CN106248690A - Defect detecting method and flaw detection apparatus - Google Patents

Abstract

The invention relates to a defect inspection method and a defect inspection device. In order to inspect the defects of the protruding lines (92) formed on the surface of the workpiece (9) as a linear pattern, there are: an imaging process of sequentially photographing inspection images (P) along the protruding lines (92); and an inspection process for inspecting The image (P) is sequentially inspected for defects, and in the inspection process, defects are judged based on the continuity of the outline of the front end surface of the protrusion (92) included in the inspection image (P) and the image area. In the photographing process, the camera (3) is moved based on the shape data of the protrusion (92) obtained from the machining program of the workpiece (9).

Description

Defect inspection method and defect inspection device

technical field

The present invention relates to a defect inspection method and a defect inspection device for inspecting defects of a workpiece on which a linear pattern is formed on the surface.

Background technique

Conventionally, there is a work (work) in which a linear pattern is formed on the surface.

For example, a punching die for stickers for display forms a closed ring-shaped continuous ridge on the surface of the mold body, and the front end of the ridge is used as a cutter head (refer to document 1: Japanese Patent Application Laid-Open No. 2014-193966 ).

When manufacturing the punching die, cutting or etching is performed on the flat surface of the die main body to form protrusions of a predetermined pattern corresponding to the punching outline of stickers or the like. Then, a quenching treatment or the like is performed on the front end of the formed ridge to form a punching blade.

In the punching blade, in order to accurately perform punching, especially half-cutting, which requires precision in cutting depth, the width and height of the blade are required to be predetermined values over the entire pattern.

In the inspection of such a bit, a pattern inspection device based on image processing is used.

As a pattern inspection device, an image is taken of the surface of the punching die as the workpiece, and the shape of the tip of the linear pattern appearing in the image is compared with the reference pattern, and if there is a difference, it is detected as a defect (refer to Document 2: JP-A-8-184570, Document 3: JP-A-2006-184037).

In inspection of the shape of the tip, the width of the tip, that is, the distance between the two side edges is also checked in addition to the contour shape of both side edges of the tip.

In the above-mentioned pattern inspection device, only the region including the linear pattern on the surface of the workpiece is selected, and the inspection range is limited to suppress the processing time and data amount of the image to be inspected.

For example, in the inspection device of Document 2, an inspection range is designated on a screen by drawing a frame line surrounding an inspection object by operating means such as a mouse. In addition, in the inspection device of Document 3, a figure enclosing a frame line to be inspected is selected, and the frame line is arranged on a screen to designate an inspection range.

However, in any of the devices, the operator specifies the inspection range, which increases the burden on the operator and is disadvantageous to automation.

Furthermore, in the aforementioned pattern inspection apparatus, it is necessary to prepare a reference image in advance for each search section in order to compare the workpiece image with the reference image.

In particular, when the linear pattern formed on the surface of the workpiece has a fine shape, high resolution is required to recognize it, and there is a problem that the amount of data of the reference image is huge.

Therefore, if the existing pattern inspection device is used for the defect inspection of workpieces having a linear pattern such as the cutter head of the punching die described above, the above-mentioned problem of specifying the inspection range and the preparation of the reference image cannot be avoided. There is a problem that it is difficult to perform efficient defect inspection.

In addition, the same problem is not limited to the linear pattern with continuous protrusions like the above-mentioned punching die, but also to the linear pattern with continuous grooves and the linear pattern with different optical characteristics continuously formed on the surface of the workpiece. same.

Contents of the invention

An object of the present invention is to provide a defect inspection method and a defect inspection device capable of efficiently inspecting a workpiece having a linear pattern formed on its surface.

The defect inspection method of the present invention, for a workpiece having a linear pattern formed on the surface, inspects the defect of the linear pattern, and is characterized in that it includes: a photographing step of sequentially photographing inspection images along the linear pattern; and an inspection step For the inspection image, defect inspection is sequentially performed, and in the inspection step, the defect is determined based on at least one of the continuity of the linear pattern included in the inspection image and an image area.

In such the present invention, since the defect is determined based on at least one of the continuity of the linear pattern included in the inspection image and the image area, it is not necessary to prepare a reference image in advance. In addition, since the inspection images are sequentially captured along the line pattern, it is not necessary to perform an operation of specifying inspection divisions.

Therefore, according to the present invention, defect inspection of a workpiece having a linear pattern formed on its surface can be efficiently performed.

In the defect inspection method of the present invention, it is preferable that, in the inspection step, the inspection image is divided into a pattern region representing the linear pattern and an out-of-pattern region outside the pattern region based on a difference in optical characteristics, The area of the pattern area and the area outside the pattern is calculated and compared with the other inspection images, and if there is no change in the area, it is determined that there is no defect.

In the present invention, as the area of the pattern area and the area outside the pattern, the area and the number of pixels in the inspection image may be used, or an occupancy ratio of one of them may be used. In addition, as another inspection image to be compared, for example, the immediately previous inspection image can be used.

According to the present invention as described above, it is possible to perform defect determination as an inspection process by a simple operation such as calculating the area of the pattern area and the area outside the pattern in the inspection image. For example, if there is a continuous linear pattern with a constant width, the areas of the pattern region and the non-pattern region in the inspection image captured along the linear pattern are constant at any location. Therefore, as in the present invention, for a plurality of inspection images captured along a linear pattern, the area of the pattern region and the area outside the pattern are calculated and sequentially compared, thereby efficiently performing defect determination as an inspection process.

In the defect inspection method of the present invention, it is preferable that, in the inspection step, the inspection image is divided into a pattern region representing the linear pattern and an out-of-pattern region outside the pattern region based on a difference in optical characteristics, A boundary line between the pattern area and the pattern-outside area is detected as the outline of the linear pattern, and if the outline is continuous, it is determined that there is no defect.

In the present invention, whether or not the contours are continuous can be determined by detecting the displacement or inclination angle of any part of the contour relative to its adjacent part in the width direction in one inspection image. In addition, the same continuity can also be determined by comparison with the contours of other adjacent inspection images.

In addition, the contour continuity means that there is no expected discontinuity corresponding to a defect in the contour. The discontinuity corresponding to a defect refers to the case where there is a prominent bend or bend in the contour line, the case where there is a portion with a sharply increased curvature compared with the surrounding area, and the presence of a corner where the contour lines intersect front and rear at an arbitrary point. Ministry situation etc. Regarding the corner portion of the linear pattern, when the bend is formed of a curved line, it may be determined that the outline is continuous. In addition, in the present invention, the linear pattern is not limited to the case of being constituted by a straight line portion, and may be constituted by a part or the whole by a curved line.

In the present invention, the degree of discontinuity determined may be appropriately set based on the degree of the defect to be inspected.

According to such the present invention, it is possible to perform defect determination as an inspection process by simple operations such as detection of the outline of a linear pattern in an inspection image and determination of its continuity. For example, if there is a defect in the continuous linear pattern, the contour of the part is discontinuous, and it can be easily determined by comparing with the adjacent part.

In the defect inspection method of the present invention, it is preferable that the inspection image is specified by an inspection frame of a predetermined shape, and after the first inspection frame is assigned to any part of the linear pattern in the imaging step, the A plurality of inspection images along the linear pattern are captured by sequentially assigning the next inspection frame to adjacent portions of the linear pattern.

In such the present invention, a plurality of inspection images along the linear pattern can be captured for the entire linear pattern or an arbitrary section, and at least an image suitable for the continuity of the linear pattern and the image area in the inspection process can be obtained. An inspection image for determining a defect by one side.

In the defect inspection method of the present invention, it is preferable that the workpiece has a linear pattern formed on its surface by a processing device operating according to a processing program, and in the imaging step, based on the linear pattern included in the processing program, The shape data moves the parts where the inspection images are sequentially captured.

According to the present invention as described above, it is possible to set a movement route during an operation of sequentially capturing inspection images with reference to a machining program for machining a workpiece. Therefore, there is no need to separately photograph the workpiece or give instructions by manual operation in order to set the moving path.

In the defect inspection method of the present invention, preferably, a camera is installed in the processing device that processes the workpiece, and the inspection image is captured by the camera.

According to such the present invention, defect inspection of the linear pattern can be performed immediately after the linear pattern is processed on the workpiece. Therefore, operations such as transferring a workpiece for post-inspection of processing are not required. In addition, the processing device can also be used for defect inspection, and the equipment cost and equipment space can be reduced.

In the defect inspection method of the present invention, it is preferable that in the photographing step, the linear pattern is illuminated intermittently with a strobe while the camera is moved along the linear pattern, During this period, the inspection image is captured by the camera.

According to such the present invention, it is possible to obtain a clear inspection image by flash photography without stopping the camera moving along the linear pattern. Therefore, shooting can be performed in a short time compared with the case of repeatedly moving and stopping to perform shooting. In addition, since high-speed imaging can be performed without a special camera such as a high-speed camera, equipment costs can be reduced.

The defect inspection device of the present invention inspects a workpiece having a linear pattern formed on its surface for defects in the linear pattern, and is characterized by comprising: an imaging unit that sequentially captures inspection images along the linear pattern; and an inspection unit. The defect inspection is sequentially performed on the inspection image, and the inspection unit determines the defect based on at least one of the continuity of the linear pattern included in the inspection image and an image area.

In the present invention, as the imaging unit, for example, a video camera mounted on a machine tool that processes a workpiece can be used, and as the inspection unit, a control device that controls the machine tool can be used. As the control device, an NC device (numerical control device) connected to the machine tool, an external computer system connected to the NC device, or the like can be used.

In such a defect inspection device of the present invention, the same effects as those of the above-mentioned defect inspection method of the present invention can be obtained.

According to the present invention, it is not necessary to prepare a reference image in advance, and it is not necessary to perform an operation of specifying an inspection region when capturing an inspection image. Therefore, according to the present invention, defect inspection of a workpiece having a linear pattern formed on its surface can be efficiently performed.

Description of drawings

FIG. 1 is a schematic diagram showing the configuration of an apparatus according to an embodiment of the present invention.

FIG. 2 is a perspective view showing the workpiece of the embodiment.

Fig. 3 is a partially cutaway enlarged perspective view showing a workpiece according to the embodiment.

FIG. 4 is a perspective view illustrating an imaging step of the embodiment.

5 is (A) plan view and (B) cross-sectional view showing an imaging process of the embodiment.

FIG. 6 is a plan view showing a normal state in the inspection process of the embodiment.

FIG. 7 is a plan view showing defect detection in the inspection process of the embodiment.

8 is a plan view showing another defect detection in the inspection process of the embodiment.

FIG. 9 is a perspective view showing another linear pattern that can be inspected in the embodiment.

FIG. 10 is a perspective view illustrating an imaging process of another embodiment of the present invention.

detailed description

Hereinafter, one embodiment of the present invention will be described based on the drawings.

[Defect inspection device 1]

In FIG. 1 , a defect inspection device 1 is configured by attaching a camera 3 to a spindle head 26 of a machine tool 2 .

The machine tool 2 three-dimensionally processes the workpiece 9 with the tool 4 attached to the spindle 27 .

The machine tool 2 has a table 22 and a door-shaped pillar 23 on the upper surface of a machine tool 21 . A spindle head 26 is supported on a beam 24 of the pillar 23 via a saddle 25 , and a spindle 27 is supported on the spindle head 26 .

The table 22 is movable in the X-axis direction relative to the machine tool 21 , the saddle 25 is movable in the Y-axis direction along the beam 24 , and the spindle head 26 is movable in the Z-axis direction relative to the saddle 25 .

Through such three-axis movement, it is possible to three-dimensionally move the camera 3 and the tool 4 relative to the workpiece 9 placed on the table 22 .

In order to control the three-dimensional movement in the machine tool 2, the machine tool 2 is connected to a control device 5, and the control device 5 is connected to a computer system 6 for control.

The control device 5 is a conventional numerical control device, and controls the operation of the machine tool 2 based on control commands from the computer system 6 . A machining program 51 describing the operation of the machine tool 2 for machining the workpiece 9 is recorded in the control device 5 .

The machining program 51 is executed by operating the control device 5 or by instructing the control device 5 via the computer system 6 to control the operation of the machine tool 2 . Thus, in the machine tool 2 , the workpiece 9 is processed by the tool 4 , and a predetermined shape can be formed on the workpiece 9 .

An inspection program 62 for causing the machine tool 2 to execute operations based on the defect inspection method of the present invention is recorded in the computer system 6 .

In addition, the same machining program 61 as the machining program 51 recorded in the control device 5 is recorded in the computer system 6 and referred to by the inspection program 62 .

In the defect inspection device 1 according to the present embodiment, the camera 3 mounted on the machine tool 2 is an imaging unit, and the computer system 6 operating according to the inspection program 62 is an inspection unit.

[Workpiece 9]

In FIG. 2 , the workpiece 9 according to the present embodiment is used as a punching die for stickers, and a protrusion 92 as a punching blade is formed on a flat upper surface 91 .

The ridges 92 are annular continuous convex linear patterns, and are processed by cutting the surface of the base material 90 of the workpiece 9 with a tool 4 such as an end mill.

As shown in FIG. 3 , the front end surface 93 of the protrusion 92 is flat, and the front end surface 93 is processed to have the same height over the entire length of the protrusion 92 .

The protruding bar 92 is used as a punching knife, so the width of the front end surface 93 is small, and the width of the root part, that is, the base material 90 side, is formed into a mountain shape with a triangular cross-section. Each is formed as an inclined surface.

〔Defect inspection procedure〕

In the present embodiment, the workpiece 9 having the aforementioned protrusions 92 (line patterns) is formed by executing the machining program 51 by the control device 5 in a state where the tool 4 is attached to the aforementioned machine tool 2 .

Then, after the processing, the computer system 6 executes the inspection program 62 using the camera 3 mounted on the working machine 2 to perform the shape of the front end surface 93 of the protruding bar 92 of the linear pattern, which functions as a punching blade. defect inspection.

The defect inspection based on the inspection program 62 is performed by the following imaging process and inspection process.

〔Photography process〕

In the photographing process, the protrusion 92 of the workpiece 9 is photographed by the camera 3 mounted on the machine tool 2 .

In FIG. 4 , imaging by the camera 3 is performed on the area surrounded by the inspection frame Pf on the surface of the workpiece 9 . Then, the images captured by the camera 3 are transmitted from the control device 5 to the computer system 6 and sequentially recorded as inspection images P. As shown in FIG.

The resolution of the inspection image P is set to be sufficiently high for inspection of the front end surface 93 of the protrusion 92 having a finer structure than the entire workpiece 9 . Therefore, the size of the inspection image P cannot be large enough to capture the entire protrusion 92 at a time.

Here, the computer system 6 executing the inspection program 62 operates the machine tool 2, moves the camera 3 along the protrusion 92, and sequentially moves the inspection frame Pf of the camera 3 to repeatedly perform imaging.

In this way, the entire protrusion 92 can be imaged by making the plurality of inspection images P photographed along the protrusion 92 sequentially continuous.

In the photographing process, in order to move the camera 3 along the protrusion 92, the computer system 6 executing the inspection program 62 refers to the processing program 61 held in the computer system 6, and obtains the position of the protrusion 92 from the content of the processing command of the processing program 61. location and shape data. Then, based on the acquired shape data of the protrusion 92 , the moving route of the camera 3 is set so as to track the protrusion 92 .

Specifically, the first inspection frame Pf is assigned to an arbitrary portion of the protrusion 92, and after the inspection image P is acquired by shooting in the inspection frame Pf, the inspection frame Pf is moved along the protrusion 92 by the amount of the inspection frame Pf. , the next inspection frame Pf is assigned to the portion adjacent to the inspection frame Pf that has just been photographed, and the inspection image P is captured. Hereinafter, by sequentially repeating movement and imaging while tracking the protrusion 92 , a plurality of inspection images P covering the entire protrusion 92 can be obtained.

As shown in FIG. 5 , the region of the inspection frame Pf and the sequentially captured inspection images P has a rectangular shape extending along the width direction of the protrusion 92 .

As shown in parts (A) and (B) of FIG. 5 , the inspection frame Pf is set so that, with the front end surface 93 of the protrusion 92 as the center, both sides go over the inclined side surfaces 94 and 95 to reach the workpiece 9 . 91 on the upper surface.

Then, the plurality of inspection images P are arranged along the protrusion 92 with their respective long sides adjacent to each other in sequence. The aforementioned movement of the inspection frame Pf is performed so that the plurality of inspection images P are arranged in this way.

〔Inspection process〕

The inspection process starts before the aforementioned imaging process is completed for the entire protrusion 92 . That is, defect inspection is performed on the imaged inspection image P while performing an imaging process at another location.

In the inspection process, two inspections, an area inspection related to the area of the region in the inspection image P and a continuity inspection related to the continuity of the outline appearing in the inspection image P, are performed.

In the inspection step, the inspection image P is divided into a pattern region representing a linear pattern and an out-of-pattern region outside the pattern region based on a difference in optical characteristics.

In FIG. 6 , in the adjacent inspection images P1 , P2 , P3 , longitudinally extending bright areas AC1 , AC2 , AC3 appear at the center in the width direction thereof, and dark areas AL1 , AL2 , AC3 appear on both sides thereof. AL3 and dark areas AR1, AR2, AR3.

Bright areas AC1 , AC2 , and AC3 are areas (pattern areas) in which front end surfaces 93 of protrusions 92 as linear patterns are imaged. The front end surface 93 is flat and faces the camera 3, thereby reflecting most of the illumination light to enter the camera 3, so that it becomes a bright area in the inspection images P1, P2, and P3.

Dark areas AL1 , AL2 , and AL3 and dark areas AR1 , AR2 , and AR3 are areas in which side surfaces 94 and 95 of protrusions 92 that are linear patterns are imaged (pattern-out areas). These side surfaces 94 and 95 are surfaces inclined with respect to the camera 3 , and the reflected light of the illuminating light incident on the camera 3 is small, so that they are dark areas in the inspection images P1 , P2 , and P3 .

〔Area inspection〕

In area inspection, in each inspection image P1-P3, the number of each pixel of bright area|region AC1-AC3 and dark area AL1-AR3 is calculated, and a defect is judged based on the change.

Specifically, in the inspection image P1, the bright area AC1 is 20 pixels, the dark area AL1 is 40 pixels, and the dark area AR1 is 40 pixels.

In FIG. 6 , in the next inspection image P2 and the next inspection image P3 , the bright areas AC2 and AC3 are 20 pixels each, the dark areas AL2 and AL3 are 40 pixels each, and the dark areas AR2 and AR3 are 40 pixels each.

That is, in the inspection images P1 to P3, the number of pixels in each of the bright areas AC1 to AC3 and the dark areas AL1 to AR3 does not vary.

Therefore, it is determined that the front end surface 93 of the ridge 92 at the site where the inspection images P1 to P3 are imaged is not defective.

In FIG. 7 , inspection images P1 and P2 are the same as those in FIG. 6 described above, but in inspection image P3 , dark area AL3 on the left side protrudes into a part of bright area AC3 .

In such a state, the dark area AL3 exceeds 40 pixels, and the bright area AC3 is less than 20 pixels. That is, in the inspection image P3, since the area of the bright area (pattern area) and the dark area (out-of-pattern area) fluctuates with respect to the inspection images P1 and P2, it is determined that there is a defect.

Such a defect in FIG. 7 is detected when the edge of the front end surface 93 is chipped or the like.

In addition, when the front end surface 93 is deformed in the width direction as in the inspection image P3 shown in FIG. Since the number of pixels of AR3 fluctuates, it can be detected as a defect.

〔Continuity check〕

In FIG. 8, as described above, when the front end surface 93 is deformed in the width direction as in the inspection image P3, the continuity of the outline of the bright area (pattern area), that is, the boundary with the dark area (pattern area) is checked. The continuity of the shape of the line can judge the defect.

In the continuity inspection, in each inspection image P1 to P3, the boundary lines EL1 to EL3 between the bright areas AC1 to AC3 and the dark areas AL1 to AL3 and the boundary lines ER1 between the bright areas AC1 to AC3 and the dark areas AR1 to AR3 -ER3 is detected as the contour of the front end surface 93 of the ridge 92 which is a linear pattern.

These boundary lines EL1 to EL3 and boundary lines ER1 to The position of ER3.

Here, in the inspection images P1, P2, the boundary lines EL1, EL2 extend in a straight line. However, in the inspection image P3, the dark area AL3 enters the bright area AC3, and the boundary line EL3 is curved in an arc shape. As a result, the boundary line EL3 is largely deviated from the extension of the boundary lines EL1 and EL2 (indicated by a dotted line in the inspection image P3), and it is determined to be a defect.

Similarly, the boundary line ER3 is curved in an arc shape with respect to the boundary lines ER1 and ER2, and the boundary line ER3 deviates greatly from the extension line of the boundary lines ER1 and ER2 (shown by a dotted line in the test image P3), and it is judged as defect.

[end of inspection]

The imaging process and inspection process as described above are sequentially carried out, and no defects are found in the area inspection and continuity inspection, and after the inspection is carried out throughout the linear pattern, that is, the protrusion 92, it is considered that the inspected workpiece 9 is "defect-free", and the defect is terminated. examine.

On the other hand, if a defect is found in any of the area inspection and the continuity inspection, the sequential execution of the imaging step and the inspection step is stopped at that point, and the defect is recorded for the workpiece 9 .

[Effects of Embodiment]

In the present embodiment, in the inspection process, the defect can be determined based on at least one of the continuity of the ridges 92 as a linear pattern appearing in the inspection image P and the image area, and in the defect inspection, it is not necessary to A reference image of the protrusion 92 is prepared in advance.

In addition, in the photographing process, the inspection images P are sequentially photographed along the protrusions 92 which are linear patterns, so the operator does not need to designate the inspection division.

Therefore, defect inspection of the workpiece 9 on which the protrusions 92 as a linear pattern are formed on the surface can be efficiently performed.

In the present embodiment, since the area inspection is adopted in the inspection process, the simple calculation of the areas of the pattern areas (bright areas AC1 to AC3 ) and the non-pattern areas (dark areas AL1 to AR3 ) in the inspection image P is performed. operation, it is possible to perform defect determination as an inspection process.

In particular, in the present embodiment, since the number of pixels in each region in each inspection image P1 to P3 is accumulated and compared with the number of pixels in other inspection images, defect determination as an inspection process can be performed more efficiently.

In this embodiment, since the continuity inspection is also performed in the inspection process, it is possible to perform defect inspection more reliably.

In the continuity inspection, the detection of the outline of the linear pattern in the inspection image P (the detection of the boundary line between the bright area and the dark area) and the judgment of its continuity can be performed as an inspection process. Defect determination.

In the present embodiment, in the imaging process, the inspection frames Pf are sequentially assigned along the protrusions 92 which are linear patterns, and a plurality of inspection images P are captured. Therefore, the above-mentioned area inspection and continuity suitable for the inspection process can be obtained. Inspection image for inspection.

In particular, in the present embodiment, it is possible to set a movement route during the operation of sequentially capturing the inspection images P with reference to the machining program 61 which is the same as the machining program 51 for machining the workpiece 9 . Therefore, it is not necessary to separately photograph the workpiece 9 or to give an instruction by manual operation in order to set the moving path.

Furthermore, in this embodiment, the camera 3 is mounted on the machine tool 2 that processes the workpiece 9, and the inspection images P are sequentially photographed by the camera 3, so that the workpiece 9 can be processed immediately after the protrusions 92 that are linear patterns are processed. Run a defect check on it.

Therefore, operations such as transferring the workpiece 9 for inspection after processing are not required. In addition, the machine tool 2 can also be used as the defect inspection device 1, and the equipment cost and equipment space can be reduced.

[Other Embodiments]

The present invention is not limited to the above-described embodiments, and modifications and the like within the range that can achieve the object of the present invention are included in the present invention.

In the above-described embodiment, both the area inspection and the continuity inspection are performed in the inspection process, but only one of them may be performed.

The areas of the pattern area (bright areas AC1 to AC3 ) and the two non-pattern areas (dark areas AL1 to AL3 and dark areas AR1 to AR3 ) in the inspection image P are not limited to those in the above embodiment.

For example, two out-of-pattern regions (dark regions AL1 to AL3 and dark regions AR1 to AR3 ) may be combined to calculate the areas of the bright and dark out-of-pattern regions.

In particular, when processing two regions, the inspection image P is divided into two regions. Therefore, only a certain area (number of pixels) of a bright pattern region and a dark non-pattern region can be calculated, and obtained from The other area can be calculated by subtracting the calculated area from the entire area (total number of pixels) of the inspection image P known by inspecting the frame Pf.

Also in such an area inspection, it is possible to detect a defect such as a curved outline on one side of the bright area AC3 as shown in FIG. 7 .

The continuity check is not limited to the inspection of the boundary lines between the pattern areas (bright areas AC1 to AC3 ) and the non-pattern areas (dark areas AL1 to AL3 and dark areas AR1 to AR3 ) in the image P as in the above-mentioned embodiment. To compare with the adjacent inspection image P.

For example, in the same inspection image P, the continuity of the boundary line appearing in the image may be inspected to detect a discontinuity corresponding to a defect.

In the above embodiment, the rectangular protrusion 92 shown in FIG. 2 is used as the linear pattern to be inspected. The inspection object in the above embodiment is not limited to a rectangle whose parts are made of straight lines, but may be a linear pattern whose part or whole is made of curves, for example, a circular protrusion 92A shown in FIG. 9 can be used.

In the above-described embodiment, in the imaging step, imaging is performed using the inspection frame Pf of a predetermined shape, and a plurality of inspection images P of a fixed size are acquired.

On the other hand, without using the inspection frame Pf at the time of imaging, an image of the shape and size corresponding to the linear pattern such as the protrusion 92 may be acquired and inspected in the inspection step.

However, by setting the plurality of inspection images P to have a constant shape and size as in the present embodiment, the inspection in the inspection process can be efficiently processed.

In the above-described embodiment, the imaging process is performed on a part of the protrusion 92 that is a linear pattern, and the inspection process is performed on the portion where the inspection image P has been imaged, so that the imaging process and the inspection process are performed in parallel, and the processing time can be shortened. time.

However, the imaging process may be performed sequentially for the protrusions 92 which are linear patterns, and the inspection process may be performed collectively after the inspection image P is obtained for the entire protrusions 92 .

In the above-described embodiment, in the imaging step, imaging is performed using the inspection frame Pf while moving the camera 3 along the protrusion 92 (linear pattern), and a plurality of inspection images P are acquired. In order to obtain a clear image in a normal video camera 3, it is necessary to stop at the shooting position. That is, the camera 3 repeats the operations of "moving-stopping-photographing-moving", which may increase the processing time.

On the other hand, as another embodiment of the present invention, flash photography can be used.

In FIG. 10 , this embodiment basically has the same structure as the embodiment of FIGS. 1 to 9 described above. However, a strobe 3A is provided adjacent to the camera 3 to intermittently illuminate the region Lf covered by the inspection frame Pf of the camera 3 with strong light.

In such this embodiment, the camera 3 is continuously moved along the protruding bar 92 , and the protruding bar 92 is illuminated intermittently by the strobe 3A during this period, and images are taken by the camera 3 during the lighting period.

Thereby, the protruding bar 92 is illuminated by the strong light of the strobe 3A, and the image captured by the camera 3 is clear. Furthermore, since the camera 3 does not need to be stopped, the processing time for shooting can be shortened. As a result, a clear image can be obtained at high speed without using a high-speed camera, and equipment cost can be reduced.

In the above-described embodiment, the machine tool 2 for processing the workpiece 9 is used as the defect inspection device 1 by attaching the tool 4 to the spindle 27 and attaching the camera 3 to the spindle head 26 .

However, instead of installing the camera 3 and the tool 4 at the same time on the machine tool 2, an interchangeable type may be used. For example, after the tool 4 is attached to the main shaft 27 and the workpiece 9 is processed, the tool 4 may be removed and the camera 3 may be attached to the main shaft 27 to image the workpiece 9 .

When the camera 3 is attached to the spindle 27 in this way, it is not necessary to attach the camera 3 to the spindle head 26, and the camera 3 can be easily attached using an automatic tool changer.

However, it is not efficient to frequently replace the tool 4 and the camera 3, so it is limited to a process in which the protruding lines 92 (line patterns) are sequentially photographed by the camera 3 after the tool 4 finishes processing the workpiece 9.

On the other hand, if the camera 3 and the tool 4 are installed on the working machine 2 as in the above-mentioned embodiment, then it is also possible to process a part of the workpiece 9, photograph the processed part, process the next part, and then perform Shooting, such a process.

In the above-described embodiment, the machine tool 2 is equipped with the camera 3 and used also as the defect inspection device 1 , but a dedicated defect inspection device 1 different from the machine tool 2 that processes the workpiece 9 may be used.

In this case, it is preferable to take out the processing program 51 executed by the control device 5 of the machine tool 2 or the shape (line pattern) of the protrusion 92 specified here, and store it in the dedicated defect inspection device 1 so that used in the process.

In the above-described embodiment, the same processing program 61 as the processing program 51 recorded in the control device 5 for processing the workpiece 9 is also recorded in the computer system 6. Referring to this processing program 61, the imaging process is performed as a linear process. The traces of the protrusions 92 of the pattern. However, if there is shape data of a linear pattern in addition to the machining programs 51 and 61 for machining the workpiece 9, this shape data can also be used.

Claims (8)

1. a defect detecting method, for being formed with the workpiece of linear pattern on surface, checks described The defect of linear pattern, it is characterised in that including:

Shooting operation, shoots check image successively along described linear pattern；And

Check operation, for described check image, carry out defect inspection successively,

In described inspection operation, based on the described linear pattern comprised in described check image continuous Property and image area at least defect is judged by certain.

2. defect detecting method as claimed in claim 1, it is characterised in that

In described inspection operation, the difference of optically-based characteristic, described check image is divided into table Show the area of the pattern of described linear pattern and the pattern exterior domain in the outside of described area of the pattern,

Calculate described area of the pattern and the area of described pattern exterior domain, and with check image other described Compare, if area does not change, be then judged to there is no defect.

3. defect detecting method as claimed in claim 1, it is characterised in that

In described inspection operation, the difference of optically-based characteristic, described check image is divided into table Show the area of the pattern of described linear pattern and the pattern exterior domain in the outside of described area of the pattern,

The boundary line detecting described area of the pattern and described pattern exterior domain is used as described linear pattern Profile,

If described profile is continuous print, then it is judged to there is no defect.

4. defect detecting method as claimed in claim 1, it is characterised in that

Described check image is specified by the check box of regulation shape,

In described shooting operation, any part of described linear pattern is assigned with initial described inspection After looking into frame, the adjacent regions of described linear pattern is sequentially allocated next described check box, thus claps Take the photograph the multiple described check image along described linear pattern.

5. defect detecting method as claimed in claim 4, it is characterised in that

Described workpiece is, is formed with wire by carrying out the processing unit (plant) of action according to processor on surface Pattern,

In described shooting operation, shape based on the described linear pattern comprised in described processor Data, make the position shooting described check image successively move.

6. defect detecting method as claimed in claim 5, it is characterised in that

The described processing unit (plant) processing described workpiece assembles video camera, is clapped by described video camera Take the photograph described check image.

7. the defect detecting method as according to any one of claim 1～6, it is characterised in that

In described shooting operation, while making video camera move along described linear pattern, by dodging Described linear pattern is illuminated by light modulation off and on, leads in the period being illuminated by described flash lamp Cross described video camera and shoot described check image.

8. a flaw detection apparatus, for being formed with the workpiece of linear pattern on surface, checks described The defect of linear pattern, it is characterised in that possess:

Shoot part, shoots check image successively along described linear pattern；And

Inspection portion, for described check image, carries out defect inspection successively,

Described inspection portion seriality based on the described linear pattern comprised in described check image and image Area at least defect is judged by certain.