JP5441764B2 - Dot set detection apparatus and dot set detection method - Google Patents

Description

  The present invention relates to a dot set detection device and a dot set detection method for detecting a clustered dot set in a dot image, and in particular, a clustered dot in the dot image based on the dot density gradient of each unit region in the dot image. The present invention relates to a dot set detection apparatus and a dot set detection method for detecting a set.

  2. Description of the Related Art Conventionally, a method for identifying ring / bulk distribution defects from defect distribution images scattered on a semiconductor substrate is known (see, for example, Patent Document 1).

  This method derives a degree of coincidence between a binarized image obtained by binarizing the distribution image and each of a plurality of pattern images registered in advance to identify ring / bulk distribution defects, A defect included in the pattern portion of the pattern image having the highest degree of coincidence is detected as a ring / bulk distribution defect.

JP 2006-352173 A

  However, even if the method described in Patent Document 1 can detect a ring / bulk distribution defect that matches a pre-registered pattern, a ring / bulk distribution defect that does not match a pre-registered pattern, It is impossible to detect a mass distribution defect of a cluster-like aggregate (dot aggregate).

  In view of the above, the present invention is a dot set detection device and a dot set detection method for detecting a clustered dot set in a dot image, and a dot set detection device and a dot set detection capable of dealing with an irregular dot set. It aims to provide a method.

  In order to achieve the above object, a dot set detection apparatus according to an embodiment of the present invention is a dot set detection apparatus that detects a clustered dot set from a dot image, and each dot of a unit region in the dot image. A dot density acquisition unit for acquiring a density; a dot density gradient vector derivation unit for deriving a dot density gradient vector corresponding to each of the unit regions based on the dot density of each unit region in the dot image; and A central point temporary setting unit that temporarily sets a central point that can be a center, an association unit that associates each of the dot density gradient vectors with one of the central points, and the central point and each of the dot density gradient vectors A center point correction unit that corrects the coordinates of the center point based on the correspondence between the center point and the dot density Characterized in that it comprises a dot grouping display unit for displaying information about the dot set detected based on the correspondence relationship between each distribution vector.

  In addition, it is preferable that the dot set display unit displays a circle or an ellipse surrounding a plurality of unit regions corresponding to the dot density gradient vector belonging to the center point, with the coordinate of the center point being the center.

  The dot set display unit may display a circle or an ellipse in which a value obtained by dividing the number of dots included in the circle or the ellipse by the area of the circle or the ellipse is equal to or greater than a predetermined value. preferable.

  Further, the center point temporary setting unit is configured such that the distance between the center points is equal to or greater than a predetermined distance, the coordinates of the unit area where the dot density is equal to or greater than a predetermined value, the coordinates of the unit area selected at random, Preferably, a predetermined number of coordinates including at least one of the coordinates of the unit area selected according to the rule are temporarily set as the coordinates of the center point.

  In addition, the dot set display unit, when the distance between the coordinates before correction of all the center points corrected by the center point correction unit and the coordinates after correction is equal to or less than a predetermined distance, the dot set It is preferable to display the information regarding.

  A dot set detection method according to an embodiment of the present invention is a dot set detection method for detecting a clustered dot set from a dot image, and acquires dot density of each unit area in the dot image. A step, a dot density gradient vector derivation step for deriving a dot density gradient vector corresponding to each unit area based on the dot density of each unit area in the dot image, and a center point that can be the center of the dot set. A center point temporary setting step to be set; a first association step for associating each of the dot density gradient vectors with one of the temporarily set center points; and the center point temporarily set in the center point temporary setting step And the coordinates of the center point based on the correspondence between the dot density gradient vector A first center point correcting step to be corrected; a second associating step for associating each of the dot density gradient vectors with one of the corrected center points; and the center point associated in the second associating step A second center point correction step for correcting the coordinates of the center point based on the correspondence relationship between the dot density gradient vector and the dot density gradient vector, and before correction of all the center points corrected in the second center point correction step When the distance between the coordinates and the corrected coordinates is equal to or less than a predetermined distance, or when the number of corrections reaches a predetermined number, a plurality of unit regions corresponding to the dot density gradient vector belonging to the center point are Among the surrounding circles or ellipses, a dot set display step for displaying a circle or ellipse in which the value obtained by dividing the number of dots contained in the circle or ellipse by the area of the circle or ellipse is a predetermined value or more. Characterized in that it comprises a flop, a.

  By the means described above, the present invention is a dot set detection device and a dot set detection method for detecting a clustered dot set in a dot image, and a dot set detection device and a dot set detection method capable of dealing with an irregular dot set. Can be provided.

It is a block diagram which shows the structural example of the dot set detection apparatus which concerns on the Example of this invention. It is a figure which shows an example of a process target image (dot image). It is a figure which shows the state which divided | segmented the process target image of FIG. 2 for every unit area | region. It is a figure explaining the method for deriving a dot density gradient vector. It is a figure which shows the dot density gradient vector corresponding to each of the unit area | region in FIG. FIG. 6 is a diagram illustrating a state in which a dot density gradient vector having a magnitude equal to or less than a predetermined value is removed from the dot density gradient vector illustrated in FIG. 5. It is a figure which shows the temporarily set center point. FIG. 3 is a diagram (part 1) illustrating a state in which a dot set is represented by an ellipse. FIG. 6 is a diagram (part 2) illustrating a state in which a dot set is represented by an ellipse. It is a flowchart which shows the flow of a dot set detection process.

  Hereinafter, the best mode for carrying out the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram illustrating a configuration example of a dot set detection apparatus 100 according to an embodiment of the present invention. The dot set detection apparatus 100 is an apparatus for detecting a clustered dot set in a dot image, A control unit 1, an image acquisition unit 2, and an output unit 3 are included.

  The dot image is, for example, an image in which each of the unique portions distributed on the surface of the product having a flat surface is represented by dots (dots). The product having the flat surface includes, for example, a plurality of transport rollers. It is a long thing conveyed in the manufacturing process.

  In this case, each of the unique parts distributed on the surface of the product is a part having different characteristics (for example, the degree of brightness, roughness, unevenness, etc.) from the other parts of the product surface. When a part of the surface of the product being transported is not cleaned due to a malfunction of the nozzle for cleaning water that discharges cleaning water for cleaning the product, it is formed by a large number of dusts distributed on the surface of the product, In general, the density increases where the cause of the failure (in this case, the failure of the washing water nozzle) is located, and the density decreases as the distance from the location where the cause of the failure is located.

  The unique portion on the surface of the product is not limited to what is formed as described above, but may be a pattern that is intentionally formed.

  In addition, the dot image is not limited to an image in which each of the singular parts on the surface of the product as described above is represented by dots, and any image composed of a group of dots having a non-uniform distribution density may be used. Shall be included.

  The control unit 1 is a computer including a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and the like. For example, the processing target image generation unit 10, the dot density acquisition unit 11, the dot While storing programs corresponding to each of the density gradient vector deriving unit 12, the central point temporary setting unit 13, the associating unit 14, the central point correcting unit 15, the dot set display unit 16, and the blockiness evaluation unit 17, in the ROM, The CPU is caused to execute processing corresponding to each unit. Each unit may be configured by hardware such as an electronic circuit.

  The image acquisition unit 2 is a device that acquires an image to be processed by the control unit 1, for example, a camera including an imaging element such as a complementary metal oxide semiconductor (CMOS) or a charge coupled device (CCD), A part or all of the surface of the product conveyed by the conveying roller so as to pass in front of the camera is imaged intermittently or continuously, and an image of the imaged product surface (hereinafter referred to as “surface image”). ) Is output to the control unit 1.

  The output unit 3 is a device for outputting various types of information. For example, a display for displaying the detection result of the feature related to the distribution of the singular part on the product surface by the control unit 1 or the sound output of the detection result Speaker or the like.

  Next, each element of the control unit 1 will be described.

  The processing target image generation unit 10 generates a processing target image (dot image) including a pixel (hereinafter referred to as “single pixel”) representing a specific portion of the product surface from the surface image acquired by the image acquisition unit 2. For example, by performing various image processing on the surface image and extracting only singular pixels, a processing target image that is a target of processing to be described later for detecting a massive set formed by the singular pixels (Dot image) is generated.

  Specifically, the processing target image generation unit 10 performs a gray scale process based on the luminance on the surface image, and further performs a binarization process using a predetermined threshold. In this binarization process, for example, the value of a pixel having a luminance value representing the normal surface of the product is indicated by “0” (white), and the luminance value representing a specific portion of the surface of the product (for example, the The pixel value having a luminance value other than the luminance value representing the normal surface of the product is executed as indicated by “1” (black).

  Note that the size of the processing target image (dot image) may correspond to the entire surface image, or may correspond to a part of the surface image.

  Further, the size of the unit pixel in the processing target image (dot image) may be the same as the size of the unit pixel in the surface image, and a plurality of unit pixels (for example, 3 vertical pixels × 3 horizontal pixels) in the surface image. The size may be a set of 9 pixels).

  FIG. 2 is a diagram illustrating an example of a processing target image (dot image) generated by the processing target image generation unit 10, where the unit pixel size in the original surface image is the same as the unit pixel size in the original surface image (dot image). If any of the 9 pixels in the surface image is a unique pixel, the value of the unit pixel in the processing target image (dot image) is “1” (black). If all the nine pixels in the surface image are not singular pixels, the value of the unit pixel in the processing target image (dot image) is assumed to be “0” (white). The unit pixel having a pixel value “1” (black) in the processing target image (dot image) is referred to as a secondary singular pixel so that it can be distinguished from the singular pixel in the surface image. FIG. 2 includes three massive dot sets C1, C2, and C3 to be detected by the dot set detection apparatus 100.

  The dot density acquisition unit 11 is a means for acquiring the dot density in each of the unit areas of the processing target image (dot image). For example, the dot density acquisition unit 11 corresponds to nine unit pixels (that is, the unit pixel in the processing target image (dot image)). The dot density (for example, the number of secondary singular pixels included in the unit region) in each of the unit regions corresponding to 81 pixels of the surface image) is acquired.

  FIG. 3 shows a state where the processing target image (dot image) of FIG. 2 is divided for each unit area, and the numerical value given to the center of each unit area including the secondary singular pixel is the dot density (the unit pixel). The number of secondary singular pixels included in (1). Note that the display of the dot density value (zero) for each unit region not including the secondary singular pixel is omitted. 3 to 9, the X axis extends in the left-right direction, and the Y axis extends in the up-down direction.

  FIG. 3 shows, for example, three secondary singular pixels (unit pixels shown in black) in a unit region composed of nine unit pixels when the numerical value attached to the center of the unit region is “3”. ) Is included.

  The dot density gradient vector deriving unit 12 is based on the dot density corresponding to each of the unit regions in the processing target image (dot image) acquired by the dot density acquiring unit 11, and the dot density gradient vector corresponding to each of these unit regions. It is a means for deriving.

  The “dot density gradient vector” is a vector representing the magnitude of fluctuation (for example, increase or decrease) of the dot density and the direction of the fluctuation (increase or decrease).

  FIG. 4 is a diagram for explaining a method for deriving a dot density gradient vector. FIG. 4A shows one unit region of interest specified by hatching in FIG. 3 and eight units surrounding the unit region of interest. It is a figure which shows the calculation object area | region comprised by one unit area (unit area of the 1st vicinity), FIG.4 (B) is FIG.4 (B) of nine unit areas E1-E9 of 3 rows 3 columns in the calculation object area | region. Each dot density gradient value is shown. The calculation target area is a unit of interest greater than or equal to one interest unit area identified by hatching and a second neighborhood (24 unit areas) or third neighborhood (48 unit areas) surrounding the interest unit area. It may be composed of a region.

  Specifically, in FIG. 4B, the dot density gradient value of the unit region E1 is a value “−1” obtained by subtracting the dot density “1” of the unit region of interest E5 from the dot density “0” of the unit region E1. Similarly, the dot density gradient value of the unit region E6 is a value “+2” obtained by subtracting the dot density “1” of the unit region of interest E5 from the dot density “3” of the unit region E6.

  FIG. 4C is an operator (hereinafter referred to as “X operator”) for deriving the X component of the dot density gradient vector of the unit region of interest, and has values arranged in 3 rows and 3 columns. Then, each value in the rightmost column is “+1”, and each value in the leftmost column is “−1”. In this case, the dot density gradient in the right direction (when the dot density increases toward the right) is represented by “+”, and the dot density gradient in the left direction (when the dot density increases toward the left). Is represented by “−”.

  Similarly, FIG. 4D shows an operator (hereinafter referred to as “Y operator”) for deriving the Y component of the dot density gradient vector of the unit region of interest, as in the case of the X operator. It has values arranged in a column, the value in the top row is “+1”, and the value in the bottom row is “−1”. In this case, an upward dot density gradient (when the dot density increases upward) is represented by “+”, and a downward dot density gradient (when the dot density increases downward). Is represented by “−”.

  4E is based on the dot density gradient values of the nine unit areas E1 to E9 in the calculation target area shown in FIG. 4B and the X operator shown in FIG. 4C. The value group of 3 rows and 3 columns is derived from the above. Specifically, the value group of 3 rows and 3 columns is the value group of 3 rows and 3 columns of FIG. 4B and 3 rows of FIG. It is calculated by multiplying the values corresponding to the row position and the column position in the X operator of three columns. Then, the magnitude and direction of the X component of the dot density gradient vector of the unit region of interest are derived by summing the calculated 9 rows of 3 rows and 3 columns (in this case, “+1” + “0 “+” 0 ”+“ + 1 ”+“ 0 ”+“ + 2 ”+“ + 1 ”+“ 0 ”+“ + 1 ”becomes“ +6 ”).

  Similarly, FIG. 4F shows the dot density gradient values of the nine unit areas E1 to E9 in the calculation target area shown in FIG. 4B and the Y operator shown in FIG. A value group of 3 rows and 3 columns derived based on it, and the magnitude and direction of the Y component of the dot density gradient vector of the unit area of interest are derived by summing these 9 values groups of 3 rows and 3 columns. (In this case, “−1” + “− 1” + “0” + “0” + “0” + “0” + “+ 1” + “− 1” + “− 1” becomes “−3”. .)

  FIG. 4G is a diagram showing a dot density gradient vector V corresponding to the unit region of interest specified by hatching in FIG. 3, and the dot density gradient vector V is shown in FIG. 4E and FIG. As described with reference to F), it has a rightward “+” X component Vx of size “6” and a downward “−” Y component Vy of size “3”. In FIG. 4, the unit of the size of the dot density gradient vector V is shown as being irrelevant to the size of the unit pixel in the processing target image (dot image) (for example, a size of “6”). The X component Vx does not correspond to the size of 6 pixels in the processing target image (dot image)), and may correspond to the size of the unit pixel in the processing target image (dot image).

  FIG. 5 is a diagram showing dot density gradient vectors corresponding to each of the unit regions in FIG. 3, and each of the dot density gradient vectors is a method described by the dot density gradient vector deriving unit 12 with reference to FIG. It shall be derived using

  FIG. 6 is a diagram showing a state in which the dot density gradient vectors whose magnitudes are equal to or smaller than a predetermined value are removed from the dot density gradient vectors shown in FIG.

  In this way, the dot density gradient vector deriving unit 12 prevents the dot density gradient vector having a magnitude equal to or smaller than a predetermined value from affecting the processing described later (that is, a dot group that does not constitute a clustered dot set is The dot density gradient vector having a magnitude equal to or smaller than a predetermined value is removed so that the dot density gradient vector is not included as a constituent element of the clustered dot set.

  The center point temporary setting unit 13 is a means for temporarily setting a predetermined number of center points that can be the center of a clustered dot set. For example, a predetermined number of unit regions are selected in descending order of dot density, and the selected points are selected. The coordinates of the central pixel of the unit area are temporarily set as the coordinates of the central point that can be the center of the dot set.

  Specifically, the center point temporary setting unit 13 temporarily sets center points in descending order of dot density, and continues the temporary setting until the number of temporarily set center points reaches a predetermined number. The order in which the center points are temporarily set for a plurality of unit areas having the same dot density value may be random using random numbers, and a certain rule (for example, upper left corner to upper right corner, lower left corner) It is the order from the corner to the lower right corner.)

  When the distance between the two coordinates to be temporarily set is equal to or less than the predetermined distance, the center point temporary setting unit 13 can center only one of the two coordinates as the center of the dot set. Temporarily set as point coordinates. By temporarily setting a plurality of center points at close positions, it is possible to prevent a dot set that should be detected as a single clustered dot set from being detected as two or more dot sets. is there. Therefore, the predetermined distance is determined to be, for example, a size that is a predetermined multiple of the reference value when the maximum value or the average value of the sizes of each known dot set group to be detected is used as the reference value. A distance (for example, a predetermined multiple of the major axis) larger than the average dot group size (for example, the major axis in the case of an elliptical dot group) known from experience is selected. The The predetermined distance may be updated whenever the dot set detection apparatus 100 attempts to detect a dot set, or may be updated on a moving average basis every predetermined number of times of detection.

  In addition to the temporary setting of the central point based on the dot density as described above, the central point temporary setting unit 13 randomly or regularly covers a predetermined number of centers over the entire processing target image (dot image). Temporarily set a point. When only the center points temporarily set in descending order of the dot density are used, the dot set detection apparatus 100 also includes dots (single pixels) that should not belong to the dot set represented by these center points in the dot set. It is inevitable that it will be treated as belonging, and as a result, the dot set is detected after moving the coordinates of the center point of the dot set from the position where it should originally be (the temporarily set center point This is because the coordinates are finally determined after being displaced according to the information regarding the dot group recognized as belonging to the center point, as will be described later.

  On the other hand, when the center point temporarily set according to a random rule or according to a certain rule is additionally used, the center point temporary setting unit 13 creates a dot set represented by the center points temporarily set in descending order of dot density. A dot that should not belong (for example, a dot located far from the center point) is located near the center point when viewed from the dot, and is randomly or temporarily set according to a certain rule It is handled as belonging to the dot set represented by the center point, and the dots that should not belong to the dot set represented by the center point in descending order of the dot density are temporarily set in descending order of the dot density It is possible to prevent improper displacement of the coordinates of the center point.

  Note that the method of detecting a set of dots using only a randomly set center point temporarily sets the center point even in a place where the dot density is not large. In this case, the coordinates of these center points are greatly displaced, and it takes time to converge. Even when the same processing (processing for temporarily setting the center point at random) is applied to the same dot image, the coordinates of the temporarily set center point differ each time, so the final dot set The detection result also varies depending on the process, and the method of detecting the dot set using only the randomly set center point detects the dot set using the center point temporarily set based on the dot density. Compared with the method, the dot set detection result is less stable.

  In addition, a method of detecting a dot set using only center points temporarily set according to a certain rule may temporarily set center points on both sides of a place where the dot density is large, for example. In some cases, a dot set that should be detected as one block-like dot set is divided into two or more dot sets and detected.

  Therefore, the most stable method is to detect a set of dots by using a center point temporarily set in descending order of dot density with a certain interval and a center point temporarily set at random or according to a certain rule. It can be said that this is a method that can detect a dot set.

  Note that the center point group that is temporarily set at random includes, for example, a center point group that is arranged at coordinates on the processing target image (dot image) selected using a random number, and that is temporarily set according to a certain rule. The set center point group includes, for example, a center point group arranged at a certain interval in each of the vertical direction and the horizontal direction of the processing target image (dot image). In this way, the center point temporary setting unit 13 temporarily sets a predetermined number of center points randomly or according to a certain rule, in addition to the predetermined number of center points temporarily set based on the dot density.

  The number of center points set based on the dot density by the center point temporary setting unit 13, the number of center points set regardless of the dot density by the center point temporary setting unit 13, and the total number thereof are respectively It is arbitrarily determined according to the nature of the dot set that is the detection target. For example, the number of detected dot sets that are empirically known (detected by one processing target image (dot image)) A larger number (for example, a predetermined multiple of the detected number) is selected. The number may be updated whenever the dot set detection apparatus 100 attempts to detect a dot set, or may be updated on a moving average basis every predetermined number of detections.

  Further, the temporary center point setting unit 13 may temporarily set all of the predetermined number of central points randomly or according to a certain rule without temporarily setting the central point based on the dot density.

  FIG. 7 is a diagram showing a center point temporarily set by the center point temporary setting unit 13, NA1 is a center point temporarily set according to a random rule or according to a certain rule, and NB1 and NC1 are dot densities. This is the center point temporarily set based on this.

  The associating unit 14 is a means for associating each of the dot density gradient vectors derived by the dot density gradient vector deriving unit 12 with one of the central points temporarily set by the central point temporary setting unit 13. The center point closest to the starting point coordinates of the dot density gradient vector (the coordinates of the central pixel of the corresponding unit area) is associated as the center point to which the dot density gradient vector belongs. Preferably, each time the associating unit 14 calculates the distance between the dot density gradient vector and the center point, a predetermined weighting based on the magnitude and direction of the dot density gradient vector is applied to the calculated distance. To do. This is to make it easier for the center point in the direction indicated by the dot density gradient vector to be handled as being closer than the center point in the other direction. As a result, when there are a plurality of center points at the same distance from the dot density gradient vector, the associating unit 14 sets the dot density gradient to the center point that is closer to the direction indicated by the dot density gradient vector. It will be selected as the center point to which the vector belongs.

  The center point correction unit 15 is a means for correcting the coordinates of the center point based on the correspondence relationship between the dot density gradient vector and the center point associated by the association unit 14. The center point is displaced so that the total or average value of the distances between the starting point coordinates of the dot density gradient vector that currently belongs to one center point and the corrected center point is minimized.

  The center point correction unit 15 uses weights corresponding to at least one of the magnitude and direction of each dot density gradient vector instead of using the simple sum of the distances for calculating the total or average of the distances. You may make it calculate the sum total or average value of these distances, after performing with respect to distance.

  In this case, for example, the center point correction unit 15 increases the weight as the direction of the dot density gradient vector is closer to the direction of the center point. This is because the dot density gradient vector tends to face the center of the dot set to which it belongs.

  The control unit 1 corrects all the center points by the center point correcting unit 15 and then determines whether the displacement amount between the coordinates before correction and the coordinates after correction of each center point is equal to or less than a predetermined value. If the displacement amount of any one of the center points exceeds the predetermined value, the association unit 14 again executes association between each dot density gradient vector and one of the corrected center points. Then, the association by the association unit 14 and the correction by the center point correction unit 15 are repeated until the displacement amount of all the center points becomes equal to or less than the predetermined value.

  The dot set display unit 16 is a dot set represented by the dot density gradient vector and the center point associated with each other by the association unit 14, and the center point correction unit 15 displaces the center point to an appropriate coordinate. For example, the center and range of the dot set are displayed on the output unit 3 by a circle or an ellipse centered on the center point.

  Specifically, the dot set display unit 16 performs ellipse estimation of the dot set corresponding to each center point by principal component analysis, for example, and the average value and Y component of the X component of each start point coordinate of the dot density gradient vector And a major axis and a minor axis orthogonal to each other at the center point (the major axis and minor axis can form an angle with respect to the X axis or the Y axis), respectively. The center point of the ellipse having the minor axis and the major axis (meaning that about 95% of the dot group constituting the dot set is included in the ellipse region) and the boundary thereof, which are the length of the second standard deviation Highlight the line.

  FIG. 8 shows that after the dot set detection apparatus 100 displaces the center points NA1, NB1, and NC1 temporarily set in FIG. 7 to the center points NA2, NB2, and NC2 by the association unit 14 and the center point correction unit 15, It is a figure which shows the state which expressed the dot set C1, C2, C3 (refer FIG. 2) corresponding to each of these center points NA2, NB2, NC2 by the dot set display part 16 by the ellipse.

  Further, FIG. 9 shows that the dot set detection apparatus 100 temporarily sets the center points NB1 and NC1 based on only the dot density without temporarily setting the center point NA1 (see FIG. 7) by the center point temporary setting unit 13. After setting and displacing the center points NB1 and NC1 to the center points NB3 and NC3 by the associating unit 14 and the center point correcting unit 15, the dot set display unit 16 corresponds to each of the center points NB3 and NC3. It is a figure which shows the state which expressed set C4, C3 by the ellipse.

  Comparing FIG. 8 and FIG. 9, the center point NC2 in FIG. 8 and the center point NC3 in FIG. 9 have the same coordinates, and a dot set C3 represented by an elliptical region composed of the same dot density gradient vector group. Although the dot sets C1 and C2 represented by the two elliptical areas in FIG. 8 (the ellipse having the center point NA2 and the ellipse having the center point NB2) in FIG. It can be seen that this is expressed by a dot set C4 expressed by an ellipse having

  As a result, the coordinates of the center point NB3 in FIG. 9 are different from the coordinates of the center points NA2 and NB2 in FIG. 8, and the dot density gradient vector group belonging to the center point NA2 and the dots belonging to the center point NB2 in FIG. The density gradient vector group constitutes a dot density gradient vector group belonging to the center point NB3 in FIG.

  This is because a dot density gradient vector group (dot density gradient vector group that should belong to the center point NA2 in FIG. 8) that should not belong to the center point NB1 in FIG. 7 belongs to the center point NB3 in FIG. 9, the coordinates of the center point NB3 are located to the left of the coordinates of the center point NB2 in FIG. 8, so the dot set detection apparatus 100 inherently detects a dot set centered on the center point NB2. This results in an impossible result (a false detection is made as a dot set C4 centered on the center point NB3 in FIG. 9).

  In this way, in addition to the temporary setting of the center point based on the dot density, the temporary setting of the center point according to a random or fixed rule can be used to eliminate noise (specify a dot density gradient vector that should not belong to a specific center point). Operation that belongs to other center points other than the center point of the dot set), and the influence of noise in determining the center point and range of the dot set can be eliminated.

  The dot set display unit 16 may finally select a dot set to be displayed to the operator from the detected dot set. Specifically, the dot set display unit 16 displays a dot set in which a value (described later) representing the lumpness corresponding to each dot set is equal to or greater than a predetermined value, and a value (described later) indicating the lumpness is predetermined. You may make it exclude the dot set which becomes less than a value from a display target.

  The blockiness evaluation unit 17 is a dot set represented by the dot density gradient vector and the center point associated with each other by the association unit 14, and the center point is displaced to appropriate coordinates by the center point correction unit 15. For example, the number of dots included in the circle or ellipse generated by the dot set display unit 16, the number of singular pixels, the number of secondary singular pixels Alternatively, a value obtained by dividing the number of dot density gradient vectors having a predetermined size or more by the area of the circle or ellipse (hereinafter referred to as “bulk evaluation value”) is calculated.

  The blockiness evaluation value means that the larger the value (for example, the larger the number of unique pixels included in the ellipse), the higher the blockiness (blockiness) is.

  The dot set display unit 16 displays a circle or an ellipse in which the blockiness evaluation value calculated by the blockage evaluation unit 17 is equal to or greater than a predetermined value among the generated circles or ellipses, and the blockage evaluation value is a predetermined value. By excluding circles or ellipses that are less than the display target, it is possible to selectively display a set of dots that are highly clumped.

  Next, a flow of processing in which the dot set detection apparatus 100 detects a dot set (hereinafter referred to as “dot set detection processing”) will be described with reference to FIG. FIG. 10 is a flowchart showing the flow of the dot set detection process.

  First, the control unit 1 of the dot set detection apparatus 100 uses the processing target image generation unit 10 to generate a processing target image (dot image) based on the surface image captured by the image capturing unit 2 (step S1) (FIG. 1). 2).

  Thereafter, the control unit 1 uses the dot density acquisition unit 11 to acquire the dot density of each unit region (region composed of nine pixels of the surface image) in the processing target image (dot image) (step S2). (See FIG. 3).

  Thereafter, the control unit 1 uses the dot density gradient vector deriving unit 12 to derive the dot density gradient vector corresponding to each of the unit regions based on the dot density of the unit regions according to the above-described method (step S3) ( 4 and FIG. 5), the magnitude is set to be equal to or smaller than the predetermined value so as to exclude the dot density gradient vector whose magnitude is equal to or smaller than a predetermined value (for example, “1”) from the target of subsequent processing. The dot density gradient vector is deleted (see FIG. 6).

  Thereafter, the control unit 1 temporarily sets the coordinates of the central pixel of the unit region as a predetermined number of central points in descending order of the dot density by the temporary center point setting unit 13 (step S4) (center points NB1, NC1 in FIG. 7). reference.).

  Each of the center points is provisionally set so that the distance between the two center points is a predetermined distance or more. For example, in FIG. 3, unit areas having a dot density of “3” are adjacent to each other. In such a case, unless restrictions on the distance between two center points are added, This is because all the coordinates of the central pixel are temporarily set as the central point, and one dot set is detected as a plurality of dot sets.

  Specifically, when a central point is temporarily set in a unit area having a dot density of “3” and another central point is temporarily set in an adjacent unit area having a dot density of “3”, The center point temporary setting unit 13 omits the temporary setting of the new center point, assuming that the distance between the center point already temporarily set and the center point to be newly set is less than the predetermined distance, The process proceeds to a similar determination process for the next unit area in which a point is to be temporarily set.

  Thereafter, the control unit 1 further provisionally sets a predetermined number of coordinates selected at random or according to a certain rule as a center point by the center point temporary setting unit 13 (step S4) (center point in FIG. 7). See NA1.)

  Also in this case, as described above, each of the center points is provisionally set so that the distance between the two center points is equal to or greater than a predetermined distance, and the number of center points that satisfy the condition does not match the predetermined number. The center point temporary setting unit 13 adjusts the predetermined value (threshold regarding the magnitude of the dot density gradient vector) so that the number of the center points temporarily set matches the predetermined number.

  Thereafter, the control unit 1 is a dot density gradient vector for each unit region derived by the dot density gradient vector deriving unit 12 by the associating unit 14, and each dot density gradient vector having a magnitude equal to or larger than a predetermined value is centered. Corresponding to any one of the center points temporarily set by the temporary point setting unit 13 (step S5).

  This association is performed by, for example, performing a predetermined weighting based on the size and direction of the dot density gradient vector for each distance from the starting point coordinates of the dot density gradient vector to each center point. This is realized by executing the process of assigning the density gradient vector to the center point that yields the minimum value of the divided distances for all dot density gradient vectors.

  Thereafter, the control unit 1 uses the center point correction unit 15 to temporarily set the center temporarily set by the center point temporary setting unit 13 based on the correspondence relationship between the dot density gradient vector and the center point associated by the association unit 14. The coordinates of the points NA1, NB1, and NC1 are corrected (step S6), and the center points NA1, NB1, and NC1 are displaced to the corrected coordinates NA2, NB2, and NC2 (see FIG. 8).

  In this correction, for example, the total or average value of the distances between the start point coordinates of each dot density gradient vector that currently belongs to one center point to be corrected and the new center point after correction (after displacement) is minimized. This is realized by determining the coordinates of the new center point after correction.

  Thereafter, the control unit 1 determines whether or not the displacement amount between the coordinates before correction and the coordinates after correction of all the center points temporarily set by the center point temporary setting unit 13 is equal to or less than a threshold value TH. (Step S7).

  When even one displacement amount of the center point temporarily set by the center point temporary setting unit 13 exceeds the threshold value TH (NO in step S7), the control unit 1 performs the association (step S5) and the center by the association unit 14 The correction by the point correction unit 15 (step S6) is executed again, and the processing from step S5 to step S7 is repeated until the displacement amount of all the center points is equal to or less than the threshold value TH.

  When the displacement amount of all the center points set by the center point temporary setting unit 13 is equal to or less than the threshold value TH (YES in step S7), the control unit 1 centers the center point after correction by the block property evaluation unit 17. The block property evaluation value of each dot set is calculated (step S8).

  Thereafter, the control unit 1 causes the dot set display unit 16 to display each dot set centered on the corrected center point on the output unit 3 (step S9) (see FIG. 8).

  This display is center points NA1, NB1, NC1 temporarily set by the center point temporary setting unit 13 using principal component analysis, for example, and the center points NA2, NB2, This is realized by generating an ellipse centered on NC 2 and displaying the ellipse on the output unit 3.

  At this time, the control unit 1 selectively excludes a dot set having a strong lump by excluding an ellipse representing a dot set whose lump evaluation value calculated by the lump evaluation unit 17 is less than a predetermined value from the display target. You may make it display.

  Alternatively, when the number of repetitions of the processing of step S5 to step S7 reaches a predetermined number, the control unit 1 uses the block property evaluation unit 17 to evaluate each block property evaluation value of the dot set centered on the center point at that time. Then, the dot set display unit 16 may display the dot set centered on the center point at that time on the output unit 3.

  Similarly, at this time, the control unit 1 excludes an ellipse representing a set of dots whose blockiness evaluation value calculated by the blockage evaluation unit 17 is less than a predetermined value from the display target, whereby a dot set having a strong blockiness is obtained. You may make it display selectively.

  In this way, the dot set detection apparatus 100 can detect a clustered dot set in the processing target image (dot image) and present the characteristics (center coordinates and range) of the dot set to the operator in an easily understandable manner. it can.

  If the image to be processed (dot image) is based on the surface image of a long product that is transported in the manufacturing process using a plurality of transport rollers, the operator can place the image on the surface of the product. Installation position (for example, corresponding to the X coordinate of the center point) of the cause of formation of the distributed singular part (for example, the nozzle for the washing water that has failed), formation start time of the singular part (for example, within the elliptical range) Corresponding to the Y coordinate of the point located at the upper end in the Y-axis direction.), The formation end time of the singular part (for example, corresponding to the Y coordinate of the point located at the lower end in the Y-axis direction within the ellipse range), or The duration of the singular part (for example, corresponding to the width between the upper end and the lower end in the Y-axis direction of the elliptical range) can be estimated. As a result, the operator can identify the cause of formation and formation timing of these unique portions at an early stage, improve the product yield, and minimize the production of defective products.

  In addition, if the processing target image (dot image) is based on the surface image of the product on which the pattern is formed, the operator intended the pattern based on the detection result of the dot set. It can be determined whether or not it has been formed.

  Further, since the dot set detection apparatus 100 detects a dot set using a dot density gradient vector representing the magnitude and direction of dot density variation, the average dot density (processing target image (processing target image (dot image) of the processing target image (dot image)) is detected. Regardless of whether the distribution of singular pixels (dots) in the dot image) is generally dense or coarse, the dot set can be detected accurately.

  In addition, the dot set detection apparatus 100 includes a condition (for example, the size of a unit area) when the processing target image generation unit 10 generates a processing target image (dot image), and the center point temporary setting unit 13. Are temporarily set according to the conditions for temporarily setting the center point (for example, the minimum distance between two temporarily set center points, the minimum dot density of the unit area in which the center points are temporarily set, random or a fixed rule) The number of center points to be determined, the number of center points temporarily set based on the dot density, and the total number of center points to be temporarily set). For example, it can be arbitrarily set according to the property of the surface image), so that the dot set can be accurately detected regardless of the difference in the property of the basic image.

  Although the preferred embodiments of the present invention have been described in detail above, the present invention is not limited to the above-described embodiments, and various modifications and substitutions can be made to the above-described embodiments without departing from the scope of the present invention. Can be added.

  For example, in the above-described embodiment, the dot set detection apparatus 100 displays a dot set corresponding to each of the center points temporarily set by the center point temporary setting unit 13 by the dot set display unit 16. May be combined and displayed as one dot set.

  Specifically, the dot set display unit 16 is configured when at least one of the center point, shape, arrangement, and overlapping area of two ellipses representing the two dot sets satisfies a predetermined condition (for example, the overlapping area). When these two dot sets are combined and the coordinates of the new center point (for example, the midpoint between the two center points) are determined, An ellipse representing one dot set may be displayed.

DESCRIPTION OF SYMBOLS 1 Control part 2 Image acquisition part 3 Output part 10 Process target image generation part 11 Dot density acquisition part 12 Dot density gradient vector derivation part 13 Center point temporary setting part 14 Correlation part 15 Center point correction part 16 Dot set display part 17 Lump shape Evaluation unit 100 dot set detection device

Claims (6)

A dot set detection device for detecting a clustered dot set from a dot image,
A dot density acquisition unit that acquires the dot density of each unit region in the dot image;
A dot density gradient vector deriving unit for deriving a dot density gradient vector corresponding to each of the unit regions based on the dot density of each of the unit regions in the dot image;
A central point temporary setting unit that temporarily sets a central point that can be the center of the dot set;
An association unit that associates each of the dot density gradient vectors with one of the center points;
A center point correction unit that corrects the coordinates of the center point based on the correspondence between the center point and each of the dot density gradient vectors;
A dot set display unit that displays information about the dot set detected based on the correspondence between the center point and each of the dot density gradient vectors;
A dot set detection apparatus comprising: The dot set display unit displays a circle or an ellipse surrounding a plurality of unit regions corresponding to a dot density gradient vector belonging to the center point, with the coordinates of the center point being the center.
The dot set detection apparatus according to claim 1. The dot set display unit displays a circle or ellipse in which a value obtained by dividing the number of dots included in the circle or ellipse by the area of the circle or ellipse is a predetermined value or more among the circle or ellipse.
The dot set detection apparatus according to claim 2. The center point temporary setting unit is configured such that the distance between the center points is equal to or greater than a predetermined distance, the coordinates of the unit area where the dot density is equal to or greater than a predetermined value, the coordinates of the unit area selected at random, and a predetermined rule. A predetermined number of coordinates including at least one of the coordinates of the selected unit region are temporarily set as the coordinates of the center point;
The dot set detection apparatus according to any one of claims 1 to 3, wherein The dot set display unit is information regarding the dot set when the distance between the coordinates before correction and the coordinates after correction of all the center points corrected by the center point correction unit is equal to or less than a predetermined distance. Display,
The dot set detection device according to any one of claims 1 to 4, wherein A dot set detection method for detecting a clustered dot set from a dot image,
A dot density acquisition step of acquiring the dot density of each of the unit areas in the dot image;
A dot density gradient vector derivation step for deriving a dot density gradient vector corresponding to each of the unit areas based on the dot density of each of the unit areas in the dot image;
A temporary center point setting step for temporarily setting a central point that can be the center of the dot set;
A first associating step of associating each of the dot density gradient vectors with one of the temporarily set center points;
A first center point correction step of correcting the coordinates of the center point based on the correspondence between the center point temporarily set in the center point temporary setting step and the dot density gradient vector;
A second associating step of associating each of the dot density gradient vectors with one of the corrected center points;
A second center point correction step of correcting the coordinates of the center point based on the correspondence between the center point and the dot density gradient vector associated in the second association step;
When the distance between the coordinates before correction and the coordinates after correction of all the center points corrected in the second center point correction step is less than a predetermined distance, or when the number of corrections reaches a predetermined number Further, among circles or ellipses surrounding a plurality of unit regions corresponding to the dot density gradient vector belonging to the center point, a value obtained by dividing the number of dots contained in the circle or ellipse by the area of the circle or ellipse is a predetermined value or more A dot set display step for displaying a circle or an ellipse,
A dot set detection method comprising:

Images