JP2012078195A - Inspection method for foreign matter in solution - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an inspection method for a foreign matter in a solution capable of quickly and accurately discriminating air bubbles and a foreign matter included in the solution.SOLUTION: A foreign matter detection method for detecting a foreign matter in a solution by imaging the foreign matter in the solution by an imaging means includes (a) a first arithmetic operation means for comparing an image acquired by the imaging means with a preset first threshold to extract bright part data and dark part data, (b) a second arithmetic operation means for generating a convex hull pattern of the dark part data extracted by the first arithmetic operation means to extract it, and (c) a third arithmetic means for calculating the ratio between the area of the dark part data extracted by the first arithmetic operation means and the area of the convex hull pattern of the dark part data extracted by the second arithmetic operation means, comparing it with a preset second threshold and discriminating air bubbles and a foreign matter.

Description

  The present invention relates to a foreign matter inspection method in a solution that can be used for a solution containing foreign matters, and more particularly to a foreign matter inspection method in a solution that can quickly and accurately discriminate bubbles and foreign matters. .

  For example, in the polymerization solution of polyacrylonitrile-based polymer for carbon fiber precursor fiber, foreign matter mixed from the outside after the production of the polymer solution or gel in which a part of the solution is hardened due to fluctuation of thermal history is passed to the subsequent process. If it is leaked, it can be a factor of degrading quality. For this reason, it is important to accurately detect the size and number of foreign matters mixed in the polymer solution and judge the quality.

  In addition, air bubbles often occur in the solution due to air entering from the outside, but the bubbles disappear without flowing out to the subsequent process, so there is no effect on the quality degradation in the subsequent process and the manufacturing process. It is not regarded as a problem. Therefore, it is also important to discriminate bubbles and foreign matters that do not need to be detected.

  Conventionally, there has been proposed a method of detecting a foreign substance in a polymer solution by irradiating light to a pipe for sending a polymer solution by a light projecting means and receiving light transmitted through the pipe by a light receiving means (Patent Document). 1). However, since the method described in Patent Document 1 does not have a function of discriminating foreign objects and bubbles, bubbles that do not need to be detected may be erroneously detected as foreign objects.

  In addition, a method for discriminating foreign objects and bubbles by calculating image data obtained by an imaging means has been proposed (see Patent Documents 2 and 3). However, since the bubbles generated in the solution have various shapes and the direction in which the inside of the solution is imaged by the imaging means is often unique, the image in which the bubbles are imaged is an image in which the central part is hollow Besides, it can be in various shapes. At that time, in the methods described in Patent Documents 2 and 3, bubbles may still be erroneously detected as foreign matters.

JP 2009-31181 A JP 2004-354100 A JP 2006-10612 A

  The present invention relates to a method for inspecting foreign matter in a solution that improves the drawbacks of the conventional technique and can be used for a solution containing foreign matter, and in particular, can quickly and accurately discriminate foreign matter and bubbles. An object is to provide a method for inspecting foreign matter in a solution.

In order to achieve the above object, the present invention employs the following method. That is, a foreign matter detection method for detecting foreign matter in a solution by imaging foreign matter in a solution by an imaging means,
(A) a first computing means for extracting bright part data and dark part data by comparing with a first threshold value provided in advance for an image obtained by the imaging means;
(B) second computing means for generating and extracting a convex hull figure of the dark part data extracted by the first computing means;
(C) A ratio of the area of the dark portion data extracted by the first calculating means and the area of the convex hull figure of the dark portion data extracted by the second calculating means is calculated, and a second preliminarily provided It is a foreign matter inspection method in a solution which has the 3rd calculating means which distinguishes a bubble and a foreign substance compared with a threshold.

  In the foreign matter inspection method in the solution of the present invention, the ratio of the area of the dark portion data extracted by the first calculation means and the contour length is calculated, and compared with a preliminarily provided third threshold value, bubbles and foreign matters are detected. It is preferable to include a fourth calculation means for determining.

  In the foreign matter inspection method in the solution of the present invention, the solution used as a detection target is preferably a polymerization solution used for polymerization of a polyacrylonitrile polymer for a carbon fiber precursor fiber.

  ADVANTAGE OF THE INVENTION According to this invention, in the solution containing a foreign material, the bubble and foreign material which generate | occur | produced in the solution can be discriminate | determined accurately, and the improvement of the quality control of a solution can be aimed at.

It is a structure perspective view of the foreign material detection apparatus used for implementation of the foreign material inspection method in the solution of this invention. It is the image of the foreign material and bubble which become the detection targets of the present invention. Note that (a) is a gel, (b) is a fibrous foreign material, and (c) to (e) are bubbles. It is the schematic of the data processing means used for implementation of the foreign material inspection method in the solution of this invention. It is a conceptual diagram of the convex hull figure produced | generated by the 2nd calculating means of this invention. In addition, (a) is an original figure and (b) is a convex hull figure. It is a specific example of the convex hull figure produced | generated by the 2nd calculating means of this invention. In addition, (a) is an original figure and (b) is a convex hull figure.

  Hereinafter, an embodiment of an apparatus suitably used for a foreign matter detection method in a solution according to the present invention will be described with reference to the drawings. FIG. 1 shows a foreign object detection apparatus suitably used in the foreign object detection method of the present invention. The foreign object detection apparatus processes an image obtained by the imaging unit, an illuminating unit 22 that irradiates light to the transparent pipe 21 that supplies the solution, an imaging unit 23 that images the solution illuminated by the illuminating unit, and the imaging unit. And data processing means 24 for detecting foreign matter.

  As the illumination means that is preferably provided in order to make the effects of the present invention more prominent, a lighting unit that can uniformly illuminate the transparent pipe for feeding the solution is preferable. Specific examples include high-frequency lighting fluorescent lamps, metal halide lamps, halogen lamps, and LEDs.

  The image pickup means in the present invention is not particularly limited as long as it can pick up foreign substances in the solution, but the light irradiated by the illumination means can be received by an image pickup element (pixel), and digitally output and recorded as an image. Those are preferred. Specific examples include a digital camera, a small CCD camera, and a digital video camera.

  Moreover, the image which imaged the foreign material, gel, and bubble used as a detection target with the imaging means is shown in FIG. Here, (a) is a gel, (b) is a fibrous foreign material, and (c) to (e) are bubbles. Thus, it can be seen that the foreign substances, gels and bubbles have various shapes such as round shapes and fiber shapes depending on the generation conditions.

  As shown in FIG. 3, the data processing means in the present invention includes a first computing means 31, a second computing means 32, and a third computing means 33.

  The first calculating means is a calculating means for comparing the first obtained threshold value with respect to the image obtained by the imaging means and extracting the bright portion data and the dark portion data.

  The second calculating means is a calculating means for generating and extracting a convex hull figure of the dark part data extracted by the first calculating means.

  The third calculation means calculates the ratio of the area of the dark portion data extracted by the first calculation means and the area of the convex hull figure of the dark portion data extracted by the second calculation means, and is provided in advance. Comparing with a second threshold value, it is a computing means for discriminating between bubbles and foreign matters.

  Here, each calculation means in the data processing means will be described in detail.

  First, the first computing means extracts bright part data and dark part data from an image obtained by the image pickup means by comparing with a first threshold value provided in advance. Since the image obtained by the imaging means here may include a foreign object other than the foreign object to be detected (for example, the background of the imaging region), only the foreign object to be detected from this image. The process to extract is executed. Such extraction processing is generally binarized, and if the pixel in the image is brighter than a preset brightness (threshold), it is extracted as bright portion data, and if it is dark, it is extracted as dark portion data. In the foreign matter detection method in the solution according to the present invention, since the foreign matter can be imaged darker than the background by imaging the light emitted from the illumination unit as transmitted light, it is possible to apply binarization processing. it can. By performing such processing on all the pixels in the image, the background is extracted as bright portion data, and foreign matters and bubbles are extracted as dark portion data.

  In addition, when an image includes noise (for example, tumbling noise called sesame salt noise), before performing binarization processing, an averaging filter, a Gaussian filter, and a median are used. It is preferable to reduce noise so that foreign matter can be more easily extracted as dark part data by performing filter processing such as filtering.

  The second calculation means generates a convex hull figure for the dark part data extracted by the first calculation means, and extracts the generated convex hull figure. A convex hull figure here is a figure surrounded by a polygon having only an acute angle, an obtuse angle, or a right angle at each vertex, or a curve having no inflection point. For example, when the dark part data has a shape as shown in FIG. 4A, only the inner angle of the first vertex 41 is neither an acute angle, an obtuse angle, or a right angle. Therefore, by connecting the second vertex 42 and the third vertex 43 in FIG. 4A, as shown in FIG. 4B, each vertex has only an acute angle, an obtuse angle, or a right angle. A polygon, that is, a convex hull figure can be obtained. Specifically, when the dark portion data extracted by the first calculation means has a shape having a dent as shown in FIG. 5A, the convex hull figure of the dark portion data is filled with the dent, and FIG. ). Thus, the convex hull figure of the dark part data extracted by the first computing means is generated and extracted.

  The third calculation means calculates the ratio of the area of the dark portion data extracted by the first calculation means and the area of the convex hull figure of the dark portion data extracted by the second calculation means, Compared with the threshold value of 2, bubbles and foreign matters are discriminated. Here, the area refers to the number of pixels constituting the convex hull figure of the dark portion data extracted by the first calculation means and the dark portion data extracted by the second calculation means. Therefore, the area can be obtained by counting the number of pixels constituting the dark portion data.

Further, the ratio of the area of the dark portion data extracted by the first calculating means and the area of the convex hull figure of the dark portion data extracted by the second calculating means is called convexity, and is calculated by the following equation (1). can do.
T = Fo / Fc (1)
Here, T: convexity, Fo: area of dark part data, Fc: area of convex hull figure.

  The convexity T is 1 if the dark portion data is a simple shape having no dents or holes (for example, a rectangle or a circle), and is less than 1 if the dark portion data is a complicated shape having dents or holes. The degree of convexity T decreases as the distance from the shape increases.

  Therefore, since the convexity T varies depending on the shape of the extracted dark part data, a threshold is set in advance. For example, if the convexity T of the dark part data is larger than this threshold, the dark part data is foreign. It is possible to determine whether it is a bubble or a bubble.

  In addition, after the first computing means, the second computing means, and the third computing means are implemented, the following fourth computing means may be implemented.

  The fourth computing means computes the ratio of the area of the dark area data extracted by the first computing means and the contour length, and compares it with a third threshold value provided in advance to discriminate bubbles and foreign matters. It is a calculation means.

  Here, the fourth computing means will be described in detail.

  The fourth computing means computes the ratio of the area of the dark portion data extracted by the first computing means to the contour length, and compares it with a third threshold value provided in advance to discriminate bubbles and foreign matters. Here, the contour length is the length of a portion (contour portion) obtained by extracting pixels adjacent to the pixels constituting the surrounding bright portion data among the pixels constituting the dark portion data extracted by the first calculating means. Say. For example, there is contour tracking as a method of extracting pixels constituting the contour portion from the extracted dark portion data. A pixel in the vicinity of the surrounding 8 adjacent to one pixel to be extracted is examined, and if a dark pixel is found, the pixel moves to that pixel. Then, by examining the neighboring pixels in the vicinity of the surrounding 8 with respect to one pixel of the movement destination and finding a dark pixel, the procedure of moving to that pixel can be repeated to extract the contour portion. Therefore, the contour length can be obtained by counting the number of extracted pixels.

The ratio of the area of the dark portion data extracted by the first calculation means to the contour length is called circularity and can be calculated by the following equation (2).
C = L ² / (4 · Fo · π) (2)
Here, C: circularity, L: contour length, Fo: area of dark part data, and π: circumference ratio.

  The circularity C is 1 when the dark portion data is circular, is greater than 1 when the dark portion data is not circular (for example, a square, a rectangle, a polygon, etc.), and the circularity C increases as the distance from the circle increases. Become.

  Therefore, since the circularity C differs depending on the shape of the extracted dark part data, by providing a threshold value in advance, for example, the dark part data such as a foreign object if the circularity C of the dark part data is larger than the threshold value, and a bubble if it is smaller. Can be discriminated whether it is a foreign object or a bubble.

  By implementing the fourth calculation means after the first calculation means, the second calculation means, and the third calculation means described above, it becomes possible to discriminate foreign substances and bubbles more accurately and finely.

  In the foreign matter inspection method in the solution of the present invention, the solution used as a detection target is not particularly limited, but is preferably a polymer polymerization solution from the viewpoint of discriminating gels and foreign matters contained in the polymer polymerization solution from bubbles. The polymer solution used for the polymerization of the polyacrylonitrile-based polymer for the carbon fiber precursor fiber is used to discriminate the gel and foreign matter and bubbles present in the solution to grasp the generation state of the gel and foreign matter, This is particularly preferable in terms of quality control.

  Examples of the present invention will be described below. The configuration of the foreign matter detection apparatus in the solution used in this example is shown below.

Illumination means: Metal halide lamp (Surface light source type light guide)
Imaging means: area camera (2 million pixels, manufactured by Keyence Corporation, CV-200C).

  FIG. 2 is an image obtained by capturing foreign substances and bubbles in the solution with the above configuration. The image processing library software HALCON (Ver. 8.0, MVTec Co., Ltd.) uses the calculation means of the present invention for the image of the foreign matter and bubbles shown in FIG. Made).

  Note that FIG. 2 is an image obtained by performing the first calculation unit on the image obtained by capturing the foreign object by the imaging unit and extracting the foreign object or the bubble as dark part data.

  First, with respect to FIGS. 2A to 2E, the second computing means is implemented to generate and extract a convex hull figure of dark part data. Table 3 shows the result of calculating the convexity T in FIGS. 2A to 2E after performing the third calculating means.

  Here, (a) is a gel, (b) is a fibrous foreign material, and (c) to (e) are bubbles.

  Therefore, by setting the threshold value of the convexity T to 0.75 (foreign matter if the convexity T is larger than 0.75, bubbles if it is smaller), whether the dark part data extracted by the first calculation means is a bubble or foreign matter. It became possible to determine whether it was.

  Table 2 shows the results of calculating the circularity C in FIGS. 2A to 2E after performing the third calculation means and then performing the fourth calculation means. Here, it is completed to determine whether the dark part data is a foreign substance or a bubble by the third calculation means, but when it is desired to further determine what the foreign substance is (in this embodiment, it can be said). For example, when it is desired to discriminate between the gel in FIG. 2A and the fibrous foreign matter in FIG. 2B, the threshold of the circularity C is 2.30 (if the circularity C is greater than 2.30, the fibrous It is possible to determine whether the dark portion data extracted by the first calculation means is a gel or a fibrous foreign matter by setting the foreign matter to be a foreign matter or a gel if it is small.

  Thus, it is possible to discriminate bubbles and foreign matters quickly and accurately without performing complicated processing such as differentiation as compared with the conventional method.

21: Transparent piping 22: Illuminating means 23: Imaging means 24: Data processing means 31: First computing means 32: Second computing means 33: Third computing means 41: First vertex 42: Second vertex 43: Third vertex

  According to the present invention, it is possible to quickly and accurately discriminate bubbles and foreign matters from an image obtained by imaging a solution containing foreign matters (for example, a polymerization solution of a polyacrylonitrile-based polymer for carbon fiber precursor fibers) by an imaging means. Since it has the characteristic that it can be used, it can be suitably used for foreign matter inspection and quality control in the polymer solution manufacturing process, but its application range is not limited to this.

Claims (3)

A foreign matter detection method for detecting foreign matter in a solution by imaging foreign matter in a solution by an imaging means,
(A) a first computing means for extracting bright part data and dark part data by comparing with a first threshold value provided in advance for an image obtained by the imaging means;
(B) second computing means for generating and extracting a convex hull figure of the dark part data extracted by the first computing means;
(C) A ratio of the area of the dark portion data extracted by the first calculating means and the area of the convex hull figure of the dark portion data extracted by the second calculating means is calculated, and a second preliminarily provided A foreign matter inspection method in a solution having a third computing means for discriminating bubbles and foreign matters as compared with a threshold value. A fourth calculation unit that calculates a ratio between the area of the dark portion data extracted by the first calculation unit and the contour length and compares the third threshold value with a predetermined third threshold, A method for inspecting foreign matter in a solution according to claim 1. The foreign matter inspection method in the solution according to claim 1 or 2, wherein the solution is a polymerization solution used for polymerization of a polyacrylonitrile-based polymer for carbon fiber precursor fibers.

Images