JPH083842B2 - Image processing method - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method for enhancing a color image.

[Conventional technology]

Conventionally, when processing a black-and-white image, the histogram (frequency distribution) of the image is checked to perform image enhancement processing such as Histgram Modification. Also, in the case of a color image represented by three-dimensional variables such as red (R), green (G), and blue (B), image enhancement is performed for each plane of R, G, B. .

[Problems to be solved by the invention]

However, if the color image represented by R, G, B is image-enhanced for each plane of R, G, B, it should be processed by ignoring the correlation of distribution among R, G, B. Therefore, there is a problem that the emphasis processing is not efficient.

The present invention has been made to solve the above problems, taking into consideration the multidimensional frequency distribution of a color image,
An object of the present invention is to provide an image processing method capable of effectively enlarging a subtle difference in color or the like which cannot be sufficiently effectively processed by the conventional processing method.

[Means and Actions for Solving Problems]

In order to solve the above problems, the present invention provides an image processing method for displaying a color image based on color image data composed of multidimensional variables, wherein the color image data composed of the multidimensional variables is displayed as the color image. The color image based on a statistic defining the distribution of the color image data composed of the multidimensional variables transferred to the second coordinate system different from the first coordinate system at the time of transfer. Is emphasized based on the emphasizing formula in which the parameter is determined, and the color image is displayed. The display is moved to the first coordinate system at that time.

In this way, the color image data composed of multidimensional variables is moved to the second coordinate system different from the first coordinate system when displaying the color image, and the enhancement is performed based on the statistic that defines the color image data distribution. By performing the processing, unlike the conventional method for performing image processing for each plane, it is possible to easily and effectively make a subtle difference in color taking into consideration the multidimensional frequency distribution of the color image. Can be expanded to.

〔Example〕

Examples will be described below. FIG. 1 is a block diagram showing a configuration example of hardware of an image processing system according to the present invention. In this configuration example, a color TV camera 1 for inputting an image having color information, an image input device 3 for converting analog data input from the color TV camera 1 into digital data and sending the digital data to an image bus 2, an image bus 2 An image memory 4 for accumulating the above data, or for transmitting already accumulated data to the image bus 2, and an image processor 5 for performing various operations on the data on the image bus 2 and transmitting again to the image bus 2. , A D / A converter 6 for D / A converting the data on the image bus 2 and outputting it to the color display monitor 7, and a pointing / pointing device capable of designating a parameter for an arbitrary partial area on the color display monitor 7. The device unit 8, the A / D conversion image input unit 3, the image processor 5, the image memory 4, the D / A conversion unit 6,
System bus 9 such as pointing device 8
And a microcomputer 10 for sending instructions for various processes and controls to the device connected to.

By the way, generally, multidimensional images such as color images Has a strong correlation between the colors, so the covariance matrix (Covariance Matri
x) S, average value Are often used. At this time, using the inverse matrix S ⁻¹ of S, in the orthogonal matrix A such that A ^t A = S ⁻¹ , With the orthogonal matrix A, that is, A = (a _ij ), A new coordinate system Move on to. For example, in the case of a color image defined by R (red), G (green), B (blue), Becomes At this time The coordinate system is a coordinate system normalized by the distribution. This is shown in Fig. 2 (A). Distribution in space, of Fig. 2 (B) It becomes like a distribution in space.

By the way, the operation to perform image enhancement processing for this new coordinate system is Can be expressed as as a result Is In the coordinate system, It is represented by. For example, In the case of processing by quasi-linear transformation (B is a matrix), That is, If you say, It becomes possible to process by the following quasi-linear transformation. For example, Then, the process is to expand the distribution.

Next, the operation of performing the enlargement processing by the system configuration shown in FIG. 1 will be described. Color in Figure 1
The data input from the TV camera 1 is temporarily stored in the image memory 4 and displayed on the color display monitor 7 via the image output D / A conversion device 6. With respect to the image data displayed on the color display monitor 7, the observer uses the pointing device device 8 to specify a standard partial region to be a base of a certain distribution. There are many methods to specify, but for example, in the case of a rectangular area, the coordinates of two points at the diagonal ends that define the rectangle may be specified, and in the case of a circular area, the coordinates of one point on the center and on the circumference. Should be specified, and in the case of a region of arbitrary shape, the coordinates of the point cloud on the outer circumference should be specified. The parameters relating to these partial areas are given from the pointing device unit 8 to the microcomputer 10 via the system bus 9, and the microcomputer 10 reads out the image data in the image memory 4 storing the corresponding areas via the system bus 9. , Mean of distribution Compute the variance-covariance matrix S. And the enhancement function On the basis of the above, the parameters of the conversion formulas represented by the formulas (3) and (6) are calculated, and the parameters are sent to the image processor 5 via the system bus 9. The image processor 5 processes the data stored in the image memory 4 according to the parameters while sequentially operating the data, stores the result again in the image memory 4, and then displays it on the color display monitor 7.

Also the average value for the image And the information of the variance-covariance matrix S is known in advance, the partial area is specified for the image and the average value of the distribution is calculated. The process of calculating the variance-covariance matrix S can be eliminated. In this case, when the exchange type is relatively simple as shown in the formula (6), the image input processor 5 processes the data input in real time from the color TV camera 1 in real time, and the color display monitor 7 displays in real time. It is also possible to configure the system to display at. FIG. 3 is a flow chart showing the above processing procedure.

FIG. 4 is a block diagram showing an example of the inside of the image processor 5. This block configuration diagram particularly shows a configuration for executing the equation (3). As shown in the figure, this image processor includes a register 11 for taking in three-dimensional data x ₁ , X ₂ , X ₃ on the image bus 2, and a _ji1 x _i (i, j = 1) for each register 11 data. , 2,3) by a look-up table conversion, a look-up table memory (LUT) 12, a register 13 for temporarily holding the output of the LUT 12, and a _ji x ₁ stored in the register 13. i (i
= 1,2,3), that is, arithmetic logic unit ALU14 that adds a _ji x ₁ + a _j2 x ₂ + a _j3 x ₃ and stores the output of ALU14 in registers 15 and 15 Look-up.
A table memory (LUT) 16, a register 17 that holds the output of the LUT 16, registers 18 and 20 that respectively hold a command for selecting one of a plurality of table conversions of the LUT 12 and LUT 16, and an arithmetic instruction of the ALU 14. Register that temporarily holds
It is composed of 19 etc.

The LUT12 and LUT16 are composed of a random access memory, and can be read and written by the microcomputer 10 via the system bus 9, and should be set according to the processing before the processing starts. Is configured.

FIG. 5 is a block diagram showing a detailed configuration of the LUT 12. In the figure, a look-up table memory (LUT) 12-i is composed of a random access memory and can be accessed from the microcomputer 10 via the system bus 9. In the present invention, the LUT 12-i has a plurality of output conversion tables, and in order to actually select one of them, it is possible to perform conversion by rewriting the data in the register 18. Has become.

Next, access to the LUT 12-i will be described. L
The register 11, the buffer 21, and the register 18 are connected to the address line of the UT12-i, and the address of either the pair of the register 11 and the register 18 or the buffer 21 is selected by the LUT access control signal 23. It will be. On the other hand, the data line of the LUT 12-i is connected to the buffer 22 so that the buffer 22 is enabled when the microcomputer 10 performs write access.
The LUT access control signal 23 is controlled.

By adopting the method described above, it becomes possible for the microcomputer 10 to easily access the LUT 12. In addition to LUT12, LUT16 has the same structure.

Next, an example of the processing operation by the image processor 5 having the LUT having such a configuration will be described. In the image processor 5, pipeline processing is performed via the registers 11, 13, 15, and 17. As for the image data, the rectangular area stored in the image memory 4 is scanned pixel by pixel as in the prior art, so that the image data sequentially arrives at the image processor 5 every pipeline cycle.

The processing of Expression (3) is performed by passing the data twice through the image processor 5. However, after the first processing, the intermediate progress is once written in the image memory 4 and the next scan is performed. After the processing is again performed in the processor 5 by the above, the processing is ended by being written in the image memory 4 again.

Two types of tables are written in LUT12 and LUT16, respectively. The first table of LUT12 is a table for converting each element a _ji for matrix A into x _i → a _ji (x _i -c _i ). , The second table is configured such that each element b _{ji with} respect to the inverse matrix A ⁻¹ of the matrix A is converted into x _i → b _ji x _i .

On the other hand, the LUT 16 corresponds to the equation (2) in the first table. The table is written so that it is converted to After adding, it is applied to the correction of overflow data and negative data for the output of ALU14. For example ALU14
If the range of the image data is x _{L ≤x i} + x _{oi ≤x H} for the output x _i of x, when x _L > x _i + c _i , x _i + c _i = x _L x _H <x _{When i} + c _i , a table should be organized so that x _i + c _i = x _H. This data correction conversion will be referred to as T.

In addition, the ALU14 is set to a mode for adding three input data. After setting the table in advance in this way, the first table is selected during the first scanning and the second table is selected during the second scanning. Now the data that is sequentially read from the image memory 4 On the other hand, in the first scan And in the second scan Will be calculated.

A timing chart for each block is shown in FIG. Ie data As for the data read from the image memory 4 before step t ₀ The value in the LUT 12 is passed in step t _1, undergo conversion by said LUT 12, the data is added in ALU14 in step t _2, at LUT16 in step t ₃ Receive the conversion of. Below data ... are sequentially processed while being delayed by one pipeline cycle.

7 and 8 show what kind of calculation is performed in each block of the image processor 5 in the first scanning and the second scanning. As described above, the parameters related to the distribution, that is, the covariance matrix S and the mean value Parameters derived from A = (a _ji ), A ^-1 = (b _ji ), The image enhancement processing can be performed by writing the table information for the above into LUT12 and LUT16.

FIG. 9 shows the processing by the equation (6),
It is possible to perform processing with the same hardware configuration as the above processing operation example. In the figure, P = (P _ji ), and T represents the data correction table conversion described above. In this example, all the processing can be completed by one scanning.

In the above example, for ease of explanation, color image Although the case has been mainly described, the image processing method according to the present invention can be applied not only in the case of R (red), G (green), and B (blue). For example, CIEL ^* a ^* b ^* space, CIEL ^* u ^* which is adapted to match human color perception from R, G, B spaces ^.
It is needless to say that the image enhancement parameter can be determined by once obtaining the parameter of the statistical distribution on the v ^* space or the like and then calculating the parameter of the statistical distribution on the space.

〔The invention's effect〕

As described above based on the embodiment, according to the present invention, the color image data composed of multidimensional variables is moved to the second coordinate system different from the first coordinate system when displaying the color image, The multidimensional frequency distribution of the color image is obtained by determining the parameters of the emphasis formula based on the statistics that define the color image data distribution, and performing the emphasis process based on the emphasis formula determined by the parameters. It is possible to effectively expand a subtle difference in color taking into consideration the above.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of a hardware configuration of an image processing method according to the present invention, and FIG. 2 (A) shows a multidimensional image. A diagram showing the distribution in space, FIG. 2 (B), FIG. 3 is a diagram showing the distribution in space, FIG. 3 is a flowchart showing the processing procedure by the image processing system shown in FIG. 1, FIG. 4 is a block diagram showing an example of the configuration of the image processing processor, and FIG. FIG. 6 is a block diagram showing a configuration example of a look-up table memory, FIG. 6 is a timing chart of processing of each image data in the image processing processor, and FIG. 7 is a calculation mode at the first scanning in the image processing processor. FIG. 8 is a diagram showing a calculation mode during the second scan in the image processor, and FIG. 9 is a diagram showing another processing mode in the image processor. In the figure, 1 is a color TV camera, 2 is an image bus, and 3 is
A / D conversion image input device, 4 image memory, 5 image processor, 6 D / A conversion image output device, 7 color display monitor, 8 pointing device device, 9 system bus, 10 micro Indicates a computer.

Claims (3)

[Claims] 1. An image processing method for displaying a color image based on color image data consisting of multidimensional variables, wherein the color image data consisting of said multidimensional variables is converted into a first coordinate system for displaying the color image. And a parameter of an emphasis formula for emphasizing the color image based on a statistic that defines a distribution of the color image data composed of the multidimensional variable that has been moved to a second coordinate system different from And the color image data consisting of the multidimensional variables in the second coordinate system are enhanced based on the enhancement equation in which the parameters are determined, and the first coordinate system when displaying the color image is determined. An image processing method, characterized in that the image is transferred to and displayed. 2. When a statistic defining the distribution of color image data composed of the multidimensional variables has already been calculated, a parameter of an enhancement formula for enhancing the color image is determined using the statistic, The image processing method according to claim 1, wherein the enhancement processing is performed. 3. A parameter of an emphasis formula for emphasizing a color image is determined by using a variance-covariance matrix and an average value as a statistic defining the distribution of color image data consisting of the multidimensional variables. The image processing method according to claim 1.