Surface inspection device and surface inspection method

Abstract

To accurately detect uneven defects without requiring much labor, even when there is a large thickness variation in a measurement object, or even when there is influence caused by warpage of the measurement object, or vibration during conveyance.SOLUTION: A surface inspection device acquires surface height distribution data of a measurement object in a width direction of the measurement object being a direction perpendicular to a conveyance direction of the plate-shaped measurement object during conveyance, acquires the surface height distribution data at predetermined intervals along a longitudinal direction of the measurement object being the conveyance direction of the measurement object, extracts and stores the surface height distribution data of the measurement object in the longitudinal direction of the measurement objects at different positions in the width direction of the measurement object from two-dimensional height distribution data in which the width direction and the longitudinal direction are integrated and stored, extracts reference height distribution data of the measurement object in the longitudinal direction of the measurement object through common feature extraction of the surface height distribution data of the measurement object in the longitudinal direction of the measurement objects, and detects uneven defects on the surface of the measurement object on the basis of the reference height distribution data and the surface height distribution data in the longitudinal direction of the measurement objects.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a surface inspection device and a surface inspection method for detecting irregular defects on the surface of an object to be measured. [Background technology]

[0002] When a defect occurs on a strip, such as a steel plate, during the manufacturing process, it is necessary to quickly detect the defect and change the manufacturing conditions for the strip or eliminate the equipment abnormality that caused the defect to prevent its impact on subsequent strips. For this reason, for example, in steel plate production lines, steel plates are inspected for surface defects while they are moving. Optical inspection methods are widely used as a non-contact method for detecting surface defects. Optical inspection methods include a method that captures an image of the surface of the object to be measured and detects surface defects based on the brightness of the image (see Patent Document 1), and a light-section method that detects surface defects by obtaining unevenness data of the object using a triangulation method using laser light and a camera (see Patent Document 2). Light-section optical inspection methods are suitable for detecting unevenness defects because they directly obtain unevenness data of the object to be measured, and they can also determine the grade of the unevenness defect based on the depth of the unevenness. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6741173 [Patent Document 2] Japanese Patent Application Laid-Open No. 2003-207319 Summary of the Invention [Problem to be solved by the invention]

[0004] In optical inspection methods using the light-section method, if the thickness change of the transported object is significantly greater than the depth of the irregularity defect, it becomes difficult to detect irregularity defects from the absolute value of the obtained irregularity data. Therefore, to detect irregularity defects, reference irregularity data is required for comparison with the obtained irregularity data. However, because factors other than thickness change of the object also affect the irregularity data, such as warpage of the object and unsteadiness during transport, it is not possible to prepare reference irregularity data in advance. For this reason, reference irregularity data is generated by applying a moving average filter or a weighted moving average filter to the obtained irregularity data to remove singular points such as scratches. However, this requires manual optimization of the number of data points to be averaged and the weighting of each data point. Furthermore, if the width or length of the object changes, optimal conditions corresponding to these changes must be separately set. Therefore, there has been a need for a technology that can accurately detect irregularity defects without requiring significant effort, even when the object's thickness changes significantly or when the object is affected by warpage or unsteadiness during transport.

[0005] The present invention has been made to solve the above-mentioned problems, and its purpose is to provide a surface inspection device and a surface inspection method that can accurately detect uneven defects without requiring much effort, even when the thickness of the object to be measured varies greatly or when the object to be measured is affected by warping or rattles during transportation. [Means for solving the problem]

[0006] The surface inspection device according to the present invention comprises an acquisition means for acquiring surface height distribution data of a plate-shaped object being transported in a width direction of the object, which is a direction perpendicular to the transport direction of the object; a first storage means for acquiring the surface height distribution data at predetermined intervals along the longitudinal direction of the object, which is the transport direction of the object, and storing data obtained by integrating the acquired surface height distribution data as two-dimensional height distribution data; a second storage means for extracting and storing surface height distribution data of a plurality of objects in the longitudinal direction of the object that are different in position in the width direction of the object from the two-dimensional height distribution data; an extraction means for extracting reference height distribution data of the object in the longitudinal direction of the object by extracting common features of the surface height distribution data of the plurality of objects in the longitudinal direction; and a detection means for detecting uneven defects on the surface of the object based on the reference height distribution data and the surface height distribution data of the plurality of objects in the longitudinal direction of the object.

[0007] The extraction means may generate the reference height distribution data using principal component vectors obtained by performing principal component analysis on surface height distribution data of the plurality of objects in the longitudinal direction of the objects.

[0008] The extraction means may generate the reference height distribution data so as to minimize an error with all surface height distribution data used in the principal component analysis.

[0009] A surface inspection method according to the present invention includes the steps of acquiring surface height distribution data of a plate-shaped object being transported in a width direction of the object, which is a direction perpendicular to the transport direction of the object; acquiring the surface height distribution data at predetermined intervals along the longitudinal direction of the object, which is the transport direction of the object, and saving data obtained by integrating the acquired surface height distribution data as two-dimensional height distribution data; extracting and saving surface height distribution data of a plurality of objects in the longitudinal direction that are different in position in the width direction of the object from the two-dimensional height distribution data; extracting reference height distribution data of the objects in the longitudinal direction by extracting common features of the surface height distribution data of the plurality of objects in the longitudinal direction; and detecting irregularity defects on the surface of the object based on the reference height distribution data and the surface height distribution data of the plurality of objects in the longitudinal direction. [Effects of the Invention]

[0010] The surface inspection device and surface inspection method according to the present invention can accurately detect uneven defects without requiring much effort, even when there is a large change in the thickness of the object to be measured or when the object is affected by warping or rattles during transport. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a schematic diagram showing the configuration of a surface inspection system according to one embodiment of the present invention. [Figure 2] FIG. 2 is a block diagram showing the configuration of the surface inspection apparatus shown in FIG. [Figure 3] FIG. 3 is a flowchart showing the flow of a surface defect detection process according to one embodiment of the present invention. [Figure 4] FIG. 4 is a diagram showing an example of surface height distribution data. [Figure 5] FIG. 5 is a diagram showing an example of two-dimensional height distribution data. [Figure 6] FIG. 6 is a diagram for explaining the process of step S3 in FIG. [Figure 7] FIG. 7 is a diagram showing the difference between the reference height distribution data and the surface height distribution data in the y direction. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, the configuration and operation of a surface inspection system according to one embodiment of the present invention will be described with reference to the drawings.

[0013] 〔composition〕 First, the configuration of a surface inspection system according to one embodiment of the present invention will be described with reference to Figures 1 and 2. Figure 1 is a schematic diagram showing the configuration of a surface inspection system according to one embodiment of the present invention. Figure 2 is a block diagram showing the configuration of a surface inspection device 4 shown in Figure 1.

[0014] As shown in Fig. 1, a surface inspection system 1 according to one embodiment of the present invention is a system for detecting uneven defects on the surface of a steel sheet S, and includes as its main components a linear light source 2, an imaging device 3, and a surface inspection device 4. Note that, although this embodiment is an application of the present invention to a process for detecting uneven defects on the surface of a steel sheet S, the present invention is not limited to this embodiment and can also be applied to a process for detecting uneven defects formed on the surface of other plate-like objects to be measured, such as steel slabs.

[0015] The linear light source 2 is a device that irradiates a linear laser beam (linear light) parallel to the width direction (hereinafter abbreviated as the x direction, a direction perpendicular to the y direction) of the steel sheet S that is transported along the longitudinal direction (hereinafter referred to as the y direction) of the steel sheet S.

[0016] The imaging device 3 is a device that captures images of light-section lines (linear light extending in the x-direction) on the surface of the steel sheet S caused by linear light irradiated from the linear light source 2 at predetermined time intervals, and outputs image data of multiple light-section lines at different positions in the y-direction (hereinafter referred to as surface height distribution data) to the surface inspection device 4. The interval between image captures is limited by the size of the defect to be detected, and should be determined so that images are captured at intervals that are no greater than 1 / 5 of the size of the defect. Therefore, the interval between image captures should be determined depending on the conveyance speed of the steel sheet S and the size of the defect to be detected. The x-direction resolution of the image sensor that constitutes the imaging device 3 should be no greater than 1 / 5 of the size of the defect to be detected.

[0017] The surface inspection device 4 is configured by an information processing device such as a computer. When an arithmetic processing device such as a CPU in the information processing device executes a computer program, the surface inspection device 4 functions as a surface height distribution data acquisition unit 41, a two-dimensional height distribution data storage unit 42, a longitudinal height distribution data storage unit 43, a reference height distribution data extraction unit 44, and a defect detection unit 45. The functions of each of these units will be described later.

[0018] The surface inspection system 1 having such a configuration executes the surface defect detection process described below, thereby accurately detecting uneven defects without requiring much effort, even when there is a large change in thickness of the steel sheet S or when there is influence from warping of the steel sheet S or wobbling during transportation. Below, the operation of the surface inspection system 1 when executing the surface defect detection process will be described with reference to the flowchart shown in Fig. 3.

[0019] [Surface defect detection process] Fig. 3 is a flowchart showing the flow of surface defect detection processing according to one embodiment of the present invention. The flowchart shown in Fig. 3 starts when an execution command for the surface defect detection processing is input to the surface inspection device 4, and the surface defect detection processing proceeds to step S1.

[0020] In the processing of step S1, the surface height distribution data acquisition unit 41 acquires a plurality of surface height distribution data at different positions in the y direction output from the imaging device 3. An example of the surface height distribution data is shown in Fig. 4. As shown in Fig. 4, the surface height distribution data represents a one-dimensional surface height distribution for each x coordinate of the steel plate S. This completes the processing of step S1, and the surface defect detection processing proceeds to the processing of step S2.

[0021] In the process of step S2, the two-dimensional height distribution data storage unit 42 generates two-dimensional distribution data of the surface height for one steel plate S by integrating multiple pieces of surface height distribution data at different positions in the y direction acquired in the process of step S1. Then, the two-dimensional height distribution data storage unit 42 stores the generated two-dimensional distribution data of the surface height for one steel plate S in a memory unit as two-dimensional height distribution data. FIG. 5 shows an example of two-dimensional height distribution data. Note that the region for which two-dimensional height distribution data is generated is not limited to the entire steel plate S, and two-dimensional height distribution data may be generated for any region within the steel plate S. This completes the process of step S2, and the surface defect detection process proceeds to the process of step S3.

[0022] In the process of step S3, the longitudinal height distribution data storage unit 43 divides the two-dimensional height distribution data stored in the process of step S2 into one-dimensional surface height distribution data in the y direction for multiple steel sheets S at different positions in the x direction, as shown in Fig. 6. This process obtains (width of steel sheet S) / (x-direction resolution of image sensor) pieces of one-dimensional surface height distribution data as vector data with the surface height distribution in the y direction at predetermined intervals in the x direction as a variable. Regarding the number of data points to be acquired, all data may be acquired depending on the x-direction resolution of the image sensor, or data may be acquired at intervals of 1 / 5 or less of the size of the defect to be detected.

[0023] As shown in FIG. 6, the acquired multiple one-dimensional surface height distribution data have a common feature in the y direction. For example, if the steel sheet S has a warped or wavy shape, such a shape becomes a common feature in the y direction. Furthermore, if the steel sheet S has a wobbly shape when the image is captured, the surface height distribution data is recompiled as one-dimensional surface height distribution data in the y direction, so that the common error in surface height due to the wobbly shape is reflected in each data as a common feature in the y direction. The longitudinal height distribution data storage unit 43 stores the multiple one-dimensional surface height distribution data in the memory unit as multiple longitudinal height distribution data. This completes the process of step S3, and the surface defect detection process proceeds to step S4.

[0024] In step S4, the reference height distribution data extraction unit 44 performs principal component analysis on the multiple longitudinal height distribution data saved in step S3, and determines the principal component vectors to be used in subsequent processing based on a predetermined number of principal components or total contribution rate. While the number of principal components and total contribution rate differ for each facility, it is generally desirable for the number of principal components to be around five and for the total contribution rate of the principal components to be around 80-90%. This completes step S4, and the surface defect detection process proceeds to step S5.

[0025] In step S5, the reference height distribution data extraction unit 44 uses the principal component vectors employed in step S4 to generate y-direction reference height distribution data that reproduces common features in the y direction while removing local information such as defects. Specifically, the reference height distribution data extraction unit 44 generates the reference height distribution data by approximating the longitudinal height distribution data using a linear combination of the principal component vectors employed in step S4. Specifically, the reference height distribution data can be generated by determining the combination coefficients of the linear combination of the principal component vectors using the least squares method or the like so as to minimize the differences between all longitudinal height distribution data and the reference height distribution data. Note that approximately 5,000 pieces of data are required to generate accurate reference height distribution data. In this case, the number of pieces of data is determined by (width of the steel sheet S) / (x-direction resolution of the image sensor). Therefore, if the x-direction resolution is approximately 0.2 mm, the width of the steel sheet S to be processed should be 1,000 mm or greater. This completes step S5, and the surface defect detection process proceeds to step S6.

[0026] In the process of step S6, the defect detection unit 45 calculates the difference between the reference height distribution data generated in the process of step S5 and the surface height distribution data in the y direction in the two-dimensional height distribution data, thereby calculating a distribution of differences from the reference height distribution for each y direction surface height distribution data as shown in Fig. 7. Then, the defect detection unit 45 uses the calculated difference as a depth score to calculate a depth score for the entire steel sheet S. This completes the process of step S6, and the surface defect detection process proceeds to the process of step S7.

[0027] In the process of step S7, the defect detection unit 45 detects the presence or absence of uneven defects by comparing the depth score calculated in the process of step S6 with a preset threshold value corresponding to the surface defect to be detected. For example, the defect detection unit 45 sets a threshold value for the feature amount of the surface defect (volume, x-direction length, y-direction length, etc.), and detects a recessed portion where the feature amount calculated from the depth score is above the set threshold value as an uneven defect. This completes the process of step S7, and the series of surface defect detection processes ends.

[0028] As is clear from the above description, in the surface defect detection process according to one embodiment of the present invention, first, the surface height distribution data acquisition unit 41 acquires surface height distribution data of the steel plate S in the width direction of the steel plate S. Next, the two-dimensional height distribution data storage unit 42 acquires surface height distribution data at predetermined intervals along the longitudinal direction of the steel plate S and stores data combining the acquired surface height distribution data as two-dimensional height distribution data. Next, the longitudinal height distribution data storage unit 43 extracts and stores surface height distribution data of the steel plate S in the longitudinal direction of multiple steel plates S that are at different positions in the width direction of the steel plate S from the two-dimensional height distribution data. Next, the reference height distribution data extraction unit 44 extracts reference height distribution data of the steel plate S in the longitudinal direction of the steel plate S by extracting common features of the surface height distribution data of the multiple steel plates S in the longitudinal direction of the steel plate S. Then, the defect detection unit 45 detects uneven defects on the surface of the steel plate S based on the reference height distribution data and the surface height distribution data of the multiple steel plates S in the longitudinal direction. With this configuration, there is no need to manually perform optimization processing or the like to obtain reference height distribution data, so even if there is a large change in the thickness of the steel sheet S or if the steel sheet S is affected by warping or rattles during transportation, unevenness defects can be accurately detected without requiring much effort.

[0029] Although the present inventors have described embodiments of the present invention, the present invention is not limited to the descriptions and drawings that form part of the disclosure of the present invention according to the present embodiments. For example, in the present embodiments, height data of the object to be measured is acquired using a light-section method, but height data of the object to be measured may also be acquired by scanning a rangefinder such as a laser rangefinder. As such, all other embodiments, examples, and operational techniques made by those skilled in the art based on the present embodiments are included in the scope of the present invention. [Explanation of symbols]

[0030] 1. Surface Inspection System 2 Linear light source 3. Imaging device 4. Surface inspection equipment 41 Surface height distribution data acquisition unit 42 2D height distribution data storage section 43 Longitudinal height distribution data storage section 44 Reference height distribution data extraction unit 45 Defect detection section S steel plate

Claims

1. an acquisition means for acquiring surface height distribution data of a plate-shaped object to be measured in a width direction of the object, which is a direction perpendicular to a conveyance direction of the object to be measured; a first storage means for acquiring the surface height distribution data at predetermined intervals along a longitudinal direction of the object to be measured, which is a transport direction of the object to be measured, and storing data obtained by integrating the acquired surface height distribution data as two-dimensional height distribution data; a second storage means for extracting and storing surface height distribution data of the object in the longitudinal direction of a plurality of objects that are different in position in the width direction of the object from the two-dimensional height distribution data; an extraction means for extracting reference height distribution data of the objects to be measured in the longitudinal direction of the objects by extracting common features of surface height distribution data of the objects to be measured in the longitudinal direction of the objects; a detection means for detecting irregular defects on the surface of the object to be measured based on the reference height distribution data and the surface height distribution data of the plurality of objects to be measured in the longitudinal direction; A surface inspection device comprising:

2. 2. The surface inspection device according to claim 1, wherein the extraction means generates the reference height distribution data using principal component vectors obtained by performing principal component analysis on surface height distribution data of the plurality of objects in the longitudinal direction of the objects.

3. 3. The surface inspection apparatus according to claim 2, wherein said extracting means generates said reference height distribution data so as to minimize an error with all surface height distribution data used in said principal component analysis.

4. acquiring surface height distribution data of the plate-shaped object being transported in a width direction of the object, which is a direction perpendicular to a transport direction of the object; acquiring the surface height distribution data at predetermined intervals along a longitudinal direction of the object to be measured, which is a transport direction of the object to be measured, and storing data obtained by integrating the acquired surface height distribution data as two-dimensional height distribution data; extracting and storing surface height distribution data of a plurality of objects in the longitudinal direction of the objects, the objects being positioned differently in the width direction of the objects, from the two-dimensional height distribution data; extracting reference height distribution data of the objects to be measured in the longitudinal direction of the objects by extracting common features of surface height distribution data of the objects to be measured in the longitudinal direction of the objects to be measured; detecting irregular defects on the surface of the object to be measured based on the reference height distribution data and surface height distribution data of the plurality of objects to be measured in the longitudinal direction; A surface inspection method comprising:

Images