JP2010266366A - Characteristic extraction method of image, tool defect inspection method, and tool defect inspection device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a tool defect inspection method and a tool defect inspection device for viewing the whole blade surface with fixed brightness, and inspecting efficiently and accurately. <P>SOLUTION: In this characteristic extraction method for imaging a workpiece 12 which is an inspection object, and extracting a characteristic line of its shape, imaged image data are used as they are, and a pixel row of one pixel×n pixels (n is a natural number) is taken in a vertical axis direction, in a horizontal axis direction or in an oblique direction of a pixel array of an imaging element, and a pixel having the smallest value or the largest value of brightness of each pixel is determined in the row. The pixel is extracted as a part of the shape of an inspection object, and in the extraction processing, the pixel having the smallest value or the largest value is determined, while changing successively the position of the pixel row to a direction orthogonal to a direction of the pixel row. Each extracted pixel is connected, to thereby obtain the characteristic line of the inspection object. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention uses a method of capturing an image of an object to be inspected by an optical imaging device such as a CCD camera, and uses a throw-away tip that is a blade tool of various rotary cutting machine tools, and a blade of a blade tool such as a drill or a tap. The present invention relates to an image feature extraction method, a tool defect inspection method, and a tool defect inspection apparatus that can be used for shape inspection and defect inspection.

  Conventionally, for example, a throwaway tip that forms a cutting edge of a cutting tool such as a drill blade, a tap, a cutting tool, a milling cutter, etc., various types of structures formed in a substantially rhombus, a substantially square, or a substantially equilateral triangle having an appropriate thickness Exists. Furthermore, many throw-away tips are manufactured by a powder sintering method. For example, defects such as fine chips and cracks of about several tens of μm may occur partially or over the entire blade. In this case, it was necessary to reliably identify these. Therefore, since the inspection process requires complicated and delicate determination, conventionally, defective products are extracted and products are sorted by visual inspection.

  In visual inspection by humans, for inspection objects with small changes in brightness between the defective part and the peripheral part, the change in brightness is recognized while moving and rotating the inspection object, and the part where the change in brightness is large is defective. It is determined that the portion or the shape of the inspection object changes. In addition, in appearance inspection using light, the brightness of the defective part and the peripheral part of the defective part on the flat part or the side part with a steep angle is reflected by light reflection or light scattering in a specific direction. Using the fact that the change tends to be large, defect inspection is performed using the difference in brightness.

  However, with the visual extraction of defective products for each type, it is difficult to classify defective products having minute chippings or deformations and different types of products in various types of products that are slightly different in shape. For this reason, fine defects and defective products may be overlooked during sorting. In addition, with regard to a defective part in a part having a gradual shape change such as an arc, since there is light reflection in all directions regardless of the presence or absence of a defect, the brightness change between the defective part and the peripheral part tends to be small. And it is difficult to identify the defective part. In addition, for the part where the defect itself has a gradual shape change, the brightness change between the defective part and the peripheral part tends to be small regardless of the part having a gradual shape change such as a flat part or an arc. It was difficult to identify.

  In view of this, the defect inspection of the throw-away tip and other blade tools includes inspection by image processing in addition to visual inspection. For example, in the throw-away tip inspection devices disclosed in Patent Documents 1 to 3, a laser beam is scanned to detect a shape defect of the cutting edge of the throw-away tip, and a cutting edge defect or the like is discriminated. Further, in Patent Document 4, a plurality of illumination devices that illuminate the edge edge of a workpiece from different angles and are located on both sides with respect to the imaging direction of the camera, and a camera that is formed so as to be capable of imaging the illuminated workpiece, An image processing device that performs image processing on the inspection image of the workpiece imaged by the camera, arranges the camera for each inspection process of the workpiece, moves the workpiece on the stage, images the surface, side surface, and blade edge of the workpiece, A tool defect inspection method for determining a defect by comparing each predetermined reference image and each inspection image is disclosed.

JP-A-10-202476 JP-A-10-202477 Japanese Patent Laid-Open No. 10-202478 JP 2007-278915 A

  However, in the case of an inspection method in which an appearance inspection is performed using an automatic machine for defect inspection, the blade surface portion of a blade tool such as a throw-away tip has an arc shape and has a gradual shape change. For this reason, when light is irradiated onto the blade surface portion from a specific direction, only a part of the blade surface shines, and it is difficult to see the entire blade surface with a constant brightness. Therefore, in order to see the entire blade surface with a constant brightness, it is necessary to irradiate the inspection object with light from various directions.

  In addition, when acquiring an image by an optical imaging method such as a CCD camera, the camera position is a specific position / angle with respect to the blade surface portion, so that the entire blade surface can be viewed at a constant brightness. A method of covering the inspection range with one image by arranging the illumination in a multifaceted manner and the reflected light of the entire blade surface is incident on the camera, or moving and rotating the inspection object in a three-dimensional direction. Any one of methods of covering the inspection range by combining images of a plurality of blade surface portions by imaging while changing the reflection surface is used. However, the former method is expensive due to the large illumination mechanism, and the latter method is expensive and requires a long inspection time because the conveyance mechanism for the inspection object is complicated and the number of images to be captured increases.

  Also, when viewing the entire blade surface at a certain brightness, even if there is a difference in the shape of the object to be inspected or a difference between a non-defective part and a defective part, the brightness difference is small, so it is difficult to identify only the defective part. is there.

  In addition, in the inspection apparatus disclosed in Patent Documents 1 to 3, the blade line of the throw-away tip is inspected by scanning with laser light, the apparatus is complicated, and the spot diameter is less than the spot diameter of the laser light. In other words, it was not possible to detect minute defects, and the inspection time by scanning was long, and there were few inspection items. Moreover, since the inspection apparatus disclosed in Patent Document 4 irradiates the workpiece inspection surface with the irradiation light irradiated from the illumination device from a plurality of directions, a relatively bright and uniform workpiece image can be obtained. Since the defect is determined by comparing the reference image with each inspection image, the image processing becomes complicated, the apparatus becomes expensive, and inspection time is required.

  The present invention has been made in view of the above problems of the background art, and provides an image feature extraction method capable of easily extracting image features at a constant brightness with an inexpensive and simple method. With the goal.

  Furthermore, the present invention provides a tool defect inspection method and a tool defect inspection apparatus which can efficiently and accurately inspect the entire blade surface with a constant brightness using the image feature extraction method. The purpose is to do.

  The present invention relates to an image feature extraction method, a tool defect inspection method, and a tool defect in which an image is acquired by an imaging device using a CCD camera or the like using uniform illumination, and a defect inspection is performed by examining a change in luminance of the image. Inspection equipment.

  When an image is acquired by an imaging method such as a CCD camera, the conditions for good imaging are that the position of the camera is a specific position and angle with respect to the blade surface of the tool. In order to see the lights, the lights must be arranged in multiple ways. Considering an image captured under uniform illumination, the difference between the luminance value at a certain pixel position and the luminance value of an adjacent pixel is small. That is, the luminance changes moderately in a small area in the image. If there is a part with a different shape in this image, the tendency of the luminance change changes. The present invention utilizes this phenomenon to determine in advance the characteristics of the luminance change with respect to the shape characteristics of the inspection object, and to extract a portion that has a luminance change different from the characteristic of the luminance change. The shape and size of the defective part are detected.

  The present invention relates to a feature extraction method for picking up an image of an inspection object and extracting a feature line of the shape, and using the picked-up image data as it is without performing image processing or the like, in the vertical axis direction of the pixel array of the image sensor. Also, a pixel row of 1 pixel × n pixels (n is a natural number) is taken in the horizontal axis direction or the oblique direction, and a pixel having the minimum value or the maximum value of the luminance of each pixel is obtained in the row, and the pixel is subjected to the inspection. Extracted as a part of the shape of the object, this extraction process is performed by obtaining the pixel having the minimum value or maximum value while sequentially changing the position of the pixel column in a direction orthogonal to the direction of the pixel column, and extracting the pixel Is a feature extraction method of an image that is extracted as a feature line of an inspection object by joining together.

  A plurality of illuminations are arranged on a plane including an optical axis of a lens of a camera lens provided with the image sensor and on a plane orthogonal to the plane, and the inspection object is illuminated from a plurality of angles, and the inspection object Is picked up by the image pickup device, and a characteristic line of the inspection object is extracted. The plurality of illuminations are arranged such that a plurality of illuminations are arranged at equal angles on both sides of the camera lens obliquely facing the edge of the inspection object, and the inspection object is formed by the plurality of illuminations. In light of this, it is possible to capture bright and uniform images over a wide range of the imaging range of the imaging device.

  Further, the image picked up by the image pickup device is divided into a plurality of image ranges, and at that time, the image is divided into image ranges having relatively uniform brightness, and the feature line of the inspection object is extracted for each image range. It is.

  The pixel row is defined in the image range, and the luminance of the pixel at the current position and the pixel at the next position is obtained for each pixel in a direction orthogonal to the pixel row in order from the pixel at the end of the image range. A pixel whose luminance is less than a predetermined luminance threshold and the luminance of the pixel at the next position is larger than the luminance threshold, and the luminance of the pixel at the current position larger than the luminance threshold is higher than the luminance threshold. Obtain a smaller pixel, and obtain all pixels that satisfy these conditions while moving this process in a direction that intersects the direction of the pixel column by a predetermined pixel, for example, one pixel at a time. This is a feature extraction method for an image in which a pixel range between line segments is extracted as a boundary portion.

  The luminance threshold value is a value obtained by adding a predetermined luminance value to the minimum luminance value or a value obtained by subtracting a predetermined luminance value from the maximum luminance value as the luminance threshold value.

  In addition, when imaging the inspection object and extracting the feature line of the shape, if the pixel position of the feature line is shifted from the extending direction of the arrangement and the next pixel of the feature line exists, the shape abnormality It is judged that there is a possibility of.

  The number of pixels in the interval between the boundary portion and the feature line is obtained, and the feature of the boundary line is extracted based on the change in the number of pixels, and the boundary portion at the pixel position where there is a possibility of the shape abnormality When the number of pixels between the feature lines exceeds a predetermined number of thresholds, it is determined that the shape is abnormal.

  A difference between the number of pixels between the boundary lines in the pixel range on the feature line determined to be the abnormal shape portion and an average value of the number of pixels between the boundary lines in the image range is obtained, and the difference is a predetermined value. If it is larger, the shape abnormal portion is determined as a defective portion.

  The present invention is also a defect inspection method for inspecting a tool blade surface using the image feature extraction method. Furthermore, the present invention is a tool defect inspection apparatus that performs a defect inspection of a tool blade surface using the image feature extraction method.

  According to the image feature extraction method, the tool defect inspection device, and the tool defect inspection method of the present invention, it is possible to perform feature extraction of the image using a change in luminance of the image by a simple imaging device and illumination device. . Thereby, it is possible to extract the feature of the shape of the inspection object and the feature of the defect with a small shape change that have been difficult to inspect. Furthermore, by applying this feature extraction method to a blade tool, it is possible to easily capture and extract an image of a blade surface portion and a defective portion. In particular, when the tool blade portion is illuminated by the illumination device used in the present invention, the illumination range that is uniformly illuminated is wide, and it is possible to capture an image of uniform brightness over a wide range with the maximum brightness. Then, it is possible to realize an inspection apparatus having a feature extraction function having a shape having a small shape change, which has a function that is lower in price and higher in versatility than conventional and has been difficult to detect automatically in the past.

1 is a schematic overall configuration diagram showing a tool defect inspection apparatus according to an embodiment of the present invention. It is a schematic perspective view which shows a part of illuminating device of the tool defect inspection apparatus of this embodiment. It is a graph which shows the light intensity part distribution of the illumination used for this embodiment. It is the elements on larger scale of the blade surface of the throw away tip inspected by the tool defect inspection apparatus of this embodiment. It is the partial enlarged image of the edge part of the throw away tip imaged with the tool defect inspection apparatus of this embodiment. It is a schematic diagram of the image of the tool blade surface inspected by the tool defect inspection apparatus of this embodiment. It is a graph which shows the brightness | luminance of a one part pixel row | line of the image imaged with the tool defect inspection apparatus of this embodiment. It is a schematic diagram which shows the pixel of the image part judged to be a defective part by the tool defect inspection apparatus of this embodiment. It is a flowchart of the feature extraction process in the tool defect inspection method of this embodiment. It is a flowchart which detects a defect part by the feature extraction process in the tool defect inspection method of this embodiment. It is a flowchart which fixes a defect part by the feature extraction process in the tool defect inspection method of this embodiment. It is an image which shows the boundary extracted by the feature extraction process in the tool defect inspection method of this embodiment. It is an image which shows the defect part detected by the feature extraction process in the tool defect inspection method of this embodiment.

  Hereinafter, an embodiment of an image feature extraction method, a tool defect inspection method, and a tool defect inspection apparatus according to the present invention will be described with reference to the drawings. The tool defect inspection apparatus 10 and the inspection method of this embodiment are used for inspection of a workpiece 12 that is a throw-away tip that forms a cutting edge of a cutting tool such as a drill blade, a tap, a cutting tool, and a milling cutter used for cutting or the like.

  As shown in FIG. 1, the tool defect inspection apparatus 10 is provided with various illuminating devices 20 to be described later, and a camera 14 such as a CCD for imaging the illuminated workpiece 12 and a lens 15 are not shown. It is fixed by a holding member and attached at a predetermined position. Further, the camera 14 is connected to an image processing device 16 such as a computer provided with a processing program for performing predetermined image processing on the captured image, and the image processing device 16 displays the output of the display device 18. Is provided.

  Here, the edge portion 12b of the blade surface 12a, which is the outer peripheral portion of the throw-away tip, which is the workpiece 12, has a small arc shape when enlarged as shown in FIG. Therefore, the illuminating device 20 which can illuminate the whole blade surface uniformly is required, and is arranged as described below.

  First, the camera 14 and the lens 15 are arranged at an angle of 45 ° with respect to the blade surface 12 a that is the surface of the workpiece 12. And the illuminating device 20 is arrange | positioned on the vertical surface (plane orthogonal to a horizontal surface) which passes along the straight line which is the optical axis of the lens 15. FIG. Here, an example in which six individual lights are arranged in the lighting device 20 is shown. The angle formed by the direction of the optical axis of each illumination of the illumination device 20 includes illumination 20-1 and illumination 20-2, illumination 20-2 and illumination 20-3, illumination 20-3 and camera 14, lens 15, camera 14, and The angles between the lens 15 and the illumination 20-4, the illumination 16-4 and the illumination 16-5, and the illumination 16-5 and the illumination 16-6 are set to be equal. Further, as shown in FIG. 2, the illumination 20-7 and the illumination 20-8 are arranged in the horizontal direction on both sides of the illumination 20-6 at the same angle as the angle formed by the optical axes in the vertical direction of the illumination device 20. That is, the angle formed between the optical axis of the illumination 20-6 and 20-7 and the blade surface 12a is equal to the angle formed between the optical axis of the illumination 20-6 and 20-8 and the blade surface 12a. Then, by adjusting the lighting state and brightness from the illumination 20-1 to the illumination 20-8, illumination with uniform brightness can be made corresponding to various types of blade surface shapes.

  Here, in general, as shown in FIG. 3, the illumination has the maximum light intensity on the optical axis of the illumination, and the light intensity decreases as the distance from the optical axis increases. Therefore, it is preferable to use an illumination with a wide range and a small difference in brightness between the optical axis and the vicinity of the optical axis so that the light intensity is high, that is, bright and uniform illumination. For example, as shown in FIG. 3, when using a light intensity range equal to or greater than the broken line h in the light intensity distributions 1 and 2 of the illumination light, the width A of the light intensity distribution 1 is wider than the width B of the light intensity distribution 2. For this reason, it is preferable to use illumination with a light intensity distribution 1.

  Next, a method for extracting features and defect portions of the edge portion 12b of the blade surface 12a of the workpiece 12 using the images captured by the device shown in FIGS. 1 and 2 of the tool defect inspection device 10 of this embodiment will be described. To do. In general, as shown in FIG. 5, the captured image of the blade surface 12 a of the workpiece 12 has a dark edge portion 12 b and a defective portion of the blade surface 12 a and a bright peripheral portion. Therefore, an inspection range is provided in this image, and the features of the blade surface portion and the defect portion are extracted within this range.

  First, as shown in FIG. 9, the image captured by the camera 14 is divided into a predetermined inspection range A by the processing of the computer 16 (S1). For the image (FIG. 6) in the divided inspection range A, the minimum value Bmin of the luminance B of each pixel in the Y direction which is the vertical axis direction in the image plane for each line L of each pixel of the image data. And its pixel position. This is detected and stored while moving one pixel at a time in the X-axis direction, which is the horizontal direction within the inspection range A (S2). When all pixel positions of the minimum value Bmin of the brightness B are connected, a curve Lb indicated by a broken line in FIG. 6 is formed. The curve Lb is a pixel position corresponding to the edge portion 12b of the blade surface 12a, and the curve Lb is a characteristic line of the edge portion 12b of the blade surface 12a.

  Further, a predetermined brightness is added to each brightness minimum value Bmin to determine a threshold value Bth (S3). Then, the brightness B is examined for each pixel from the top to the bottom for each line in the Y-axis direction, and the brightness threshold Bth is interposed between the pixel position B1 that changes from the brightness higher than the brightness threshold Bth to the brightness lower than the brightness threshold Bth. A pixel position B2 that changes from low luminance to high luminance is obtained (FIG. 7). Moreover, those positions are used as the boundary part of the blade surface 12a part and a defect part. This is performed for the inspection range A, and when the pixel position is moved up and down from the pixel position on the curve of the minimum luminance value Bmin shown in FIG. As will be described later, this width is the vertical width of the white portion shown in FIG. These intervals are obtained as the boundary line width Bw (S4). Further, the abnormal value of the boundary line width Bw is removed, and the average value of the boundary line width Bw in the inspection range A is obtained.

  Next, as shown in FIG. 10, the pixel positions on the curve of the minimum luminance value Bmin in FIG. 6 are examined in order from the left end, and as shown in FIG. A difference in pixel position in the axial direction is obtained (S5). If this difference is greater than or equal to a predetermined number of pixels n, for example, 2 or more, it is determined that the pixel position is “possibly defective” (S6). In this state, the number of pixels in the vertical direction is compared while moving the pixel position, and the process is performed up to the end of the inspection range A (S7).

  Further, the pixel position in the Y-axis direction of the minimum luminance value Bmin at the position “possibly defective” and the Y-axis direction of the pixel position of the movement destination, for example, the minimum luminance value Bmin of the pixel column at the end in the X-axis direction Are compared (S8). When pixel rows having a difference in pixel position in the Y-axis direction of a predetermined number of pixels m or more, for example, two pixels or more are continuous, or when the number of consecutive pixels is a predetermined number p, for example, p> 3 pixels And “defect portion” are determined (S9). Note that the pixel position of the movement destination can be set as appropriate, and adjacent pixel columns may be compared, and the pixel position in the Y-axis direction of the minimum luminance value Bmin to be used as a reference is extracted in advance. Also good.

  Further, as shown in FIG. 11, in order from the leftmost pixel position of the pixel range on the curve of the minimum luminance value Bmin determined as the “defective part”, the boundary line width and the average value of the boundary line width at the pixel position are Is obtained (S10). If this difference is larger than a predetermined value q (for example, 3 pixels) (S11), it is determined as “defective part” (S12), and the boundary line width is set as the length in the Y-axis direction. Further, this calculation is performed within the range previously determined as “defective part”, and the number of pixels in which the part determined to be “defective part” continues in the X-axis direction is defined as the length in the horizontal direction (X-axis direction). By this method, it is possible to determine the presence / absence of a defect and the size of the defective portion. In this way, the position and size of successive defects can be detected in pixel units, as shown in the white areas in FIG.

  According to the image feature extraction method, the tool defect inspection apparatus, and the tool defect inspection method of this embodiment, the simple camera 14 and lens 15, the illumination apparatus 20, and the image processing apparatus 16 using a computer can perform high-speed and accurate image inspection. It can be performed. In particular, since the feature extraction of the image is performed using the change in luminance of the image, it is possible to extract the feature of the inspection object and the feature of the defect with a small shape change, which has conventionally been difficult to inspect. Further, this feature extraction method can be applied to a blade tool to easily and accurately capture and extract an image of a blade surface portion and a defect portion.

  Note that the present invention is not limited to the above-described embodiment, and the camera to be used can use an image pickup device other than a CCD, and a plurality of pixels can be used even if the detection in pixel units is not in units of one pixel. The above processing may be performed as one unit. In the feature extraction, the minimum value of the luminance of the image is detected and used as the feature line. However, depending on the illumination, the maximum shape of pixels may be connected to form a feature shape.

DESCRIPTION OF SYMBOLS 10 Tool defect inspection apparatus 12 Work piece 12a Blade surface 14 Camera 15 Lens 16 Image processing apparatus 18 Display apparatus 20 Illumination apparatus

Claims (11)

  In a feature extraction method for picking up an image of an inspection object and extracting a feature line of the shape, the picked-up image data is used as it is, and 1 pixel × n in the vertical, horizontal, or diagonal directions of the pixel array of the image sensor. Taking a pixel column of pixels (n is a natural number), obtaining a pixel having the minimum or maximum luminance value of each pixel in the column, extracting the pixel as a part of the shape of the inspection object, In the extraction process, the pixel having the minimum value or the maximum value is obtained while sequentially changing the position of the pixel row in a direction orthogonal to the direction of the pixel row, and the extracted pixels are connected to be extracted as a characteristic line of the inspection object. An image feature extraction method characterized by:   A plurality of illuminations are arranged on a plane including an optical axis of a lens of a camera lens provided with the image sensor and on a plane orthogonal to the plane, and the inspection object is illuminated from a plurality of angles, and the inspection object The image feature extraction method according to claim 1, wherein a feature line of the inspection object is extracted by imaging the image with the imaging element.   The plurality of illuminations are arranged such that a plurality of illuminations are arranged at equal angles on both sides of the camera lens obliquely facing the edge of the inspection object, and the inspection object is formed by the plurality of illuminations. 3. The image feature extraction method according to claim 2, wherein a bright and uniform image can be captured over a wide range of the imaging range of the image sensor.   The image picked up by the image pickup device is divided into a plurality of image ranges, and at that time, the image is divided into image ranges having relatively uniform brightness, and the characteristic line of the inspection object is extracted for each image range. Image feature extraction method.   The pixel row is defined in the image range, and the luminance of the pixel at the current position and the pixel at the next position is obtained for each pixel in a direction orthogonal to the pixel row in order from the pixel at the end of the image range. A pixel whose luminance is less than a predetermined luminance threshold and the luminance of the pixel at the next position is larger than the luminance threshold, and the luminance of the pixel at the current position larger than the luminance threshold is higher than the luminance threshold. Find a smaller pixel, move this process in the direction intersecting the direction of the pixel column by a predetermined pixel, find all the pixels that satisfy these conditions, and connect the obtained pixels to the line segment The image feature extraction method according to claim 1, wherein the feature extraction is performed using the pixel range as a boundary portion.   6. The image feature extraction method according to claim 5, wherein the luminance threshold value is set as a luminance threshold value by adding a predetermined luminance value to the minimum luminance value or by subtracting a predetermined luminance value from the maximum luminance value. .   When imaging the inspection object and extracting the feature line of the shape, if the pixel position of the feature line is deviated from the extending direction of the arrangement and there is a pixel of the next feature line, a shape abnormality is possible The image feature extraction method according to claim 5, wherein it is determined that the image has a characteristic.   The number of pixels in the interval between the boundary portion and the feature line is obtained, and the feature of the boundary line is extracted based on the change in the number of pixels, and the boundary portion at the pixel position where there is a possibility of the shape abnormality The image feature extraction method according to claim 5 or 6, wherein when the number of pixels at an interval from the feature line exceeds a predetermined threshold, it is determined as a shape abnormality portion.   A difference between the number of pixels between the boundary lines in the pixel range on the feature line determined to be the abnormal shape portion and an average value of the number of pixels between the boundary lines in the image range is obtained, and the difference is a predetermined value. 7. The image feature extraction method according to claim 5 or 6, wherein the shape abnormal portion is determined as a defective portion when the difference is larger.   A defect inspection method for inspecting a tool blade surface using the image feature extraction method according to claim 1. A tool defect inspection apparatus for performing a defect inspection on a tool blade surface using the image feature extraction method according to claim 10.