JPH11250247A - Isolated area judging device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a highly sensitive isolated area judging device for operating high speed isolated area judgement without any erroneous judgment due to the influence of the edge of an object to be measured. SOLUTION: Each picture element of a gradation picture inputted from an input terminal 1 is operated by a first filter arithmetic means 2 so that a first output value can be calculated, and the first output value is compared with a first judgement reference value by a first judging means 5 so that whether or not the picture element is an isolated area can be judged. The first output value is compared with a second judgement reference value by a second judging means 6 so that whether or not the picture element is an isolated area candidate can be judged, and only the picture element which is judged to be the isolated area candidate is operated by a second filter arithmetic means 7, and the output value is compared with a third judgment reference value so that whether or not the picture element is the isolated area can be judged. The logical sum of the the results of the first and third judging means is obtained as a judged result.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an isolated area judging apparatus for judging whether each pixel in a grayscale image is an isolated area having a density higher than that of an adjacent area and an isolated area. An isolated area for digital signal processing of a real-time image used for a foreign object detector that detects minute foreign matter such as metal, stone, glass, etc. mixed into an object such as food using an X-ray transmission image It relates to a determination device.

[0002]

2. Description of the Related Art Various image processing methods have been proposed as methods for determining the presence or absence of foreign matter in an object to be measured using an X-ray transmission image. Foreign matter can be regarded as an isolated area in the X-ray transmission image. Therefore, most methods use the X
A difference value between the X-ray transmission image and the image obtained by performing a smoothing filter process on each pixel of the X-ray transmission image is calculated, and the difference value is compared with a predetermined determination reference value. This is a method of determining whether or not there is.

FIG. 10 shows a configuration of an isolated area determining apparatus using a moving average filter as a smoothing filter. The moving average filter is a process of setting a region average value of a local region of each pixel of the grayscale image around the pixel of interest as an output value of the pixel of interest, and is one of low-pass filters. Since the moving average filter can perform processing at high speed, it is often used when real-time processing is required.

[0004] The operation of the isolated area determination device using the moving average filter will be described below. With respect to the X-ray transmission image input to the input terminal 1, the average value of the local area around the pixel is calculated for each pixel by the average value calculation unit 3. This operation removes high-frequency components from the image output from the average value calculation means 3 and makes the image smooth. Next, a difference value between the density value of each pixel of the X-ray transmission image and the average value of the area of the pixel is calculated by the difference value calculation means 4. By this operation, the high-frequency component of the transmission image removed by the average value calculation means 3 is extracted, and an isolated region or the like where a steep change in density is emphasized. Finally, the determination means 5 compares the difference value with a predetermined determination reference value to determine whether or not the area is an isolated area, and outputs the determination result from the output terminal 10.

[0005]

However, in an isolated area determination apparatus using a moving average filter, when the size of a local area for which an average value is obtained is increased, not only the isolated area but also the edge of the DUT included in the background are increased. Are also emphasized, and this is a factor of erroneous determination. If a determination reference value is set so that an edge or the like is not determined as an isolated area and the presence or absence of an isolated area is determined, there is a problem that the detection sensitivity of a foreign substance that is an isolated area is reduced.

For example, in order to emphasize only an isolated area without emphasizing an edge, the smoothing filter of the isolated area determination device applies the center value of the local area centered on the target pixel to the output value of the target pixel. Median filter,
Alternatively, a non-linear filter may be used, such as searching for a local region that does not include an edge around the target pixel, and using a local average of the local region as an output value of the target pixel. However, the processing of a nonlinear filter is more complicated than that of a linear filter such as a moving average filter, and the processing time is often longer.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and has almost the same processing time as that of an isolated area determination apparatus using a moving average filter, without erroneous determination due to the influence of edges and the like. It is an object of the present invention to provide a highly sensitive isolated region determination device capable of performing isolated region determination.

[0008]

An isolated area determination apparatus according to the present invention performs a product-sum operation on a density value of a pixel in a local area around a pixel of interest in a grayscale image by using a predetermined coefficient matrix. First filter calculating means for calculating a filter output value;
Receiving the first filter output value, comparing the first filter output value with a predetermined first determination reference value, the pixel of interest is isolated higher in density than the neighborhood in the grayscale image. First determining means for determining whether or not the area is an isolated area; receiving the first filter output value, determining whether the first filter output value is equal to a predetermined second determination reference value; A second determination unit that determines whether the target pixel is an isolated region candidate that is a region having a higher density than the vicinity in the grayscale image, and a determination result from the second determination unit. Receiving data of each pixel of the grayscale image, among the pixels, for a pixel determined to be an isolated region candidate, a filter operation is performed by using a density value of a pixel in a local region around the pixel. Calculate the filter output value of 2 Receiving the second filter output value, and comparing the second filter output value with a predetermined third determination reference value to determine whether the pixel of interest is an isolated region. And a third determination means for determining whether or not the target pixel is an isolated area based on a determination result of the first determination means and a determination result of the third determination means. Is determined.

[0009]

The isolated region determination apparatus of the present invention calculates a first filter output value by performing a product-sum operation on a density value of a pixel in a local region around a target pixel of a grayscale image using a predetermined coefficient matrix. Comparing the first filter output value with the first determination reference value to determine whether or not the pixel of interest is an isolated area, and comparing the first filter output value with the second
, It can be determined at a high speed whether or not the pixel of interest is an isolated region candidate. afterwards,
For only pixels determined as isolated region candidates, a second filter output value is calculated by performing a filter operation using the density value of the local region around the pixel of interest, and for pixels not determined as isolated region candidates, A predetermined value is output as a second filter output value. By comparing the second filter output value with the third determination reference value to determine whether or not the pixel of interest is an isolated region, it is determined that the target pixel is an isolated region based on the first determination reference value. An isolated region to be detected that has not been detected can be determined as an isolated region.

Regarding the processing time, for example, it is assumed that 1% of the pixels in a grayscale image of one frame are isolated region candidates, and the second filter operation means is five times as long as the first filter operation means. It is processing time, and the case where the first filter operation means is a filter using a conventional moving average filter is considered. At this time, assuming that the processing time of one frame of the isolated region determination device of the present invention can ignore the time of the determination process, etc., the conventional isolated region determination device using the moving average filter has the following relationship: 1 + 1/100 × 5 = 1. 05 (times), which is almost the same as the processing time of the isolated region determination device using the conventional moving average filter, and is high speed.

[0011]

FIG. 1 shows a first embodiment of the present invention.
This will be described with reference to FIG. The grayscale image input to the input terminal 1 is subjected to the first filter operation means 2 to perform a product-sum operation on the density values of the pixels in the local area around the pixel using a predetermined coefficient matrix. A filter output value is calculated.

Here, the first filter calculating means 2 using a moving average filter will be described. In the grayscale image input to the input terminal 1, the average value of the local area around the pixel is calculated by the average value calculation means 3 for each pixel.
Next, a difference value between the density value of the pixel of the grayscale image and the average value of the area of the pixel is calculated by the difference value calculation means 4 and output as a first filter output value.

As an example of the processing, FIG. 4 shows an X-ray transmission image in which a foreign substance is mixed in food. 4 (a), 4 (b) and 4 (c) show foreign matters, FIG. 4 (d) shows overlapping food products, and FIG. 4 (e) shows the food edges. FIG. 5 shows a grayscale image obtained by the first filter output value. Here, the negative value is set to 0, the maximum value in the grayscale image is normalized to 255, 255 is represented by black, and 0 is represented by white. The first filter output value becomes the output value of the high-pass filter, and FIG.
Pixels having a sharp change in density, such as the foreign matter in (b) and (c), the overlap of the food in FIG. 5 (d), and the edge of the food in FIG. 5 (e) are emphasized.

Next, the first judgment means 5 compares the first filter output value of each pixel with a predetermined first judgment reference value to determine whether the pixel is an isolated area. Is determined and output. The first determination reference value is set so that an edge or the like is not erroneously determined. Also,
The second determination means 6 compares the first filter output value for each pixel with a predetermined second determination reference value to determine whether the pixel is an isolated area candidate, Is output. An edge or the like may be erroneously determined based on the second determination reference value.

The first determination criterion value is set, for example, in the foreign matter detector, by processing the image to be determined in which no foreign matter is present by the first filter calculating means 3, and setting the largest among the processing result values of all the pixels. There is a method of setting a value having a margin to the determination reference value. Further, the setting of the second determination reference value is performed by processing the image to be determined in which only foreign matter is present by the first filter calculating means, and setting a value obtained by giving a margin to the minimum value among the processing results of the foreign matter pixels. There is a method of setting a judgment reference value.

As an example of the processing, FIG. 6 shows an image in which the grayscale image of FIG. 5 is determined as an isolated area by the first determining means. Originally, it is only necessary to determine whether or not the pixel is an isolated region. However, pixels that are determined to be isolated regions are set to black for easy understanding visually, and among the pixels that are not determined, the first filter output value is Positive ones were gray and negative ones were white. Since the first determination reference value is set so as not to be erroneously determined, there is a foreign substance that cannot be detected as shown in FIG. FIG. 7 shows an image obtained as a result of the isolated area candidate determination of the grayscale image of FIG. 5 by the second determination means. As in FIG. 5, pixels determined as isolated region candidates are set to black, and among the pixels not determined, those having a positive first filter output value are set to gray and those having a negative first filter output value are set to white. In FIG. 7, all foreign substances that could not be detected in FIG. 6 as shown in FIG. 6B can be detected as in FIG. 7B. However, the overlapping of foods as shown in FIG. 7D and the edge of the food as shown in FIG. 7E are erroneously determined.

Next, the grayscale image input to the input terminal 1 is processed by the second filter operation means 7 for each pixel determined as an isolated area candidate by the second determination means 6 for the vicinity of the pixel. A second filter output value is calculated by performing a filter operation using the density values of the pixels in the local area,
For a pixel that is not determined as an isolated region candidate by the second determination unit 6, a predetermined value that is not determined to be an isolated region by the third determination unit 8 described below is output as a second filter output value. Is done.

FIG. 2 shows a configuration of a filter using a median filter as an example of the second filter operation means 7. The grayscale image input to the input terminal 11 receives the determination result of the isolated area candidate determined by the second determination unit 6 input to the input terminal 12 for each pixel, and the second determination unit 6 The pixel determined to be a region candidate is sent to the path 13a by the selecting means 13, and the median value of the local area around the pixel is calculated by the median value calculating means 14. Next, a difference value between the density value of the pixel of the grayscale image and the central value of the area of the pixel is calculated by a difference value calculating unit 15.
At the output terminal 17 as the second filter output value.
Is output to The pixel which is not determined as an isolated area by the second determination unit is sent to the path 13 b by the selection unit 13, 0 is output by the zero value output unit 16, and is output to the output terminal 17. 0 is the third determination means 8 described below.
Is a specific example of a value that is not determined to be an isolated area.

As an example of the processing, the X-ray transmission image shown in FIG. 4 is input to the filter input terminal 11 in FIG. 8 and the foreign matter candidate determination result shown in FIG. The resulting gray image is shown. Here, the negative value is set to 0, the maximum value in the grayscale image is normalized to 255, 255 is represented by black, and 0 is represented by white. When the second filter output value is compared with the grayscale image based on the first filter output value in FIG. 5, foreign matters as shown in FIGS. 8A and 8B are similarly emphasized. It can be seen that the edge of the food as in e) is not emphasized.

Next, the third judgment means 8 compares the output value of the second filter with a predetermined third judgment reference value to judge whether or not the pixel is an isolated area. Is output. The third determination criterion value is, for example, that the foreign object detector processes an image to be determined in which no foreign object is present by the second filter operation means 7 so that the maximum value among the processing results of all pixels has a margin. There is a method of setting the determined value as a determination reference value.

As an example of the processing, FIG. 9 shows an image obtained as a result of the third determining means 8 determining the shaded image of FIG.
Pixels determined to be isolated areas were set to black, and pixels not determined to have the second filter output value positive were set to gray, and negative pixels were set to white. First determining means 5 in FIG.
The foreign matter in FIG. 6B that could not be detected in FIG.
It can be detected as shown in FIG. However, the foreign matter shown in FIG. 6C which can be detected by the first determination means 5 cannot be detected as shown in FIG. 9C.

Finally, the logical sum calculating means 9 calculates the logical sum of the isolated area judgment result of the first judgment means 5 and the isolated area judgment result of the third judgment means 8 for each pixel. Is output. Since the pixel determined to be an isolated area by either one of the first determining means 5 and the third determining means 8 is determined to be an isolated area, the isolated area is determined as compared with performing the isolated area determination by one determining means. Is improved.

A second embodiment of the present invention will be described with reference to FIG. The difference from the first embodiment is that the determination result of the isolated region candidate by the second determination means 6 is not only sent to the second filter operation unit 7 but also to the third determination means 8 and the OR calculation means 9. Is to be sent. For the pixel determined as the isolated region candidate by the second determination means 6, a filter operation is performed using the density values of the pixels in the local region around the pixel, and a second filter output value is calculated. Determination means 8
The second filter output value is compared with a predetermined third determination reference value to determine whether or not the pixel is an isolated area. The logical sum of the isolated area determination result of the means 5 and the third isolated area determination result is calculated and output to the output terminal 10. The second filter operation unit and the third determination unit 8 do not need to perform processing on the pixels that are not determined as an isolated region by the second determination unit 6, and the first OR determination unit 9 performs the first determination. The result of the determination by the means 5 is output as it is and output to the output terminal 10.

In the above description, the first filter calculating means using the moving average filter has been described. However, any filter that can emphasize all the isolated regions and can perform the calculation at high speed can be similarly realized. Since a simple moving average filter has a known algorithm for calculating the average value at a high speed, the operation can be performed at a higher speed than an ordinary linear filter that performs a convolution operation.

In the above description, the second filter operation means using a median filter has been described. However, any filter that enhances an isolated area can be similarly realized. For example, as an isolated region extraction filter by a morphological operation, there are a top-hat transform, a loop filter (Quoit filter), and the like. Even if a filter in which the coefficient matrix for performing the filter operation of the first filter operation unit is changed has a different filter characteristic from that of the first filter operation unit, even if the filter is used as the second filter operation unit, the determination of the isolated region can be performed. Sensitivity can be improved.

[0026]

According to the present invention, the first filter output value calculated by the high-speed first filter operation means is compared with a predetermined first determination reference value for each pixel of the grayscale image. Then, while determining whether or not it is an isolated area, comparing the first filter output value and a predetermined second determination reference value, to determine whether or not it is an isolated area candidate, A second filter output value is calculated by the second filter operation means only for pixels determined as isolated region candidates, and the second filter output value is compared with a predetermined third determination reference value. It is determined whether or not the pixel is an isolated area, and a pixel determined to be an isolated area by either one of the first determination reference value and the third determination reference value is determined to be an isolated area. Using a moving average filter In the process similar to the standing area determination device fast, the sensitivity can be realized more sensitive isolation region judging device.

[Brief description of the drawings]

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

FIG. 2 is a diagram showing a configuration of a second filter operation means 7;

FIG. 3 is a diagram showing a configuration of a second exemplary embodiment of the present invention.

FIG. 4 is a diagram showing an X-ray transmission image of an object to be measured including a foreign substance.

FIG. 5 is a diagram showing an output result of the first filter operation means 2.

FIG. 6 is a diagram showing a result of determination of an isolated area by a first determination unit 5;

FIG. 7 is a diagram showing a result of determination of an isolated area candidate by a second determination unit 6;

FIG. 8 is a diagram showing an output result of the second filter operation means 7;

FIG. 9 is a diagram showing a determination result of an isolated area by a third determination unit 8;

FIG. 10 is a schematic configuration diagram of a conventional isolated region determination device.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 input terminal 2 first filter calculating means 3 average value calculating means 4 difference value calculating means 5 first determining means 6 second determining means 7 second filter calculating means 8 third determining means 9 OR calculating means Reference Signs List 10 output terminal 11 input terminal 12 input terminal 13 selecting means 14 median value calculating means 15 difference value calculating means 16 zero value output means 17 output terminal

Claims (1)

[Claims] A first filter calculating means for calculating a first filter output value by performing a product-sum operation on a density value of a pixel in a local area around a pixel of interest in a grayscale image using a predetermined coefficient matrix; 2) receiving the first filter output value, comparing the first filter output value with a predetermined first determination reference value, and determining whether or not the pixel of interest has a higher density than the neighborhood in the grayscale image. A first determining means (5) for determining whether or not an isolated area is a highly isolated area; receiving the first filter output value and determining the first filter output value as a predetermined value A second determination unit (6) that compares the target pixel with the second determination reference value to determine whether the target pixel is an isolated region candidate that is a region having a higher density than the vicinity in the grayscale image; The determination result from the second determination means and the grayscale image Receiving the data of each of the pixels, the pixels determined as the isolated region candidates among the pixels are subjected to the filter operation using the density values of the pixels in the local region around the pixel, and the second filter A second filter calculating means (7) for calculating an output value, receiving the second filter output value, and comparing the second filter output value with a predetermined third determination reference value. A third determining unit (8) for determining whether the target pixel is an isolated region, and based on a determination result of the first determining unit and a determination result of the third determining unit. An isolated area determination device that determines whether the pixel of interest is an isolated area.