JPH0676075A - Alias removing system in three-dimensional picture generation - Google Patents

Abstract

PURPOSE:To provide more real pictures by detecting the area of a picture buffer corresponding to the contour part of a pattern of/on an object using the differentiated values of the picture buffer and a Z buffer and filtering it. CONSTITUTION:When picture generation is completed, a differential filter 6 reads the picture buffer 3, detects a part where the fluctuation of the brightness is large and raises a bit 1 at a position inside a mask buffer 7 corresponding to the part. Then, the differential filter 6 reads the Z buffer 4, similarly detects the part where the fluctuation of a Z value is large and raises the bit 1 at the position inside the mask buffer 7 corresponding to the part. When contour detection is completed, a controller 1 selects a frame buffer 10 by controlling a switch 9 and an averaging filter 8 reads picture data from the picture buffer 3 and mask data from the mask buffer 7, exchanging them with a weighing average value with the adjacent picture elements of the picture data for which a mask value is '1' and writes a picture element value in the frame buffer 10.

Description

Detailed Description of the Invention

[0001]

[Industrial field of application] Due to advances in computer image generation technology, its application fields include various designs, visualization of analysis results,
Creation of presentation materials, animation, and even art has spread. Generally, in these application areas:
Commonly, there is a demand for high-speed and realistic image generation with small hardware.

[0002]

2. Description of the Related Art When a computer-generated image is displayed on a raster scan CRT, an image buffer (also called a frame buffer) that stores pixel data corresponding to all the pixels normally displayed on the screen of the CRT is one-to-one. ) Is installed, first, all pixel data to be displayed on the CRT is generated on the computer side and is temporarily stored in this image buffer. And CR
The display for T is performed by reading the contents of the image buffer in synchronization with the scanning of the CRT screen and sending it to the CRT screen. The reading of the image buffer for displaying on the CRT is performed independently of the operation of the computer. It is generally known that such a display method has a problem in that a staircase-like display appears on a diagonal line of a screen or a diagonal outline of a figure. Even when the image of the three-dimensional object generated in the image buffer by the computer is displayed on the raster scan type CRT, the contour line of the object image or the contour line of the pattern on the object may be displayed in a stepwise manner.

The phenomenon in which this step-like display appears is called aliasing. On the other hand, dealing with this phenomenon is called anti-aliasing.

Although some methods are known for this antialiasing, they are roughly classified into a supersampling method and a filtering method.

In the super-sampling method, when the color or brightness of one pixel stored in the image buffer is determined, the pixel is divided into smaller pixels, the color or brightness of each pixel is calculated, and the average value thereof is calculated. And obtain the value as the pixel value.

In the filtering method, the value of a pixel is temporarily calculated in the unit of the pixel stored in the image buffer, and then the weighting calculation with the surrounding pixels is performed to obtain the actual value of the display pixel.

[0007]

As described above, in the supersampling method, when the number of pixel divisions is n, 2 of n is obtained.
The value of the power of pixels will be calculated. For example, n
Even if = 2, four times as many values are calculated, and there is a problem that the calculation cost increases significantly.

On the other hand, in the filtering method, what is needed is a product-sum calculation between pixels, and the calculation cost for the calculation is small. However, since the entire image buffer is processed, there is a problem that the entire image is blurred.

Therefore, in the filtering method, only a necessary portion is filtered to prevent the entire image from being blurred (1) a method of tracing the contour of an image on the Z buffer, or (2) differentiating the Z buffer. Japanese Patent Laid-Open No. 3-139774 discloses a method of detecting the contour portion of the image stored in the image buffer by any one of the methods and filtering only the portion.

However, the method (1) is disclosed in JP-A-3-139.
As pointed out in Japanese Patent No. 774, there is a problem that contour tracking requires a large calculation cost. On the other hand, the method (2) has a problem that when there is no significant difference in Z value, for example, when two objects are very close to each other and overlap each other, the contour of the overlapping portion cannot be detected. Further, since the contour of the pattern on the object is not reflected in the Z buffer, there is a problem that both the methods (1) and (2) cannot be detected.

Therefore, as another method, it is conceivable to differentiate the image buffer to detect the contour, but this method alone does not cause a significant difference in color or brightness between overlapping objects or the background. However, there is a problem that the contour cannot be detected.

In view of such a problem, the present invention is a method of more reliably detecting a target area of antialiasing processing and obtaining a more realistic image without blurring the whole and without increasing the calculation cost too much. The purpose is to provide.

[0013]

In order to achieve the above object, the present invention provides an object and an object on the object by using a differential value of the Z buffer and a differential value of the image buffer in an image generation method using a Z buffer. The feature is that an area of the image buffer corresponding to the contour portion of the pattern is detected, and the area is filtered.

[0014]

By differentiating the Z buffer, the contour portion of the object having a significant difference in Z value can be detected. Further, by differentiating the image buffer, it is possible to detect the contour portion of the pattern on the object having a significant difference in color or brightness. Further, there is no significant difference in Z value, and the outline portion of the overlapping portion, which cannot be detected by the differentiation of the Z buffer, also has a large difference in color or brightness. It can be detected by differentiation. In this way, by filtering only the detected part instead of filtering the whole part, antialiasing is executed at a low calculation cost for all parts where the whole is not blurred and aliases are conspicuous. A method can be provided.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described below with reference to FIG. 1 and a modified second embodiment will be described with reference to FIG.

In FIG. 1, the graphic processor 2 is instructed by a start signal of the controller 1 from Japanese Patent Laid-Open No. 3-139.
The Z buffer 4 is used to generate an image of a hidden surface-erased three-dimensional figure by the method described in Japanese Patent No. 774, etc.

When the image generation is completed, the controller 1 controls the switch 5 to select the image buffer 3 and activate the differential filter 6. The differential filter 6 is formed, for example, by the method described in Japanese Patent Laid-Open No. 3-139774.
Of the red, green, and blue colors in the case of a color image and the white color in the case of monochrome are detected, and a part having a large change in brightness is detected, and bit 1 is set at the position in the mask buffer 7 corresponding to the part. Stand up. Next, the controller 1 controls the switch 5 to select the Z buffer 4 and activates the differential filter 6 again. The differential filter 6 reads the Z buffer 4 and processes the Z buffer 4 in the same way as for the image buffer 3 to detect the part where the change in Z value is large, and the mask buffer 7 corresponding to that part.
Set bit 1 in the position. The process of setting bit 1 is performed in the additional mode. That is, bit 1 is set in the mask buffer 7 at the position determined as the contour portion by the processing of either the image buffer 3 or the Z buffer 4.

When the contour detection is completed, the controller 1 controls the switch 9 to select the frame buffer 10,
The averaging filter 8 is activated. The averaging filter 8 reads the image data from the image buffer 3 and the mask data from the mask buffer 7 by the method described in Japanese Patent Laid-Open No. 3-139774, and weights the image data having a mask value of 1 with its neighboring pixels. The average value is calculated, and the average value is replaced with the calculated value to write the pixel value in the frame buffer 10.

When the writing to the frame buffer 10 is completed, the controller 1 controls the switch 12 to switch the output to the display device 13 from the frame buffer 11 to 10.

As described above, the frame buffer 10
A three-dimensional image in which the contour portion is anti-aliased is generated and stored, and the result is displayed on the display device 13.

In the configuration of FIG. 1, the image buffer 3 and the frame buffers 10 and 11 have different configurations. The size of the mask buffer 7 is the same as that of the image buffer 3 and the frame buffers 10 and 11. Therefore, there is a problem that the scale of hardware is large. The configuration of FIG. 2 improves on this point.

2, the controller 1, the graphic processor 2, the image buffer 3, the Z buffer 4, and the display device 13 are the same as those in FIG. In FIG. 2, unlike the case of FIG. 1, the entire image buffer 3 and the Z buffer 4 are not processed at once, but are sequentially processed for some parts. Also,
The result of processing for each part is the image buffer 3 or 3 sequentially.
The value is returned to 1 and stored, and is displayed on the display device 13.
Therefore, although the image buffer 31 and the partial image buffer 100 are newly added, the frame buffer 10 of FIG.
11 can be omitted.

The controller 1 controls the switch 91 to connect the graphic processor 2 to the image buffer 3 and activate it. The graphic processor 2 uses the Z buffer 4 to generate image data in the image buffer 3.

Next, upon activation from the controller 1, the differential filter 61 reads out the data of the first processing target portion divided into m portions from the image buffer 3 and the Z buffer 4 and performs the differential processing. Is output to the mask buffer 71. At this time, a copy of the image data read from the image buffer 3 is written in the partial image buffer 100.

The averaging filter 81 is the partial image buffer 1
00 and the mask buffer 81 are read out, the averaging process of the portion is performed, and the result is returned to the image buffer 3.

This is sequentially repeated for the second and third parts,
Perform up to the last m-th part.

When the above processing is completed, the controller 1
Controls the switch 121 to switch the output to the display device 13 from the image buffer 31 to 3.

Here, the unit of partial processing is the image buffer 3,
If the number of lines is 31, the number of lines in the image buffers 3 and 31 is usually about 1000 lines, so the capacities of the mask buffer 71 and the partial image buffer 100 are extremely smaller than those. Also, the image buffer 3
By adding 1 and omitting the frame buffers 10 and 11, the hardware for one subtraction frame buffer can be omitted. Therefore, as a whole, the configuration of FIG. 2 is greatly effective in reducing the scale of hardware.

[0029]

As described above, according to the present invention, it is possible to reliably detect the contour portion of an object or a pattern on the object with a small calculation cost. Therefore, by performing the anti-aliasing process on that portion, it is possible to reliably obtain an image in which the alias is removed from the object and the contour portion of the pattern on the object at a small calculation cost without blurring.

[Brief description of drawings]

FIG. 1 is a system configuration diagram showing a first embodiment of the present invention.

FIG. 2 is a system configuration diagram showing a second embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Controller, 2 ... Graphic processor, 3 ... Image buffer, 4 ... Z buffer, 5 ... Switch, 6 ... Differential filter, 7 ... Mask buffer, 8 ... Averaging filter, 9 ... Switch, 10 ... Frame buffer, 11 ... Frame buffer, 12 ... Switch, 13 ... Display device, 31 ... Image buffer, 61 ... Differential filter, 71 ... Mask buffer, 81 ... Averaging filter, 91 ... Switch, 121 ... Switch.

Claims (2)

[Claims] 1. In a three-dimensional image generation system using a Z buffer, the discontinuities are detected by referring to the contents generated in the image buffer and the Z buffer, and the discontinuity of the image buffer corresponding to this discontinuity is detected. An antialiasing method in three-dimensional image generation, characterized in that for each pixel, a weighted average value with its neighboring pixels is calculated and used as a new pixel value. 2. In a three-dimensional image generation method using a Z buffer, the discontinuities of these are detected by referring to the contents generated in the image buffer and the Z buffer, and the outline portion of the image generated in the image buffer is detected. A method for finding the corresponding pixels.