JP2025043484A - Image processing method, lens inspection method, and lens inspection device - Google Patents

Abstract

[Problem] To provide a technology that can accurately identify the center coordinates of a lens in a captured image. [Solution] A captured image Ia includes a thin line portion 911 that follows the contour of the lens, and a wide portion 912 that is wider in the radial direction than the thin line portion 911. First, of the thin line portion 911 and the wide portion 912, a thin line erased image is generated by erasing the thin line portion 911. Next, a difference image between the captured image Ia and the thin line erased image is generated. After that, the center coordinates of the lens are identified based on the thin line portion 911 included in the difference image. This makes it possible to accurately identify the center coordinates of the lens in the captured image Ia while suppressing the influence of the wide portion 912. [Selected Figure] Figure 6

Description

The present invention relates to an image processing method for processing captured images of a lens, a lens inspection method, and a lens inspection device.

Conventionally, there is known an inspection device that photographs a transparent lens and inspects the lens for scratches and foreign matter adhering to the lens based on the photographed image. With this type of inspection device, because the lens is transparent, it is difficult to clearly photograph all areas of the lens in one photograph. Therefore, there are cases where the lens is rotated relative to the camera and lighting unit, photographs are taken multiple times, and defects are detected based on a composite image of the multiple images obtained.

In this case, in order to accurately synthesize the lenses, it is necessary to accurately identify the center coordinates of the lenses in each captured image. A technique for determining the center coordinates of the lenses in a captured image is described, for example, in Patent Document 1. The technique in Patent Document 1 uses a Hough transform to estimate the center coordinates of a circle (paragraph 0056 of Cited Document 1).

JP 2015-18324 A

However, the lens to be inspected may have a support, chuck, or other member connected to part of its contour. In such cases, the part of the lens contour in the captured image will have a shape that deviates from a circular arc, making it difficult to accurately identify the center coordinates of the lens using a Hough transform or the like.

The present invention has been developed in consideration of these circumstances, and aims to provide a technology that can accurately identify the center coordinates of a lens in a captured image.

The first invention of the present application is an image processing method for processing a captured image of a lens, and includes a thin line erasing step for generating a thin line erased image by erasing thin line portions that follow the contour of the lens in the captured image and wide portions that are wider in the radial direction than the thin line portions; a difference image generating step for generating a difference image between the captured image and the thin line erased image; and a center identifying step for identifying the center coordinates of the lens based on the thin line portions included in the difference image.

The second invention of the present application is the image processing method of the first invention, in which the center identification step identifies the center coordinates by applying a Hough transform to the difference image.

The third invention of the present application is the image processing method of the first or second invention, in which the thin line erasing step performs contraction and expansion on the thin line portions and the wide portions in the captured image.

The fourth invention of the present application is an image processing method according to any one of the first to third inventions, further comprising an image synthesis step of executing the thin line erasing step, the difference image generating step, and the center identifying step for each of the plurality of captured images, and synthesizing the plurality of captured images with the center coordinates identified in the center identifying step aligned.

The fifth aspect of the present invention is the image processing method of the fourth aspect, in which the image synthesis step generates a maximum value image of the multiple captured images.

The sixth invention of the present application is a lens inspection method, which includes an inspection step for detecting defects in the lens based on the maximum value image generated by the image processing method of the fifth invention.

The seventh invention of the present application is the lens inspection method of the sixth invention, further comprising a ring-shaped figure arrangement step of arranging a ring-shaped figure centered on the central coordinates and following the contour of the lens in the maximum value image, a first area calculation step of calculating a first area, which is the area of the overlapping area between the ring-shaped figure and the high brightness area of the maximum value image, a second calculation step of calculating a second area, which is the area of the overlapping area between the high brightness area and the ring-shaped figure while gradually decreasing the radius of the ring-shaped figure, and a contour mask creation step of creating a contour mask area that follows the contour of the lens based on the ring-shaped figure when the ratio of the second area to the first area becomes approximately the same as a predetermined reference value, and the contour mask area is excluded from inspection in the inspection step.

The eighth invention of the present application is a lens inspection method according to the sixth or seventh invention, further comprising an outer mask creation step of creating an outer mask area by taking the logical sum of the high brightness area in the maximum value image and the outer area of a circular ring centered on the central coordinates and following the contour of the lens, and the outer mask area is excluded from inspection in the inspection step.

The ninth invention of the present application is the lens inspection method of the eighth invention, in which in the outer mask creation process, a non-masked area that is isolated within the outer mask area and has a predetermined area or less is changed to an outer mask area.

The tenth invention of the present application is a lens inspection method according to the eighth or ninth invention, in which the image synthesis step further generates a minimum value image of the multiple captured images, and the inspection step includes a first inspection step of detecting defects in the lens in an area other than the outer mask area in the maximum value image, and a second inspection step of detecting defects in the lens inside a circular ring shape centered on the central coordinates in the minimum value image and following the contour of the lens.

The eleventh invention of the present application is a lens inspection device for inspecting lenses, comprising a camera for photographing a lens to obtain a photographed image, and a computer for processing the photographed image, the computer executing a thin line erasing process for generating a thin line erased image by erasing thin line portions that follow the contour of the lens in the photographed image and wide portions that are wider in the radial direction than the thin line portions, a difference image generating process for generating a difference image between the photographed image and the thin line erased image, and a center identifying process for identifying the center coordinates of the lens based on the thin line portions included in the difference image.

According to the first to eleventh aspects of the present application, by utilizing the thin line portions contained in the difference image, the influence of the wide portions can be suppressed and the center coordinates of the lens in the captured image can be identified with high accuracy.

In particular, according to the fourth aspect of the present invention, by aligning the center coordinates that are determined with high precision, multiple captured images can be synthesized with high precision.

In particular, according to the fifth aspect of the present invention, a maximum value image of multiple captured images can be generated with high accuracy.

In particular, according to the sixth aspect of the present invention, lens defects can be detected with high accuracy based on a maximum value image that is generated with high accuracy.

In particular, according to the seventh aspect of the present invention, the outline of the lens in the maximum value image can be excluded from inspection.

In particular, according to the eighth aspect of the present invention, the area outside the lens in the maximum value image and the chuck portion that grips the peripheral edge of the lens can be excluded from inspection.

In particular, according to the ninth aspect of the present invention, the entire chuck portion in the maximum value image can be excluded from the inspection target.

In particular, according to the tenth aspect of the present invention, by performing inspection using a maximum value image and an inspection using a minimum value image, it is possible to reduce the number of defects that go undetected.

FIG. 1 is a diagram showing a configuration of a lens inspection device. FIG. 2 is a top view of a lens product held in a holder. 11 is a flowchart showing a flow of an inspection process. FIG. 13 is a diagram showing an example of a captured image. 10 is a flowchart showing a process for identifying the center coordinates of a lens in a captured image. FIG. 13 is a diagram showing an example of a captured image from which unnecessary portions have been erased. FIG. 13 is a diagram showing an example of a thin line erased image in which thin lines in a captured image have been erased. FIG. 13 is a diagram showing an example of a difference image. 11 is a flowchart showing a flow of a process for combining photographed images. 13 is a flowchart showing a procedure for creating a contour mask region. FIG. 13 is a diagram showing a state in which a circular ring is arranged in a maximum value image. FIG. 13 is a diagram showing an example of a contour mask region created on a maximum value image. 13 is a flowchart showing a procedure for creating an outer mask region. FIG. 13 is a diagram showing an example of a binarized maximum value image. FIG. 13 is a diagram showing an example of an outer mask region. FIG. 13 is a diagram showing an example of an outer mask region after the process of step S74 is performed. FIG. 13 is a diagram showing an example of an outer mask region after the process of step S75 is performed. 1 is a flowchart showing a flow of an inspection process. FIG. 13 is a diagram showing an example of a maximum value image to which a contour mask region and an outer mask region are applied. FIG. 13 is a diagram showing an example of a minimum value image.

A preferred embodiment of the present invention will be described below with reference to the drawings.

1. Configuration of lens inspection device
Fig. 1 is a diagram showing the configuration of a lens inspection device 1 according to an embodiment of the present invention. This lens inspection device 1 is an appearance inspection device that inspects a lens product 9, which is an object to be inspected, based on an image captured by a camera 40. The lens product 9 is, for example, a lens that is attached to the human body, such as an intraocular lens or a contact lens. The lens product 9 may also be a lens used in an optical instrument. The case where the lens product 9 is an intraocular lens will be described below.

As shown in FIG. 1, the lens product 9 is held by a holder 90. FIG. 2 is a top view of the lens product 9 held by the holder 90. As shown in FIG. 2, the lens product 9 has a lens 91 and two support parts 92. The lens 91 is a transparent member that can transmit light. The lens 91 has a substantially circular shape when viewed from above. The support parts 92 are members for fixing the lens 91 in the eye. The support parts 92 extend outward from the peripheral edge of the lens 91. The two support parts 92 are located on opposite sides of the center of the lens 91.

As shown in Figs. 1 and 2, the holder 90 has a flat holder body 901 and three chuck portions 902. The holder body 901 has a first through hole 903. The first through hole 903 penetrates the holder body 901 in the vertical direction. The first through hole 903 has a circular shape when viewed from above. The three chuck portions 902 are provided at intervals around the first through hole 903 on the upper surface of the holder body 901. Each chuck portion 902 grips the peripheral portion of the lens 91. This holds the lens 91 in a position above the first through hole 903.

As shown in FIG. 1, the lens inspection device 1 includes a stage 10, a rotation mechanism 20, an illumination unit 30, a camera 40, and a computer 50.

The stage 10 is a platform that supports a holder 90 that holds a lens product 9. The holder 90 is placed on the upper surface of the stage 10. The stage 10 also has a holding mechanism 11 that holds the holder 90. The holding mechanism 11, for example, clamps the holder 90 from both sides in the horizontal direction. However, the holding mechanism 11 may also be one that adsorbs the lower surface of the holder 90 by negative pressure. By holding the holder 90 on the stage 10, the position of the lens product 9 relative to the stage 10 is fixed.

As shown in FIG. 1, the stage 10 has a second through hole 12. The second through hole 12 passes through the stage 10 in the vertical direction. When inspecting the lens product 9, the holder 90 is positioned on the stage 10 so that the first through hole 903 of the holder 90 overlaps with the second through hole 12 of the stage 10.

The rotation mechanism 20 is a mechanism that rotates the stage 10 around an axis 8 that extends in the vertical direction. The rotation mechanism 20 has a motor 21. The stage 10 is connected to the motor 21 via a power transmission mechanism such as a gear. When the motor 21 is driven, the stage 10 rotates around the axis 8 that extends in the vertical direction. This allows the lens 91 to rotate around the axis 8 relative to the camera 40 and the lighting unit 30. The rotation mechanism 20 is capable of rotating the stage 10 by a predetermined angle θ, which will be described later, based on a signal supplied from the computer 50.

The illumination unit 30 is a light source that irradiates the lens 91 with illumination light for photographing. For example, an LED illumination is used for the illumination unit 30. The illumination unit 30 is disposed below the second through-hole 12 of the stage 10. The illumination unit 30 emits illumination light upward. The illumination light emitted from the illumination unit 30 passes through the second through-hole 12 and the first through-hole 903 and is irradiated to the lens 91.

The illumination unit 30 may be composed of a light source arranged to the side of the stage 10 and an optical system such as a mirror that guides the light emitted from the light source to the lens 91.

The camera 40 is a unit that photographs the lens 91 on the holder 90 supported by the stage 10. The camera 40 is positioned above the stage 10 at a distance. The camera 40 has an image sensor with an imaging element such as a CCD or CMOS, and an optical system that focuses the light incident on the camera 40 on the image sensor. The camera 40 photographs the lens 91 to obtain a multi-tone photographed image Ia composed of a large number of pixels. The camera 40 outputs the obtained photographed image Ia to the computer 50.

The computer 50 is a unit that controls the operation of each part of the lens inspection device 1 and inspects the lens 91 based on the captured image Ia. As shown in FIG. 1, the computer 50 has a processor 51 such as a CPU, a memory 52 such as a RAM, and a storage unit 53 such as a hard disk drive. The storage unit 53 stores a control program P1 for controlling the operation of each part of the lens inspection device 1 and an inspection program P2 for inspecting the lens product 9 based on the captured image Ia. The control program P1 and the inspection program P2 may be stored in a storage medium such as an optical disk.

As shown in FIG. 1, the computer 50 is electrically connected to the rotation mechanism 20, the lighting unit 30, and the camera 40. The processor 51 of the computer 50 controls the operation of the rotation mechanism 20, the lighting unit 30, and the camera 40 by executing a control program P1. This causes the lens 91 to be photographed. The processor 51 of the computer 50 also inspects the lens 91 by processing the captured image Ia output from the camera 40 based on an inspection program P2.

<2. Inspection procedure>
Next, an explanation will be given of the inspection of the lens product 9 by the above-mentioned lens inspection device 1. Fig. 3 is a flowchart showing the flow of the inspection process. The process in Fig. 3 is executed by the computer 50 operating in accordance with the control program P1 and the inspection program P2.

<2-1. Shooting with lenses>
When inspecting the lens 91, first, the lens product 9 is set in the holder 90. Then, the holder 90 holding the lens product 9 is set on the stage 10. Thereafter, the lens inspection device 1 irradiates the lens 91 with illumination light from the illumination unit 30 while photographing the lens 91 with the camera 40 (step S1). This results in a photographed image Ia of the lens 91 using transmitted light. The obtained photographed image Ia is input from the camera 40 to the computer 50.

Figure 4 is a diagram showing an example of the captured image Ia. As shown in Figure 4, the captured image Ia includes low-luminance areas (parts shown in gray in Figure 4) and high-luminance areas (parts shown in white in Figure 4) that are brighter than the low-luminance areas. Because the lens 91 is transparent, most of the lens 91 is a low-luminance area. However, the outline of the lens 91 and a defect d on the lens 91 are high-luminance areas. In addition, the support portion 92 and the peripheral portion of the chuck portion 902 are also high-luminance areas.

When the image capturing of the lens 91 is completed, the computer 50 determines whether the number of images captured has reached a predetermined number N (step S2). If the number of images captured has not reached the predetermined number N (No in step S2), the rotation mechanism 20 rotates the stage 10 by a predetermined angle θ around the axis 8. When the stage 10 rotates, the lens 91 rotates by the predetermined angle θ together with the stage 10 and the holder 90 (step S3).

The above-mentioned predetermined angle θ is set so that one rotation (360°) is completed by repeating the rotation the above-mentioned predetermined number N. For example, if the predetermined number N is 8 times, the predetermined angle θ is set to 45°.

After rotating the lens 91 by the predetermined angle θ, the process returns to step S1, and the lens inspection device 1 again photographs the lens 91 using the camera 40. The lens inspection device 1 repeats the above-described processes of steps S1 to S3 until the number of photographs reaches the predetermined number N.

When the number of times the lens 91 has been photographed reaches the predetermined number N (Yes in step S2), the lens inspection device 1 ends photographing the lens 91. This results in a predetermined number N of photographed images Ia (one revolution's worth of photographed images) obtained by changing the rotation angle of the lens 91 by the predetermined angle θ.

It is difficult to obtain a uniform captured image Ia in a single capture due to biased illumination light, etc. For this reason, in this lens inspection device 1, as described above, a predetermined number N of captured images Ia are obtained by repeatedly capturing images while rotating the lens 91 relative to the illumination unit 30 and camera 40. The predetermined number N of captured images Ia are stored in the memory unit 53 of the computer 50 with the rotation angles aligned. The computer 50 inspects the lens 91 by performing image processing based on the predetermined number N of captured images Ia.

<2-2. Identifying the center coordinates>
The axis 8 of rotation of the stage 10 by the rotation mechanism 20 does not necessarily coincide with the center of the lens 91. Therefore, in order to accurately synthesize the captured images Ia, which will be described later, the computer 50 first identifies the center coordinates O of the lens 91 in each of the predetermined number N of captured images Ia (step S4). Fig. 5 is a flowchart showing the flow of a process for identifying the center coordinates O of the lens 91 in one captured image Ia.

First, the computer 50 erases unnecessary parts in the captured image Ia (step S41). That is, parts of the captured image Ia that are not necessary for identifying the center coordinates O are erased. FIG. 6 is a diagram showing an example of the captured image Ia from which unnecessary parts have been erased. In the subsequent processing of steps S42 to S45, the center coordinates O of the lens 91 are identified using a thin line portion 911 that follows the contour of the lens 91. Therefore, in step S41, as shown in FIG. 6, the parts that are sufficiently inside the contour of the lens 91 and the parts that are sufficiently outside the contour of the lens 91 are erased as unnecessary parts. Specifically, all unnecessary parts are made into low-brightness areas.

As shown in FIG. 6, the high-brightness area in the captured image Ia includes a thin line portion 911 and a wide portion 912 that follow the contour of the lens 91. The thin line portion 911 corresponds to the portion of the contour of the lens 91 to which the support portion 92 and the chuck 902 are not connected. The wide portion 912 corresponds to the portion of the contour of the lens 91 to which the support portion 92 or the chuck 902 is connected. The radial width of the wide portion 912 is wider than the radial width of the thin line portion 911.

Next, the computer 50 erases the thin line portion 911 from the above-mentioned thin line portion 911 and wide portion 912 in the captured image Ia (step S42: thin line erasing process). The thin line portion 911 can be erased, for example, by an opening process. Specifically, the computer 50 performs a contraction (thinning) process on the high-luminance region in the captured image Ia, and then performs an expansion (thickening) process. This makes the thin line portion 911 a low-luminance region. As a result, as shown in FIG. 7, a thin-line erased image Ib is obtained in which the thin line portion 911 in the captured image Ia has been erased.

Then, the computer 50 generates a difference image Ic by taking the difference between the thin line erased image Ib and the captured image Ia immediately before erasing the thin line portion 911 (step S43: difference image generation process). Figure 8 is a diagram showing an example of the difference image Ic. When the difference between the thin line erased image Ib and the captured image Ia is taken, the thin line portion 911 erased in step S42 appears, as shown in Figure 8. Therefore, in the difference image Ic, only the thin line portion 911 becomes a high brightness area, and the other parts become low brightness areas.

Then, the computer 50 identifies the center coordinates O of the lens 91 based on the thin line portion 911 included in the difference image Ic (step S44: center identification process). The center coordinates O can be identified, for example, by applying a Hough transform to the difference image Ic. That is, the center coordinates O are identified by performing voting processing in Hough space for multiple points on the thin line portion 911. By using the difference image Ic, it is possible to eliminate the wide portion 912 and perform a Hough transform based only on the thin line portion 911. Therefore, the influence of the wide portion 912 can be suppressed and the center coordinates O of the lens 91 in the captured image Ia can be accurately identified.

The computer 50 executes the above steps S41 to S44 for each of the predetermined number N of captured images Ia. This identifies the center coordinates O of the lens 91 for each captured image Ia.

<2-3. Combining captured images>
Next, the computer 50 synthesizes a predetermined number N of photographed images Ia (step S5: image synthesis step). Fig. 9 is a flow chart showing the flow of processing for synthesizing the photographed images Ia.

As shown in FIG. 9, in step S5, the computer 50 generates one maximum value image Id based on a predetermined number N of captured images Ia (step S51). The maximum value image Id is an image composed of the maximum brightness values of the predetermined number N of captured images Ia. In step S8, which will be described later, basically, the detection of defect d is performed based on the maximum value image Id.

In step S51, the computer 50 first aligns the center coordinates O of a predetermined number N of captured images Ia. Then, with the center coordinates O aligned, the computer 50 calculates the maximum luminance value of the predetermined number N of captured images Ia for each coordinate. Then, the computer 50 generates a maximum value image Id by applying the maximum luminance value to each coordinate of the image.

In a single capture, there may be unclear parts in the captured image Ia due to biased illumination light, etc. However, as described above, if a maximum value image Id is generated from a predetermined number N of captured images Ia with different rotation angles, the entire image becomes clear. Therefore, the detection of defect d in step S8 described below can be performed with high accuracy.

In addition, in this embodiment, in step S4, the center coordinates O of the lens 91 are accurately identified for each captured image Ia, and maximum value images Id of a predetermined number N of captured images Ia are generated with the identified center coordinates O aligned. This allows the maximum value images Id to be generated with high accuracy.

As shown in FIG. 9, in this embodiment, the computer 50 also generates one minimum value image Ie based on the predetermined number N of captured images Ia (step S52). The minimum value image Ie is an image composed of the minimum luminance values of the predetermined number N of captured images Ia. The minimum value image Ie is used in step S82, which will be described later, to suppress missed detection of defects d.

In step S52, the computer 50 first aligns the center coordinates O of a predetermined number N of captured images Ia. Then, with the center coordinates O aligned, the computer 50 calculates the minimum luminance value of the predetermined number N of captured images Ia for each coordinate. Then, the computer 50 applies the minimum luminance value to each coordinate of the image to generate a minimum value image Ie.

In this embodiment, in step S4, the center coordinates O of the lens 91 are accurately identified for each captured image Ia, and a minimum value image Ie of a predetermined number N of captured images Ia is generated with the identified center coordinates O aligned. This allows the minimum value image Ie to be generated with high accuracy.

Note that in the flowchart of FIG. 9, the maximum value image Id is generated first, followed by the minimum value image Ie, but these steps may be performed in the reverse order, or simultaneously.

<2-4. Creating a contour mask area>
When detecting defect d based on maximum value image Id, erroneous detection of defect d is likely to occur near the contour of lens 91. Therefore, computer 50 creates contour mask region M1 for excluding the vicinity of the contour of lens 91 in maximum value image Id from inspection (step S6). Fig. 10 is a flowchart showing the procedure for creating contour mask region M1.

As shown in FIG. 10, in step S6, the computer 50 first places a circular ring C in the maximum value image Id (step S61: circular ring placement process). The circular ring C is a ring-shaped figure centered on the central coordinate O and following the contour of the lens 91. The radius of the circular ring C that is placed first may be a value that is preset according to the design value of the lens 91. FIG. 11 is a diagram showing the state in which the circular ring C has been placed in the maximum value image Id. In FIG. 11, the circular ring C is indicated by a dashed line.

The computer 50 calculates the area of the overlapping area between the high brightness area in the maximum value image Id and the initially placed annular figure C as the first area A1 (step S62: first area calculation step). Then, the computer 50 stores the calculated first area A1 in the storage unit 53.

Next, the computer 50 calculates the area of the overlapping area between the high brightness area in the maximum value image Id and the annular figure C as the second area A2 while gradually decreasing the radius of the annular figure C (step S63: second area calculation step). The computer 50 then calculates the ratio A2/A1 of the second area A2 to the above-mentioned first area A1. As the annular figure C moves inward from the contour of the lens 91, the above-mentioned ratio A2/A1 gradually decreases.

A reference value of the above-mentioned ratio A2/A1 suitable for the contour mask region M1 is stored in advance in the memory unit 53 of the computer 50. The computer 50 gradually reduces the size of the annular shape C until the ratio A2/A1 of the second area A2 to the first area A1 becomes substantially the same as the reference value stored in the memory unit 53.

The computer 50 determines the annular shape C when the ratio A2/A1 is substantially equal to the reference value as the final annular shape C. More specifically, the computer 50 determines the previous annular shape C when the ratio A2/A1 is equal to or less than the reference value for the first time as the final annular shape C. The computer 50 creates a contour mask region M1 based on the final annular shape C (step S64: contour mask creation process).

Figure 12 is a diagram showing an example of a contour mask region M1 created on a maximum value image Id. As shown in Figure 12, the contour mask region M1 is a circular mask region that follows the contour of the lens 91. The portion of the maximum value image Id that is covered by the contour mask region M1 is not subject to inspection. The computer 50 creates the contour mask region M1, for example, by increasing the radial width of the final annular figure C.

As described above, the computer 50 does not create a contour mask region M1 for each of the predetermined number N of captured images Ia, but instead creates one contour mask region M1 for the combined maximum value image Id. In this way, the processing burden on the computer 50 can be reduced while excluding the area near the contour of the lens 91 from the inspection target. Also, in this embodiment, in step S4, the center coordinates O of the lens 91 are identified for each captured image Ia, and a contour mask region M1 is created for the maximum value image Id that is created with high accuracy based on the identified center coordinates O. This makes it possible to more accurately mask the area near the contour of the lens 91.

<2-5. Creating outer mask area>
In the maximum value image Id, there is no need to inspect the portion outside the lens 91. Therefore, the computer 50 creates an outer mask region M2 for excluding the portion outside the lens 91 in the maximum value image Id from the inspection target (step S7: outer mask creating step). Fig. 13 is a flowchart showing the procedure for creating the outer mask region M2.

As shown in FIG. 13, in step S7, the computer 50 first binarizes the maximum value image Id (step S71). Specifically, the maximum value image Id is binarized into a low brightness region and a high brightness region using a predetermined brightness value as a threshold. At that time, the brightness value of the high brightness region is set according to the brightness value of the outer mask region M2 to be created. FIG. 14 is a diagram showing an example of the binarized maximum value image If.

Next, the computer 50 places an annular figure C in the binarized maximum value image If (step S72). The placement of the annular figure C is performed in the same manner as in step S61 described above. However, the placed annular figure C may be the final annular figure C determined in steps S63 to S64 described above.

Then, the computer 50 creates an outer mask region M2 by taking the logical sum of the high-luminance region (black region in FIG. 14) of the binarized maximum value image If and the region outside the annular figure C (step S73). FIG. 15 is a diagram showing an example of the outer mask region M2. By taking the logical sum as described above, the chuck portion 902 that grips the peripheral portion of the lens 91 can also be included in the outer mask region M2.

Then, the computer 50 changes any isolated non-masked areas in the outer mask area M2 that are equal to or smaller than a predetermined area to the outer mask area M2 (step S74). For example, as shown in FIG. 15, if an isolated non-masked area remains inside the chuck portion 902, the computer 50 changes the non-masked area to the outer mask area M2. This makes it possible to exclude the entire chuck portion 902 in the maximum value image If from the inspection target. FIG. 16 is a diagram showing an example of the outer mask area M2 after the processing of step S74 has been performed.

Then, the computer 50 leaves only the area of the outer mask region M2 that is continuous with the outer end of the maximum value image If, and excludes the other areas from the outer mask region M2 (step S75). This makes it possible to exclude areas such as defect d that are located inside the contour of the lens 91 from the outer mask region M2. Figure 17 is a diagram showing an example of the outer mask region M2 after the processing of step S75 has been performed.

As described above, the computer 50 does not create an outer mask region M2 for each of the predetermined number N of captured images Ia, but instead creates one outer mask region M2 for the combined maximum value image Id. In this way, the processing burden on the computer 50 can be reduced while excluding the area outside the lens 91 from the inspection target. Also, in this embodiment, in step S4, the center coordinates O of the lens 91 are identified for each captured image Ia, and an outer mask region M2 is created for the maximum value image Id that is accurately created based on the identified center coordinates O. This makes it possible to more accurately mask the area outside the lens 91.

<2-6. Inspection>
Thereafter, the computer 50 detects a defect d in the lens 91 (step S8: inspection process). The defect d is, for example, a scratch on the lens 91 or a foreign matter attached to the lens 91. Fig. 18 is a flow chart showing the flow of the inspection process.

As shown in FIG. 18, in step S8, the computer 50 first detects defects d in the lens 91 based on the maximum value image Id (step S81: first inspection process). FIG. 19 is a diagram showing an example of a maximum value image Id to which the above-mentioned contour mask region M1 and outer mask region M2 are applied. The computer 50 excludes the contour mask region M1 and outer mask region M2 in the maximum value image Id from inspection. In other words, the computer 50 excludes the areas other than the contour mask region M1 and outer mask region M2 in the maximum value image Id from inspection in the maximum value image Id.

The computer 50 detects a defect d in the lens 91 in the above-mentioned inspection target area of the maximum value image Id. Specifically, the computer 50 detects a portion in the inspection target area in which the brightness value is higher than a predetermined threshold as a defect d.

Next, the computer 50 detects defects d in the lens 91 based on the minimum value image Ie (step S82: second inspection process). FIG. 20 is a diagram showing an example of the minimum value image Ie. The computer 50 places an annular figure C in the minimum value image Ie. The placement of the annular figure C is performed in the same manner as in step S61 described above. The computer 50 then determines the area inside the annular figure C as the inspection target area in the minimum value image Ie.

The computer 50 detects a defect d in the lens 91 in the above-mentioned inspection target area of the minimum value image Ie. Specifically, the computer 50 detects a portion in the inspection target area in which the brightness value is higher than a predetermined threshold as a defect d.

Some small defects d may not appear in the minimum value image Ie. For this reason, it is desirable to basically detect defects d based on the maximum value image Id. However, when inspecting the maximum value image Id in step S81, the outer mask region M2 may be set wide near the support portion 92, masking the defects d. This may result in defects d going undetected.

For example, in the maximum value image Id of FIG. 19, defect d in the portion surrounded by the dashed line is part of the outer mask region M2 and is therefore excluded from inspection. However, in the minimum value image Ie of FIG. 20, the support portion 92 is smaller than in the maximum value image Id, so the defect d can be included in the inspection target region. Therefore, defect d that could not be detected in the maximum value image Id can be detected in the minimum value image Ie. In this way, by using inspection using the maximum value image Id and inspection using the minimum value image Ie in combination, it is possible to reduce the number of defects d that go undetected.

The computer 50 integrates the inspection results based on the maximum value image Id and the inspection results based on the minimum value image Ie to output the final detection result of defect d (step S83). Specifically, the computer 50 determines the union of the defect d detected in the maximum value image Id and the defect d detected in the minimum value image Ie as the final detection result of defect d. The computer 50 stores the final detection result of defect d in the memory unit 53 and displays it on a display unit connected to the computer 50.

3. Modifications
Although one embodiment of the present invention has been described above, the present invention is not limited to the above embodiment.

In the above embodiment, the computer 50 identifies the center coordinates O by applying a Hough transform to the thin line portion 911 in the difference image Ic. However, the computer 50 may identify the center coordinates O by other methods, such as fitting or template matching, based on the thin line portion 911 in the difference image Ic.

In the above embodiment, the computer 50 generates a maximum value image Id and a minimum value image Ie from a predetermined number N of captured images Ia, and performs inspection using the maximum value image Id and inspection using the minimum value image Ie. However, the computer 50 may generate only the maximum value image Id from the predetermined number N of captured images Ia, and perform only inspection using the maximum value image Id.

In the above embodiment, the computer 50 sets both the contour mask region M1 and the outer mask region M2 for the maximum value image Id. However, the mask region set for the maximum value image Id may be only the outer mask region M2. Furthermore, the computer 50 may set a mask region other than the contour mask region M1 and the outer mask region M2 for the maximum value image Id.

In addition, to the extent that no inconsistencies arise, some of the elements appearing in the above embodiments may be deleted, or other publicly known elements may be added.

REFERENCE SIGNS LIST 1 Lens inspection device 9 Lens product 10 Stage 20 Rotation mechanism 30 Illumination unit 40 Camera 50 Computer 90 Holder 91 Lens 92 Support unit 902 Chuck unit 911 Thin line unit 912 Wide unit A1 First area A2 Second area C Ring shape Ia Photographed image Ib Thin line erased image Ic Difference image Id Maximum value image Ie Minimum value image M1 Contour mask area M2 Outer mask area O Center coordinates P1 Control program P2 Inspection program d Defect

Claims (11)

An image processing method for processing an image captured by a lens, comprising:
a thin line erasing step of erasing a thin line portion along a contour of the lens in the captured image and a wide portion having a radial width wider than that of the thin line portion to generate a thin line erased image;
a differential image generating step of generating a differential image between the captured image and the thin line erased image;
a center identifying step of identifying center coordinates of the lens based on the thin line portion included in the difference image;
An image processing method comprising the steps of: 2. The image processing method according to claim 1,
In the center specifying step, the center coordinates are specified by applying a Hough transform to the difference image. 3. The image processing method according to claim 1, further comprising:
In the thin line erasing step, shrinking and expanding are performed on the thin line portions and the wide portions in the captured image. 3. The image processing method according to claim 1, further comprising:
performing the thin line erasing step, the difference image generating step, and the center identifying step for each of the plurality of captured images;
The image processing method further comprises an image synthesis step of synthesizing the plurality of captured images while aligning the center coordinates identified in the center identification step. 5. The image processing method according to claim 4,
In the image synthesis step, a maximum value image of the plurality of captured images is generated. A lens inspection method comprising an inspection step of detecting defects in the lens based on the maximum value image generated by the image processing method according to claim 5. 7. A lens inspection method according to claim 6, comprising:
a ring-shaped figure arrangement step of arranging a ring-shaped figure along a contour of the lens, the ring-shaped figure being centered on the central coordinates, in the maximum value image;
a first area calculation step of calculating a first area which is an area of an overlapping region between the annular figure and a high luminance region of the maximum value image;
a second calculation step of calculating a second area, which is an area of an overlapping region between the high luminance region and the annular figure, while gradually decreasing a radius of the annular figure;
a contour mask creating step of creating a contour mask area along a contour portion of the lens based on the annular shape when a ratio of the second area to the first area becomes substantially equal to a predetermined reference value;
and
The lens inspection method, wherein the contour mask region is excluded from inspection in the inspection step. 7. A lens inspection method according to claim 6, comprising:
an outer mask creating step of creating an outer mask region by taking a region of a logical sum of a high luminance region in the maximum value image and an outer region of a circular ring shape centered on the central coordinates and following the contour of the lens,
The lens inspection method, wherein the outer mask area is excluded from inspection in the inspection step. 9. A lens inspection method according to claim 8, comprising:
In the outer mask creating step, an isolated non-masked area having a predetermined area or less in the outer mask area is changed to an outer mask area. 9. A lens inspection method according to claim 8, comprising:
In the image synthesis step, a minimum value image of the plurality of captured images is further generated,
The inspection step includes:
a first inspection step of detecting defects of the lens in a region other than the outer mask region in the maximum value image;
a second inspection step of detecting defects of the lens inside a circular ring shape centered on the central coordinate in the minimum value image and along a contour of the lens;
A lens inspection method comprising: A lens inspection apparatus for inspecting a lens, comprising:
a camera for capturing an image by photographing a lens;
A computer for processing the captured image;
Equipped with
The computer includes:
a thin line erasing step of erasing a thin line portion along a contour of the lens in the captured image and a wide portion having a radial width wider than that of the thin line portion to generate a thin line erased image;
a differential image generating step of generating a differential image between the captured image and the thin line erased image;
a center identifying step of identifying center coordinates of the lens based on the thin line portion included in the difference image;
A lens inspection device that performs the above steps.

Images