JPH01118969A - Picture processor - Google Patents

Abstract

PURPOSE:To increase the picture processing speed by obtaining a distance picture in the optional direction of a 3-dimensional binary picture via the 2-dimensional gradation picture processing. CONSTITUTION:The frame memories 15-1-15-n store the binary pictures w1-wn of each section (n sections) of a 3-dimensional binary picture in the form of the 8-bit picture data whose value is equal to 0 or 1 for each picture element. Then a distance picture wd is obtained in the optional direction of the 3-dimensional binary picture against a 2-dimensional gradation picture in case the binary pictures of hitherto sections, e.g., the picture wd for w1-wi-1 are obtained in a frame memory 16-5. Therefore a 2-dimensional picture processor can be applied to said picture processing. Then the picture processing speed is increased.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention (Industrial Application Field) The present invention relates to an image processing device suitable for obtaining a distance image of a three-dimensional binary image from any direction.

(Prior Art) In recent years, attempts have been made to construct a three-dimensional image (three-dimensional 211 images) from continuous two-dimensional tomographic images obtained by CT or the like. In order to display the shape of this type of 3D binary image when observed from any direction with a three-dimensional effect on the display screen by shading, it is necessary to obtain a distance image in that direction. be. This distance image is created from each point on the desired plane in three dimensions! The distance to the surface of the target portion shown in the I image is expressed as a density value, that is, as grayscale image data. However, a three-dimensional image has a larger data size than a two-dimensional image and requires an enormous amount of processing, so there is a problem in that it takes a long time to obtain one distance image.

(Problems to be Solved by the Invention) As described above, conventionally, it has been difficult to quickly obtain a distance image from an arbitrary direction of a three-dimensional binary image.

This invention was made in view of the above circumstances, and its purpose is to create a three-dimensional 21i! An object of the present invention is to provide an image processing device that can easily and quickly perform processing for obtaining a distance image from an arbitrary direction of an image using two-dimensional grayscale image processing.

[Structure of the Invention] (Means for Solving the Problems) This invention performs affine transformation on a binary image of one cross section of a three-dimensional binary image to a specified projection plane, and converts the binary image into a two-dimensional image including a density value of 0.
A process of obtaining a two-dimensional first image expressed by the density value of a species, and an affine transformation of a two-dimensional second image whose density value changes in one direction in units of minimum density increments to the above projection plane to create a two-dimensional image. The process of obtaining the third image is repeatedly executed for each cross-section, and each time the first and third images are obtained by this affine transformation, the concatenated data of the density values of each corresponding pixel of the first and third images is If the density value of the pixel on the first image side is O, it is converted to the maximum density value, and the density value of the pixel on the first image side is O.
If not, it is converted to the density value of the pixel on the third image side, and a fourth image having this converted density value as the density value of the corresponding pixel is obtained, and each time this fourth image is obtained, this fourth image and Select the density value with the smaller value among the corresponding pixels of the fifth image that has the maximum density value in the initial state, have the selection result as the density value of the corresponding pixel, and use it as the new fifth image. In this example, the sixth image is obtained.

(Operation) According to the above configuration, by repeatedly obtaining the sixth image to be used as the fifth image by sequentially switching the sections of the three-dimensional binary image,
Dimension 21! A distance image from a designated projection plane of the image can be obtained.

Moreover, the processing to obtain this distance image is a two-dimensional one.
ji Since the process can be performed on light images, a two-dimensional image processor can be applied, and therefore the processing speed can be increased.

(Embodiment) FIG. 2 shows a block configuration of an image processing apparatus according to an embodiment of the present invention. In the figure, CPLI 11 controls the entire device, and CPLI 12 is a main memory.

13 is a CRT monitor with a keyboard for displaying images, etc., and 14 is a display controller for the CRT monitor 13. 15-1.15-2. ...l5-n, 16-1
．． 16-2°...16-6 is, for example, a 512 x 512 bit two-dimensional frame memory with 8-bit gradation, 1
Reference numeral 7 denotes a table memory used to store a data conversion table for 12-bit to 8-bit data conversion.

21 is a frame memory (here, frame memory 15-
1 to 15-n and 16-2) are affine-transformed and transferred to another frame memory (here, frame memory 16-2).
An affine transformation processor (hereinafter referred to as AFN processor) 22 is a two-dimensional image processor that outputs a 12-bit image (12-bit image data) to an 8-bit image (12-bit image data) using the table memory 16. A data conversion processor (hereinafter referred to as DCV processor), which is a two-dimensional image processor (hereinafter referred to as DCV processor), converts both images in two frame memories (frame memories 16-4 and 16-5 here). The minimum value processor (hereinafter referred to as MIN processor) is a two-dimensional image processor that selects the smaller density value of each corresponding pixel and outputs it to the corresponding pixel position in another frame memory (frame memory 16-6 in this case). ). 24
25 is a control bus used by the CPU 11 to control the display controller 14, frame memories 15-1 to 15-n, 16-1 to 16-6, table memory 17, and processors 21 to 23; 25 is a display controller; 14,
Frame memories 15-1 to 15-n, 16-1 to 16
-6, table memory 17 and processors 21 to 23
This is an image bus used for image data transfer between

Next, the operation of the configuration shown in FIG. 2 will be described with reference to the flowchart shown in FIG. 1 in the case of obtaining a distance image from an arbitrary direction of a three-dimensional binary image. Here, the following processes 1 to 2 are performed for each cross section of the three-dimensional binary image in order from the cross section furthest from an arbitrarily given projection plane.

Now, the frame memories 15-1 to 15-n contain binary images W1 to wn of each section (n section in this case) of the three-dimensional binary image (8-bit image data with a value of 0 or 1 for each pixel). The binary images of the cross sections so far, for example, the distance images w for W1 to wil, are stored in the frame memory 16-5.
Assume that d is required.

■ First, the CPU 11 selects any three points that are not aligned on the same straight line on the binary image 1wi of the cross-section that is next to the projection plane after the previous cross-section, for example, the four vertices of Wl as shown in FIG. Select three points Pi, P2, P3, and select points P1, P2
, P3 and the intersection points Q1, Q2, Q3 of the projection plane and straight lines perpendicular to the projection plane, that is, the projection points 01 to Q3 from the points P1 to P3 to the projection plane (here, the points P1 to P3
Find the affine transformation point) from to the projection plane. And C.P.
U11 is the three points P1 to P3 on the image wi that is the input image.
and the projection plane of Wl which becomes the output image @ (affine transformation image)
From the relationship with the three points 01 to Q3 on wp, well-known affine transformation coefficients are determined, and the AFN processor 21 performs affine transformation processing on the image wi using these coefficients via the control bus 24.
instruct.

As a result, the AFN processor 21 controls the frame memory 1
Binary image 1g+ stored in 5-i (1≦i≦n)
wi is affine-transformed using the affine transformation coefficients given from the CPU 11, and an affine-transformed image (1wp) is obtained in the frame memory 16-1.The image data (density value) of each pixel of this wp has a value of O. or 1 8-bit data.

■ The AFN processor 21 uses the frame memory 16-1
When the affine transformed image @Wp is found in the image, the CPU 11 is notified of this via the control bus 24.

Upon receiving the notification from the AFN processor 21, c p u ii selects the three points p1 on the image wi as shown in FIG.
．． The distance from P2, P36 to the projection plane, that is, Pi, P
2, P3 and its affine transformation point Ql.

Find the distances d1, d2, and d3 between Q2 and Q3.

The CPU 11 calculates the distances dl, d2, d
For example, 3 is an 8-bit density value (81 light distance image data)
Convert to DI, D2, and D3 as follows.

Here, in the 3D binary image area (cuboid),
The distance of the closest point from the projection plane is d l! ! In, and the distance of the farthest point is d wax, then the density value of the distance image at the point with distance d 1n is the minimum value ○ (all “0”), and the density value of the distance image at the point with distance @ d max is The distance I! is set so that the value becomes the maximum value 255 (all “1”). Il
A density value for d is determined.

Therefore, if the value of d max - d min is Δd, this D is obtained according to the following formula % formula % (11).Then, the CPU 11 sets d to d.
By calculating the above equation (1) using 1, d2, and d3, three points P1. Projection point (affine transformation point) Ql of P2 and P3 onto the projection plane.

Q2, Q3 Teno distance image a standing planting DI, D2
.

Find D3. Note that the above dmax, dmin,
and Δd are determined in advance by the CPU 11.

Now, the frame memory 16-2 has, for example, 256X2
Two-dimensional grayscale image WQ with a size of 56 bits
is stored. As shown in FIG. 5, the density value of this image wg varies from density value O to 25
It changes linearly by 1 up to 5. c p u ii
is the density value Di, D2 on the upper side of the grayscale image @WQ
point with < DI , D2 point) and the grayscale image w
A point having a density value D3 (point D3) on the lower side of Find the affine transformation coefficients such that

Next, the CPU 11 instructs the AFN processor 21 via the control bus 24 to perform affine transformation processing on the grayscale image WQ using the affine transformation coefficients.

As a result, the AFN processor 21 controls the frame memory 1
The gray scale image wg stored in 6-2 is
Affine transformation is performed using the affine transformation coefficients given from P U 11, and the image wg' is stored in the frame memory 16-3.
seek. This image Wq' shows a distance image of each point on the current cross section.

■ The AFN processor 21 uses the frame memory 16-3
When the image w o + is obtained within the control bus 24,
The CPU 11 is notified via the CPU 11. The CPU 11 is
Upon receiving the notification from the AFN processor 21, a data conversion table is set in the table memory 17, and then the DCV processor 22 is activated. table memory 17
In the data conversion table set to , the value of the upper 4 bits is O("0000") as shown in Figure 6.
has an entry specified by a 12-bit density value (input density) of 1 (“0001”), and among the entries, the entries whose upper 4 bits are indicated by O all have a value of 255. 8-bit output image data (output density) with the same value as the lower 8 bits (0 to 255) is shown in the entry whose upper 4 bits are 1. Each is set.

Note that the above data conversion table may be set in the table memory 17 in advance.

Now, when the DCV processor 22 is activated by the CPU 11, the DCV processor 22 processes the image wp in the frame memories 16 to 1.
The lower 4 bits (O or 1 for fields) of the 8-bit image data (31 degree value) of each pixel of are used as the upper entry address,
8 of the corresponding pixels of the image Wg' in the frame memory 16-3
Refer to the data conversion table in the table memory 17 using bit image data (12 bits 1 to 1 with rear entry address cJ as the lower entry address, and convert the lower 4 bits of the 8-bit degree value from image wp and image Wg' A 12-bit density value concatenated with 8-bit density values from
The setting value (8-bit density value) of the corresponding entry in the data conversion table in the table memory 17 is converted.

Then, the DCV processor 22 writes this conversion a open chain into the corresponding pixel position of the frame memory 16-4. As is clear from the contents of the data conversion table shown in FIG. 6, this converted density value becomes the maximum density value 255 when the density value on the wp side is O, and when the direct a value on the wp side is 1, Wg' The density value on the side is the same as that on the side.

Therefore, by the AFN processor 21 repeating the above processing for all pixels of the images wp and wo', the distance image &W l)' for the binary image wi of the current cross section is stored in the frame memory 16-4. Memory 16-
Required within 4.

■ The DCV processor 22 uses the frame memory 16-4
When the image w p L is obtained within the time limit, the CP tJ 11 is notified of this via the control bus 24. CPU 1
1 receives the notification from the DCV processor 22, M
Start the IN processor 23. As a result, the MIN processor 23 generates a distance image Wp for the binary image wi of the current cross section stored in the frame memory 16-4.
′ and the 8-bit density value of each pixel of the distance image @wd for the binary image (iW1 to W1-1) of the cross section up to now stored in the frame memory 16-5, find the minimum value of both (i.e. (determine the density value with the smaller value) and fill the determined density value into the corresponding pixel position in the frame memory 16-6.Then, the MIN processor 23 performs the above minimum value calculation processing on the images wp' and wd. By repeating this process for all pixels, the distance image wd' for the binary images W1 to wi up to the current cross section is obtained in the frame memory 1G-6. Note that the 8 bits in the initial state of each pixel of the image @Wd All density values are 255, which indicates the maximum distance.

(2) When the distance image wd' is obtained in the frame memory 16-6, image data is transferred from the frame memory 16-6 to the frame memory 16-5. As a result, the image wd' obtained in the frame memory 16-6 in the process (2) becomes the distance image wd up to the latest cross-sectional binary image, instead of the previous image wcj.

The above processing of ■ to ■ is performed on the frame memories 15-1 to 15
By repeating the binary images of each section of the 3-dimensional binary image stored in -n in order from the section farthest from the given projection plane, the 3-dimensional binary image from any given direction is A distance image wd can be obtained. Then, by using this finally determined distance image Wd, a shading image of the three-dimensional binary image from any direction can be determined.

In the above embodiment, the frame memory 16-5 is
Although it has been described that the frame memory 16-6 is used to store the image wd' and the frame memory 16-6 is used to store the image wd', it is also possible to alternately use the frame memory 16-6 to store the image wd'. In this case, there is no need to transfer the image from frame memory 16-6 to frame memory 16-5. Furthermore, in the embodiment described above, the binary images of each section of the three-dimensional binary image are stored in different frame memories, but the present invention is not limited to this.

For example, if the frame memory has a depth of 8 pits, in order to effectively utilize the memory capacity, two
Value images may be stored in one frame memory. Furthermore, in the embodiment described above, the binary images of each section of the three-dimensional binary image are processed in order from the section farthest from the given projection plane, but the invention is not limited to this. The processing may be performed in order from the closest cross section.

[Effects of the Invention] As detailed above, according to the present invention, the process of obtaining a distance image from an arbitrary direction of a three-dimensional binary image can be performed on a two-dimensional grayscale image. It becomes possible to apply a two-dimensional image processor to the processing, and it is possible to speed up the processing.

[BRIEF DESCRIPTION OF THE DRAWINGS] FIG. 1 is a flowchart showing an embodiment of distance image acquisition processing directly related to the present invention, FIG. Figure 3 is a diagram explaining the relationship between a point on a cross section of a 3D binary image and an affine transformation point on a given projection plane, and Figure 4 is a diagram explaining the relationship between a point on a cross section of a 3D binary image and a point on the projection plane. Figure 5 is a diagram explaining the distance to the affine transformation point, Figure 5 is a diagram explaining the position on the grayscale image corresponding to the density value of the distance image at the affine transformation point, Figure 6 is the table of Figure 2. FIG. 3 is a diagram showing a data conversion table set in memory. 11, CPU, 15-1 to 15-n, 16-1
~16-6-... Frame memory, 17... Table memory, 21... Affine transformation (AFN) processor, 22... Data conversion (DCV) 70ttt, 23
...minimum value (MIN) processor. Applicant's representative Patent attorney Takehiko Suzue No. 2 Figure 3 Figure 4 Figure 5 Figure 6

Claims (1)

[Claims] Three predetermined points on one cross section of a three-dimensional binary image and three
An affine transformation coefficient is determined based on each projection point onto the specified projection plane, and the binary image of the cross section is affine transformed to the projection plane using the affine transformation coefficient, and the binary image including the density value 0 is a first affine transformation processing means for repeatedly performing a process for obtaining a two-dimensional first image expressed by the density value of the species for each section of the three-dimensional binary image; 3 points on the two-dimensional second image that change at , and correspond to density values indicating distances from the predetermined three points on the cross section to each projection point on the projection plane, and each of the projection points. The process of obtaining an affine transformation coefficient based on the affine transformation coefficient and performing affine transformation of the second image onto the projection plane using the affine transformation coefficient to obtain a third image is repeatedly performed for each section of the three-dimensional binary image. Second
affine transformation processing means; each time the third image is obtained by the second affine transformation processing means and the first image is obtained by the first affine transformation processing means, each correspondence between the first and third images; If the density value of the pixel on the first image side is 0, the concatenated data of density values of pixels is converted to the maximum density value, and if the density value of the pixel on the first image side is 0, it is converted to the maximum density value on the third image side. an image data conversion processing means for converting into density values of pixels and obtaining a fourth image having the converted density values as density values of corresponding pixels; and each time the fourth image is obtained by the image data conversion processing means; , select the density value of the smaller value among the corresponding pixels of this fourth image and the fifth image having the maximum density value in the initial state, and use the selection result as the density value of the corresponding pixel and create the new above. minimum density value selection processing means for obtaining a sixth image to be used as the fifth image; and by repeatedly obtaining the sixth image by switching the cross section of the three-dimensional binary image, An image processing device characterized in that an image is obtained as an image.