JPH09145466A - Spark burst region detecting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately detect a burst image by expansion-processing the binary original image of a spark fixed times, shrinkage-processing it more times, temporarily identifying the remaining picture element lump as a burst region, obtaining the size, shape, and position of the region, and integrating overlapped regions. SOLUTION: A spark image photographed by a CCD camera 6 is converted in a binary value by a binary section 11, it is expansion-processed multiple times (e.g. five times) by an expansion process section 12, it is shrinkage- processed more times (e.g. seven times) by a shrinkage process section 13, and an image having a small picture element lump at the existing position of a burst is obtained. It is labeled by a labeling section 14, the area and lateral and longitudinal lengths of the picture element lump are obtained by a burst region calculation section 15, the burst area is determined from the area, the lateral and longitudinal lengths of the burst region are determined from the lateral and longitudinal lengths, the obtained position of the center of gravity is set as the center of the burst region, and it is reported to an integration process/burst region determination section 16. The burst image of the spark image can be accurately and surely detected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to detection of a spark rupture area for detecting a rupture point of a spark for the purpose of observing, for example, a spark generated when a steel material is rubbed in order to determine the amount of elements contained in the steel material. Regarding the device.

[0002]

2. Description of the Related Art It is conventionally known that the state of sparks produced when a steel material is rubbed varies depending on the amount of elements (carbon or the like) contained in the steel material. Using such a property, the inspector visually recognizes the state of sparks generated when a steel material is rubbed by a grinder, and it is conventionally possible to judge the composition of the steel material (carbon content etc.) or the type of steel material. Is done more.

However, in the above visual judgment, the carbon content of the steel material and the like are judged based on the inspector's empirical rules and intuition, so a skilled inspector is required and the judgment result is inaccurate. There was a risk of becoming.

Therefore, it has been considered to accurately determine the carbon content and the like of the steel material by capturing and inputting sparks with an electronic camera such as a CCD camera and inputting the image and using a computer or the like. In this case, there is a rupture in the spark that has a large feature, and by monitoring the rupture state, the carbon content and the like of the steel material can be determined with high accuracy.

[0005]

However, the position where the burst occurs is indefinite, and the burst exists in various positions in the spark image obtained by the electronic camera. Therefore, it is difficult to accurately detect the rupture in the spark image, and the rupture state cannot be reliably monitored.

Further, not only is it difficult to specify the position of the rupture, but also the number, position and length of the bright lines constituting the rupture,
That is, since the size and shape of the rupture vary depending on the rupture, it is difficult to recognize where and where one rupture is.

The present invention has been made in view of the above circumstances, and an object thereof is to provide a spark rupture area detection device capable of accurately and reliably detecting individual rupture areas in a spark image. Especially.

[0008]

In order to achieve the above object, the present invention provides a binary image by binarizing an original image obtained by capturing an image of a spark containing rupture by an image capturing unit with a predetermined threshold value. The binarizing means for generating, the expanding means for expanding the binary image generated by the binarizing means a predetermined number of times to generate an expanded image, and the expanding processing by the expanding means. A contraction unit that performs a contraction process a large number of times on the expansion image generated by the expansion unit to generate a contraction image, and a lump of each pixel that remains in the contraction image, in the rupture region where the rupture occurs in the spark. Rupture area identification means for once identifying as corresponding, and for each cluster of pixels once identified, calculate the size of the burst area based on the area of the cluster of pixels, and calculate the centroid of the cluster of pixels, Vertical length, horizontal length, etc. A rupture region calculation means for calculating the shape and position of the rupture region using at least one of the parameters, and a plurality of regions that overlap each other for each rupture region for which the size, shape and position have been once identified. Is a spark burst area detection device characterized by including an integrated processing / burst area determination means for determining the other areas as a burst area by integrating the plurality of areas into one burst area. is there.

Therefore, first, in the spark rupture area detection device of the present invention, the original image of the spark containing the rupture imaged by the image pickup means such as the CCD camera is binarized by the binarization means by a predetermined threshold value. It

Further, the dilation means subjects the binary image to dilation processing a predetermined number of times to generate an dilated image. Further, the contraction unit performs the contraction process a predetermined number of times larger than the number of expansion processes by the expansion unit, and the contracted image is generated by the contraction unit.

At this time, since the number of contraction treatments is larger than the number of expansion treatments, the coarse area of the bright line in the spark disappears by the contraction treatment, but in the dense area of the bright line in the spark, that is, the ruptured portion, Since the bright lines are fused into one lump by the expansion process, the lump does not disappear even when contracted, and remains as a lump of pixels corresponding to the size and shape of the rupture on the contracted image.

Next, the rupture area identifying means temporarily identifies the lump of each pixel remaining in the contracted image as corresponding to the rupture area in the spark where the rupture has occurred. What is supposed to be once identified here is the integrated processing /
This is because the final burst area may be changed later by the burst area determination means.

Further, the rupture area calculation means calculates the size of the rupture area based on the area of the lump of each pixel once identified, and the center of gravity and the vertical length of the lump of pixels are calculated. , The shape and position of the rupture region are calculated using at least one of the parameters such as the horizontal length.

As for the shape and position of the region at this time, for example, a rectangle having the same ratio as the ratio of the vertical length to the horizontal length may be arranged at the center of gravity. Then, the integrated processing / rupture area determination means integrates a plurality of areas that overlap each other and determines one area as one burst area, and the other areas are directly determined as the burst area. In this way, the burst area in the spark image is accurately and reliably detected and determined.

[0015]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a main configuration of a steel material type determination device configured by applying a spark rupture region detection device according to an embodiment of the present invention.

In the figure, reference numeral 1 is a spark burst area detecting device, and 2 is an image processing device. The spark burst area detection device 1 and the image processing device 2 are each provided with a spark image obtained by capturing an image of a spark 5 generated by rubbing the steel material 3 to be inspected with a grinder 4 with a CCD camera 6.

The spark burst area detection device 1 detects the area occupied by the burst portion in the spark image provided by the CCD camera 6 and notifies the image processing device 2 of the detection. This spark burst area detection device 1 includes a binarization unit 1
1, an expansion processing unit 12, a contraction processing unit 13, a labeling unit 14 as a rupture region identifying means, a rupture region calculation unit 15, and an integration process / rupture region determination unit 16 are connected in series in order.

The image processing device 2 extracts a rupture from the spark image given from the CCD camera 6 based on the rupture region notified from the spark rupture region detection device 1, and from the state of the rupture, the steel material 3 to be inspected is extracted. Perform processing to determine the type.

Next, the operation of the steel material type determination device configured as described above will be described. First, when the grinder 4 is rotated at a high speed to rub the steel material 3 to be inspected, a spark 5 is generated, and the spark 5 is imaged by the CCD camera 6. As a result, the spark images obtained by the CCD camera 6 are given to the spark burst area detection device 1 and the image processing device 2, respectively.

In the spark rupture area detection device 1, the binary image conversion unit 1 uses the spark image provided by the CCD camera 6 as an original image.
Enter 1 The binarization unit 11 binarizes the original image to generate a binary image.

By the way, in the case of sparks, generally, the vicinity of the rupture is brighter than the streamline portion, and in the original image, the vicinity of the rupture has a higher brightness level than the streamline portion. Therefore, the threshold value used in the binarization unit 11 is set to a relatively high value that removes the streamline portion having a significantly lower brightness than the vicinity of the burst from the binary image.

Thus, for example, when the original image is as shown in FIG. 2, the low-luminance streamline portion is removed as shown in FIG. 3, and the rupture portion and the high-luminance streamline portion remain. A value image is generated.

The binary image generated by the binarizing unit 11 in this manner is supplied to the dilation processing unit 12. The dilation processing unit 12 repeatedly performs dilation processing on the binary image supplied from the binarization unit 11 a plurality of times (for example, five times).

As a result, for example, the image shown in FIG. 4 existing in the binary image is converted into the image shown in FIG. 5, and the dilated image shown in FIG. 6 is converted to the image shown in FIG. 3 based on the binary image shown in FIG. Is generated.

This expanded image is given to the contraction processing unit 13. Here, the contraction processing unit 13 performs contraction processing on the expanded image provided from the expansion processing unit 12,
It is repeatedly applied over a plurality of times (for example, 7 times), which is larger than the number of times of the expansion processing in step S. As a result, since the contraction process is performed more times than the expansion process, most of the images existing in the expanded image are removed.

However, since a plurality of bright lines are present close to each other at the ruptured portion, the images of the plurality of bright lines stick to each other when the expansion process is performed, and as shown in FIG. 5 in the expanded image. It exists as a large image (lump of pixels).

Such an image is not completely removed even after the contraction process, and remains as a small image as shown in FIG. 7, for example. Thus, based on the expanded image shown in FIG. 6, a contracted image having a small image at the position where the rupture exists is obtained as shown in FIG.

This contracted image is given to the rupture area calculation section 15 via the labeling section 14. Labeling section 14
Performs a labeling process for each pixel block in the contracted image provided from the contraction processing unit 13.

Further, the rupture region calculation unit 15 obtains the center of gravity and the rupture region as shown in FIG. 9 based on the position and the shape of the pixel cluster subjected to the labeling processing by the labeling unit 14.

Here, the method of calculating the rupture area will be described in more detail. First, as shown in FIG. 9, the area of a pixel block, the vertical length (vertical length) of the pixel block, and the horizontal length (horizontal length) are calculated as parameters.

Next, a predetermined coefficient is applied to the area of the lump of pixels to obtain a rupture area area indicating the size of the rupture area.
Further, the vertical length of the rupture region and the horizontal length of the rupture region are determined so that the product thereof becomes the rupture region area while maintaining the ratio of the vertical length and the horizontal length. The center position of the rupture area is the center of gravity shown in FIG.

The rectangular rupture area is calculated as described above, and is notified to the integrated processing / rupture area determination unit 16.
The individual rupture regions are calculated in this way. However, when the individual rupture shapes are relatively simple as shown in FIGS. 2 and 3, the rupture region calculation unit 15 calculates the rupture regions. The individual bursts are correctly identified and extracted at. FIG. 10 shows the rupture areas A1 and A calculated in such a case.
2 and A3 are shown.

However, for the actual rupture, as shown in FIG.
As shown in (c), there are ruptures of various sizes and shapes from simple shapes or small ones to complicated shapes or large ones.

Among the bursts of various sizes and shapes as described above, if the burst has a relatively small and simple shape as shown in FIG. 10 and FIG. One rupture area is set and the rupture is extracted correctly.

However, in the case of a rupture of a relatively large and complicated shape as shown in FIGS.
In the contraction processing, one rupture may be a cluster of a plurality of different pixels, and in that case, each cluster of pixels is set as a separate burst area.

FIG. 12 is a diagram showing a state in which a region is set for the rupture shown in FIG. 11 (b). FIG. 11B
When the expansion / contraction process is performed on one burst shown in (4), clusters L4 and L5 of two pixels are formed as shown in FIG. This is because the rupture is large as a whole, and the main luminescent lines constituting the rupture are divided into two, and smaller luminescent lines are generated from each of these two luminescent lines.

In this case, as shown in FIG. 12, the rupture area is divided into two different areas, the rupture area A4 and the rupture area A5. That is, what is originally one rupture is set separately for two rupture regions.

Therefore, in the present embodiment, in such a case, the two rupture regions A4 and A5 are integrated into one. That is, the integrated processing / rupture area determination unit 16 checks whether or not each of the burst areas (temporary) calculated by the burst area calculation unit 15 has an overlapping portion, and detects a plurality of burst areas having overlapping portions. Are integrated into one burst area.

FIG. 13 is a diagram showing a state in which the two rupture regions A4 and A5 shown in FIG. 12 are integrated into one rupture region A45. In this way, the rupture regions (temporary) calculated by the rupture region calculation unit 15 are integrated as overlapping, and are determined as rupture regions indicating one rupture. Then, the integrated processing / rupture area determination unit 16 notifies the image processing apparatus 2 of each confirmed rupture area.

When the image processing apparatus 2 receives the notification of the burst area from the integrated processing / rupture area determination unit 16, the image processing apparatus 2 extracts the burst area from the spark image given from the CCD camera 6,
Image processing is performed, and the image in the rupture area is the image corresponding to the rupture,
That is, it is converted into a burst image.

The burst image converted here is subjected to appropriate expansion and thinning processing to reproduce the burst shape, and
It is a processed image in which a feature amount processed so that a bright line of a burst is not broken can be easily extracted.

Then, the rupture state obtained from the rupture image is used as a steel material type determining material to perform a process of determining the type of the inspection steel material 3. As described above, according to the spark rupture area detection device according to the embodiment of the present invention, binarization, expansion / contraction processing is performed on the obtained spark screen,
The area of the remaining pixel mass, the size of the burst area from the center of gravity, etc.
Since the shape and position are calculated, and the overlapping rupture regions are integrated to determine the rupture region, the individual ruptures in the spark can be accurately and surely taken out.

Further, the position of the burst in the spark image can be accurately detected. Therefore, the image processing device 2
In, it is possible to reliably extract the burst portion from the spark image based on the burst area set by the spark burst area detection device 1, and, based on the burst thus extracted, the state of the spark, that is, the steel material to be inspected. The type of can be accurately determined.

The present invention is not limited to the above embodiment. For example, the spark image input device of the present invention is not limited to the steel material type determination device, and the present invention can be applied to other devices such as a device for inspecting the carbon content of the steel material 1 to be inspected. Is.

Further, in the present embodiment, as a method of calculating the burst area, the size of the area is determined by the area of the cluster of pixels in the contracted image, and the center of gravity is taken as the center of the area, and the vertical length of the cluster of pixels and Although the aspect ratio of the area is determined by the horizontal length, the area calculation method of the present invention is not limited to this, and various methods can be considered.

As a first example, first, in the method of this embodiment, the rupture region is determined in proportion to the area of the block of pixels, the vertical length and the horizontal length. Whether the size of the burst area is large, medium, or small, and prepared in advance as shown in FIG. 14, and using the table in which the size and shape pattern of the burst area are set according to each parameter, The shape and size of the area may be determined.

By doing so, the calculation load of the spark area detecting device 1 can be reduced, and the spark area detecting device 1 can contribute to high-speed operation.
Also, based on experience, the size of the rupture area in the table
Since the shape can be set, it is possible to determine the rupture region according to the actual condition.

As a second example, with the center of gravity as the center of the area, the size of the area is determined by the area of the lump of pixels in the contracted image, and the shape of the rupture area is a circular area, and the area of the lump of pixels is Is multiplied by a predetermined value to obtain the radius of the rupture area.

In this way, the burst region can be set easily. As a third example, the position shape of the rupture region is determined only by the vertical length and the horizontal length of the block of pixels. That is, the rectangle formed when calculating the vertical length and the horizontal length of the block of pixels is enlarged at that position by a predetermined magnification. At this time, the enlargement ratio of the rectangle may be determined based on the area of the pixel block. In this way, the burst area can be easily set. In addition, various modifications can be made without departing from the spirit of the present invention.

[0050]

As described in detail above, according to the present invention, the burst area is calculated by the area of the block of pixels obtained by image-processing the spark image, and the overlapping burst areas are integrated. It is possible to provide a spark rupture area detection device capable of accurately and surely detecting the rupture area of the.

[Brief description of the drawings]

FIG. 1 is a diagram showing a main configuration of a steel material type determination device configured by applying a spark burst area detection device according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of an original image input to the spark rupture area detection device of the embodiment.

FIG. 3 is a diagram showing a binary image obtained by binarizing the original image in the same embodiment.

FIG. 4 shows 1 existing in a binary image in the same embodiment.
The figure which shows typically an example of the image of the vicinity of one burst.

FIG. 5 is a diagram schematically showing an image obtained by performing expansion processing in the same embodiment.

FIG. 6 is a diagram showing an expanded image generated by performing expansion processing on a binary image in the same embodiment.

FIG. 7 is a diagram schematically showing an image obtained by performing contraction processing in the same embodiment.

FIG. 8 is a diagram showing a contracted image generated by performing contraction processing on the expanded image in the embodiment.

FIG. 9 is a diagram illustrating a calculation state of a rupture region based on a block of pixels existing in a contracted image according to the same embodiment.

FIG. 10 is a diagram showing an example of a rupture region when a rupture portion is extracted from a spark image in the same embodiment.

FIG. 11 is a view showing an example of rupture.

FIG. 12 is a diagram showing a state when a region is set for a rupture in the same embodiment.

FIG. 13 is a diagram showing how two rupture regions are integrated in the same embodiment.

FIG. 14 is a diagram showing an example of a table for burst region calculation according to another embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Spark rupture area | region detection apparatus, 11 ... Binarization part, 12 ... Expansion processing part, 13 ... Contraction processing part, 14 ... Labeling part, 1
5 ... Burst area calculation unit, 16 ... Integrated processing / rupture area determination unit, 2 ... Image processing device, 3 ... Steel to be inspected, 4 ... Grinder, 5 ... Spark, 6 ... CCD camera.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yuki Enokido 66-2 Horikawa-cho, Sachi-ku, Kawasaki City, Kanagawa Prefecture, Toshiba Engineering Co., Ltd. 21-11

Claims (1)

[Claims] 1. A binarizing unit that binarizes an original image obtained by capturing a spark containing a burst by an image capturing unit with a predetermined threshold value to generate a binary image, and the binarizing unit. And an expansion unit that performs an expansion process on the binary image a predetermined number of times to generate an expansion image, and a contraction process that is generated by the expansion unit a predetermined number of times greater than the number of expansion processes performed by the expansion unit. Shrinking means for applying the dilated image to generate a shrunken image, and burst area identifying means for once identifying a block of each pixel remaining in the shrunken image as corresponding to a burst area in which a burst in the spark has occurred. And for each of the once identified clusters of pixels, calculating the size of the rupture region based on the area of the cluster of pixels, and
A rupture area calculation means for calculating the shape and position of the rupture area using at least one of the parameters such as the center of gravity, the vertical length, and the horizontal length of the block of pixels; And, for each rupture region whose position has been calculated, a plurality of regions that overlap each other are:
A spark burst area detection device, characterized by comprising an integrated process / rupture area determination means for integrating the plurality of areas into one burst area and determining the other areas as burst areas as they are.