated and displayed based on the area that the triangle covers the pixel.

On the other hand, in order to apply the area-sampling algorithm, it is necessary to find a contact point where a triangle segment is in contact with the x-axis and the y-axis in each pixel, and multiplication operation is required to calculate the area. There is a problem that requires a lot of computing power. A case of rendering a triangle composed of the coordinates "(3, 1), (1, 6), (7, 8)" will be described as an example. In this case, the mathematical triangle may be represented as shown in FIG. 5. To obtain pixel intensity of the outer portion of the triangle by applying the region-sampling to the triangle, each segment of the triangle is traced to overlap the line segment. Find the pixels and find out how much each pixel is covered by a triangle. That is, for the line segment proceeding from "(3, 1)-> (1, 6)", the areas shown in FIG. 6 are obtained, and the calculation made in one pixel is shown in FIG. It is calculated as in [Equation 1] below.

((L1 + L2) * pixel height) / 2

At this time, if the horizontal and vertical size of one pixel is "1", L1 and L2 are represented by a decimal point smaller than 1 in the figure of FIG. 7, which has a problem in that floating point operation is required. Therefore, in most practical algorithms, the horizontal and vertical sizes of one pixel are expressed by up-scaling, for example, "256". Thus, the triangle coordinates "(3, 1), (1, 6), (7, 8)" mean "(3 * 256, 1 * 256), (1 * 256, 6 * 256) while the program is processing them. , (7 * 256, 8 * 256) ". In the final step, the down-scaling process of dividing the coordinate value by " 256 " is performed again to return to the original value. At this time, the expression of multiplying or dividing " 256 " can be processed very quickly by shifting 8 bits.

However, even if the floating-point operation is eliminated in this way, the conventional zone-sampling algorithm that calculates the area as described above must find the intersection point where the X, Y axis and the line segment at each pixel intersect. Since multiplication and division are included in calculating the area, there is a problem that a lot of time is consumed as a result.

Accordingly, the present invention has been made to solve the above-mentioned conventional problems, and an object of the present invention is to show a mathematically described figure on a display device, thereby enabling a high speed antialiasing process to enable a faster display of the figure. To provide a way.

In order to achieve the above object, the antialiasing processing method of the present invention uses the figure information stored in the figure information storage unit to be applied to a display device having a figure information storage unit, a subpixel information storage unit, and a display buffer. A tenth step (S10) of extracting subpixels overlapping each line segment of the figure and storing the extracted subpixels in a subpixel information storage unit; A twentieth step (S20) of aligning the subpixels stored in the subpixel information storage unit with respect to X and Y; And a thirty step (S30) of calculating the intensity of each pixel through the subpixels arranged in the subpixel information storage unit and storing the result in the display buffer so that the pixel intensity is reflected in the physical display (S30). It is composed.

Further, in the antialiasing processing method of the present invention, the tenth step S10 may be performed by scanning along the line segment to extract the subpixels overlapping along the line segment of the figure using the figure information stored in the figure information storage unit. An eleventh step (S11) of extracting the X-coordinate and Y-coordinate of the data and storing the extracted X-coordinate and Y-coordinate in the subpixel information storage unit; In step 12, it is determined whether the remaining line segment of the figure stored in the figure information storage unit exists, and if so, the process proceeds to the eleventh step S11, and if not, the twelfth step S12. It is configured to include.

In addition, the anti-aliasing processing method of the present invention operates to record the coverage and intensity together when storing the X- and Y-coordinate values of the subpixels in the subpixel information storage unit in the eleventh step S11.

Hereinafter, a high speed antialiasing processing method according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

8 is a diagram illustrating a configuration of a display apparatus to which a fast antialiasing method is applied according to an embodiment of the present invention. The subpixel information extraction module 11, the subpixel information alignment module 12, and the pixel color value calculation module ( A calculator 10 having 13); A memory 20 having a figure information storage unit 21, a sub pixel information storage unit 22, and a display buffer 23; And a display unit 30.

First, the configuration of the calculator 10 will be described. The subpixel information extraction module 11 extracts the subpixels overlapping the figure along each line segment of the figure by using the coordinate information of the figure stored in the figure information storage unit 21 of the memory 20 and subtracts the result value. It serves to store in the pixel information storage unit 22. The subpixel information aligning module 12 aligns the subpixels stored in the subpixel information storage unit 22 with respect to the X and Y coordinates. It can be configured to sort in ascending-ascending order. The pixel color value calculation module 13 calculates the intensity of each pixel based on the subpixel information arranged in the subpixel information storage 22 and stores the result in the display buffer 21 to display the result in the display buffer 21. The role of the to be reflected in the physical display unit 30.

Next, the configuration of the memory 20 will be described. The figure information storage unit 21 is a memory that holds information about a figure, that is, information about a line segment forming a figure or information about a Bezier curve, in which the Bezier curve can be easily approximated with a line segment. The subpixel information storage unit 22 is a memory for retaining subpixel coordinates extracted along line segments constituting a figure, and additionally holds various information tables used in the progress of an algorithm for antialiasing processing according to the present invention. do. The display buffer 23 is a memory that maintains the intensity of each pixel to be shown in the display unit 30.

An embodiment of a high speed antialiasing processing method of the present invention applied to an apparatus having the above configuration will be described with reference to FIG. Conceptually, the algorithm according to the present invention does not calculate the area of the overlapping part in the process of extracting the cells overlapping with the line segments of the figure. Instead, one pixel is divided into 4, 16, 64, or 256 subfields. Perform the calculation by dividing by the pixel.

The subpixel information extraction module 11 extracts a subpixel overlapping the figure along each line segment of the figure by using the coordinate information of the figure stored in the figure information storage unit 21, and then stores the subpixel information storage unit 22 in the subpixel information storage unit 22. Save (S10).

In the tenth step S10, the subpixel information extraction module 11 scans along the line segment to extract the subpixels overlapping along the line segment of the figure using the figure coordinate information stored in the figure information storage unit 21. An eleventh step S11 of extracting x, y coordinates of the subpixel and storing the extracted subpixel in the subpixel information storage unit 22; The subpixel information extraction module 11 determines whether there is a remaining line segment of the figure stored in the figure information storage unit 21, and if so, proceeds to the eleventh step S11 again. A twelfth step S12 proceeds to step S20.

In relation to the twentieth step S20, the method of arranging the subpixels stored in the subpixel information storage unit 22 with respect to the X and Y coordinates may vary depending on the selection as described above. In FIG. 9, as an embodiment of the sorting method, FIG. 9 illustrates a case of arranging Y coordinates in ascending order and X coordinates in ascending order.

The process of the tenth step S10 will be described in more detail with reference to FIGS. 10 to 35 as follows.

FIG. 10 divides one pixel into 16 subpixels, and calculates only the subpixels in which a triangle segment overlaps. In other words, the area overlapping with the line segment is not calculated. In this case, if one pixel is divided into 16 subpixels, each coordinate is up to "(3 * 16, 1 * 16), (1 * 16, 6 * 16), (7 * 16, 8 * 16)". -Scaled. Here, the operation of multiplying 16 as described above is processed very quickly by shifting 4 bits.

When the subpixel overlapping with the line segment is found in FIG. 10, the result as shown in FIG. 11 may be obtained. As such, for example, a "Bresenham Line Drawing" algorithm or a Digital Differential Analyzer (DDA) algorithm may be used to track the subpixel overlapping with the line segment.

In order to determine the pixel intensity of FIG. 10, the number of subpixels covered by the figure must be counted, and the subpixels covered by the figure are the same as those of FIG. 12, which indicates the lower 4 bits of each subpixel coordinate obtained from FIG. 11. It can know by confirming. After all, pixel intensity is calculated as 12/16, or 75%. At this time, a higher quality can be obtained depending on how many subpixels are divided into pixels, and when divided into 256 subpixels, a value of 256 steps can be obtained, which enables a very high quality antialiasing process. do.

Meanwhile, a process of drawing a triangle as shown in FIG. 13 using 16 subpixels will be described below.

First, as shown in FIG. 14, multiplication of 16 coordinates of triangle coordinates is performed to up-scale the triangle coordinates. Accordingly, the triangle coordinates are “(1, 0) → (16, 0), (0, 1) → ( 0, 16), (2, 2) → (32, 32) ”. The intensity and coverage of each pixel are recorded while tracking the subpixels overlapping the figure along each line segment. For this purpose, a modified line drawing algorithm, that is, a "Bresenham Line Drawing" algorithm or a digital differential analyzer (DDA) algorithm may be used. That is, the intensity or coverage of actual pixels are recorded while finding each subpixel.

At this time, the intensity is recorded differently for the line segment going downward and the line segment going upward, and the coverage is recorded how much the height of one pixel is covered by the line segment, which also goes downward. The sign is recorded differently for the line segment and the line segment traveling upward. Therefore, when tracking each subpixel, the intensity and coverage are recorded while finding the subpixels as shown in FIG.

16 to 35 show a process in which the above calculation is performed. FIG. 16 shows the first table entry being added, and when the table entry is added, the strength and the coverage value are initialized to "0". At this time, the intensity and coverage have a negative value when the line segment progresses downward, and the strength and coverage value have a positive value when the line segment progresses upward. (This may be recorded in the opposite way depending on the processing method in the pixel color value calculation module. In the present specification, it is assumed to have a negative number when going down and a positive number when going up, but the spirit of the present invention is described herein. It is not limited.)

16 is currently in progress from " (16, 0) → (0, 16) ", so the intensity and coverage have negative values. Subpixels in the same sub-Y-scanline within a pixel have an intensity of "0, 1, 2, 3" from the left, respectively, so the subpixel shown in FIG. 16 is located at the leftmost within a pixel, so the intensity "0", and the coverage is "-1" because the coverage always increases "1" or decreases by "1" depending on the direction of the line segment. The result of reflecting the intensity "0" and the coverage "-1" for the subpixel shown in FIG. 16 in the table is shown in the table of FIG.

Next, the subpixel shown in Fig. 17 is located at the pixel coordinate (0, 0), which is different from (1, 0), which is the last entry coordinate of the table, so that a new entry is added. In this case, the initial strength and the coverage value are initialized to "0" as shown in the table of FIG. Meanwhile, since the intensity and the coverage of the subpixels shown in FIG. 17 are "-3" and "-1", the results are obtained by reflecting the results.

Since the pixel coordinates for the subpixels of FIG. 18 are the same as the last entry coordinates of the table entry, no new entry is added. In this case, the strength and the coverage value are added to the immediately preceding table entry as shown in the table of FIG. 18.

19 and 20 show the last subpixel of " (16, 0) → (0, 16) ", and the intensity and coverage values in this case are as shown in the table of FIG.

FIG. 21 shows a state of starting edge tracking for a line segment proceeding from the next line segment "(0, 16) to (32, 32)". At this time, the subpixel at the point where two line segments meet is always detected twice, and the subpixel (0, 16) is detected once when proceeding from "(16, 0) to (0, 16)" (the previous step). Is recorded again), and is detected again when proceeding from "(0, 16) to (32, 32)". The cell at the point where the line segments overlap is recorded if the direction of the line segment changes, while not recorded if the direction of the line segment is the same. In the situation of FIG. 21, since the directions of "(16, 0)-> (0, 16)" and "(0, 16)-> (32, 32)" both advance downward, the recording is not performed.

22 to 24 show that a new table entry is added to “(0, 16) → (32, 32)”, and FIG. 25 shows the last sub of “(0, 16) → (32, 32)”. In this case, the intensity and coverage values are shown in the table of FIG.

Fig. 26 shows the start of edge tracking for the line segment proceeding from "(32, 32) to (16, 0)". At this time, the first subcell (32, 32) has a direction in which the previous line segment "(0, 16) → (32, 32)" and the current line segment "(32, 32) → (16, 0)" are mutually different. The opposite is written as shown in the table of FIG.

27 to 33 show that a new table entry is added to "(32, 32)-> (16, 0)", and FIG. 34 shows "(32, 32)-> (16, 0)". The last subpixel is shown, in which case the intensity and coverage values are shown in the table of FIG. 34. At this time, whether to record the first subpixel in the above should be determined depending on whether the first and last line segments have the same advancing direction. In the above process, the judgment part is partially omitted for convenience.

By the above-described process, a table as shown in FIG. 35 is finally obtained. The table value is stored in the subpixel information storage unit 22 of the memory 20.

Thereafter, the subpixel information alignment module 12 aligns the subpixels stored in the subpixel information storage 22 with respect to the X and Y coordinates (S20). The order of sorting will generally be in the order of Y coordinate-ascending first and X coordinate-ascending second, but the direction of drawing (top to bottom or bottom to top of the screen) and the direction of drawing each line (of screen). It can be selected differently from left to right or right to left), and the sorting method can be a "Quick Sort" sorting algorithm or a "Shell Sort" sorting algorithm.

Subsequently, the pixel color value calculation module 13 sums the intensity and the coverage value of the same pixel coordinates. For example, there are two table entries with coordinate values (1, 0) in the table of FIG. 36, but the coverage and intensity values of the two table entries are summed, so that one entry with coordinate values (1, 0) is shown in FIG. Became. Completion of this process through the pixel color value calculation module results in the result as shown in FIG. Subsequently, the pixel color value calculation module from the finally obtained table 37 reflects the color values of the pixels in the display buffer 23 so that the contents of the display buffer 23 are reflected in the physical display unit 30 (S30). . Therefore, each Y-pixel line is drawn with an appropriate intensity from top to bottom through the information as shown in FIG.

The pixel color value calculation module 13 may calculate the intensity of each pixel by using a pseudo code shown below.

The table below refers to the table of FIG. 37, and accesses the information of each table entry through a While loop to draw Span (each row image to be drawn to output a picture to a span-raster device). Serve The X and Y coordinates of the current pixel are maintained by the $ CurrX and $ CurrY coordinates, and the color intensity of the pixel is represented by $ C. In addition, $ Coverage maintains the coverage value of the entry currently being accessed. The following process assumes that the picture is drawn from the top of the screen to the bottom and from the left to the right of the screen, but this may be implemented vice versa. In addition, the code below assumes that one pixel is divided into 16 subpixels.

   Function Rasterize Select the first element in the table While Current element is not exceed table $ CurrY = Current element's Y $ Coverage = 0; Do If $ Coverage <> 0 $ C =-$ Coverage * 4 'There are 4 subpixels in row in one pixel $ CurrX = $ CurrX + 1 While $ CurrX <Current element's X Draw pixel at ($ CurrX, $ CurrY) with $ C intensity. $ CurrX = $ CurrX + 1 End While End If $ CurrX = Current element's X $ C =-(Current element's intensity) +-$ Coverage * 4 Draw pixel at ($ CurrX, $ CurrY) with $ C intensity Select the next element . While $ CurrY == Current element's Y; End While End Function

Although the present invention has been described in more detail with reference to some embodiments, the present invention is not necessarily limited to these embodiments, and various modifications can be made without departing from the spirit of the present invention.

As described above, according to the anti-aliasing processing method according to the present invention, when extracting a pixel overlapping each line segment of the figure, only the subpixel overlaps with the line segment and does not calculate the coverage area. By eliminating multiplication and division operations, it is possible to show mathematically described figures faster, which maximizes system efficiency and improves service quality.

Claims (3)

An antialiasing processing method for applying to a display device having a figure information storage unit, a subpixel information storage unit, and a display buffer, A tenth step (S10) of extracting subpixels overlapping with each line segment of the figure by using the figure information stored in the figure information storage unit and storing the extracted subpixels in the subpixel information storage unit; A twentieth step (S20) of aligning the subpixels stored in the subpixel information storage unit with respect to X and Y coordinates; And And a thirty step (S30) of calculating the intensity of each pixel through the subpixels arranged in the subpixel information storage unit and storing the result value in the display buffer so that the pixel intensity is reflected in the physical display (S30). A high speed antialiasing processing method, characterized in that. The method of claim 1, wherein the tenth step (S10), Scanning along the line segment to extract the subpixels overlapping along the line segment of the figure using the geometry information stored in the figure information storage unit, extracting the X- and Y-coordinates of the subpixel and storing the subpixels in the subpixel information storage unit. An eleventh step (S11); In step 12, it is determined whether the remaining line segment of the figure stored in the figure information storage unit exists, and if so, the process proceeds to the eleventh step S11, and if not, the twelfth step S12. A high speed antialiasing processing method comprising a. The method of claim 2, The method of claim 11, wherein the coverage and intensity are recorded together when the X- and Y-coordinate values of the subpixels are stored in the subpixel information storage unit in the eleventh step (S11).

Images