JPS59148470A - Method and apparatus for processing picture - Google Patents

Abstract

PURPOSE:To eliminate a noise component and to adjust a profile by extracting an m-frame of sharp picture data V11-V1m and n-frame of unsharp picture data V21-V2n from an object and subtracting the total sum of the latter from the total sum of the former. CONSTITUTION:A control circuit 37 adjusts an image pickup device 30 to the 1st sharpness by a control signal DF to start the scanning of the object. A picture signal AV11 at each scanning point is converted into a sharp picture data V1i by an A/D converter 31 and data V11-V1m for m-frame's share are added by an operating device 32 and a picture memory 33. Then, the image pickup device 30 is adjusted to the 2nd sharpness to obtain the sum of the unsharp picture data V11-V1n for n-frame's share similarly, and the difference of the sum of the former is operated and the result is stored in the picture memory 33. A data for one frame's share corrected by a gamma correcting circuit 34 is converted into an analog video signal ATV at a D/A converter 35, and a picture with rejected noise and an emphasized profile is displayed on a television monitor 36.

Description

[Detailed Description of the Invention] 1' Technical Field of the Invention] The present invention d relates to an image processing method and apparatus for enhancing contours and removing noise components in images of objects extracted by an imaging device such as a television camera. It is. J [Prior Art J] With recent advances in digital signal processing technology, in the field of medical equipment, for example, noise components included in X-ray images are removed and contours are emphasized to significantly improve the image quality of large-line images. Image processing apparatuses have been put into practical use that greatly contribute to improvements in diagnostic accuracy.

However, conventional image processing devices emphasize contours and remove noise components using spatial differential operations and filtering operations that involve many multiplications and divisions, which requires a long processing time and is not suitable for purposes other than research. The problem was that it was very impractical. In other words, in order to put it into practical use, extremely high-speed processing equipment such as a processing unit for scientific calculations is required, and there is a problem in that it is inferior in versatility.

[Object and structure of the invention]

The present invention has been made in view of the above circumstances, and its purpose is to provide a highly versatile image processing method and device that can remove noise components of an object image and enhance contours in a short period of time. Our goal is to provide the following.

For this purpose, the image processing method of the present invention uses an imaging device that can vary the spatial frequency characteristic (sharpness) of an image extracted from a target object, The sharp image data corresponding to the light and shade of is Vl, and the number of extracted frames is m (m≧2)
, the unsharp image data corresponding to the shading of the object at each scanning point is v2, and the number of extracted frames is n (m≧n>l
), the following calculation is performed on each scanning point image data in one frame scanning area by changing the spatial frequency characteristics of the imaging device to remove noise components included in the object image and create an image with a predetermined spatial frequency or higher. The components are not emphasized and the object image is extracted. , An image processing device according to the present invention includes an imaging device, a control device that selectively sets the spatial frequency characteristic of the imaging device to a first sharpness and a second sharpness, and the above-mentioned item (1). It consists of an arithmetic unit that executes the formula. .

In this specification, the spatial frequency characteristic of an imaging device is defined as the response sensitivity to spatial frequency convergence of an object (original image), and a sharp image is defined as an image that includes even high spatial frequency components, and the so-called focal point is defined as an image that includes high spatial frequency components. An image with low sharpness that does not contain high spatial frequency components, such as the Kenoichi Island image, is called a non-sharp image. □When scanning an object with an imaging device, an image scanning area of a predetermined area consisting of a large number of sequential scanning points is defined as one frame scanning area. A set of image data corresponding to the shading of the object at each sequential scanning point is called one frame of image data.

Hereinafter, the present invention will be described in detail based on Examples.

〔Example〕

FIG. 1 is a flowchart showing an embodiment of the image processing method according to the present invention. First, in step 1°, the sharpness SHP of the imaging device is set to high sharpness (first sharpness). Ru. This sharpness adjustment is performed using either a method of manually controlling the electron beam focus of the image pickup tube or a method of optically controlling the image formed on the image pickup tube.

When the sharpness setting is completed, the number of frames is set to [ll] in order to extract m frames of sharp image data V1 from the object (step 11), and then step 1
2, the object is scanned in an image scanning area of a predetermined area.

Here, the scanning point in one frame scanning area is "xxy"
The coordinates of each scanning point are (1, I), C1, 2)...
Expressed as (x, y), as shown in Figure 2, sharp image data Vli corresponding to the shading of the object are sequentially extracted for each scanning point. In the next step 13, the sharp image data V'11 (t=1) of the first frame consisting of rxXy, , one sharp image data extracted in this way is extracted from rxXy one scanning point (1 , 1) to (x, y) are stored in the first image memory M1 having memory addresses corresponding to i.

Next, it is determined whether the number of frames i is larger than a predetermined number of frames m (step 14), and if mouth < ml, the sharp image data V+i of the (iJ-1)th frame is extracted in step 15. The number of frames i should be Ji+1j
After the update, step 12 is executed again. As a result, the second frame sharp image data V+1 (i=
2) The sharp image data v+ i (i=2') of the second frame extracted from the image data 5 is stored in the first image memory M1.
It is added to the sharp image data Vli (i=i) of the first frame already stored in the first frame, and then written to the first image memo IJMI again. In this case, sharp image data V
The addition of 11 and V]2 is performed for the same pixel.

That is, 1-Xyy'' scanning points (1,l'1~(x
, y), ■](1, 1, 1= ``li(+, tl 4'■12(
1,II=V ■112.1+ 11(2,1+ +V12n,,
I +VITX, 1l-v114X, +1 +v12(
X, 11V, (l, H = Vll('1.21 + V
INI, 21v] (1, yl 11N, yl”V1
2(1,y+=V ■+(2,yl'=Vt1(2,y) +2(
z,ylvlfX,)I) IN! ,3rl”
12(K, y+=V...(2) The following calculation is executed.

As a result, the first image memo IJMI contains one frame's worth of integral image data Σ representing an integral image equivalent to superimposing the object images obtained in the 1!1st frame and the 2nd frame, respectively. (1) 11 is stored at t=1. , Such processing is performed when the number of frames i becomes l''i - m j.As a result, the first image memo IJ
Integral image data representing the integral image of the object image obtained from the first frame to the first frame in the MI ← Next,
In order to extract non-sharp image data with low sharpness, the sharp integral image data Σv11, which is the storage content of the first image memo IJMI, is transferred to the second image memo IJM2 and temporarily saved (step 16).

After this, in step 17, the sharpness SHP of the imaging device is
is set to low sharpness (second sharpness), and then the frame number 1 is set to 1i = 11 in order to extract n frames of non-sharp image data Vi+1 (i-1 to n). (step 18.), then steps 19-22 are formed.

At this stage, the second image memo 1Jly12 stores non-sharp integral image data ΣV2i representing the integral image of the object image obtained from the first frame to the nth frame.

i-+ After this, in step 23, the first image memo IJM
Sharp integral image data that is the memory content of I This calculation of the difference is performed for the same scanning point as in the case of forming integral image data.
Regarding J scanning points (1,]) to (x, y), the following calculations are performed: ...(:3).

This makes it possible to remove noise components and obtain an image with enhanced contours.

That is, assuming that the sharp image is Gl and the non-sharp image is G2, if the spatial frequency characteristics for obtaining these images Ql and G2 are set as a boundary around 0.5 cycles/l, steps 10 to 15 In Figure 3(a)
The sharp image Gl distributed up to the high spatial frequency range as shown in
are sequentially extracted, and in steps 17 to 22, the third
As shown in Figure (b), unsharp images G2 containing only low spatial frequency components having a maximum value around 05 cycles/cycle are sequentially extracted.

Therefore, the sharp image Gi for one frame and the unsharp image It
Calculating the difference from the +i image G2, the two images i'Q l -G2 The spatial frequency component is removed. In other words, an image G consisting only of high spatial frequency components such as the outline or d of the object image and line segments is extracted. By the way, this image G is obtained by calculating the difference between a sharp image Gl and a non-sharp image Qz for one frame, but in this invention, integration processing for m frames is performed, so the contour or line segment is It is emphasized m times. That is, the number m of extraction frames of the sharp image G1 is set to, for example, [m=71. If the number n of extraction frames of the non-sharp one-plane f-image G2 is set to [-61, for example, the rabbit.

As shown in FIG. 4(b), the frequency component around one cycle/key is emphasized seven times. In this case, the image with enhanced contours contains 137 noises such as discontinuous line segments, but this noise component is caused by the process of integrating the sharp image f and the unsharp image G2, respectively. will be removed.

As a result, it is possible to obtain an image in which noise components are removed and the contour is emphasized.

Therefore, the integration operation of images G1 and G2 will be in high space. Therefore, by selecting the number m of extraction frames of image G1 and the extraction frame n of image G2 according to the shading of the object and the mixing ratio of noise components, noise components in the high spatial frequency θ range can be removed. In other words, by performing the calculation of equation (1) above, noise components can be removed and contours can be emphasized. 1, and the processing for this is only d, adjustment of the spatial frequency characteristics of the imaging device, and addition/subtraction calculations, so bt
It is possible to obtain an image with good visibility in a short poem (1 u) even with a yokei q machine.

In the embodiment shown in FIG. 1, the sharp image Gl for m frames is
and the unsharp image G2 for n frames I norm.
- The spatial frequency characteristic ff of the imaging device is changed for each frame scan, and the sharp image G
1 and non-sharp image G2 are taken out alternately,
) may be calculated based on the formula 7. Also, image data V+i, V2i (1=1-n', l
If the calculations based on equation (1) are executed sequentially while extracting the image, it is sufficient to provide an image memory for one frame, and the configuration can be simplified.

FIG. 5 is a block diagram showing an embodiment of the image processing device according to the present invention. In FIG. The spatial frequency characteristic (sharpness) is varied by a control signal DF generated from a control circuit 3γ, which will be described later, and a sharp image signal AVI and a non-sharp image signal AV2 are selectively extracted according to the shading of the object. 1,31 sharp image signal AV+ and non-sharp image signal AV2
32 is an AD converter that converts digital sharp image data 1 corresponding to tK and non-sharp image data v2;
This is an arithmetic unit that performs an addition/subtraction operation between the sharp image data V+4 unsharp image data (■2) output from the D converter 31 and the reproduced image data V read out from the image memory 33, which will be described later. Whether or not to perform the operation is instructed by a control signal As generated from the control circuit 37. 1 33 has memory addresses corresponding to [xyy 1 pixel in one frame scanning area, and the arithmetic unit 32
An image memory stores the image data V outputted from the control circuit 3γ in a first memory address indicated by the address signal ADRK generated from the control circuit 3γ; This is a gamma correction circuit that corrects Ila lightness when displaying an image using the generated image data V, and the correction coefficient γ is 1 given from the control circuit 37, and 35 is the reproduced image data V that has undergone gamma correction.
An AD i converter 37 converts γ into a corresponding analog video signal ATV K and displays it on a television monitor 36, and 37 is a control circuit that generates a synchronization signal φ and control signals for each of the circuit sections described above.

In such a configuration, the control circuit 31 and the imaging device 30
After generating a control signal DF for adjusting the spatial frequency characteristic (sharpness) of the object to the first characteristic (first sharpness), scanning of the object is started. As a result, sharp image signals AV+iN, tl- corresponding to the shading of the object are transmitted from the imaging device 30.
AV+Hx, yl are sequentially extracted for each scanning point. , the sharp image signal AVI + (] , l for each sequential scanning point
l-AVI Hx, y + is converted into sharp image data VliN, 1l-Vli (x, yl) by the AD converter 31 and then input to the arithmetic unit 32. On the other hand, the control circuit 37 sequentially scans the object. At the same time, an address signal MAR that changes in synchronization with the change in the scanning point is applied to the image memory 33. , iJ, and address signal M are applied to the image memory 33 within a time period in which a specific address is specified. is switched to read mode and write mode, and in read mode the stored data V (=-
VN, +I -Vfx, yl are read and the output data V'(-V'(+
, , tl -V'(x,yl) is controlled to be stored at that specific address. Further, the control circuit 37 outputs a control signal As instructing the arithmetic unit 32 to perform the addition a operation until the extraction operation of the sharp side chair data Vi + to V1m for m frames is completed.

Therefore, when scanning of the object is started in the imaging device 30, the arithmetic unit 32 generates reproduced image data V regarding the same scanning point and image data V extracted from the object.
In the image memory 33, the image data V stored in the memory address corresponding to the current scanning point is converted to new image data V calculated by the operator n' 32.
′ is executed each time the scanning point changes. Such an operation is performed using m frames of image data Vn.
- It is repeatedly executed until the extraction of Vtm is completed.

As a result, after this, m frames are stored in the image memory 33.
After the spatial frequency characteristic of the imaging device 30 is adjusted to the second sharpness by the control signal DF of the control circuit 37, scanning of the same object is started again. .

At the same time, the above-mentioned address signal ADR for the image memory 33 is generated, and the control signal As instructing the arithmetic unit 32 to perform a subtraction operation is sent to complete the extraction of n frames of non-sharp image data Vlt to Vln. Output up to.

As a result, the unsharp image signal AV2i(1,
1l-AVzlfx, yl are sequentially extracted. This unsharp image signal A'l/2i(1,] l -AVI,L(
x, yl are converted into non-sharp image data V by the AD converter 31.
21f+. 1) After being converted to H-Vzi(x, yl,
It is input to the computing unit 32. On the other hand, in the image memory 33, when one memory address, . Therefore, n7 rail
, from the start to the end of extraction of the unsharp image data V21 to V2n, the computing unit 32
) operations are executed sequentially.

As a result, the image memory 33 has Image data V for one frame shown by is stored.

That is, as described above, image data V of an image from which noise components have been removed and image components having a predetermined spatial frequency or higher are emphasized is stored.

The image data V for one frame stored in the image memory 33 is then sequentially read out under the control of the control circuit 37.
The 1° signal is supplied to the gamma correction circuit 34, and after gamma correction is performed in the gamma correction circuit 34 based on the correction coefficient γ, it is converted into an analog video signal ATV in the DA converter 35. As a result, an image with noise components included in the object image removed and whose contours are emphasized is displayed on the television 1. In this case, the characteristics of the coefficient γ of the gun correction in the gamma correction circuit 34 are as shown in FIG.
By continuously changing , it is possible to prevent false images that are likely to occur in the contour enhancement process as described above, and it is possible to display an image with even better image quality.

Note that in this embodiment, the numbers m and n of extraction frames of the sharp image G] and the non-sharp image G2 may be manually set in the control/circuit 37, or may be set manually in the control/circuit 37.
It may be automatically set based on the average value of the analog image signal AV obtained from 0.

〔Effect of the invention〕

As explained above, the present invention extracts m frames of sharp image data Vr+ to V1m and n frames of non-sharp image data V21 to V2n from an object, and performs calculations for each scanning point. be. Therefore, it is now possible to remove noise components and emphasize contours by simply adding and subtracting, and it has excellent versatility, allowing you to obtain images with good visibility in a short time using a general-purpose computer etc. . Therefore, if applied to various image processing such as X-ray image processing, an extremely practical device can be provided.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing an embodiment of the image processing method according to the present invention, FIG. 2 is a diagram showing the configuration of an image scanning area,
FIG. 3 is a diagram showing the spatial frequency characteristics of a sharp image and a non-sharp image, 413&l is a diagram showing the spatial frequency characteristics of the image obtained after calculation processing, and FIG. The block diagram shown in FIG. 6 is a diagram showing the characteristics of the gamma correction coefficient. 30... Imaging device, 31... AD converter, 32
... Arithmetic unit, 33... Image memory, 34...
・Gamma correction circuit, 35...DA converter, 36...
...Television monitor, 37...Control circuit. Patent applicant: Nippon Avionics Co., Ltd. Agent: Masaki Yamakawa (and one other person)

Claims (1)

[Claims] (11 In one frame scanning area consisting of a plurality of scanning points, Vl is the sharp image data corresponding to the shading of the object at each scanning point, and the number of frames to be extracted is m (m≧2 at each scanning point. The unsharp image data corresponding to the shading of the object
2. When the extracted frame value is n (m≧n〉]), the spatial frequency characteristics of the imaging device are changed and the following operations are performed for each scanning point image data in one frame scanning area, and the target object is An image processing method characterized by removing noise components contained in an image and extracting an object image in which image components having a predetermined spatial frequency or higher are emphasized. (2) The image processing method according to claim 1, characterized in that the number m of extraction frames of sharp image data and the number n of extraction frames of non-sharp image data are changed depending on a and the food of the object. (3) An imaging device that scans an object in a one-frame scanning area and extracts image data corresponding to the shading of the object, and the spatial frequency characteristics of this imaging device! illr! 11 and the second sharpness, and sharp image data V+ and non-sharp image data V9. il, m (m≧2
) frame and n frames (m≧n〉1); and a control device for extracting only frames (m≧n〉1); components are removed and image components having a predetermined spatial frequency or higher are emphasized.
Further, an image processing device includes an arithmetic unit for extracting an object image. (4) The number m of extraction frames of the sharp image data vl and the number n of extraction frames of the non-sharp image data V2 are changed according to the shading of the object.
The image processing device described in Section 1. . (5) The image processing device 1 according to claim 3 or 4, wherein the calculation in the calculation device is performed sequentially every time sharp image data Vl and non-sharp image data v2 are extracted.