JP2008304312A - Image processing program and image processor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To detect a stripe-like flaw in a waveform noise regardless of the thickness of a stripe at a high speed with high precision in performing the evaluation of a flaw. <P>SOLUTION: An image processing program is characterized in that the profile calculation step (steps S1-S2) for obtaining the profile in the imaging of an object and the low-frequency removing step (step S3) for removing a low-frequency component from the profile according to a noise removing width parameter are executed by a computer. An image processor is equipped with a profile calculation part for obtaining the profile in the imaging of the object and a low frequency removing part for removing the low-frequency component from the profile obtained in the profile calculation part according to the noise removing width parameter. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an image processing program and an image processing apparatus.

  In a conventional inspection method for inspecting a streak defect when an image such as a toner image is formed on a predetermined medium, a filter is generally applied to emphasize a defective portion from an image of an object. In order to accurately extract a streak defect using a normal filter, it is necessary to apply a filter suitable for the width of the streak.

  However, it is not possible to know the width of the stripe generated in the object before the examination. Therefore, sufficient performance could not be obtained. In order to solve this, Patent Document 1 is proposed. In this method, the line width is dealt with by using a first-order differential filter.

JP 2006-189293 A

  The present invention is intended to detect a streak defect in waveform noise at a high speed and with high accuracy regardless of the thickness of the streak defect when the object is imaged to remove noise and evaluate the streak defect. It is.

  The invention according to claim 1 of the present application causes a computer to execute a profile calculation step for obtaining a profile in imaging of an object and a low frequency removal step for removing low frequency components from the profile according to a noise removal width parameter. Is an image processing program.

  The invention according to claim 2 of the present application is a profile calculation step for obtaining a profile in imaging of an object, a low frequency removal step for removing low frequency components from the profile according to a noise removal width parameter, and a defect score calculation width parameter. And a defect score calculating step of calculating a defect score profile according to the image processing program.

  Further, the invention according to claim 3 of the present application is a profile calculation step for obtaining a profile in imaging of an object, a low frequency removal step for removing low frequency components from the profile according to a noise removal width parameter, and a defect score calculation width parameter. A defect score calculation step for calculating a defect score profile in accordance with the defect score detection step and a defect position detection step for determining a defect position from the defect score profile are executed by a computer.

  Further, the invention according to claim 4 of the present application is a profile calculation step for obtaining a profile in imaging of an object, a low frequency removal step for removing low frequency components from the profile according to a noise removal width parameter, and a defect score calculation width parameter. A defect score calculating step for calculating a defect score profile according to the defect score, a defect position detecting step for determining a defect position from the defect score profile, and an evaluation value calculating step for calculating an evaluation value of the defect at the detected defect position This is an image processing program that is executed by.

  The invention according to claim 5 of the present application is characterized in that, in the low frequency removal step, a low frequency component is removed by subtracting a neighborhood average in a noise removal width parameter from a signal value of a profile in imaging of the object. The image processing program according to any one of claims 1 to 4.

  Further, in the invention according to claim 6 of the present application, in the defect score calculation step, the average of the neighborhood in the defect score calculation width parameter is subtracted from the signal value of the target position of the profile after removal of the low frequency component, and the standard deviation is obtained. The image processing program according to any one of claims 1 to 4, wherein the defect score is calculated by division.

  The invention according to claim 7 of the present application is the image processing program according to claim 3 or 4, wherein the defect position is determined by threshold processing of the defect score profile in the defect position detection step.

  The invention according to claim 8 of the present application is characterized in that, in the defect score calculation step, the defect score width parameter is determined according to the noise removal width parameter. The image processing program described in the above.

  The invention according to claim 9 is characterized in that, in the defect score calculation step, the defect score width parameter obtained by multiplying the noise removal width parameter by a predetermined coefficient is used. The image processing program according to any one of the above.

  In the invention according to claim 10 of the present application, in the defect score calculation step, a value equal to or smaller than the noise removal width parameter is used as the defect score width parameter. The image processing program described in the above.

  The invention according to claim 11 is characterized in that, in the defect position detecting step, the defect position is detected from a plurality of defect scores by a combination of the plurality of noise removal width parameters and the defect score width parameters. The image processing program according to claim 3 or 4.

  The invention according to claim 12 of the present application is the image processing program according to claim 4, wherein, in the evaluation value calculating step, an evaluation value is calculated from a defect score or a low frequency removal profile at the defect position. .

  The invention according to claim 13 of the present application is a profile calculation unit that obtains a profile in imaging of an object, and a low frequency removal unit that removes low frequency components from the profile obtained by the profile calculation unit according to a noise removal width parameter. And an image processing apparatus.

  The invention according to claim 14 is a profile calculation unit that obtains a profile in imaging of an object, and a low frequency removal unit that removes a low frequency component from the profile obtained by the profile calculation unit according to a noise removal width parameter. And a defect score calculation unit that calculates a defect score profile corresponding to a defect score calculation width parameter for the profile from which the low frequency component has been removed by the low frequency removal unit.

  Further, the invention according to claim 15 of the present application is a profile calculation unit that obtains a profile in imaging of an object, and a low frequency removal unit that removes low frequency components from the profile obtained by the profile calculation unit according to a noise removal width parameter. A defect score calculation unit that calculates a defect score profile corresponding to a defect score calculation width parameter for a profile from which a low frequency component has been removed by the low frequency removal unit, and a defect from the defect score profile calculated by the defect score calculation unit An image processing apparatus comprising: a defect position detection unit that determines a position.

  Further, the invention according to claim 16 of the present application is a profile calculation unit that obtains a profile in imaging an object, and a low frequency removal unit that removes low frequencies from the profile obtained by the profile calculation unit according to a noise removal width parameter; A defect score calculation unit that calculates a defect score profile corresponding to a defect score calculation width parameter for a profile from which a low frequency component has been removed by the low frequency removal unit; and a defect position from the defect score calculated by the defect score calculation unit An image processing apparatus comprising: a defect position detection unit to be determined; and an evaluation value calculation unit that calculates an evaluation value of a defect at a defect position detected by the defect position detection unit.

  According to the first aspect of the present invention, the object can be imaged to remove noise and emphasize the streak defect. When evaluating the defect, the streak defect in the waveform noise is Regardless, it is possible to detect at high speed and with high accuracy.

  According to the invention according to claim 2 of the present application, it is possible to remove noise by imaging an object, and to emphasize the streak defect by the defect score profile. When evaluating the defect, the streak defect in the noise is detected. It becomes possible to detect at high speed and with high accuracy regardless of the thickness of the streak.

  According to the invention according to claim 3 of the present application, it is possible to remove noise by imaging an object, and to emphasize the streak defect by the defect score profile. When detecting the position of the defect, the streak defect in the noise Can be detected at high speed and with high accuracy regardless of the thickness of the line.

  According to the invention of claim 4 of the present application, noise can be removed by imaging an object, and the streak defect can be emphasized by the defect score profile. When calculating the evaluation value of the defect, the streak in the noise The defect can be calculated at high speed and with high accuracy regardless of the thickness of the line.

  According to the invention of claim 5 of the present application, it is possible to remove the noise and emphasize the streak defect by imaging the object and subtracting the neighborhood average in the noise removal width parameter from the signal value of the profile. In performing the evaluation, it becomes possible to detect a streak defect in noise at high speed and with high accuracy regardless of the thickness of the streak.

  According to the invention of claim 6 of the present application, noise is removed by subtracting the neighborhood average in the defect score calculation width parameter from the signal value of the target position of the profile after imaging the object and removing the low frequency component of the profile. The streak defect can be emphasized, and when calculating the defect score, the streak defect in the noise can be calculated at high speed and with high accuracy regardless of the thickness of the streak.

  According to the seventh aspect of the present invention, the defect position can be detected at high speed and with high accuracy by the threshold processing of the defect score profile.

  According to the invention of claim 8 of the present application, the defect score width parameter is determined according to the noise removal width parameter, so that the position of the defect can be detected at high speed and with high accuracy.

  According to the invention of claim 9 of the present application, it is possible to detect the position of the defect at high speed and with high accuracy by using the noise removal parameter obtained by multiplying the noise removal parameter by a predetermined coefficient.

  According to the invention of claim 10 of the present application, the position of the defect can be detected at high speed and with high accuracy by using a value equal to or smaller than the noise removal parameter as the noise removal width parameter.

  According to the invention of claim 11 of the present application, the defect position is detected at high speed and with high accuracy by detecting the defect position from a plurality of defect scores based on a combination of the plurality of noise removal width parameters and the defect score width parameter. Can be detected.

  According to the invention of claim 12 of the present application, it is possible to evaluate the defect at high speed and with high accuracy by calculating the evaluation value from the defect score at the defect position or the low frequency removal profile.

  According to the invention of claim 13 of the present application, the object can be imaged to remove noise and emphasize the streak defect. When evaluating the defect, the streak defect in the noise is reduced to the thickness of the streak. Regardless of this, it becomes possible to detect at high speed and with high accuracy.

  According to the invention of claim 14 of the present application, the object is imaged to remove noise, and the streak defect can be emphasized by the defect score profile. When the defect is evaluated, the streak defect in the noise is detected. It becomes possible to detect at high speed and with high accuracy regardless of the thickness of the streak.

  According to the invention of claim 15 of the present application, it is possible to remove noise by imaging an object, highlight a streak defect by a defect score profile, and detect a streak defect in noise when detecting the position of the defect. Can be detected at high speed and with high accuracy regardless of the thickness of the line.

  According to the invention according to claim 16 of the present application, it is possible to remove noise by imaging an object, emphasize a streak defect by a defect score profile, and calculate a streak in noise when calculating an evaluation value of a defect. The defect can be calculated at high speed and with high accuracy regardless of the thickness of the line.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<Description of Image Processing Program According to this Embodiment>
FIG. 1 is a flowchart for explaining the flow of an image processing program according to this embodiment. This image processing program is executed by a computer (personal computer, workstation, etc.), or executed by the image processing apparatus of this embodiment described later, and is stored in a predetermined recording medium or via a predetermined network. Delivered. In addition, each step of the image processing program may be executed by the same computer or by different computers via a network. Further, information input / output of the image processing program of the present embodiment may be performed on the client side, and arithmetic processing may be performed on the server side by a client / server system.

  First, image data for capturing an object is acquired. Here, the object is a printed matter of a uniform image of a halftone level printed on a sheet. For example, in a printed matter formed by xerography, if a defect such as dust is present in an image forming unit such as a photosensitive drum, it will appear as a streak defect along the paper feeding direction. By detecting the position and size of the streak defect, it is possible to obtain information for determining that there is some trouble in the image forming unit, for example.

In order to acquire image data in imaging of an object, an image of the object is captured by a scanner or the like, and color conversion processing is performed (step S1). The color conversion is, for example, dRGB, L ^* a ^* b ^* .

Next, for example, a profile of a predetermined signal is calculated from the image information after the color conversion process (step S2). In this embodiment, for example, an L ^* profile is calculated. The profile is obtained by adding and averaging signal values of image data along a direction orthogonal to the paper feed direction along the paper feed direction. When the photosensitive drum of an image forming apparatus (for example, a copying machine) on which an object is printed corresponds to the four colors CMYK, a brightness profile for each color may be obtained.

  Next, low frequency removal processing is performed from the calculated profile (step S3). The low frequency removal process uses a value obtained by subtracting the moving average of the profile from the calculated profile. The moving average is obtained by using a preset moving average width parameter as a noise removal width parameter.

  Subsequently, Z value calculation is performed about the program which performed the low frequency removal process (step S4). In this embodiment, the Z value is obtained by (profile value−neighbor average value) / standard deviation, and is used as a defect score profile in this embodiment. Here, as a neighborhood average value, a neighborhood width parameter is used as a defect score calculation width parameter.

  Next, threshold processing is performed on the obtained Z value (step S5), and defect position extraction and defect evaluation are performed.

<Configuration of Image Processing Device According to this Embodiment>
FIG. 2 is a block diagram illustrating the configuration of the image processing apparatus according to the present embodiment. An image processing apparatus 10 according to the present embodiment includes a CPU (Central Processing Unit) 11, a RAM (Random Access Memory) 12, a color conversion unit 101, a profile calculation unit 102, a low frequency removal unit 103, a defect score calculation unit 104, a defect A position detection unit 105 and an evaluation value calculation unit 106 are provided.

  In the image processing apparatus 10 of the present embodiment, the color conversion unit 101, the profile calculation unit 102, the low frequency removal unit 103, the defect score calculation unit 104, the defect position detection unit 105, and the evaluation value calculation unit 106 are configured by program processing. It is realized by a computer or a dedicated device.

  The image processing apparatus 10 acquires image data to be inspected from the image input unit 20 and performs defect inspection by a process described later.

  The image input unit 20 includes an optical image input device such as a line sensor such as a flatbed scanner, an area sensor such as a CCD (Charge Coupled Devices) camera, or a storage device that stores image data.

  The CPU 11 is a central processing unit that controls each unit. The RAM 12 is composed of a nonvolatile memory such as a hard disk or a flash memory, and has a buffer for storing image data to be processed and a work area required during data processing.

  The image input unit 20 and the RAM 12 are connected to a CPU 11 that controls processing via a network or a bus.

The image data input by the image input unit 20 is converted into a color independent of the input device by the color conversion unit 101. If the input is an RGB signal, it is generally converted to a visual lightness by a LUT (Look Up Table). The image data after color conversion is sent to the profile calculation unit 102. In the present embodiment, as an example, RGB is converted to L ^* a ^* b ^* .

  The profile calculation unit 102 obtains an average brightness variation in the short side direction of the streak defect by averaging or integrating the image data in the long side direction of the streak defect. That is, in the image data obtained by reading the print on the paper, streak defects are generated along the scanning direction (for example, the paper transport direction) during image formation, but the image data along the direction orthogonal to the scanning direction. A value at each position (position) is added for each position along the scanning direction, and an average is calculated to obtain a profile.

  The average or direction of integration can be easily determined when the direction of the streak defect is known in advance. If it is not known in advance, the processing may be applied in two vertical and horizontal directions of image data, or in four directions including ± 45 degrees.

  The low frequency removing unit 103 performs an operation for removing the low frequency component from the profile calculated by the profile calculating unit 102 according to the noise removal width parameter. The noise removal width parameter is mainly a moving average width parameter. That is, the moving average corresponding to each position is obtained from the profile, and the moving average value is subtracted from the profile value to remove the low frequency component. Here, the profile after removing the low frequency component is referred to as a “noise removal profile”.

  The defect score calculation unit 104 performs a process of calculating a defect score profile from the noise removal profile according to the defect score calculation width parameter. The defect score profile is mainly a Z value corresponding to each position of the profile. The Z value can be calculated by (noise removal profile value−neighbor average value) / standard deviation. The size of the neighborhood width for obtaining the average value of the neighborhood becomes the defect score calculation width parameter. The defect score calculation width parameter is determined according to the noise removal width parameter.

  The defect position detection unit 105 detects a defect position from the defect score profile calculated by the defect score calculation unit 104 or the process of detecting the defect position from the profile (noise removal profile) from which the low frequency component is removed by the low frequency removal unit 103. Or a process of detecting a defect position from the defect score profile calculated by the noise removal profile and defect score calculation unit 104. Which process is detected depends on the setting by the user, for example.

The evaluation value calculation unit 106 calculates the evaluation value of the defect at the defect position detected by the defect position detection unit 105. The evaluation value is a value for finally determining a defect from the detected defect position. For example, a brightness difference (ΔL ^* ) from the vicinity is used. The final defect is determined based on this evaluation value.

<Defect detection method to which this embodiment is applied>
In the present embodiment, as an example, a streak defect inspection in the case where the long side of a sample output from xerography is in the paper feed direction will be described. When intermediate density is printed over a wide range by xerography, streaks occur in the paper feed direction of the sample or in the vertical direction. In particular, streaks in the paper feed direction are caused by dust or electrical troubles, and often occur independently and have no periodicity, and processing based on periodicity such as FFT (Fast Fourier Transform) is not suitable. The location and width of defects cannot be predicted, but the direction is fixed. However, since tilt may occur during image input, it is desirable to perform tilt detection / correction as preprocessing. Processing for removing density outliers (for example, 3σ of the density distribution) in the paper feeding direction may be performed.

  An example of a xerographic brightness profile is shown in FIG. In the sample having the profile of FIG. 3A, white lines (in which the brightness increases) can be visually confirmed near 140 mm, and in FIG. 3B, near 260 mm.

  Further, when the brightness change toward the image edge occurs as shown in FIG. 4, the image edge is erroneously detected as a line. In order to prevent this, it is possible to perform processing such as calculation assuming that the image edge n-point average brightness (eg, n = 3) continues outside the image.

  Next, the low frequency removing unit 103 removes the low frequency from the profile. That is, a moving average with the noise removal width parameter is calculated and subtracted from the profile. It is preferable to prepare a plurality of noise removal widths (in pairs with a defect score calculation width parameter described later) in order to correspond to the width of the parameter streak defect.

  As an example, parameters with 30 stages of moving average width such as 4 mm, 8 mm,..., 120 mm are prepared and calculated. Since data is converted from two-dimensional image data to a one-dimensional profile, it can be processed at high speed even if calculation is performed for each parameter. In the calculation of the moving average, the point of interest may be included or excluded as a neighborhood. The calculated noise removal profile is sent to the defect score calculation unit 104.

  The defect score calculation unit 104 obtains the Z value of the point of interest in the vicinity of the defect score calculation width parameter and sets it as a defect score profile. The Z value is a value obtained by subtracting the neighborhood average from the point of interest and dividing by the standard deviation.

  Here, the defect score calculation width parameter is set according to the noise removal width parameter, so that the defect score calculation suitable for the noise removal strength is performed. For example, as the defect score calculation width parameter, a value obtained by multiplying the noise removal width parameter by a predetermined number or a value equal to or smaller than the noise removal width parameter is used.

  As a preferred example, (defect score calculation width) = (3/5) × (noise removal width). As described above, when the noise removal width is set in 30 stages of 4 mm, 8 mm,..., 120 mm, the defect score calculation width is 2.4 mm, 4.8 mm,. When this embodiment is applied to xerography with each parameter setting, a guideline of the line width at which the defect score becomes large (emphasized) is shown in FIG. In this way, the noise removal width parameter and the defect score calculation width parameter are prepared as table data in accordance with the standard of the width of the line defect to be detected, and the noise removal width parameter and the defect score calculation width are determined depending on the width of the line defect to be detected. Determine a set of parameters.

  The calculation of the Z value may include or exclude the point of interest when calculating the average and standard deviation. Further, when the average value of the image edge is derived, a process such as calculation that the image edge n-point average brightness (eg, n = 3) continues outside the image may be performed. As the standard deviation at the edge of the image, the closest point among points that can be calculated in the defect score calculation width parameter is used, or the average standard deviation of the entire image is used.

  FIG. 6 is a diagram for explaining an actual data processing example. 6A shows a profile of a target including a plurality of streak defects, FIG. 6B-1 shows a profile obtained by removing a low frequency component with a noise removal width of 4 mm in the low frequency removing unit 103, and FIG. ) Is a profile obtained by removing low frequency components with a noise removal width of 44 mm in the low frequency removing unit 103.

  6B-2 is obtained as a result of calculating the Z value with the defect score calculation width parameter 2.4 mm in the defect score calculation unit 104 from the low frequency removal profile shown in FIG. 6B-1. The defect score profile, FIG. 6 (c-2), is obtained as a result of calculating the Z value with the defect score calculation width parameter 26.4 mm in the defect score calculation unit 104 from the low frequency removal profile shown in FIG. 6 (c-1). The defect score profile obtained.

  The streak defects with different widths are emphasized according to each parameter. That is, in the profiles shown in FIGS. 6 (b-1) and 6 (b-2), more low frequency components are removed from the original profile (see FIG. 6 (a)), and the overall swell is reduced. Of the protruding waveform portions indicated by arrows A to D, the waveform portion indicated by arrow A is particularly emphasized, but the waveform portion indicated by arrow D is compressed.

  On the other hand, in the profiles shown in FIGS. 6 (c-1) and 6 (c-2), although the removal of low frequency components is not much from the original profile (see FIG. 6 (a)) compared to the previous example, Of the protruding waveform portions indicated by arrows A to D, the waveform portion indicated by arrow D is particularly emphasized, while the waveform portion indicated by arrow A is compressed.

  As described above, the strength of removing low-frequency components and the strength of emphasizing a specific waveform vary depending on each parameter. Therefore, in the present embodiment, for example, 30 defect score profiles are calculated using 30 steps of parameters, and are described below. It is used as a material for determining the defect position in the defect position detection unit 105 shown.

  The defect position detection unit 105 uses the defect score profile to obtain position information of a place where there is a possibility of defect, and sends this to the evaluation value calculation unit 106. For example, when 30 parameters are set, 30 defect score profiles are obtained. First, the value with the maximum absolute value of the Z value at each point (position) in each of the 30 defect score profiles is stored, and the defect score calculation width parameter at that time is recorded. In other words, 30 defect score profiles are referred to by staking the positions together, and the Z value when the absolute value of the Z value at each position is maximized (the same position among the 30 defect score profiles). The Z value when the absolute value of the 30 Z values is maximum) and the defect score calculation width parameter when the maximum Z value is reached are recorded.

  As a result, a collection of Z values having the maximum absolute value at each position in the 30 defect score profiles is obtained, and the Z value at each position has a predetermined positive threshold value from the collection of Z values having the maximum absolute value. The position information of a point exceeding (for example, 3) or falling below a predetermined negative threshold (for example, −3) and the defect score calculation width parameter at the Z value are sent to the evaluation value calculating unit 106.

  The evaluation value calculation unit 106 calculates an evaluation value from the noise removal profile near the position information sent from the defect position detection unit 105. At this time, the noise removal profile is based on the noise removal width parameter corresponding to the defect score calculation width parameter obtained from the defect position detection unit 105.

For example, if the defect score width parameter at the Z value determined as a threshold at a certain position is 2.4 mm, the noise removal width parameter corresponding to the defect score width parameter (2.4 mm) is 4 mm, and the noise removal width The value of L ^* at the corresponding position of the noise removal profile where the parameter is 4 mm is set as the evaluation value.

Since the value of the noise removal profile is a brightness difference (for example, ΔL ^* ) from the vicinity, it can be used as an evaluation value. If the evaluation value (the value of the noise removal profile) exceeds a predetermined positive threshold (for example, 1.2) or falls below a predetermined negative threshold (for example, -1.5), the final value Judged to be defective.

  In addition, the numerical value shown in the defect inspection method in the said embodiment is an example, and this invention is not limited to this. Further, the present invention can be applied to defect inspections on the surfaces of various materials and members other than xerographic defect inspections.

It is a flowchart explaining the flow of the image processing program which concerns on this embodiment. It is a block diagram explaining the structure of the image processing apparatus which concerns on this embodiment. It is a figure which shows the brightness profile example of evaluation object. It is a figure which shows the example of the brightness profile in which the false detection of an image edge generate | occur | produces. It is a figure which shows the standard of a parameter and a detection defect. It is a figure which shows the example of a defect detection of the brightness profile of evaluation object.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Image processing apparatus, 11 ... CPU, 12 ... RAM, 20 ... Image input part, 101 ... Color conversion part, 102 ... Profile calculation part, 103 ... Low frequency removal part, 104 ... Defect score calculation part, 105 ... Defect position Detection unit, 106 ... evaluation value calculation unit

Claims (16)

A profile calculation step for obtaining a profile in imaging of an object;
An image processing program that causes a computer to execute a low frequency removal step of removing low frequency components from the profile according to a noise removal width parameter. A profile calculation step for obtaining a profile in imaging of an object;
A low frequency removal step of removing low frequency components from the profile according to a noise removal width parameter;
An image processing program that causes a computer to execute a defect score calculation step of calculating a defect score profile according to a defect score calculation width parameter. A profile calculation step for obtaining a profile in imaging of an object;
A low frequency removal step of removing low frequency components from the profile according to a noise removal width parameter;
A defect score calculating step of calculating a defect score profile according to the defect score calculation width parameter;
An image processing program that causes a computer to execute a defect position detection step of determining a defect position from the defect score profile. A profile calculation step for obtaining a profile in imaging of an object;
A low frequency removal step of removing low frequency components from the profile according to a noise removal width parameter;
A defect score calculating step of calculating a defect score profile according to the defect score calculation width parameter;
A defect position detecting step for determining a defect position from the defect score profile;
An image processing program that causes a computer to execute an evaluation value calculation step of calculating an evaluation value of a defect at a detected defect position. 5. The low frequency component is removed by subtracting a neighborhood average in a noise removal width parameter from a signal value of a profile in imaging of the object in the low frequency removal step. The image processing program according to item 1. In the defect score calculation step, the defect score is calculated by subtracting the average of the neighborhood in the defect score calculation width parameter from the signal value of the target position of the profile after removing the low frequency component and dividing by the standard deviation. The image processing program according to claim 1, wherein the image processing program is one of claims 1 to 4. 5. The image processing program according to claim 3, wherein in the defect position detection step, a defect position is determined by threshold processing of the defect score profile. The image processing program according to any one of claims 2 to 4, wherein, in the defect score calculation step, the defect score width parameter is determined according to the noise removal width parameter. 5. The image processing program according to claim 2, wherein the defect score calculation step uses a value obtained by multiplying the noise removal width parameter by a predetermined coefficient as the defect score width parameter. . 5. The image processing program according to claim 2, wherein the defect score calculation step uses a value equal to or smaller than the noise removal width parameter as the defect score width parameter. 5. The image processing program according to claim 3, wherein, in the defect position detection step, the defect position is detected from a plurality of defect scores based on a combination of the plurality of noise removal width parameters and the defect score width parameters. . The image processing program according to claim 4, wherein, in the evaluation value calculating step, an evaluation value is calculated from a defect score or a low frequency removal profile at the defect position. A profile calculation unit for obtaining a profile in imaging of an object;
An image processing apparatus comprising: a low-frequency removal unit that removes a low-frequency component from the profile obtained by the profile calculation unit according to a noise removal width parameter. A profile calculation unit for obtaining a profile in imaging of an object;
A low-frequency removal unit that removes a low-frequency component from the profile obtained by the profile calculation unit according to a noise removal width parameter;
An image processing apparatus comprising: a defect score calculation unit that calculates a defect score profile corresponding to a defect score calculation width parameter for a profile from which a low frequency component has been removed by the low frequency removal unit. A profile calculation unit for obtaining a profile in imaging of an object;
A low-frequency removal unit that removes a low-frequency component from the profile obtained by the profile calculation unit according to a noise removal width parameter;
A defect score calculation unit that calculates a defect score profile according to a defect score calculation width parameter for a profile obtained by removing a low frequency component in the low frequency removal unit;
An image processing apparatus comprising: a defect position detection unit that determines a defect position from the defect score profile calculated by the defect score calculation unit. A profile calculation unit for obtaining a profile in imaging of an object;
A low frequency removing unit that removes a low frequency according to a noise removal width parameter from the profile obtained by the profile calculating unit;
A defect score calculation unit that calculates a defect score profile according to a defect score calculation width parameter for a profile obtained by removing a low frequency component in the low frequency removal unit;
A defect position detector for determining a defect position from the defect score calculated by the defect score calculator;
An image processing apparatus comprising: an evaluation value calculation unit that calculates an evaluation value of a defect at a defect position detected by the defect position detection unit.

Images