JPH05282443A - Image processing method - Google Patents

Abstract

PURPOSE:To provide an image processing method capable of calculating the area and the center of gravity of an object accurately and in a short time by using an original image from a television camera by which the object is photographed. CONSTITUTION:The minimum image area including only the image of the object can be specified (3) by binarizing (2) the original image from the television camera photographing the object by using two threshold values for brightness, and setting the threshold values on the number of accumulated picture elements in a Y-axial direction and an X-axial direction for the X-axis and the Y-axis of a binarized image. Thence, a part lost by binarization can be filled by scanning each straight line on both coordinates in a specified image area by the minimum value and the maximum value of the image area, and assuming a part between both coordinates arriving first to the image of the object as the area of the object, and also, the area and the center of gravity can be calculated (4) from the image of the object.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method for obtaining an area and a center of gravity of an object from an original image of the object output from a television camera.

[0002]

2. Description of the Related Art Conventionally, for example, in a sawmill, an image processing device for automating a saw machine has been operated, and in order to perform centering and wood cutting required for sawmilling, the center of gravity of a wood cross section The calculation of the area is carried out from the wood cross-section image projected by the TV camera.

In a conventional image processing apparatus, the brightness of an original image from a TV camera including a background and a wood cross section is represented by a numerical value, and a numerical value or a threshold value is defined as a boundary at which the wood cross section and the background are separated. By setting, a portion brighter or darker than the threshold value is set as a binarized image of the wood cross section. Then, regarding the binarized image, it is assumed that the wood cross section is included in the preset region, and the center of gravity and the area in the image are calculated using only all the bright portions distributed in this region. ing.

[0004]

However, in the original image from the TV camera, it is common to include a lighter background and a darker background than the wood section, and as described above, the wood section is separated by only one threshold value. The method of binarizing by using the method causes a problem that the binarized image of the wood cross section includes the background.

Moreover, in order to ensure that the wood cross section is included in the preset area of the binarized image,
It is necessary to set this area relatively large, but as a result, more background is included in the area, which tends to adversely affect the calculation of the area of the wood cross section and the center of gravity. Since the detection of the wood cross section and the calculation of the area and the center of gravity of the large area are performed, there arises a problem that the processing takes time.

Furthermore, when the wood is stored outdoors, the cross section of the wood is soiled and discolored by rain, sand, etc., and even in the image of a television camera which shoots this, the wood cross section is partially discolored. Therefore, in the image obtained by binarization, the wood cross section is partially missing. However, in the conventional method, this missing part is processed while it is missing, so the calculation of the center of gravity and area is performed. There is a problem that the reliability of the result is low.

The present invention has been made in view of the above problems of the prior art, and an object of the present invention is to accurately measure an image of an object from an original image from a television camera. An object of the present invention is to provide an image processing method capable of more accurately and quickly calculating the area and the center of gravity by correcting even if there is a partial omission.

[0008]

In order to achieve the above-mentioned object, in the image processing method of the present invention, an original image from a television camera which images an object is divided into two brightnesses for separating the background and the object. Threshold value is set and binarized, and the threshold value is set for the cumulative number of pixels in the Y axis direction with respect to the X coordinate and the cumulative number of pixels in the X axis direction with respect to the Y coordinate of the binarized image. And the minimum image area including only this object image is specified, and the minimum value and the maximum value of the image area with respect to the straight line on each Y coordinate and the straight line on each X coordinate in the specified image region. Areas of the object that are lost by the binarization are filled by considering the area between the two coordinates that first reach the object image by scanning from and, and the area and the area are calculated from the object image corrected by the hole filling. Calculate the center of gravity Than it is.

[0009]

According to the image processing method of the present invention, two threshold values are used when binarizing an original image from a television camera, and a portion sandwiched by these thresholds or a portion not sandwiched by the thresholds is an image of the object. Therefore, the background of the image of the object can be excluded as much as possible.

Then, for the binarized image, the cumulative number of pixels in the X-axis direction and the Y-axis direction is calculated, and only the object image is extracted from this, and the minimum image area including this is binarized in order to specify it. It is possible to suppress the adverse effect on the calculation of the area and the center of gravity due to the background that cannot be separated by
Since only the specified minimum image area is targeted for calculation, the calculation time can be shortened.

Further, in the specified image area, a straight line on each Y coordinate and a straight line on each X coordinate are scanned from the minimum value and the maximum value of the image area, respectively, and between the both sides of the object image on this straight line. Is regarded as the object region, it is possible to compensate for the part lost by the binarization, and the calculation results of the area and the center of gravity of the object become more accurate.

[0012]

EXAMPLES Examples will be described with reference to FIGS. 1 to 10. First, FIG. 2 shows a schematic configuration of an image processing apparatus for an automatic saw machine, in which 1 is wood as an object, 2 is two compact CCD cameras which image both end faces of the wood 1, respectively. TV camera 3 is a camera interface, and switching between both TV cameras 2 and shading processing, binarization of an image, superposition of a wood removal pattern and an image from the TV camera 2, and overwriting of character data are performed in accordance with instructions from the CPU 4. Etc., and converts into a TV video signal and outputs.

Reference numeral 5 is a television monitor for displaying a color image for displaying a television video signal from the camera interface 3. The monitor screen displays an image of the end of the timber captured by the television camera 2, a wood removal pattern, characters and the like. To be done.

The CPU 4 controls the operation of the entire image processing unit, writes the wood removal pattern to the camera interface 3 based on the wood removal pattern information stored in advance in the memory of the CPU 4, and outputs the wood removal data to a control device (not shown). The output is performed, and the center of gravity and the area of the wood 1 are calculated from the image.

A position detector 6 detects the position of the center of gravity of the binarized image as coordinates on the screen of the television monitor 5, and supplies coordinate data to the CPU 4. Reference numeral 7 denotes an input / output interface as a digital input / output device through which the CPU 4 checks the state of the operation switch, blinks an indicator light, or outputs print data to the printer 8. 9 is a power supply unit.

FIG. 1 is a flow chart showing the process for calculating the area and the center of gravity in the image processing apparatus. Image input, binarization, identification of a wood area, and area in the wood area, It consists of each process of calculating the center of gravity.

Image input: The end face of the wood 1 is photographed by the TV camera 2, and the image signal from the TV camera 2 is subjected to shading processing, that is, the image signal is superimposed on the image signal to correct the image signal, and as shown in FIG. Get the original image. In this original image, A and C are backgrounds and B is a wood cross section.
As shown in FIG. 4, which shows the brightness distribution on the straight line L-L, the brightness increases in the order of the background A, the wood section B, and the background C.

Binarization As shown in FIG. 4, a wood cross section B is applied to the obtained original image.
And the backgrounds A and C are separated from each other, two thresholds a and b having different brightness are set, and the portion sandwiched between the thresholds a and b is set to 1,
Other values are set to 0 and binarized so that only the brightness of the wood cross section B is extracted.

Conventionally, only one threshold value, for example, a, was set when binarizing the original image, so that the image obtained by binarizing the wood cross section B as shown in FIG.
And the background C brighter than this are included, but by setting the two threshold values a and b as described above, the background is included as much as possible in the wood cross section as shown in FIG. 5B. You can avoid it.

Here, FIG. 6 shows the relationship between the number of pixels on the screen and the brightness in the original image, where A, B, and C in the figure are the background A, the wood section B, and B in the original image, respectively. It corresponds to the background C and has two brightness thresholds a and b.
It can be seen that the wood section B can be taken out with. Note that the background C brighter than the wood cross section B has the brightness between the thresholds a and b at the periphery, so that only the outline of the background C remains as a binarized image as shown in FIG. 5B. Sometimes.

Identification of Wood Region FIG. 7 shows a binarized image obtained as described above, in which D is the whole image and E is the cross section B of the wood moving by the support table of wood. The maximum movement range, F, is a quadrangle circumscribing the wood section B, that is, the minimum wood area including only the wood section. In addition, (A) of FIG.
In the binarized image, it shows the cumulative number of pixels in the Y-axis direction with respect to the X coordinate, and FIG. 8B shows the cumulative number of pixels in the X-axis direction with respect to the Y coordinate.

In FIG. 8A, a certain threshold value c for separating the wood section B and the background C is set, and from the minimum value X ₁ to the maximum value X ₂ and the maximum value X ₂ in the X coordinate of the moving range E. Scanning is performed toward the minimum value X ₁ , and the X-coordinates X ₁ 'and X ₂ ' that _first exceed the threshold value c are set as the minimum and maximum X-coordinates of the wood region F, respectively.

Similarly, in FIG. 8B, a certain threshold value d for separating the wood section B and the background C is set, and the minimum value Y ₁ to the maximum value Y ₂ and the maximum value in the Y coordinate of the moving range E are set. Scanning is performed from Y ₂ toward the minimum value Y ₁ , and first, Y coordinates Y ₁ 'and Y ₂ ' that exceed the threshold value d are respectively detected in the wood region F.
The minimum and maximum Y-coordinates of This makes it possible to extract only the wood section B from the binarized image and specify the minimum image area including only the wood section B, that is, the wood area F. It is desirable that the threshold values c and d be as small as possible.

Calculation of Area and Center of Gravity in Wood Region FIG. 9 shows a wood region F including only the wood cross section B identified from the binarized image as described above. Includes missing portions M ₁ and M ₂ lost by binarization due to discoloration of the surface.

Each pixel P (x,
y) has a value of 1 or 0 as the weighting by binarization, and the pixel P (x, y) in the portion except the missing portions M ₁ and M ₂ in the wood cross section B, that is, the pixel P (x, y) 1, It is assumed that 0 is given to the pixels P (x, y) other than. Then, when calculating the area and the center of gravity for the wood cross section B, the missing portions M ₁ and M ₂ lost by binarization are filled.

Now, a Y coordinate y ₁ in the wood area F
On the upper straight line (X ₁ ', Y ₁ )-(X ₂ ', y ₁ ), scanning is performed from the minimum value X ₁ 'in the X coordinate of the wood region F toward the maximum value X ₂ ', and first P ( x, y) = 1, the coordinates of the point P ₁ are (x ₁ , y ₁ ), and similarly, scanning is performed from the maximum value X ₂ ′ to the minimum value X ₁ ′, and first P (x, y) The coordinates of the point P ₂ where = 1 is set to (x ₂ , y ₁ ). The straight line (x ₁ , y ₁ ) − (x ₂ ,
Let P (x, y) = 1 for all pixels in y ₁ ).

As a result, for example, a missing portion M ₁ such as a straight line (X ₁ ', y ₂ )-(X ₂ ', y ₂ ) on the Y coordinate y ₂
, Then first scan P (x, y) from both sides.
P for all pixels between points P ₃ and P _{4 where} = 1
Since (x, y) = 1, the pixel between the missing points P ₅ and P ₆ also has P (x, y) = 1, and the missing portion M ₁ can be filled in and corrected.

In this manner, the missing portions M ₁ and M ₂ in the wood cross section B are filled up, and then the area S of the wood cross section B is calculated using the following equations (1) to (3).

[0029]

[Equation 1]

[0030]

[Equation 2]

[0031]

[Equation 3]

Here, Sx shown in the equation (1) is first obtained by scanning from the minimum X coordinate value X ₁ ′ to P (x) for all Y coordinates Y ₁ ′ to Y ₂ ′ of the wood area F. , Y) = 1 from the X coordinate value X ₁₂ to the X coordinate value X ₂₁ at which P (x, y) = 1 is first obtained by scanning from the maximum X coordinate value X ₂ ′. It is the area calculated as P (x, y) = 1.

Further, Sy shown in the equation (2) is the wood region F
The minimum Y coordinate Y for all X coordinates X ₁ 'to X ₂ ' of
Y which first became P (x, y) = 1 by scanning from ₁ '
From the coordinate value Y ₁₂ to the Y coordinate value Y ₂₁ at which P (x, y) = 1 was first obtained by scanning from the maximum Y coordinate Y ₂ ′, all P (x, y) = 1 are set. This is the calculated area.
Then, the average of the areas Sx and Sy of the both is defined as the area S of the wood cross section B by the equation (3). In addition, the center of gravity W of the wood cross section B is calculated by using the following formulas 4 to 9 for the X coordinate X thereof.
Calculate the w, y coordinates Yw. Note that P (x, y) = 1 in the equations (4) to (7) in consideration of the above-mentioned filling.

[0034]

[Equation 4]

[0035]

[Equation 5]

[0036]

[Equation 6]

[0037]

[Equation 7]

[0038]

[Equation 8]

[0039]

[Equation 9]

In FIG. 9, W ₀ is the position of the center of gravity according to the conventional method in which the missing portions M ₁ and M ₂ are not filled. By the way, in the above, there is a case in which the missing portions M ₁ and M ₂ lost due to binarization exist in the wood section B, that is, the missing sections M ₁ and M ₂ are surrounded by the wood section B. In such a case, the missing portions M ₁ and M ₂ can be reliably corrected by the filling processing of the present invention.

On the other hand, as a wood cross section obtained by binarization, as shown in FIG. 10, the periphery of the wood cross section B may be chipped by the missing portions Q and R, and only the P portion may remain.

In such a case, the area Sx obtained by the equation 1 does not include the missing portion Q, and Sx = P + R, and the area Sy obtained by the equation 2 includes the missing portion R. Without this, Sy = P + Q, but as in the equation (3),
The average value of both Sx and Sy is the area S (=
Since (Sx + Sy) / 2), it is possible to calculate including the lost parts Q and R, as compared with the calculation result by the conventional method where the area S ′ = P, and a more accurate calculation result Also in the calculation of the center of gravity W, a good result is obtained for the center of gravity position W ₀ by the conventional method. In the case of FIG. 10, a more accurate area S (= Sx + Sy−S ′) can be obtained by using the area S ′ obtained by the conventional method and the areas Sx and Sy according to the present invention.

[0043]

Since the present invention is constructed as described above, it has the following effects. When binarizing the original image of the object, by setting two brightness thresholds, it is possible to obtain an object image that does not include the background as much as possible. To
Moreover, since the minimum image area including only the object image in the binarized image is specified, even if the background cannot be separated even by binarization, only the object image is extracted and the background has the area and the center of gravity. It is possible to suppress the adverse effect on the calculation and further shorten the calculation time.

Furthermore, due to partial discoloration of the object,
Even if a part of the object image is lost due to the binarization, the missing part can be filled in by calculating from the currently remaining part, and the area and the center of gravity can be accurately calculated.

[Brief description of drawings]

FIG. 1 is a flow chart for explaining an operation for explaining an embodiment of an image processing method according to the present invention.

FIG. 2 is a block configuration diagram of an image processing apparatus to which the present invention has been applied.

FIG. 3 is a diagram showing an original image from a television camera.

FIG. 4 is a brightness distribution diagram along line LL in FIG. 3;

FIG. 5 is a diagram showing a binarized image, in which (A) is a conventional example,
(B) is an example.

6 is a diagram showing a relationship between the number of pixels and brightness in the original image of FIG.

FIG. 7 is an explanatory diagram for specifying a wood region in a binarized image.

FIG. 8 shows an example of scanning for identifying a wood region in FIG. 7, (A) is an explanatory diagram of the cumulative number of pixels with respect to the X coordinate,
(B) is an explanatory diagram of the cumulative number of pixels for the Y coordinate.

FIG. 9 is a diagram showing a specified wood area.

FIG. 10 is a diagram showing another example of a wood cross section.

[Explanation of symbols]

 2 TV cameras A, C Background B Wood cross section a, b, c, d Threshold F Wood area

Claims (1)

[Claims] 1. An X-coordinate of a binarized image in which an original image from a television camera that images an object is binarized by setting two thresholds of brightness for separating the background and the object. With respect to the Y-axis direction cumulative pixel number and the Y-axis direction cumulative pixel number with respect to each of the threshold values are set, and only the object image is extracted.
A minimum image area including only the object image is specified, and a straight line on each Y coordinate and a straight line on each X coordinate in the specified image area is respectively selected from the minimum value and the maximum value of the image area. The area between the two coordinates which has reached the object image by scanning is regarded as the area of the object and the portion lost by the binarization is filled in, and the area of the object image corrected by the filling An image processing method characterized by calculating a center of gravity.