JP2006226834A - Surface inspection device and surface inspection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a surface inspection device and a surface inspection method with discrimination power enhanced between an irregularity flaw and dirt on a moving surface of a thick steel plate or a thin steel sheet. <P>SOLUTION: This surface inspection device, using a light-section method for performing detection by discriminating between an irregularity flaw and dirt on a surface of a moving object 1 under inspection, comprises: a lighting means 2 for projecting slit light in a width direction orthogonal to the moving direction of the surface of the object with the slit light inclined at a prescribed angle with respect to a direction vertical to the surface of the object; a detection camera 3 for receiving reflected light of the slit light in the vertical direction; a movement speed detection means 4 for detecting the speed of the object 1 in the moving direction to generate a unit length signal on the moving direction; and an image processor 5 for thereinto inputting a signal from the detection camera 3 and the signal from the detection means 4, picking up the characteristics of the profile of an irregularity image of the slit light, and thereby discriminating between an irregularity flaw and dirt. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a surface inspection apparatus and a surface inspection method using a light cutting method for identifying and detecting uneven scratches and dirt on the surface of a moving object to be inspected.

  Thick steel plates used for pressure vessels, tankers, building materials, etc., and thin steel plates used for automobiles, home appliances, cans, etc., are scratched or soiled due to some abnormality caused by the manufacturing process of these steel plates. Will occur.

  Scratches on the surface of thick steel plates and thin steel plates are roughly classified into two types. One of them is uneven scratches caused by cracking, peeling, or foreign matter mixing in the surface of the steel sheet, and the other is uneven oxide film or traces of rolling oil. It is dirt that occurs.

  In surface inspection equipment that inspects the scratches on the surface of these steel sheets, from the viewpoints of product quality assurance and yield improvement, it is usually distinguished from uneven scratches and dirt, and particularly harmful uneven scratches are reliably detected to remove dirt. It is required not to detect it as a scratch.

  Conventionally, as such a surface inspection apparatus, the surface of the object to be inspected is irradiated with three types of light beams having different spectra, and a detection optical system in which one light beam is bright field and the other two light beams are dark field, A method of time-sharing detection of three types of reflected light with a single camera is disclosed (for example, see Patent Document 1).

  In this method, the gradient of the surface is detected from the difference between the image signals of the two kinds of reflected light under the dark field condition irradiated symmetrically, and the image signal of the reflected light under the bright field condition and the reflected light under the dark field condition are detected. Gloss and reflectivity are respectively detected from the difference and the sum with the image signal.

  Here, the gradient detection mainly assumes the detection of uneven scratches, and the reflection detection mainly assumes the detection of dirt.

  Also, in the light cutting method, the workpiece is irradiated with slit light, and both surfaces of the workpiece at the same position are imaged with an imaging device arranged outside the irradiation optical axis, and the slit image data of the workpiece on both sides obtained by the imaging device is obtained. Based on this, a method for determining the quality of a workpiece is disclosed (for example, see Patent Document 2).

In addition, as a defect detection method based on image recognition, a normalized correlation coefficient between a good template image in a predetermined divided mesh unit and a captured image signal is obtained, and the case where the coefficient value is equal to or less than a predetermined value is defective. Is disclosed (see, for example, Patent Document 3).
JP 2001-514386 (second page, FIG. 11) JP 2000-292132 A (first page, FIG. 7) Japanese Patent Laid-Open No. 2004-177170 (third page, FIG. 2)

  Each of the gradient and reflective detection systems disclosed in Japanese Patent Application Publication No. 2001-514386 described above irradiates three types of light beams in a time-sharing manner, and uses three types of images from the reflected light. , And the sum of the images, and the feature is extracted from the three time-division signals.

  Also, the discrimination between uneven scratches and dirt is determined by the difference in the feature amount of the image generated from the camera signal imaged under each optical condition. There is no disclosed means.

  In this method, a feature is extracted from the difference between the left and right dark field images and the image signal obtained by the sum of the bright field image and the dark field image. When the degree of unevenness scratches and dirt become light, there is a problem that the difference and sum signal are not different, and the feature cannot be extracted and cannot be identified.

  Therefore, a method of statistically capturing and identifying feature values of various flaws is also adopted, but if it becomes mild, there will be no difference in the statistical features of the feature values, so there is a problem in identification performance .

  Next, in the work quality inspection method based on the optical cutting method described in Japanese Patent Laid-Open No. 2000-292132 described above, although unevenness on the surface can be detected even when the work moves up and down, unevenness scratches and other dirt scratches can be detected. A method of identifying is not disclosed.

  Next, with respect to the detection of uneven scratches according to the above-mentioned Japanese Patent Application Laid-Open No. 2004-177170, an uneven scratch is recognized by a pattern matching method for obtaining a normalized correlation value between a template image of a good product surface and a predefined uneven scratch image. In order to detect uneven scratches of unspecified patterns, it is necessary to prepare a large number of uneven scratch templates, which complicates recognition processing and takes time. There is a problem that inspection cannot be performed when moving at high speed.

  In addition, the non-defective surface of the object to be inspected is not fixed.For example, there are various patterns for the non-hazardous surface of the non-defective product. Since it is difficult to do so, it is difficult to identify uneven scratches and dirt by the pattern matching method.

  The present invention has been made to solve the above-described problems, and provides a surface inspection apparatus and a surface inspection method with improved ability to distinguish uneven scratches and dirt on moving surfaces of thick steel plates and thin steel plates. With the goal.

  In order to achieve the above object, a surface inspection apparatus according to claim 1 according to the present invention is a surface inspection apparatus that uses a light cutting method for identifying and detecting uneven scratches and dirt on the surface of a moving inspection object, Illumination means for irradiating slit light in a width direction perpendicular to the moving direction of the inspection object, tilted at a predetermined angle with respect to the vertical direction of the surface of the inspection object, and reflected light of the slit light from the slit light A detection camera that receives light outside the irradiation optical axis and images the locus of the slit light, a movement speed detection unit that detects a speed in the movement direction of the inspection object, and generates a unit length signal in the movement direction; An image signal from a detection camera and a signal from the moving speed detection means are input, and an unevenness scratch on the surface of the object to be inspected is determined by measuring a change (a change in height) of the locus of the slit light. A processing device, and the image The physical apparatus generates a unit length profile in which the trajectory of the slit light is stored for each unit length in the moving direction of the inspection object as an image with a predetermined resolution in the width direction and the height direction of the inspection object. Generating a first determination length / width profile in which the unit length profile and the unit length profile are stored a number of times of the unit length (determination length) and superimposed in the moving direction. Width direction adding means for obtaining the occurrence frequency by adding the number of occurrences for each same image position from the determination length / width profile, and the determination length / length storing the peak value of the unit length profile for the determination length Multiplication value of profile calculation means comprising movement direction addition means for generating a profile and the occurrence frequency of the corresponding width position for data of a predetermined height or more of the first determination length / width profile A width determination profile obtained by linking the sums in the width direction is obtained, the center of gravity position in the width direction and the standard deviation in the width direction are determined for the width determination profile, and the data of a predetermined height or more of the determination length / length profile is determined. , A feature extraction / shape determination processing means for obtaining a center of gravity position in the movement direction and a standard deviation in the movement direction, and obtaining a position and a shape of the concave / convex scratch from the center of gravity position and the standard deviation, and the feature extraction / shape judgment processing means Output processing means for displaying and printing the output on the input / output device.

  In order to achieve the above object, a surface inspection apparatus according to claim 3 according to the present invention is a surface inspection apparatus that uses a light cutting method for discriminating and detecting uneven scratches and dirt on the surface of a moving inspection object, Illumination means for irradiating slit light in a width direction perpendicular to the moving direction of the inspection object, tilted at a predetermined angle with respect to the vertical direction of the surface of the inspection object, and reflected light of the slit light from the slit light A detection camera that receives light outside the irradiation optical axis and images the locus of the slit light, a movement speed detection unit that detects a speed in the movement direction of the inspection object, and generates a unit length signal in the movement direction; The image signal from the detection camera and the signal from the moving speed detection means are input, and the uneven scratch on the surface of the object to be inspected is determined by measuring the change in the locus of the slit light (change in the height information). An image processing apparatus, The image processing apparatus stores a unit length profile in which the trajectory of the slit light is stored as an image having a predetermined resolution in the width direction and the height direction of the inspection object for each unit length in the moving direction of the inspection object. A second determination length in which the preprocessing means to be generated and the unit length profile are overlapped in the moving direction by a predetermined number of times of the unit length (determination length) and the peak value of the pixel unit in the width direction is stored. A profile calculation means comprising: a width direction addition means for generating a width profile; and a movement direction addition means for generating a determination length / length profile in which the peak value of the unit length profile is stored for the determination length. For data of a predetermined height or more of the second determination length / width profile, the position of the center of gravity in the width direction and the standard deviation in the width direction are equal to or less than the predetermined height of the determination length / length profile. The feature extraction / shape determination processing means for obtaining the center of gravity position in the movement direction and the standard deviation in the movement direction, and obtaining the position and shape of the concave / convex scratch from the center of gravity position and the standard deviation, and the feature extraction / shape Output processing means for displaying and printing the output of the determination processing means on the input / output device.

  In order to achieve the above object, a surface inspection method according to claim 9 of the present invention is a surface inspection method using a light cutting method for discriminating and detecting uneven scratches and dirt on the surface of a moving inspection object. Inclining at a predetermined angle with respect to the normal direction of the surface of the inspection object, irradiating slit light in a width direction perpendicular to the moving direction of the surface of the inspection object, reflected light of the slit light, The slit light is received outside the optical axis, the locus of the slit light is captured as an image, the image is linked to the moving direction of the inspection object, and the profile is generated, and the feature amount of the shape of the profile is preset. It is characterized in that the unevenness scratch of the inspection object is detected by comparing with a determination threshold value.

  As described above, according to the present invention, the unevenness of the surface of the object to be inspected is detected as distortion of the image of the slit light, and the image data of the slit light is stored for each unit length in the moving direction. A long profile is generated, and this unit length profile is overlapped in the movement direction of the object to be inspected by a predetermined judgment length to generate a judgment length / width profile, and the peak value of the unit length profile is set in the movement direction. The determination length / length profile is generated and stored, and the occurrence frequency and width direction of the same image position (width and height information) of a predetermined height or more are determined from the determination length / width profile. A width determination profile of a predetermined height is formed from the determination length / length profile in the moving direction, and the width determination profile and the length determination profile are displayed. Since the unevenness scratch is judged from the feature amount of the shape of the surface, it is configured so that the shape of the profile (height contour) of the unevenness scratch is captured and the information from the dirty portion is not detected. It is possible to provide a surface inspection apparatus having a high discrimination capability.

  In addition, since the degree of unevenness scratches is determined from the total area of the width determination profile and the length determination profile, when determining the degree of scratches, a surface inspection apparatus that can easily associate the size of the unevenness scratches with the total area. Can be provided.

  In addition, even if the object to be inspected fluctuates in the height direction, the distance detector that detects this amount of fluctuation, or the amount of fluctuation in the height direction is automatically obtained from the fluctuation of the slit light. In contrast, it is possible to provide a surface inspection apparatus and a surface inspection method that are free from erroneous detection.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a configuration diagram of a surface inspection apparatus according to the present invention. The surface inspection apparatus includes an illumination unit 2 that irradiates a surface of an inspection object 1 such as a thin steel plate that moves in a horizontal direction at a high speed with a predetermined angle with respect to the vertical direction, and an illumination unit 2 that irradiates the surface of the inspection object 1 A detection camera 3 that captures the reflected light of the slit light at a light receiving angle different from the irradiation angle, a speed detector 4 for detecting the moving speed of the inspection object 1 and generating a unit length signal in the moving direction, and The image processing apparatus 5 is configured to receive signals from the detection camera 3 and the speed detector 4 and to detect uneven scratches from the image from the distortion of the slit light.

  The image processing unit 5 receives the image signal from the detection camera 3 and the length signal from the speed detector 4 and generates a unit length profile in which the image data of the slit light for each unit length is binarized. From the profile output of the profile calculation unit 5b and the profile calculation unit 5b, the processing unit 5a calculates the characteristics of the profile in the width direction and the movement direction of the inspected object 1 in units of preset unevenness determination length from the unit length profile. A feature extraction / shape determination unit 5c that obtains the feature amount of the profile image and determines the shape, and an output processing unit 5d that processes the output of the determined unevenness flaw.

  Next, the detailed configuration of each part will be described. The inspection object 1 is moving on the transport roll 1a at a high speed in the moving direction (t-axis). The illumination unit 2 is tilted by a predetermined angle or more with respect to the vertical direction (y-axis) of the surface of the inspection object 1 so that a change in the distance between the inspection object 1 and the detection camera 3 can be imaged, and In the width direction (x-axis) of the inspection object 1, a slit light beam that covers the entire width is irradiated.

  For this slit light beam, for example, a laser beam is used as a light source, and a thin line having a predetermined width or less is formed at the surface position of the inspection object 1 through a collimator.

  The detection camera 3 is, for example, an area CCD camera, and is disposed at a light receiving angle different from that of the illuminating unit 2, for example, at the upper part in the normal direction (y-axis).

  The resolution of the area CCD camera is such that when the inspection object 1 changes up and down more than a predetermined amount in the vertical direction, the light receiving position of the slit light beam has a predetermined resolution and in the horizontal direction the slit light beam A device having the number of elements that can capture the entire width with a predetermined resolution is selected.

  Further, the speed detector 4 is constituted by, for example, a laser Doppler type speedometer that can obtain the moving speed of the inspection object 1 with a predetermined resolution without contact. This speed detector 4 can also be comprised with the pulse transmitter mechanically directly connected with the conveyance roll 1a.

  In addition, the image processing apparatus 5 is configured by each unit having the functions as described above. Generally, this realization means is configured by an image processing apparatus 5 as shown in FIG.

  This configuration includes a camera interface 51 that receives signals from the detection camera 3 and the speed detector 4, a signal from the camera interface 51, an image memory 52 that stores an image signal for each unit length, and a stored image signal. A processing program for processing, an image arithmetic processing unit 53 for executing the processing program, an input / output device for displaying and recording the processed result via the input / output interface 54, and an internal connecting them This is realized by an image processing computer including a bus 56.

  According to this configuration, the processing function of each unit shown in FIG. 1 can be handled by changing the program, which is efficient. However, here, the image processing apparatus 5 will be described as being configured from each unit shown in FIG.

  First, the measurement principle of the surface inspection apparatus configured as described above will be described with reference to FIGS.

  FIG. 2 shows that when the object 1 moves as indicated by an arrow D, the detection camera 3 captures images at the coordinates t0, ti, and tn in the moving direction, and the resulting slit light image is binarized by the preprocessing unit 5a. It is a model figure of the image output at the time of doing.

  2A, 2B, and 2C, the horizontal axis (x-axis) coordinates indicate the position in the width direction of the inspection object 1, and the vertical axis (y-axis) coordinates indicate the width positions. The height position from the reference plane position of the inspection object 1 is shown.

  For example, as shown in FIG. 2 (a), an image at the coordinate t0 of the A portion where there is no scratch on the surface of the inspection object 1 has no distortion at the A portion. It becomes.

  However, as shown in FIG. 2B, the image of the coordinate ti where the concave scratch B portion exists on the surface of the inspection object 1 is detected because the height positions of the concave and convex scratch B portion and other surfaces are different. The image of the slit beam viewed from the direction of the camera 3 is a portal-shaped image.

  On the other hand, as shown in FIG. 2C, the image of the coordinate tn where the dirt C portion exists on the surface of the inspection object 1 does not differ in height position between the dirt C portion and the other surface. Since the reflected light of the part is weak, it cannot be detected, and the image of the dirt C part is lost.

  A profile image is created by superimposing such slit beam images in the moving direction in the moving direction. For example, if each image is projected onto a yt axis plane and the peak value of the profile image at the time is linked to the moving direction, a projected image of an inclined mountain profile is obtained as shown in FIG.

  Further, when a large number of profile images t0 to tn viewed in the width direction are overlapped in the movement direction, a projection image of a plurality of portal profiles as shown in FIG.

  The information on the dirt C portion is not visible from any profile image (projected image) because there is no uneven portion. Then, the feature amount of the shape of each profile image from the projected two directions is obtained to determine the shape and size of the uneven scratch.

  Next, the operation of the present invention based on such a measurement principle will be described with reference to FIGS. FIG. 4 is a process flow diagram of uneven scratches processed by the image processing apparatus 5.

  First, in the pre-processing unit 5a, the slit light image from the detection camera 3 is determined by a predetermined threshold (search for uneven scratch candidates) every time the inspection object 1 moves a unit length in the moving direction, for example, 1 mm. And stored as a unit length profile converted to a binarized image (s1, s2).

  Images of the unit length profile image memory are shown in FIGS. 6 (a) to 6 (c). FIGS. 6A to 6C illustrate unit length profiles obtained at coordinates t0, t1, to tn in the moving direction of the inspection object 1, and the vertical axis (y axis) in each figure. Indicates the height of the inspection object 1 from the reference plane position, and the horizontal axis (x-axis) indicates the position of the inspection object 1 in the width direction.

  Since the height from the reference plane position is uniquely determined by the irradiation angle of the slit light and the optical conditions of the detection camera 3, the position imaged by the detection camera 3 is corrected in advance to obtain unit length profile data. Keep it as

  Next, in the profile calculation unit 5b, the unit length profile obtained every time the unit length is moved is projected onto the lateral direction of the moving direction, and the length of the determination unit for determining the presence of uneven scratches (hereinafter referred to as the determination length). In addition, they are overlapped for the determination length in the moving direction.

The determination length for determining the presence or absence of the uneven scratch is, for example, 500 mm for 500 times of a unit length of 1 mm, and the determination length / length as shown in FIG. Profile (d _f ) and a first determination length / width profile (d _w ) as shown in FIG. 6 (e) are generated (s3, s4).

Next, the feature extraction / shape determination unit 5c determines the determination length / length profile and the first determination length / width profile with a predetermined determination length threshold value (t _f , t _w ), and the determination. A length determination profile (P _f ) as shown in FIG. 6 ( _f ) and a width determination profile (P _w ) as shown in FIG. 6 (g), which are generated as data of uneven scratches, are obtained (s 5 ).

  The details of this unevenness determination processing will be described later with reference to the unevenness determination processing flowchart shown in FIG.

  Next, when there are multiple uneven scratches extracted by the judgment length, the distance between the coordinate positions of these uneven scratches is obtained to determine the presence or absence of connection, and when it is equal to or less than a predetermined value, the same uneven scratch is determined. (S6). Details of this will be described later with reference to FIGS.

  The detected uneven scratch is output to a display monitor such as a CRT (not shown) or an output device such as a printer (s7).

Next, details of the unevenness determination process (s6) will be described with reference to FIGS. 5 and 6 again. The first determination length / width profile d _w (x, y) of the determination length unit generated from the unit length profile shown in FIGS. 6A to 6C and the determination length / length profile d _f (y _pt , T) are shown in FIGS. 6D and 6E, respectively.

  These drawings show a state in which profile information of the height (y coordinate) at the length direction (t coordinate) and the width direction position (x coordinate) of the inspection object 1 is stored.

Since the first determination length / width profile d _w (x, y) is data obtained by superimposing a plurality of unit length profiles, a lot of data having the same height is generated in the image. Therefore, the occurrence frequency n (x, y) for each height information at the same width position is obtained for the determination length.

Also, when the projection unit length profile in y-t plane, the peak value y _pt height in the width direction, determined length, length profile what had been per unit length profile d _{f (y pt,} t ).

First, initial values of the respective threshold values for determining the feature values of the determination length / length profile d _f and the first determination length / width profile d _w and the determination length threshold values t _f and t _w are set (s61).

  Since this value is a value for determining whether or not there are uneven scratches, it is set in advance to such a value that the required minute uneven scratches can be detected.

Next, processing procedures for the width determination profile P _w (x) and the length determination profile P _f (x) will be described.

Width determination profile, as shown in FIG. 6 (g), determined length threshold t _w or more, the data up to the range of up to ye, the frequency n (x, y) for each same height information by multiplying The sum is obtained for each same width position and set as a width determination profile P _w (x) in the width direction.

The width determination profile P _w (x) is represented by the following arithmetic expression.

In this calculation, as shown in the processing flow chart, first, an initial value of the determination length threshold t _w (t _f ) is set (s61), and the width determination profile P _w (x) is sequentially updated from t _w to ye. This is achieved by repeating the product-sum calculation of (s62 to s64).

Then, from the obtained width determination profile P _w (x), the center of gravity Pcgx and the standard deviation σx are obtained (s65), and the occurrence position and width of the concavo-convex scratch are calculated from these values.

Further, the total area of the obtained width determination profile P _w (x), ΣP _w (x), is obtained as a feature amount for evaluating the size of the uneven scratch.

  Here, the occurrence position of the uneven scratch is the coordinate position of the obtained center of gravity, and the width of the uneven scratch is the coordinate position of the standard deviation. Change with.

Next, as shown in FIG. 6 (f), the length determination profile P _f (x) is generated by _linking the peak value (y _pt ) in the moving direction with respect to the data in the range from the threshold value t _{f to the} maximum ye. To do.

This calculation is executed by sequentially updating from the determination length threshold t _f to ye as shown in the process flow chart (s67 to s69). Then, the center of gravity Pcgf and the standard deviation σf are obtained from the obtained length determination profile P _f (x) (s69a), and the occurrence position and width of the concavo-convex scratch are calculated from these values (s69b).

  As described above, the obtained coordinates of the center of gravity and the standard deviation may be used as the occurrence position of uneven scratches. However, when they occur in combination, they are changed by a connection determination process described later.

The determination length thresholds t _f and _tw are normally set to the same value, but if they are set individually, the uneven scratches such as a long uneven scratch in the moving direction and a long uneven scratch in the width direction. Since it is possible to change the determination depending on the shape, there is an advantage that the degree of freedom of determination is increased.

  Next, the connection determination process will be described with reference to FIGS. For example, FIG. 7A shows a case in which the uneven scratch A1 and the uneven scratch A2 are detected continuously, and FIG. 7B shows that the uneven scratch A3 and the uneven scratch A4 are close to each other in the width direction. The case where it adjoins is shown in figure.

  FIG. 7A shows a case where the distance between the coordinates of the center of gravity in each moving direction is Td, and the minimum distance of the standard deviation is Gt. In this case, the connection determination process is performed when the value of Gt is a predetermined value or less. The uneven scratch A1 and the uneven scratch A2 are determined to be the same uneven scratch A12.

  Then, the intermediate position of the coordinates of the center of gravity is set to the position of the same unevenness scratch A12, and the length is a value obtained by adding the standard deviation σt1, the standard deviation σt2 and the minimum distance Gt.

  Similarly, in the width direction, the position in the width direction and its width are obtained from the center of gravity obtained from the width determination profile and its standard deviation.

  As shown in FIG. 8, this process first determines the presence or absence of a plurality of irregularities (s71), and then, when a plurality of irregularities scratches are detected, execute the connection determination process as described above, The coordinate position and length are corrected (s72).

  Since the same processing can be performed when the unevenness in the width direction shown in FIG. 7B is combined, the description thereof is omitted.

  Next, FIG. 9 shows an example of the output of the positions and shapes of the uneven scratches processed as described above. First, the position of the uneven scratch in the moving direction is obtained from the coordinates of the center of gravity in the moving direction, and is recorded as the distance to the center of gravity where the uneven scratch with the tip of the inspection object 1 as the origin occurs.

  Similarly, as the position in the width direction, the distance in the width direction with one end of the inspection object 1 as the origin is recorded.

The size of the uneven scratch is recorded as the total area of the width determination profile P _w (x), ΣP _w (x). This ΣP _w (x) is compared with a predetermined determination value to identify the size, and the degree of unevenness scratches is recorded.

In general, the uneven scratch has a shape that is long in the moving direction of the inspection object 1 and a shape that is long in the width direction. Therefore, the total area of the length determination profile P _t (x), ΣP _t (x) It is also possible to selectively use either ΣP _w (x) or ΣP _t (x), or a combination of both, depending on the shape of the uneven scratch.

  As described above, according to the first embodiment, since the distortion image of the slit light is captured and the entire shape of the distortion is captured from the profile image in the moving direction, it is possible to discriminate between uneven scratches and dirt. Can be done with high accuracy. Moreover, since the uneven | corrugated damage | wound is determined by catching the whole shape, the grade of a damage | wound can also be evaluated.

  A second embodiment of the present invention will be described with reference to FIG. The configuration of the second embodiment is different from the first embodiment in that when the distance between the reference plane position of the inspection object 1 and the detection camera 3 moves up and down in the normal direction, the height position of the detected image due to the variation in the distance is determined. The distance sensor 7 for correction is provided, and the function of correcting the coordinate in the height direction of the unit length profile is provided in the preprocessing unit 5a.

  About each part of this Example, the same part as each part shown in FIG. 1 attaches | subjects the same code | symbol, and abbreviate | omits the description.

  Since the inspection object 1 usually moves on the transport roll 1a, the reference plane position does not fluctuate. However, when the transport roll 1a is replaced or when an abnormality occurs in the tension state of a transport machine (not shown), May vary in direction.

  In this case, the unit length profile data moves, for example, to a position higher by Δh from the reference plane position as shown in FIG. The distance Δh is detected by the distance detector 7 to correct the height data in the y-axis direction of the unit length profile.

  If the distance detector 7 is not provided, the pre-processing unit 5c samples the unit length profile data in the width direction or the movement direction at a predetermined interval Δx to obtain and correct the difference from the reference position. You can also do so.

  According to the second embodiment, even when the distance between the inspection object 1 and the detection camera 3 varies, the variation amount can be automatically corrected.

A third embodiment of the present invention will be described with reference to FIG. The difference between the third embodiment and the first embodiment is that in the first embodiment, when the first determination length / width profile d _w (x) is obtained from the unit length profile, data of the same image position (height) is obtained. The width determination profile P _w (x) and the occurrence frequency n (x, y) are obtained by adding to the movement direction. In the third embodiment, the unit length overlapped in the movement direction is obtained. This is because the peak value of the profile is obtained for each width and the second determination length / width profile d` _w (x) is generated.

Then, the total area of in this case, the second determination length, width profile and generates a determination length threshold t _w at the determined width determined profile P` <sub>w (x),</sub> width determination profile P` _{w (x),} Let ΣP` _w (x) be the size of the uneven scratch.

  According to the third embodiment, the width determination profile is the width determination profile in which the peak value of the height information in the moving direction is the data of each width, like the length determination profile. it can.

  The present invention is not limited to the above-described embodiments, and it goes without saying that the inspection object 1 can be applied to a material other than a thick steel plate or a thin steel plate.

  Then, as a light source having a wavelength characteristic that makes it easy to obtain the contrast between the unevenness scratch and dirt of the object 1 to be inspected, for example, the illumination light source of the illumination unit 2 generates slit light by arranging high-luminance LEDs in the width direction in addition to laser light. Alternatively, any configuration may be used as long as it captures a projected image of the distortion of the surface of the object to be inspected by the light cutting method to detect the uneven scratch, and various modifications can be made without departing from the scope of the present invention. be able to.

The block diagram of the surface inspection apparatus by this invention. FIG. 3 is a diagram illustrating the principle of detecting uneven scratches according to the present invention. The block diagram of the image processing apparatus of the surface inspection apparatus by this invention. The flowchart explaining the uneven | corrugated damage | wound process of this invention. The detailed flow figure of the unevenness | corrugation determination process part of the unevenness | corrugation crack processing flow of this invention. Explanatory drawing of an uneven | corrugated crack process of this invention. The connection determination explanatory drawing of this invention. The processing flowchart of the connection determination of this invention. Processing operation explanatory drawing. The example of the determination output of the uneven | corrugated damage | wound of the surface inspection apparatus of this invention. FIG. 6 is a diagram for explaining the configuration and operation for distance correction when the measurement distance varies. The processing flow figure in the case of calculating | requiring the 2nd width | variety determination profile when the occurrence frequency of a width profile is not considered.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Measured object 1a Transport roll 2 Illumination part 3 Detection camera 4 Speed detector 5 Image processing part 5a Pre-processing part 5b Profile calculation part 5c Feature extraction / shape determination part 5d Output processing part 6 Speed detector 7 Distance detector 51 Camera Interface 52 Image memory 53 Image arithmetic processing unit 54 Input / output interface 55 Input / output device 56 Internal bus

Claims (9)

A surface inspection apparatus using a light cutting method for identifying and detecting uneven scratches and dirt on the surface of a moving inspection object,
Illumination means for irradiating slit light in a width direction perpendicular to the moving direction of the inspection object, tilted at a predetermined angle with respect to the vertical direction of the surface of the inspection object, and reflected light of the slit light from the slit light A detection camera that receives light outside the irradiation optical axis and images the locus of the slit light, a movement speed detection unit that detects a speed in the movement direction of the inspection object, and generates a unit length signal in the movement direction; An image signal from a detection camera and a signal from the moving speed detection means are input, and an unevenness scratch on the surface of the object to be inspected is determined by measuring a change (a change in height) of the locus of the slit light. A processing device,
The image processing apparatus stores a trajectory of the slit light for each unit length in the moving direction of the inspection object, and a unit length profile storing the width direction and the height direction of the inspection object as an image with a predetermined resolution. Pre-processing means for generating
The unit length profile is generated for a predetermined number of times of the unit length (determination length), and a first determination length / width profile is generated in which the unit length profile is stored in the moving direction, and further, the first determination length / width profile is generated. Width direction adding means for obtaining the occurrence frequency by adding the number of occurrences for each same image position, and a moving direction for generating a judgment length / length profile in which the peak value of the unit length profile is stored for the judgment length. Profile calculation means comprising addition means;
A width determination profile obtained by linking the sum of multiplication values of the occurrence frequencies of the corresponding width positions with respect to data having a predetermined height or more in the first determination length / width profile in the width direction is obtained. The center of gravity position in the width direction and the standard deviation in the width direction are obtained, and the center of gravity position in the movement direction and the standard deviation in the movement direction are obtained for data of a predetermined height or more of the determination length / length profile. Feature extraction / shape determination processing means for obtaining the position and shape of the uneven scratch from the center of gravity position and the standard deviation,
A surface inspection apparatus comprising: output processing means for displaying and printing the output of the feature extraction / shape determination processing means on an input / output device.   2. The feature extraction / shape determination processing means obtains a total sum obtained by adding the width determination profile data for the entire width, and determines the size (degree) of the unevenness scratch from the total sum. The surface inspection apparatus described. A surface inspection apparatus using a light cutting method for identifying and detecting uneven scratches and dirt on the surface of a moving inspection object,
Illumination means for irradiating slit light in a width direction perpendicular to the moving direction of the inspection object, tilted at a predetermined angle with respect to the vertical direction of the surface of the inspection object, and reflected light of the slit light from the slit light A detection camera that receives light outside the irradiation optical axis and images the locus of the slit light, a movement speed detection unit that detects a speed in the movement direction of the inspection object, and generates a unit length signal in the movement direction; The image signal from the detection camera and the signal from the moving speed detection means are input, and the unevenness of the surface of the object to be inspected is determined by measuring the change in the locus of the slit light (change in the height information). An image processing apparatus,
The image processing apparatus stores a trajectory of the slit light for each unit length in the moving direction of the inspection object, and a unit length profile storing the width direction and the height direction of the inspection object as an image with a predetermined resolution. Pre-processing means for generating
A width for generating a second determination length / width profile in which the unit length profile is overlapped in the movement direction by a predetermined number of times of the unit length (determination length) and the peak value of the pixel unit in the width direction is stored. Profile calculation means comprising direction addition means, and movement direction addition means for generating a judgment length / length profile in which the peak value of the unit length profile is stored for the judgment length.
For the data of a predetermined height or more of the second determination length / width profile, the center of gravity position in the width direction and the standard deviation of the width direction are set, and the data of the determination height / length profile is a predetermined height or more. About the position of the center of gravity in the moving direction and the standard deviation of the moving direction, and the feature extraction / shape determination processing means for determining the position and shape of the uneven scratch from the position of the center of gravity and the standard deviation,
A surface inspection apparatus comprising: output processing means for displaying and printing the output of the feature extraction / shape determination processing means on an input / output device.   The feature extraction / shape determination processing means calculates a total area of the second determination length / length profile and a total area of the determination length / length profile, and outputs the total area and the shape determination output of the uneven scratch. The surface inspection apparatus according to claim 3, wherein the size (degree) of the unevenness flaw is determined by corresponding to the predetermined logic.   When a plurality of outputs of the feature extraction / shape determination means are generated, the position corresponding to the width direction and the moving direction are respectively output from the center of gravity position in the width direction and the standard deviation, the center of gravity position in the movement direction, and the output of the standard deviation. The surface inspection apparatus according to claim 1, further comprising a connection determination processing unit that determines that the same unevenness scratch is present when the information distance is equal to or less than a predetermined value.   A distance detection unit that detects a distance variation between the surface of the object to be inspected and the detection camera is provided, and the preprocessing unit corrects height information of the unit length width profile. The surface inspection apparatus according to claim 1 or claim 3.   The unit length profile or the determination length / width profile data is sampled at a plurality of points in the width direction, the average value is obtained, and the inspection object is obtained from the difference between the obtained average value and preset reference position data. The distance fluctuation amount is obtained, and height information of the unit profile or the judgment length / width profile and the judgment length / length profile is corrected. The surface inspection apparatus described in 1.   A variation amount from a preset reference position is obtained for the determination length / length profile, and the unit profile or the determination length / width profile and the height information of the determination length / length profile are corrected. The surface inspection apparatus according to claim 1, wherein the surface inspection apparatus is configured as described above. A method of surface inspection using a light cutting method that identifies and detects uneven scratches and dirt on the surface of a moving inspection object,
Inclining at a predetermined angle with respect to the normal direction of the surface of the object to be inspected, and irradiating slit light in the width direction orthogonal to the moving direction of the surface of the object to be inspected,
The reflected light of the slit light is received outside the irradiation light axis of the slit light, and the locus of the slit light is captured as an image,
Concatenating the image with the moving direction of the inspection object to generate a profile thereof,
A method for surface inspection, wherein a feature amount of the profile shape is compared with a preset determination threshold value to detect uneven scratches on the inspection object.

Images