JP2008175549A - Defect detection device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a defect detection device and its method capable of detecting a linear defect on a test object with high accuracy. <P>SOLUTION: A step S301 calculates an additional value of the pixel value of each pixel arranged in a predetermined direction of an image taken of a test object. A step S304 calculates a standard deviation value as an index value that indicates uniformity in the pixel values of the respective pixels arranged in the predetermined direction of the image. A step S308 corrects the additional value based on the standard deviation value. A step S309 determines presence or absence of a linear defect on the test object based on a result of comparing the corrected additional value with a threshold value. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a defect detection apparatus and a defect detection method for detecting defects existing on the surface of an inspection object such as a liquid crystal glass substrate.

  2. Description of the Related Art Conventionally, a defect detection apparatus (defect inspection apparatus) that images an inspection object such as a liquid crystal glass substrate with a camera and inspects the presence or absence of a defect on the inspection object using image processing is known. For example, in Patent Document 1, in order to detect a defect in more detail, the inspection object is detected by detecting a light / dark defect on the inspection object based on image data obtained by imaging the inspection object and differentiating the image data. It is disclosed that edge information and minute defects are detected, and integrated image differentiation processing is performed to detect low-contrast light and dark defects to obtain comprehensive information on the defects.

The application of the resist to the liquid crystal glass substrate is performed by discharging a resist liquid onto the substrate from a plurality of nozzles while moving the substrate in one direction. At this time, a linear defect (linear irregularity, streak-like defect) that extends linearly along the moving direction of the substrate may occur due to a defective discharge nozzle or the like. There is known a detection method that uses such an efficient projection calculation that copes with such a linear defect and has a small amount of calculation (see, for example, Patent Document 2).
JP 2000-36044 A JP 2001-101388 A

  However, the conventional method for detecting a linear defect has the following problems. FIG. 13 shows an image obtained by imaging the liquid crystal glass substrate. In the image 1300 of the liquid crystal glass substrate, a linear defect 1301 and a point defect 1302 are generated. The line defect 1301 is generally a portion having higher or lower luminance than the background in the image and low contrast. The point defect 1302 is caused by dust adhering to the surface of the substrate or the like, and has higher or lower luminance than the background in the image.

  At each pixel position in the X direction of the image 1300, an addition value of pixel values of pixels arranged in a line in the Y direction is calculated, and projection data 1303 having the addition value as an element value is created. In the distribution of the projection data 1303, the horizontal axis indicates the pixel position in the X direction in FIG. 13, and the vertical axis indicates the added value of the pixel values.

  In the projection data 1303, peaks 1304 and 1305 are generated corresponding to the pixel positions where the linear defect 1301 and the point defect 1302 exist. By detecting this peak, the defect can be detected. However, since the shapes of the peaks 1304 and 1305 are very similar, if only the peak is detected, whether the peak is due to a linear defect or not. It cannot be determined whether it is due to a defect other than. Therefore, the above method has a problem that it is difficult to detect a linear defect with high accuracy.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to provide a defect detection apparatus and a defect detection method that can detect a linear defect on an inspection object with high accuracy.

  The present invention has been made to solve the above-described problem, and an addition value calculation unit that calculates an addition value of pixel values of pixels arranged in a predetermined direction of an image obtained by imaging an inspection object; Index value calculation means for calculating an index value indicating the uniformity of pixel values of the pixels arranged in the predetermined direction, addition value correction means for correcting the addition value based on the index value, and the corrected value A defect detection apparatus comprising: a determination unit that determines the presence or absence of a linear defect on the inspection object based on a result of comparing the added value with a threshold value.

  The present invention also includes a step of calculating an addition value of pixel values of pixels arranged in a predetermined direction of an image obtained by imaging an inspection object, and a uniform pixel value of pixels arranged in the predetermined direction of the image. A linear value on the inspection object based on a step of calculating an index value indicating the property, a step of correcting the added value based on the index value, and a comparison between the corrected added value and a threshold value And a step for determining the presence or absence of a defect.

  When the direction of the linear defect is substantially along the predetermined direction, if there is a linear defect in each pixel arranged in the predetermined direction, the uniformity of the pixel values of those pixels is relatively high. Further, if a defect (such as a point defect) other than a linear defect exists in each pixel arranged in the predetermined direction, the uniformity of the pixel values of these pixels is relatively low. Therefore, it is possible to reduce the influence of defects other than linear defects on the added value by correcting the added value based on the index value indicating the uniformity of the pixel value of each pixel arranged in a predetermined direction of the image. It becomes.

  According to the present invention, it is possible to increase the linear defect on the inspection object by determining the presence or absence of the linear defect based on the result of comparing the threshold value with the added value in which the influence of the defect other than the linear defect is reduced. The effect that it can detect with precision is acquired.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 shows the configuration of a defect inspection apparatus according to an embodiment of the present invention. The defect inspection apparatus according to the present embodiment includes a transport device 1, an illumination device 2, a line sensor camera 3, an image processing device 4 (corresponding to the defect detection device of the present invention), a transport control device 5, and a display 6. The inspection object 7 to be inspected by this defect inspection apparatus is a liquid crystal glass substrate.

  The transport apparatus 1 transports the inspection object 7 and moves it in a certain direction (A direction in FIG. 1). The transfer device 1 is, for example, an air levitation stage, and can fix the four corners of the inspection object 7 and blow air onto the inspection object 7 to float the inspection object 7 and move it in a certain direction. It is. The illumination device 2 illuminates the inspection object 7. The line sensor camera 3 is composed of, for example, 1024 light receiving elements arranged in a direction perpendicular to the moving direction of the inspection object 7. The line sensor camera 3 takes in reflected light or scattered light from the inspection object 7 to obtain the surface of the inspection object 7. And image signals are generated.

  The image processing device 4 generates two-dimensional image data from the imaging signal input from the line sensor camera 3 and performs defect detection processing for detecting defects on the inspection object 7 by image processing. In the defect detection process of the present embodiment, linear defects are detected, but in addition to this, image data is divided into predetermined ranges and integrated within each division, and then the difference between the integrated values of each division is calculated. Differentiation may be performed to detect low-contrast light and dark defects, or edges and minute defects may be detected by differential filter processing as microfilter processing.

The conveyance control device 5 controls the conveyance of the inspection object 7 by the conveyance device 1. The display 6 displays the result of the defect detection process by the image processing apparatus 4 and the image of the inspection object 7. In this embodiment, the image processing device 4 and the conveyance control device 5 (or including the display 6) are separate devices, but they are provided in the control device 8. The control device 8 has a function of controlling the angles of the illumination device 2 and the line sensor camera 3 (angles θ ₁ and θ _{2 in} FIG. 1).

  FIG. 2 shows a functional configuration of the image processing apparatus 4. The image processing apparatus 4 includes an image input unit 41, an image processing unit 42, a storage unit 43, and a data output unit 44. The image input unit 41 A / D-converts an image signal input for each line from the line sensor camera 3 and integrates data for a plurality of lines to generate two-dimensional image data. The image processing unit 42 detects a defect by a defect detection process using the generated image data.

  The storage unit 43 stores image data of the inspection object 7 generated by the image input unit 41 and data indicating the result of the defect detection processing by the image processing unit 42. The data output unit 44 outputs data indicating the result of the defect detection process and image data of the inspection object 7 to the display 6 in FIG.

  Next, a procedure of defect detection processing according to the present embodiment (operation procedure of the image processing unit 42) will be described. FIG. 3 shows this procedure. First, a projection calculation is performed using the input image data to generate projection data (step S301) (corresponding to the function of the added value calculation means of the present invention). The projection operation is an operation for adding the pixel values of all the pixels arranged in the same direction to the image data.

  As shown in FIG. 4, it is assumed that an image 400 formed by image data includes m rows and n columns of pixels, and a linear defect occurs along the Y direction from the upper end to the lower end of the image 400. In the projection calculation, the element values (addition values) of the projection data are calculated by adding the pixel values of all the pixels having the same pixel position in the X direction and arranged in a line in the Y direction.

That is, if the pixel value of the pixel at the position of the i-th row and the j-th column is A (i, j) and the element values of the projection data are B (1), B (2),. The i-th element value B (i) of the projection data is calculated by the following equation (1). Thus, by converting the two-dimensional image data into the one-dimensional projection data, it is possible to expect the processing speed to be increased.
B (i) = A (1, i) + A (2, i) +... + A (m, i) (1)

  Following the projection operation, shading correction processing is performed. FIG. 5 shows how the projection data changes due to the shading correction process. It is assumed that a linear defect 501 and a point defect 502 are generated in an image 500 obtained by imaging the inspection object. The projection data described above includes the influence of shading caused by the influence of illumination unevenness and the like. For example, in the projection data 510 shown in FIG. 5, peaks 513 and 514 due to the influence of shading are generated in addition to the peaks 511 and 512 caused by the linear defect 501 and the point defect 502.

  In the process of detecting a linear defect, each element value of the projection data is compared with a predetermined threshold value, and when the element value exceeds the threshold value, it is determined that a linear defect exists. For this reason, when the projection data 510 is subjected to a linear defect detection process by comparison with the threshold 515, in addition to the peak 511 due to the linear defect 501, a peak 513 due to the influence of shading is not detected. As a result, a problem that is erroneously detected occurs.

  Therefore, a smoothing process is performed on the projection data (step S302). In the smoothing process, for example, an average value of values of elements (for example, elements from i−Nth to i + Nth) within a predetermined range centering on a target element (for example, i-th element) of the projection data is used as the attention value. The element value. Thus, it becomes possible to create shading correction data 520 in which the peak portion in the projection data 510 of FIG.

  After performing the smoothing process, the difference between the projection data and the shading correction data is calculated to generate projection data in which the influence of shading is reduced (step S303). In the projection data 530 obtained by calculating the difference between the projection data 510 and the shading correction data 520 in FIG. It becomes.

  However, on the projection data 530, in addition to the peak 531 caused by the linear defect 501, a peak 532 caused by the point defect 502 is generated. Therefore, if the element value of the projection data 530 is compared with the threshold value, the peak 532 is also obtained. , It is erroneously detected as a peak due to a linear defect. Therefore, in the present embodiment, after the processes in steps S301 to S303, or in parallel with these processes, the processes in steps S304 to S306 are performed, and a standard deviation value that is a kind of index value indicating the uniformity of pixel values. The standard deviation data constituted by is generated, and false detection is reduced as will be described later.

  In step S304, the element values of the standard deviation data are calculated by calculating the standard deviation values of the pixel values of all the pixels having the same X-direction pixel position in the image and arranged in a line in the Y direction (this book). Corresponding to the function of the index value calculation means of the invention). That is, if the element values of the standard deviation data are C (1), C (2),..., C (n) under the same conditions as in FIG. 4, the i-th element value C (i ) Is calculated by the following equation (2). However, μ is an average value of A (1, i), A (2, i),..., A (m, i).

  If linear defects exist on the pixels arranged in a line in the Y direction in FIG. 4, the pixel value variation (standard deviation value) of those pixels is relatively low (that is, the uniformity is relatively high). . In addition, if point defects exist on the pixels arranged in a line in the Y direction, the variation (standard deviation value) of the pixel values of these pixels is relatively high (that is, the uniformity is relatively low). Therefore, by correcting the projection data based on the standard deviation data, it becomes possible to reduce the influence of point defects on the projection data.

  Following the generation of the standard deviation data, the standard deviation data is subjected to a smoothing process to generate shading correction data (step S305). Furthermore, by calculating the difference between the standard deviation data and the shading correction data, standard deviation data with reduced shading effects is generated (step S306). Using this standard deviation data, the projection data is corrected (referred to as weighting in this embodiment).

  FIG. 6 shows how the projection data is corrected. It is assumed that linear defects 601 and 602 and point defects 603 and 604 are generated in an image 600 obtained by imaging the inspection object. In the projection data 610, peaks 611, 612, 613, and 614 are generated due to these defects. Here, although the luminance of the linear defects 601 and 602 is slightly different from the peripheral luminance, the values of the peaks 611 and 612 in the projection data 610 are the peaks 613 and 613 caused by the point defects 603 and 604, respectively, by adding the pixel values. The value is almost equal to the value of 614.

  In the standard deviation data 620, peaks 621 and 622 corresponding to point defects 603 and 604 are generated. Therefore, it is possible to determine that a point defect has occurred at the pixel position in the X direction corresponding to the peak of the standard deviation data 620. In this embodiment, the projection data is weighted so as to reduce the size of the element value of the projection data corresponding to the peak of the standard deviation data.

  First, coefficient data composed of coefficient values used for weighting is generated from standard deviation data (step S307). Since the weighting is performed by multiplying the element value of the projection data by the coefficient value, the larger the standard deviation value, the smaller the coefficient value, and the smaller the standard deviation value, the larger the coefficient value. Since the standard deviation value at the pixel position where the point defect occurs is larger than the standard deviation value corresponding to the pixel position where there is no defect (or where a linear defect occurs), the point defect occurs. The coefficient value of the pixel position becomes smaller.

For example, the coefficient value is 1 at the pixel position where the standard deviation value is the minimum, and the coefficient value is 0 or a minute value smaller than 1 at the pixel position where the standard deviation value is the maximum (the element value of the projection data corresponding to the point defect). The coefficient value is determined so as to be a value such as 0.1 or 0.01 so as to be small. As a specific example, the following binarization process is performed on the standard deviation data. However, the element values of the coefficient data are D (1), D (2),..., D (n). D1 is a threshold value (for example, the threshold value 623 in FIG. 6).
If C (i) ≧ d1, 0 ≦ D (i) <1 (however, the element value of the projection data corresponding to the point defect is sufficiently small)
If C (i) <d2, D (i) = 1

Following the generation of the coefficient data, the projection data is weighted using the coefficient data (step S308) (corresponding to the function of the addition value correcting means of the present invention). An equation for weighting the i-th element value B (i) of the projection data and calculating the element value B ′ (i) after weighting is the following equation (3).
B ′ (i) = B (i) × D (i) (3)

  Projection data 630 in FIG. 6 is projection data after weighting. The peaks 633 and 634 due to the point defects 603 and 604 are smaller in size than the peaks 613 and 614 before weighting. Therefore, in the projection data 630, the influence of the point defects 603 and 604 is reduced. By comparing the element value of the projection data 630 with, for example, threshold values 635 and 636, only the peaks 631 and 632 due to the linear defects 601 and 602 can be detected.

  Following the weighting of the projection data, a defect detection process is performed by comparing the element value of the projection data with a threshold value (step S309) (corresponding to the function of the determination means of the present invention). If the threshold values are d2, d2 ′ (d2 <d2 ′), when the i-th element value B ′ (i) of the projection data calculated by the equation (3) is d2 or less or d2 ′ or more, It is determined that a linear defect exists at the i-th pixel position with respect to the pixel position in the X direction in FIG. If B ′ (i) is within the range from d2 to d2 ′, it is determined that there is no linear defect at the i-th pixel position.

The above thresholds d2 and d2 ′ are, for example, the following ( It is calculated by the formulas (4) and (5). However, K is an arbitrary positive number, for example, K = 3.
d2 = μ−σ × K (4)
d2 ′ = μ + σ × K (5)

  Finally, the result of the defect detection process is stored in the storage unit 43 (step S310). In particular, when it is determined that a linear defect exists, data including the coordinate value of the linear defect is stored in the storage unit 43. If necessary, the result of the defect detection process is output to the display 6 of FIG. 1 via the data output unit 44, and the result is displayed.

  According to the defect detection process described above, the linear defect on the inspection object is determined by determining the presence or absence of the linear defect based on the result of comparing the threshold value with the element value of the projection data in which the influence of the point defect is reduced. Can be detected with high accuracy.

  Next, a first modification of the present embodiment will be described. When a point defect is superimposed on a linear defect, a peak occurs in the distribution of projection data, and the standard deviation value corresponding to the pixel position where the peak occurs also increases. There is a possibility that a linear defect in which point defects are superimposed cannot be detected well. On the other hand, in the first modification, the image is divided into a plurality of regions so that a plurality of regions are arranged along the direction of the linear defect, and the defect detection result in each region is comprehensively determined to determine the line. The defect is detected with high accuracy.

  Hereinafter, the procedure of the defect detection process (the procedure of the operation of the image processing unit 42) according to the first modification will be described. FIG. 7 shows this procedure. First, the input image data is divided into partial image data corresponding to each area in the image formed by the image data (step S701) (corresponding to the function of the area dividing means of the present invention).

  For example, the image data is divided into three partial image data. FIG. 8 shows a state of this division. By dividing the image data into three partial image data, an image 800 formed by the entire image data is divided into, for example, three regions 801, 802, and 803. . The size of each area may not be equal, and the number of divisions is not limited to three.

  Following the division of the area, projection data is generated for each area (step S702). Subsequently, the shading correction data is generated by performing a smoothing process on the projection data (step S703), and the difference between the projection data and the shading correction data is calculated to obtain the projection data in which the influence of shading is reduced. Generate (step S704). The processing in steps S703 and S704 is also performed on the projection data of each region.

  In addition, after the processes in steps S702 to S704 or in parallel with these processes, the processes in steps S705 to S707 are performed to generate standard deviation data for each region. In step S705, the element value of the standard deviation data is calculated from the pixel value of the image data corresponding to each region.

  Following the generation of the standard deviation data, the standard deviation data is subjected to a smoothing process to generate shading correction data (step S706). Furthermore, by calculating the difference between the standard deviation data and the shading correction data, standard deviation data with reduced shading effects is generated (step S707). The processing in steps S706 and S707 is also performed on the standard deviation data of each region.

  Subsequently, coefficient data is generated from the standard deviation data (step S708). Further, the projection data is weighted using the coefficient data (step S709). Following this weighting, a defect detection process is performed by comparing the element value of the projection data with a threshold value (step S710). The processing in steps S708 to S710 is also performed on the data in each area.

  Subsequently, the determination result of the presence or absence of the linear defect obtained in the defect detection process of step S710 is integrated, and an integration process for comprehensively determining the presence or absence of the linear defect is performed (step S711). Finally, the result of the defect detection process is stored in the storage unit 43 (step S712).

  Hereinafter, the contents of the integration process in step S711 will be described. Projection data 810, 820, and 830 in FIG. 8 are projection data weighted in step S709. The projection data 810 is generated from the pixel values in the area 801, the projection data 820 is generated from the pixel values in the area 802, and the projection data 830 is generated from the pixel values in the area 803.

  In the image 800 of FIG. 8, linear defects 804 and 805 and point defects 806 and 807 are generated. Each projection data distribution has peaks due to these defects. Peaks 812, 822, and 832 are due to linear defects 805. Peak 813 is due to point defect 807. Peaks 821 and 831 are due to the linear defect 804. Further, the peak 811 is caused by the linear defect 804 and the point defect 806, but the influence of the point defect 806 is reduced by weighting, and thus the peak 811 is much smaller than the peaks 821 and 823.

  When the element values of the projection data 810, 820, and 830 are compared with the threshold values 814, 823, and 833 and the presence / absence of a linear defect is determined based on the comparison result, the determination result of each region regarding the pixel position P in the X direction in FIG. In any of the regions 801, 802, and 803, there is a linear defect. On the other hand, the determination result of each region regarding the pixel position Q in the X direction shows that there is a linear defect in the regions 802 and 803, but no linear defect in the region 801.

  In the integration process, the determination results of the respective areas are integrated to generate an overall determination result. For example, when the determination result that there is a line defect is associated with logic 1 and the determination result that there is no line defect is associated with logic 0, the entire determination result regarding the pixel position P is the individual of the regions 801, 802, and 803. By performing a logical operation (OR operation) on the determination result in (with any line defect), the line defect exists. Similarly, the overall determination result regarding the pixel position Q is also a logical operation (OR operation) of the determination results in each of the regions 801, 802, and 803 (two regions have linear defects and one region has no linear defects). By doing so, there is a linear defect.

  As described above, according to the first modification, the result of comparison between the element value of the projection data and the threshold value performed for each region is combined to determine the presence or absence of the linear defect, thereby allowing the linear defect to be detected. Even when the point defect is superimposed, the linear defect on the inspection object can be detected with high accuracy.

  Next, a second modification of the present embodiment will be described. In the second modification, an index value indicating the uniformity of pixel values is calculated using a density occurrence matrix used in the statistical texture analysis method. The density occurrence matrix depends on the probability that the density of the pixel S (coordinates (x2, y2)) that is separated from the pixel R (coordinates (x1, y1)) of the density i shown in FIG. 9 by the distance r in the direction of the angle θ is j. The elements of are determined.

  For example, as shown in FIG. 10A, it is assumed that the density (pixel value) of an image composed of 4 × 4 16 pixels is binarized. When r = 1 and θ = 90 ° (270 °), the element value of the density occurrence matrix shown in FIG. 10B is determined by the relationship between pixels adjacent vertically in the image. Hereinafter, a method for generating a density occurrence matrix will be described with the pixel in the m-th row and the n-th column in FIG.

  First, each element value of the concentration occurrence matrix is initialized (set to 0). Subsequently, the upper left pixel (1, 1) (density 0) is the target pixel, and i = 0 and j = 1 of the density occurrence matrix from the relationship with the lower pixel (2, 1) (density 1). Add 1 to the element value. Subsequently, the pixel (1, 2) (density 0) is set as the target pixel, and the element value of i = 0, j = 1 of the density occurrence matrix from the relationship with the pixel (2, 2) (density 1) below it. Add 1 to.

  Subsequently, the pixel (1, 3) (density 0) is set as the target pixel, and the element value of i = 0 and j = 0 in the density occurrence matrix from the relationship with the lower pixel (2, 3) (density 0). Add 1 to. Subsequently, the pixel (1, 4) (density 1) is set as the target pixel, and the element value of i = 1 and j = 1 in the density occurrence matrix from the relationship with the lower pixel (2, 4) (density 1). Add 1 to.

  Subsequently, the pixel (2, 1) (density 1) is set as the target pixel, and the element value of i = 1 and j = 1 in the density occurrence matrix from the relationship with the lower pixel (3, 1) (density 1). Add 1 to. Further, for the pixel (2, 1) (density 1), the element value of i = 1 and j = 0 of the density occurrence matrix is further set to 1 because of the relationship with the pixel (1, 1) (density 0) above it. Is added.

  Similarly, 1 is added to the element value of the density occurrence matrix while moving the target pixel from left to right and from top to bottom. When pixel (4, 1) (density 0) is the target pixel, 1 is set to the element value of i = 0, j = 1 of the density occurrence matrix from the relationship with pixel (3, 1) (density 1) above it. Is added. Similarly, when the pixel of interest is each of the pixel (4, 2), the pixel (4, 3), and the pixel (4, 4), 1 is added to the element value of the density occurrence matrix.

  As a result of the above processing, the density occurrence matrix shown in FIG. 10B is generated. In the above description, the density (pixel value) of each pixel is binary, but it may be 16 or 256, and the order of the density occurrence matrix may be determined in accordance with the density gradation.

  An index value indicating the uniformity of the pixel value is calculated from the density occurrence matrix described above. The density occurrence matrix is P (r, θ, i, j), and for example, the feature quantities shown in the following formulas (6) to (8) can be used as index values.

  Further, in the density occurrence matrix, the element value of the matrix is determined by the combination of the density (pixel value) of adjacent pixels. Therefore, when there is a linear defect with low contrast, the density occurrence matrix near the diagonal element of the density occurrence matrix. The value increases. On the other hand, when there is a point defect or the like, a large value is generated in the element values other than the vicinity of the diagonal element. From this, the sum total of the values of the diagonal elements of the density occurrence matrix may be calculated and used as an index value indicating the uniformity of the pixel value.

  Hereinafter, a procedure of defect detection processing according to the second modification (operation procedure of the image processing unit 42) will be described. FIG. 11 shows this procedure. Since the processing of steps S1101 to S1103 related to the projection data is the same as the processing of steps S301 to S303 in FIG. 3, for example, description thereof is omitted.

  The processes in steps S1104 to S1106 are processes related to the concentration occurrence matrix. In step S1104, a density occurrence matrix is generated from the pixel values of all the pixels arranged in a line in the Y direction of the image. More specifically, a density occurrence matrix is generated at each pixel position in the X direction of the image, and density occurrence matrix data having the density occurrence matrix at each pixel position as an element is generated.

  Subsequently, a feature amount is calculated from the generated density occurrence matrix, and feature amount data having the feature amount as an element value is generated (step S1105). Further, the coefficient value is calculated by converting the feature value to any value from 0 to 1, and coefficient data having the coefficient value as an element value is generated (step S1106). The coefficient value is closer to 1 as the uniformity is higher, and closer to 0 as the uniformity is lower.

  Subsequently, the coefficient data is used to weight the projection data (step S1107). Following this weighting, a defect detection process is performed by comparing the element value of the projection data with a threshold value (step S1108). Finally, the result of the defect detection process is stored in the storage unit 43 (step S1109).

  FIG. 12 shows an example of defect detection processing according to the second modification. In the image 1200, a linear defect 1201 and a point defect 1202 are generated. At each pixel position in the X direction, the element value of the projection data and the element value of the density occurrence matrix are calculated from the pixel values of the h pixels arranged in the Y direction. Further, a feature value is calculated from the density occurrence matrix, and a coefficient value is calculated from the feature value.

  As indicated by the coefficient data 1203, the coefficient value corresponding to a region with high uniformity between pixels in the Y direction (including a region where the linear defect 1201 is generated) is a substantially constant value. On the other hand, the coefficient value corresponding to the region where the uniformity of pixels in the Y direction is reduced due to the presence of the point defect 1202 is small. When the projection data is weighted using the coefficient data 1203, the size of the element value of the projection data corresponding to the area where the point defect 1202 is generated becomes small. In the determination process of the presence or absence of the linear defect, the linear defect is not erroneously detected in the region where the point defect 1202 is generated.

  As described above, according to the second modified example, similarly to the case where the standard deviation value is used as the index value indicating the uniformity of the pixel value, the element value and the threshold value of the projection data in which the influence of the point defect is reduced. By determining the presence or absence of a linear defect based on the result of comparing the above, the linear defect on the inspection object can be detected with high accuracy. Also in the second modified example, as in the first modified example, an image obtained by imaging the inspection target may be divided into a plurality of regions, and the above processing may be performed for each region.

  As described above, the embodiments of the present invention have been described in detail with reference to the drawings. However, the specific configuration is not limited to the above-described embodiments, and includes design changes and the like without departing from the gist of the present invention. .

  For example, the following may be used. The image data includes noise due to the influence of the sensitivity unevenness of the sensor used for the line sensor camera. Since the contrast of the linear defect is low, if the image data contains noise, it may cause a false detection. Therefore, smoothing filter processing for the purpose of noise removal may be performed on the image data before the projection data is generated. As a result, stable defect detection processing can be performed.

  In addition, the defect detection process as a whole is performed by an adjacent comparison method for detecting a conventional two-dimensional defect, a method for extracting a defect by binarizing a difference between an image of an inspection object and a reference image, or the like. The defect detection process according to the present invention may be performed before or after the process. Furthermore, the conventional defect detection process may be performed in parallel with the defect detection process according to the present invention.

It is a block diagram which shows the structure of the defect inspection apparatus by one Embodiment of this invention. It is a block diagram which shows the structure of the image processing apparatus with which the defect inspection apparatus by one Embodiment of this invention is provided. It is explanatory drawing for demonstrating the procedure of the defect detection process in one Embodiment of this invention. It is explanatory drawing for demonstrating the production | generation method of projection data and standard deviation data in one Embodiment of this invention. It is a reference figure which shows the mode of the change of the projection data by the shading correction process in one Embodiment of this invention. It is a reference figure which shows the mode of correction | amendment with respect to projection data in one Embodiment of this invention. It is explanatory drawing for demonstrating the procedure of the defect detection process in one Embodiment (1st modification) of this invention. It is explanatory drawing for demonstrating the content of the process which integrates the detection result of the linear defect in one Embodiment (1st modification) of this invention. It is explanatory drawing for demonstrating the density occurrence matrix in one Embodiment (3rd modification) of this invention. It is explanatory drawing for demonstrating the production | generation method of the density occurrence matrix in one Embodiment (3rd modification) of this invention. It is explanatory drawing for demonstrating the procedure of the defect detection process in one Embodiment (3rd modification) of this invention. It is a reference figure which shows an example of the defect detection process in one Embodiment (3rd modification) of this invention. It is explanatory drawing for demonstrating the detection method of a linear defect.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Conveyance apparatus, 2 ... Illumination apparatus, 3 ... Line sensor camera, 4 ... Image processing apparatus, 5 ... Conveyance control apparatus, 6 ... Display, 7 ... Inspection object 8 ... control device 41 ... image input unit 42 ... image processing unit 43 ... storage unit 44 ... data output unit

Claims (5)

An added value calculating means for calculating an added value of pixel values of pixels arranged in a predetermined direction of an image obtained by imaging an inspection object;
Index value calculation means for calculating an index value indicating the uniformity of the pixel value of each pixel arranged in the predetermined direction of the image;
Addition value correction means for correcting the addition value based on the index value;
Based on the result of comparing the added value after correction and a threshold value, determination means for determining the presence or absence of a linear defect on the inspection object;
A defect detection apparatus comprising: Further comprising region dividing means for dividing the image into a plurality of regions so that at least two regions are arranged in the predetermined direction;
2. The determination unit according to claim 1, wherein the determination unit determines the presence / absence of the linear defect based on a result of comparing the added value after correction in the individual regions arranged in the predetermined direction with the threshold value. The defect detection apparatus described.   3. The defect detection according to claim 1, wherein the index value calculation unit calculates a standard deviation value of pixel values of pixels arranged in the predetermined direction of the image as the index value. apparatus.   The index value calculation unit calculates the index value using a density occurrence matrix based on pixel values of pixels arranged in the predetermined direction of the image. Defect detection device. Calculating an addition value of pixel values of pixels arranged in a predetermined direction of an image obtained by imaging an inspection object;
Calculating an index value indicating the uniformity of the pixel value of each pixel arranged in the predetermined direction of the image;
Correcting the added value based on the index value;
Determining the presence or absence of a linear defect on the inspection object based on a result of comparing the added value after correction and a threshold;
A defect detection method comprising:

Images