CN115656192A

CN115656192A - Filter defect detection method and device

Info

Publication number: CN115656192A
Application number: CN202211673102.0A
Authority: CN
Inventors: 王志明; 邵云峰; 曹桂平; 董宁
Original assignee: Hefei Eko Photoelectric Technology Co ltd
Current assignee: Hefei Eko Photoelectric Technology Co ltd
Priority date: 2022-12-26
Filing date: 2022-12-26
Publication date: 2023-01-31

Abstract

The invention discloses a method and a device for detecting defects of a filter, comprising the following steps: the parallel light passing through the filter falls on a sensor of the camera; performing flat field correction on the camera, and storing a flat field correction coefficient; changing the relative pose between the camera or the filter and the parallel light so as to lead the position of the parallel light projected on the sensor through the filter to be relatively changed; the camera keeps the flat field correction coefficient unchanged, and acquires pictures of different relative poses between the camera or the filter and the parallel light; and preprocessing the picture, finding out an area with a changed gray value in the picture, and obtaining a flaw detection result according to the preprocessed picture. The method detects the flaw through the gray value change of the flaw position before and after flat field correction, can detect the flaw area which is difficult to be directly seen by human eyes, can detect the dust close to the background, and is suitable for the scene with higher detection precision.

Description

Filter defect detection method and device

Technical Field

The invention relates to the field of optical detection, in particular to a method for detecting defects of a filter.

Background

At present, the film-coated filter is widely applied to the fields of cameras and the like, plays roles in filtering stray light and protecting a sensor, can play a great role in the imaging effect of the camera when the surface of the film-coated filter has flaws, and can be greatly influenced in the fields of precision detection and the like. Therefore, in order to improve the imaging effect of the camera and improve the detection accuracy, accurate flaw detection needs to be performed on the filter.

Chinese patent CN115018829a proposes a method for locating a glass defect, which needs to compare a picture obtained by shooting a filter to be detected with a picture obtained by shooting a filter without a defect, and the detection precision depends on the cleanliness of the filter without a defect. Chinese patent CN115096912a proposes a method and system for imaging lens defects, which uses different image patterns to image different types of defects, and the detection accuracy is related to the image patterns, so that it is difficult to cover all types of defects.

In summary, the disadvantages of the prior art are as follows:

(1) The existing detection method usually only collects a picture at one angle and compares the picture with a picture collected by a reference flawless filter, so that the requirements on a light source, the reference filter and a collection device are high, and the problem caused by uneven illumination is difficult to completely avoid;

(2) As the detection is carried out, new dust inevitably falls on the sensor or the filter, and the existing method cannot filter out the dust or flaws;

(3) For defects close to the background, the prior art is difficult to distinguish and detect well.

Disclosure of Invention

The present invention provides a method and a device for detecting defects of a filter, which can solve at least one of the above technical problems.

In order to realize the purpose, the invention provides the following technical scheme:

a filter defect detection method, comprising:

the parallel light passing through the filter is projected on a sensor of the camera;

performing flat field correction on the camera, and storing a flat field correction coefficient;

changing the relative pose between the camera or the filter and the parallel light so as to lead the position of the parallel light projected on the sensor through the filter to be relatively changed;

the camera keeps the flat field correction coefficient unchanged, and acquires pictures of different relative poses between the camera or the filter and the parallel light;

and preprocessing the picture, finding out an area with a changed gray value in the picture, and obtaining a flaw detection result according to the preprocessed picture.

Further, the camera keeps the flat field correction coefficient unchanged, and acquires pictures of different relative poses between the camera or the filter and the parallel light, including:

adjusting the relative parallel light pose of the camera and the filter to be detected, enabling the parallel light to vertically penetrate through the filter and fall onto a camera sensor, enabling the flat field correction coefficient of the camera to be unchanged, and collecting a picture 1;

the incident angle of the parallel light is unchanged, the camera and the filter to be detected rotate left and right respectively by a set angle, the flat field correction coefficient is unchanged, and the camera respectively collects a picture 2 and a picture 3 after the rotation.

Further, the preprocessing the picture to find an area in which a gray value changes in the picture, and obtaining a flaw detection result according to the preprocessed picture include:

the picture 1 is respectively differenced with the picture 2 and the picture 3 to obtain two differential images;

and merging the defective areas on the two differential images, wherein the defective areas are areas with changed gray values, and obtaining a defect detection result.

Further, the method also comprises the following steps:

after the pictures 2 and 3 are collected, the pose of the camera and the filter to be detected relative to the parallel light is adjusted again, so that the parallel light vertically passes through the filter and falls onto a camera sensor;

taking down the filter to be tested, and collecting a picture 4;

and comparing the picture 4 with the pictures 1, 2 and 3, and judging whether new flaws are generated on the filter and the camera sensor in the detection process.

Further, the method also comprises the following steps:

before the camera is used for flat field correction, the light intensity of the parallel light and the exposure time of the camera are adjusted, so that the gray value of the picture acquired by the camera is located in a set interval.

Further, still include:

after the pictures 2 and 3 are collected, rotating the filter to be tested, wherein the rotation is that the filter rotates 180 degrees leftwards or rightwards in the plane of the filter;

the camera keeps the flat field correction coefficient unchanged, and acquires a picture 4, a picture 5 and a picture 6 at the same positions as the pictures 1, 2 and 3 respectively.

the picture 1 is respectively differenced with the picture 2 and the picture 3 to obtain two differential images, and defective areas on the two differential images are combined to obtain a picture 7;

the picture 4 is respectively differenced with the picture 5 and the picture 6 to obtain two differential images, and defective areas on the two differential images are combined to obtain a picture 8;

extracting repeated defective regions in the pictures 7 and 8, wherein the defective regions are regions with changed gray values, combining the repeated defective regions in the pictures 7 and 8 to obtain a picture 9, and obtaining defects on the sensor according to the picture 9.

Further, the method also comprises the following steps:

the picture 9 is respectively differenced with the pictures 7 and 8 to obtain two differential images, namely a picture 10 and a picture 11;

rotating the picture 10, wherein the rotation is that the picture 10 rotates 180 degrees leftwards or rightwards in the plane of the picture to obtain a picture 101;

extracting repeated defective regions in the pictures 101 and 11, combining the repeated defective regions in the pictures 101 and 11 to obtain a picture 12, and combining non-repeated defective regions in the pictures 101 and 11 to obtain a picture 13;

and obtaining the inherent flaws on the filter to be detected according to the picture 12, and obtaining the newly added flaws of the filter to be detected in the detection process according to the picture 13.

In another aspect, the present invention provides a filter defect detection apparatus, including:

the integrating sphere light source is used for generating uniform parallel light;

the camera is characterized in that a filter to be tested is arranged in front of the camera sensor; the filter to be tested is positioned between the integrating sphere light source and the camera;

the rotation angle table is used for fixing the camera and driving the camera to rotate within a set angle;

and the upper computer is used for calculating the defect detection result of the filter to be detected according to the filter defect detection method.

Further, install filter frame before the camera sensor, filter frame includes:

with camera fixed connection's framework, sliding connection has the filter support in the framework, detachably installs the filter that awaits measuring on the filter support, through the filter support slides and gets into in the framework, the corresponding position of camera is installed to the filter that will await measuring.

The invention has the following beneficial effects:

(1) The method detects the flaws through the gray value change of the flaws before and after flat field correction, can detect flaw areas which are difficult to be directly seen by human eyes, can detect dust close to the background, and is suitable for scenes with high detection precision;

(2) Because flat field correction is adopted, the requirement on a light source is reduced, and the cost of detection equipment can be reduced;

(3) The defects inherent to the filter and the defects generated in the detection process and the defects inherent to the sensor can be distinguished, and the reliability and the accuracy of detection are improved.

Drawings

FIG. 1 is a flow chart of a filter defect detection method of the present invention;

FIG. 2 is a structural diagram of a filter defect detecting apparatus according to the present invention;

fig. 3 is a view showing an installation structure of a filter in the filter defect detecting apparatus according to the present invention.

In the figure: 1-integrating sphere light source; 2-rotating the angle table; 3-large area camera; 4-a frame body; 5-a filter support; 6-a filter; 7-angular scale; 8-limiting column.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention.

Example 1

The embodiment provides a filter defect detecting device, which has an overall structure as shown in fig. 1 and includes:

integrating sphere light source 1 and rotation angle platform 2 install on the platform base, install the camera on the rotation angle platform 2, and the camera that this embodiment used is large area array camera 3, can carry out the detection of awaiting measuring filter 6 better. The filter 6 that awaits measuring is installed in front of big area array camera 3 sensor through the filter frame, and wherein, integrating sphere light source 1, filter mirror holder, big area array camera 3 are located the same axis, and the filter frame is located between integrating sphere light source 1 and the big area array camera 3.

Filter frame before big area array camera 3's the sensor is for the direction that conveniently changes filter 6 and change filter 6 and put into, a framework 4 and a filter support 5 of filtering the frame have been designed, framework 4 is fixed before big area array camera 3's sensor, it has a rectangle through-hole to hold filter 6 and make the light source pass through to have wherein, filter support 5 and framework 4 sliding connection, adopt hinged joint in this embodiment, filter 6 detachably installs on filter support 5, conveniently change filter 6 or change filter 6 and insert the direction of filter support 5. The filter support 5 drives the filter 6 to slide into the frame 4, and the filter 6 is installed in the frame 4 through a positioning structure in the frame 4. The filter frame designed by the embodiment is convenient for the installation and the taking-out operation of the filter 6 and can adapt to the detection of large batches of filters.

The integrating sphere light source 1 is used for emitting uniform light, so that the illumination stability is ensured as much as possible; and the integrating sphere light source 1 can control the intensity of a light field, adjust the gray value of an image, control light to irradiate through the lens to be detected, and ensure that the illumination of an image acquisition area is as uniform as possible. In this embodiment, uniform parallel light is required to be used as a light source to complete subsequent detection, because after the parallel light passes through the filter 6, a flaw is projected on the sensor at only one fixed position, and when the relative position relationship between the parallel light and the filter 6 is changed, the position is also changed, but the light source of this embodiment may be other light source devices emitting parallel light, and therefore other light sources capable of emitting parallel light should also be within the scope of the present invention.

The rotation angle platform 2 is used for bearing the large-area-array camera 3 and driving the large-area-array camera 3 to rotate at a small angle. The large-area-array camera 3 is used for collecting filter images according to a set collection strategy. The filter frame is used for placing the filter 6, and the disassembly and assembly machine trouble brought when the filter 6 is replaced is avoided.

The system also comprises an upper computer which is used for executing the following detection method and calculating a detection result according to the filter image collected by the large-area-array camera 3.

As shown in fig. 2 and fig. 3, the detection tool of the filter defect detection imaging device of the present embodiment includes an angle scale 7 on a rotation angle table 2, and is used for fixing the large area-array camera 3 and driving the large area-array camera 3 to rotate; the angle scale 7 and the rotation angle platform 2 are rotatably connected, a through hole is formed in the angle scale 7, the limiting column 8 on the rotation angle platform 2 penetrates through the through hole, and the large area array camera 3 can be limited to rotate leftwards and rightwards in the same angle through the limiting column 8 and the through hole. The large-area-array camera 3 is fixedly connected with the angle scale 7 through a fixing device; wherein big area array camera 3 does not install the camera lens, strains the mirror holder in the installation of camera lens adapter ring department, and the filter holder is next to big area array camera 3's sensor, strains and places the filter 6 that waits to detect on the mirror holder for the image of gathering possesses approximate even background, embodies on the image, then is: a10 x 10 pixel area is selected on the image, and the average difference of the gray value of the flaw part when the flaw part is not processed and the gray value of the image background is in the range of 2-3 pixels.

The detection principle of the embodiment is as follows:

the defects of the filter 6 include dust falling on the filter 6 and defects generated during the production, manufacture, transportation and taking processes of the filter 6, which affect the passing of light, when the incident light is parallel light, the defects on the filter 6 weaken the light intensity passing through the defect area, causing the optical response of the defect area projected onto the sensor to be small, and if the defects are small, the difference of the optical response is difficult to be distinguished by common detection means. The defect point judgment by the flat field correction method can amplify the difference of the optical response, the resolution ratio of the difference can be improved under the acquisition precision of the existing sensor, in addition, the influence of uneven illumination on the detection result can be eliminated after the flat field correction, the requirement of a light source is reduced, and the cost of the detection equipment is reduced. And the amplification principle of flat field correction of the invention is to amplify the response difference from the level of the sensor by a physical method, rather than by mathematical amplification with the aid of software, so that the detection method of the invention can identify the defect points which are very close to the background.

When the large area-array camera 3 is over against the light source, the large area-array camera 3, the filter 6 to be measured and the integrating sphere light source 1 are on the same axis, at this time, the light source passes through the filter 6 and is projected onto the sensor, and a connecting line of a flaw point on the filter 6 and a corresponding projection area on the sensor is perpendicular to the filter 6 to be measured. At this time, flat field correction is performed, the flat field correction is used for eliminating response nonuniformity of each pixel point, the flat field correction is a mature technology, the flat field correction uses a flat field correction coefficient to realize correction of response nonuniformity of different pixel points, and the correction coefficients of different pixel points are different. The light passing through the defect area is dark, so that the correction coefficient of the corresponding pixel point area on the sensor is large; at this time, the camera is rotated by a certain angle through the rotation angle table 2, a connecting line between the defect point on the filter 6 and the corresponding projection area on the sensor is no longer perpendicular to the filter 6 to be measured, the projection area is also deflected relative to the central position, the original projection area of the defect point is projected without defect, but the position is lightened relative to the background due to the unchanged correction coefficient, the new projection area of the defect point is darkened, and the image is represented as two circles with prominent light and shade. The correction coefficients of the flat field correction to the non-defective area and the defective area are different, the coefficients after correction are unchanged, the position of the defect moves in the process of rotating the camera, the more prominent gray value change occurs between the new position and the old position of the defect due to the difference of the correction coefficients, the position of the defect is found by capturing the change, and the position of the defect is judged. However, the single-angle rotation increases the risk of missing inspection, so that the detection needs to be rotated to the left and the right by the same angle respectively to prevent some special defect points from being unidentified.

The detection flow of this embodiment is:

the filter 6 to be tested is inserted into the filter support 5, the filter support 5 is rotated, and the filter 6 to be tested is mounted on the filter frame.

The square intensity of the integrating sphere light source 1 and the exposure time of the large-area-array camera 3 are adjusted to enable background illumination to be uniform, and the gray value of the image collected by the large-area-array camera 3 is in a set reasonable interval, so that subsequent flaw detection is facilitated.

The large-area-array camera 3 is over against the integrating sphere light source 1, the over-against means that parallel light emitted by the integrating sphere light source 1 vertically penetrates through the filter 6 to be measured to reach the sensor, at the moment, flat field correction is carried out on the camera, a flat field correction coefficient is stored, and the flat field correction coefficient is used by the large-area-array camera 3 behind to obtain images when shooting at different angles.

The large-area-array camera 3 respectively collects images which are just opposite to the light source, rotate by a preset angle at left and rotate by a preset angle at right, and respectively records the images as a picture 1, a picture 2 and a picture 3; in addition, when the large-area-array camera 3 is over against the camera, the filter 6 to be measured is taken out, and a filter-free picture N is acquired. The preset angle setting process comprises the following steps: after rotating a certain angle left, the image that big area array camera 3 gathered because flat field correction coefficient is unchangeable, original flaw area can brighten, and the new position of flaw area projection then can become dark, forms two circles of light and shade, and when these two circles are tangent, the angle of big area array camera 3 rotation this moment accords with the requirement, and the angle of setting to the dextrorotation is the same with the angle of setting of rotation left.

The acquired image is subjected to differential processing, and then the image is analyzed and processed through a connected domain, so that the flaws of the filter 6 and the position information of the flaws are obtained. The difference image reflects the change of the gray level of each pixel point, and the defect region and the background region can be distinguished by the change threshold, and the present embodiment aims to find the region in which the gray level value changes in the image, and obtain the defect detection result after processing, so if the change of the gray level value is analyzed by other methods to obtain the defect information, it should be considered as the range described in this specification.

Specifically, the picture 1 is subjected to difference processing with the pictures 2 and 3 respectively to obtain a difference picture 1 and a difference picture 2; combining the defective areas on the two differential images, wherein the defective areas are areas with changed gray values; further processing the image from the perspective of the connected domain to obtain a binary image, and if the binary image is obtained by other methods, considering the binary image as the range described in the present specification, analyzing to obtain the defect detection result on the filter 6, including information such as the shape, size and position of the defect; and further judging whether the filter 6 is qualified or not, and detecting dust or flaw points which are very close to the background.

Finally, comparing the picture 1, the picture 2 and the picture 3 with the picture N, and observing whether new defects are generated on the sensor and the filter 6 to be detected, wherein the new defects comprise falling dust or other new defects generated in the detection process.

Example 2

In the embodiment 1, the filter 6 to be detected and the camera are rotated only twice, so that the defect of the filter 6 can be detected, but in the detection process, new dust intervention or new defects are inevitably generated on the filter 6 to be detected and the sensor of the large-area-array camera 3. Therefore, in embodiment 2 of the present invention, on the basis of embodiment 1, an improved detection method is provided to solve the above-mentioned problems by using the same detection device.

The detection process of embodiment 2 of the invention is as follows:

after the large-area camera 3 is connected, an image is collected according to 12 bits, the image is observed through the lower 8 bits, the proper exposure time and the light field intensity of the integrating sphere light source 1 are adjusted until the difference between the flaw and the background can be obviously seen, and the flaw close to the background is difficult to see and needs further processing. When saving the image, the lower 8-bit image is saved.

Collecting images, which comprises the following steps:

(1) Collecting the same picture 1, picture 2 and picture 3 according to the same method in the embodiment 1;

(2) Taking out the filter 6 to be tested from the filter frame, rotating and then putting the filter frame into the filter frame again, wherein after the rotation, the normal vector of the plane where the filter 6 is located rotates by 180 degrees;

(3) The flat field correction coefficients of the rest of the device and the camera are unchanged except for the rotating filter 6, and then pictures 4, 5 and 6 are taken at the same positions as pictures 1, 2 and 3 described above.

The picture 1 is respectively differentiated from the picture 2 and the picture 3 to obtain two differential images, namely the differential picture 1 and the differential picture 2, the differential images reflect the change size of the gray level of each pixel point, the defect area and the background area can be distinguished through a change threshold, the defect areas on the two differential images are combined to obtain a first defect image, namely the picture 7, and in order to obtain a more obvious boundary area, the down-sampling operation is further performed on the picture 7.

The picture 4 is respectively differentiated from the picture 5 and the picture 6 to obtain two differential images, the differential picture 3 and the differential picture 4 can distinguish a flaw area from a background area through a change threshold value, the flaw areas on the two differential images are combined to obtain a second flaw image, namely the picture 8, and the down-sampling operation is also carried out on the picture 8.

And finding out the repeatedly existing defect areas on the pictures 7 and 8, and obtaining a picture 9 after being independently proposed. In the rotating process, the corresponding relation between the filter 6 and a single pixel point on the sensor changes, but the defect of the sensor does not change, so that the area without the change defect is the defect on the sensor, namely the picture 9 is the defect image on the sensor.

The pictures 7 and 8 are respectively differentiated from the picture 9, defects on the sensor are filtered to obtain a picture 10 and a picture 11, the picture 10 is rotated to obtain a picture 101, the filter 6 is rotated too much when the picture is collected, the picture is rotated back again, a normal vector of a plane where the picture 10 is located after rotation is rotated 180 degrees, and the rotation is opposite to the rotation of the filter 6. Then, the picture 101 is compared with the picture 11, repeated flaws are extracted and recorded as a picture 12, flaws represented by the picture 12 are flaws on the filter 6, and the remaining flaws that do not occur repeatedly are flaws added in the process of replacing the filter 6 and recorded as a picture 13.

And finally, analyzing whether the filter 6 detected this time is qualified or not according to a set judgment standard by combining the picture 12, the picture 13 and the picture 9.

The pictures 12, 13, and 9 obtained by the method of this embodiment can respectively reflect the inherent defects on the filter 6, the defects added to the filter 6 during the detection process, and the defects on the sensor. Compared with the detection method in the embodiment 1, the method can more accurately distinguish different types of flaws, and the information of the detection result is richer.

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting defects of a filter, comprising:

the parallel light passing through the filter is projected onto a sensor of the camera;

changing the relative pose between the camera or the filter and the parallel light so as to enable the position of the parallel light projected on the sensor through the filter to be relatively changed;

2. The method for detecting the filter defects according to claim 1, wherein the camera keeps the flat field correction coefficient unchanged, and the method for acquiring the pictures of different relative poses between the camera or the filter and the parallel light comprises the following steps:

3. The method of claim 2, wherein the preprocessing the picture to find a region with a changed gray scale value in the picture and obtaining a defect detection result according to the preprocessed picture comprises:

4. The filter defect detection method according to claim 2, further comprising:

taking down the filter to be tested, and collecting a picture 4;

5. The method of detecting filter defects according to claim 2, further comprising:

6. The filter defect detection method according to claim 2, further comprising:

after the pictures 2 and 3 are collected, rotating the filter to be detected, wherein the rotation is that the filter rotates 180 degrees leftwards or rightwards in the plane of the filter;

7. The method of claim 6, wherein the preprocessing the picture to find an area of the picture where a gray scale value changes and obtaining a defect detection result according to the preprocessed picture comprises:

extracting repeated defective areas in the pictures 7 and 8, wherein the defective areas are areas with changed gray values, combining the repeated defective areas in the pictures 7 and 8 to obtain a picture 9, and obtaining defects on the sensor according to the picture 9.

8. The filter defect detection method according to claim 7, further comprising:

9. A filter defect detecting device, comprising:

the camera is characterized in that a filter to be detected is arranged in front of the camera sensor; the filter to be tested is positioned between the integrating sphere light source and the camera;

the upper computer is used for calculating the defect detection result of the filter to be detected according to the filter defect detection method of any one of claims 1 to 8.

10. The filter defect detecting apparatus according to claim 9, wherein a filter frame is installed in front of the camera sensor, the filter frame comprising:

with camera fixed connection's framework, sliding connection has the filter support on the framework, detachably installation filter that awaits measuring on the filter support, through the filter support slides and gets into in the framework, the corresponding position of camera is installed to the filter that will await measuring.