CN111226110A - Detection method and system - Google Patents
Detection method and system Download PDFInfo
- Publication number
- CN111226110A CN111226110A CN201880067184.XA CN201880067184A CN111226110A CN 111226110 A CN111226110 A CN 111226110A CN 201880067184 A CN201880067184 A CN 201880067184A CN 111226110 A CN111226110 A CN 111226110A
- Authority
- CN
- China
- Prior art keywords
- coherent light
- beams
- glass
- detected
- partial
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
Abstract
The application provides a detection system and a method. The system comprises: a coherent light source for generating coherent light; a spectroscope for dividing the coherent light into a plurality of beams of partial coherent light for illuminating an object to be measured; a reflector for adjusting the angle of the partially or fully partially coherent light; and the collecting detector is used for collecting a speckle image formed by the plurality of beams of the partial coherent light penetrating through the object to be detected or the plurality of beams of the partial coherent light being reflected by the object to be detected, and determining whether the glass has defects according to the speckle image. The scheme can be used for rapidly detecting the bent glass.
Description
The present application relates to the field of electronics, and in particular, to a detection method and system.
During the production process, various defects such as bubble streaks and stones are generated in the glass, and the defects affect the appearance quality of the glass, reduce the mechanical strength and the thermal stability of the light transmittance of the glass and cause a large amount of waste products and defective products. At present, the glass defect detection system mainly utilizes the methods of manual online detection, laser detection, molar interference principle and machine vision. The manual detection is easily influenced by subjective factors of detection personnel, and easily causes missing detection of glass defects, especially inclusion defects with small deformation and small distortion. The laser detection is easily interfered by the outside, and the detection precision is influenced; moire interference principle requires the grating to have high contrast of light and shade because of the finer Moire fringe in the grating, otherwise the error is larger; the machine vision-based method adopts CCD imaging technology and backlight type illumination, a light source is arranged on the back of glass, light rays penetrate into a camera through the glass to be detected, and when the glass contains impurities, the emergent light rays can change, so that the light detected on the target surface of a CCD camera also changes correspondingly.
However, the above detection method cannot detect a bent glass or is not efficient.
Disclosure of Invention
The application provides a detection method and a detection system, which can quickly detect bent glass.
In a first aspect, a detection system is provided, comprising:
a coherent light source for generating coherent light;
a spectroscope for dividing the coherent light into a plurality of beams of partial coherent light for illuminating an object to be measured;
the reflector is used for adjusting the irradiation angle of the partially or fully partially coherent light;
and the collecting detector is used for collecting a speckle image formed by the plurality of beams of the partial coherent light penetrating through the object to be detected or the plurality of beams of the partial coherent light being reflected by the object to be detected, and determining whether the glass has defects according to the speckle image.
Optionally, the collecting detector is specifically configured to collect a speckle image formed by the plurality of beams of the partially coherent light passing through the object to be measured or the plurality of beams of the partially coherent light being reflected by the object to be measured in a non-imaging manner.
Optionally, the collecting detector is further configured to determine whether the object to be detected has a defect according to the speckle image and through a deep learning neural network.
Optionally, the split coherent light includes a first split coherent light and a second split coherent light, and the curved portion includes a first curved portion and a second curved portion, where the first split coherent light is used for illuminating the first curved portion and the second split coherent light is used for illuminating the second curved portion.
Optionally, the reflector includes a first reflector and a second reflector, wherein the first reflector is used for adjusting an angle of the first curved portion irradiated by the first partial coherent light, and the second reflector is used for adjusting an angle of the second curved portion irradiated by the second partial coherent light.
Optionally, the system further includes a first concave lens and a second concave lens, the first concave lens is configured to diffuse the first incoherent light, and the second concave lens is configured to diffuse the second incoherent light.
Optionally, the system further comprises a beam expander for diffusing the portion of the coherent light that illuminates the flat portion of the object to be measured.
Optionally, the system further includes a prism for expanding an irradiation area of the partially coherent light irradiating the straight portion of the object to be measured.
Optionally, the spectrum of the coherent light is in the range of 215-2000 nm.
In a second aspect, a detection method is provided, including:
the coherent light source generates coherent light;
the spectroscope divides the coherent light into a plurality of beams of partial coherent light, wherein the plurality of beams of partial coherent light are used for irradiating an object to be measured;
the reflector adjusts the irradiation angle of the partially or fully partially coherent light;
and the collecting detector collects a speckle image formed by the plurality of beams of the partial coherent light penetrating through the object to be detected or the plurality of beams of the partial coherent light being reflected by the object to be detected, and determines whether the glass has defects according to the speckle image.
Optionally, collecting a speckle image formed by the plurality of beams of the partially coherent light passing through the object to be measured or the plurality of beams of the partially coherent light being reflected by the object to be measured:
and collecting a speckle image formed by the plurality of beams of the partial coherent light passing through the object to be detected or the plurality of beams of the partial coherent light being reflected by the object to be detected in a non-imaging mode.
Optionally, the method further comprises: the acquisition detector determines whether the object to be detected has defects or not according to the speckle images and through a deep learning neural network; wherein the neural network is trained and learned using speckle images of large samples.
Optionally, the split coherent light includes a first split coherent light and a second split coherent light, and the curved portion includes a first curved portion and a second curved portion, where the first split coherent light is used for illuminating the first curved portion and the second split coherent light is used for illuminating the second curved portion.
Optionally, the reflector includes a first reflector and a second reflector, wherein the first reflector is used for adjusting an angle of the first curved portion irradiated by the first partial coherent light, and the second reflector is used for adjusting an angle of the second curved portion irradiated by the second partial coherent light.
Optionally, the system further includes a first concave lens and a second concave lens, the first concave lens is configured to diffuse the first incoherent light, and the second concave lens is configured to diffuse the second incoherent light.
Optionally, the method further comprises: and diffusing the phase of the dry light irradiating the straight part of the object to be detected through a beam expander.
Optionally, the method further comprises: and enlarging the irradiation area of the split coherent light irradiating the straight part of the object to be measured through the prism lens.
Optionally, the spectrum of the coherent light is in the range of 215-2000 nm.
In the scheme, coherent light is generated by a coherent light source, then the coherent light is divided into a plurality of sub-coherent light beams by a spectroscope, the angle of the sub-coherent light beam irradiating the curved part of the object to be detected is adjusted by a reflector, a collecting detector collects a speckle image formed by the plurality of sub-coherent light beams penetrating through the object to be detected or the plurality of sub-coherent light beams being reflected by the object to be detected, and finally, the collecting detector determines whether the object to be detected has defects according to the speckle image. It can be seen that the above solution allows for rapid inspection of curved glass.
FIG. 1 is a schematic view of a prior art variation of light transmission through curved glass;
FIG. 2 is a schematic diagram of a detection system according to the present application;
FIG. 3 is a graph comparing light intensity without and with bubbles in the present application;
FIG. 4 is a schematic diagram of a neural network proposed in the present application;
fig. 5 is a schematic flow chart of an inspection method proposed in the present application.
With the rapid development of technology, the produced glass slowly evolves from flat glass to curved glass, e.g., 3D glass, etc. The 3D glass is glass with a straight middle and gradually bent two sides. The crooked portion of 3D glass is caused after being heated and bent by plane glass, various damage appear easily, for example bump the scratch, the surface remains, sunken etc. consequently, the probability that the crooked portion goes wrong is higher than the probability that straight portion goes wrong far away, needs focus to detect to avoid appearing a large amount of bad products.
Compared with common glass, the backlight passes through a bending deformation area when penetrating through the bending edge end of the 3D glass, so that the imaging focusing difference of the area is caused, the image gray scale is greatly reduced, in addition, the white spot strip appears due to the light field deformation of the area, the characteristics all disable the traditional laser measurement and visual measurement methods, and therefore the traditional laser measurement and visual measurement can not measure the bending area. The only existing solution is to take a picture of the bending portion of the 3D glass for a plurality of rotation angles and identify a partial region of a plurality of images, but this method greatly prolongs the detection time, and the accuracy of the bending portion is significantly lower than that of the straight portion, resulting in a large amount of false detection and over-detection of the bending region of the 3D glass. As shown in fig. 1, fig. 1 is a comparison of 2D glass and 3D glass after parallel backlight detection, and C (scratch) of 3D glass at a bent position cannot be detected because the backlight penetrates too much glass and the light intensity is attenuated.
In order to solve the above problems, the present application proposes a detection system and method that can rapidly detect a bent glass and improve the accuracy of detection of the bent glass.
As shown in fig. 2, the present application provides a schematic structural diagram of a detection system. As shown in fig. 2, the detection system of the present application includes: coherent light source 110, beam splitter 120, mirror 130, and collection detector 140.
The coherent light source 110 generates coherent light and transmits the coherent light to the beam splitter 120. The beam splitter 120 splits the coherent light into a plurality of sub-coherent light beams, wherein the plurality of sub-coherent light beams are used for illuminating an object to be measured. The mirror 130 adjusts the illumination angle of the partially or fully partially coherent light. The collecting detector 140 collects a speckle image formed by the plurality of beams of the partial coherent light passing through the object to be measured or the plurality of beams of the partial coherent light being reflected by the object to be measured, and determines whether the object to be measured has a defect according to the speckle image. For simplicity, the following examples are all described with the object to be measured as glass.
In one embodiment, the coherent light generated by the coherent light source 110 is linearly polarized light with the same frequency and the same vibration direction. Due to the fact that the non-imaging speckle images are used for monitoring, the problem of chromatic dispersion of the lens of the imaging system in different spectral bands is solved, and the value range of the spectrum of the coherent light can be large, for example, the value range of the spectrum of the coherent light can be 215-2000 nm. That is, the spectrum of coherent light may range from ultraviolet light to near-infrared light. It is understood that the above-mentioned range of values of the spectrum of the coherent light is merely an example, and should not be specifically limited. The defect detection of coherent light speckles is adopted to obtain more defect information of the surface of the detected object, such as intensity information, phase information, incident angle information and the like after defect reflection, so that more defects which cannot be detected by the traditional method, such as micro-scratches, edge breakage, internal bubbles and the like, can be identified.
In a particular embodiment, the beam splitter 120 may include one or more beam splitters therein. When the optical splitter splits coherent light, a plurality of beams of the split coherent light are distributed according to the proportion corresponding to the optical power. Wherein the optical power of each beam of the partially coherent light can be determined by the area of the glass irradiated by each beam of the coherent light. For example, the ratio of the area of the glass irradiated with the first partial coherent light to the area of the glass irradiated with the second partial coherent light is 2: 1, the ratio of the optical power of the first incoherent light to the optical power of the second incoherent light is also 2: 1. it can be understood that the first and second partial coherent light beams are obtained by splitting the same coherent light beam, and therefore, it is strictly ensured that the frequencies and vibration directions of the first and second partial coherent light beams are consistent.
In one embodiment, the beam splitter 120 includes a first beam splitter 121 and a second beam splitter 122. It should be understood that the beam splitter 120 is merely an example, and in other embodiments, the number of the beam splitter 120 may be smaller or larger, and is not limited specifically herein.
In one embodiment, the reflector 130 includes a first reflector 131 and a second reflector 132. It should be understood that the above-mentioned mirror 130 is only an example, and in other embodiments, the number of the mirror 130 may be smaller or larger, and is not limited specifically herein.
In a specific embodiment, the detection system further comprises a concave lens, wherein the concave lens is used for diffusing the incoherent light. The concave lens includes a first concave lens 151 and a second concave lens 152. It should be understood that the above concave lenses are only used as an example, and in other embodiments, the number of concave lenses may be smaller or larger, and is not limited specifically herein.
In a specific embodiment, the detection system further includes a beam expander 160, wherein the beam expander 160 is configured to diffuse the portion of the dry light that irradiates the flat portion of the object to be measured. Specifically, the detection system further includes a prism for expanding an irradiation area of the partially coherent light that irradiates the straight portion of the object to be detected.
The coherent light generated by the coherent light source 110 is incident on the first beam splitter 121. The first beam splitter 121 splits the first split coherent light beam from the coherent light beam. The first concave lens 151 is disposed on the optical path of the first partial coherent light beam, and the first partial coherent light beam passes through the axis of the first concave lens 151. The first concave lens 151 diffuses the first partial coherent light to obtain diffused first partial coherent light. The first reflecting mirror 131 is disposed on the optical path of the diffused first split coherent light, and the first reflecting mirror 131 reflects the diffused first split coherent light so that the reflected first split coherent light irradiates the first curved portion (left wing portion of the 3D glass). After passing through the first beam splitter 121, the remaining coherent illumination is incident on the second beam splitter 122. The second beam splitter 122 separates the second split coherent light beam from the remaining coherent light beam. The second concave lens 152 is disposed on the optical path of the second partial coherent light beam, and the second partial coherent light beam passes through the axis of the second concave lens 152. The second concave lens 152 diffuses the second incoherent light to obtain diffused second incoherent light. The second reflecting mirror 132 is disposed on the optical path of the diffused second incoherent light, and the second reflecting mirror 132 reflects the diffused second incoherent light so that the reflected second incoherent light irradiates the second curved part (the right wing portion of the 3D glass). After passing through the second beam splitter 122, the remaining incoherent light is incident on a beam expander 160. The beam expander 160 is configured to diffuse the remaining partially coherent light to obtain a diffused partially coherent light, and to irradiate the flat portion (the middle portion of the 3D glass). In the above embodiment, the first concave lens, the second concave lens and the beam expander respectively diffuse the first partial coherent light, the second partial coherent light and the remaining partial coherent light, so that the light can be more uniformly irradiated on the first curved portion, the second curved portion and the straight portion.
It is understood that the above-mentioned detection system is only a specific embodiment, and in other embodiments, more reflectors, concave lenses, beam expanders, etc. may be included, and it is only necessary that the emergent first partial coherent light, second partial coherent light, and the rest partial coherent light are used to illuminate the glass of the curved portion and the flat portion in a transmission manner as perpendicular as possible, and this is not limited in particular.
In a particular embodiment, the collection detector 140 includes one or more photosensors, wherein the photosensors are configured to collect speckle images. In practical applications, the positions and the number of the photosensors can be set according to practical requirements, and are not particularly limited herein.
In a specific embodiment, the collecting detector 140 is specifically configured to collect a speckle image formed by the plurality of beams of the partially coherent light transmitted through the object to be measured or the plurality of beams of the partially coherent light reflected by the object to be measured in a non-imaging manner. The non-imaging mode is that whether the defects exist can be determined without performing recovery calculation on the 3D glass according to the speckle images, and the defects exist can be determined directly according to the speckle image calculation. It can be understood that when the speckle image is formed in a non-imaging form, the defect information is embodied in the speckle image, compared with the traditional optical imaging detection method, the method does not need to design very complicated illumination and imaging optics, does not need to carry out phase solving reconstruction and other real images of the reduction object on the speckle image, directly carries out discrimination and detection on the defect on the speckle image, can effectively reduce the calculated amount of data, and improves the identification speed.
In a particular embodiment, the speckle image is an image formed when light passes through or is reflected by an optically rough surface of the vibrating object. It can be understood that when light irradiates on optically rough surfaces (or transmission plates through which the optically rough surfaces pass) with average fluctuation larger than the order of wavelength, such as walls, paper, ground glass, etc., the wavelets scattered by the irregularly distributed surfaces on the surfaces are mutually superposed, so that the reflected light field (or the transmitted light field) has random spatial light intensity distribution and presents a granular structure, namely speckle. For example, as shown in FIG. 3, when a bubble occurs inside a glass curve, it causes the speckle distribution on the image collector to change. Clearly, the light intensity in the region where the bubbles are present in the glass will increase significantly. In addition to air bubbles, surface scratches, residues, pits, etc. cause variations in light intensity. Generally, these defects are generally in the range of 10 microns to 5 millimeters, and for coherent light having a wavelength of less than 1 micron, the accuracy of detection can be up to 5 microns or more, thus being sufficient to detect most defects.
Because the characteristics such as geometric shape and texture are only used for describing low-level edge information in the image, the shape of the glass defect is complex and variable, and the characteristics can not well represent the defect target. If the speckle image with the glass defect structure is directly subjected to phase inversion to obtain the restored image of the image, the restored image is greatly distorted, and the image can hardly be identified whether the glass has defects or not. The application may then employ inputting the speckle images into a neural network to determine whether a defect is present in the glass.
In a particular embodiment, the speckle image initially acquired by acquisition detector 140 is a noise-saturated speckle image. The useful information in the speckle images is buried in a large amount of noise, and therefore, the collecting detector 140 needs to process the initially collected speckle images to obtain processed speckle images to remove the speckle noise and improve the fringe contrast. The image processing method includes a phase shift method, a fringe gray scale method, a fringe central line method, a fourier transform method, a sub-pixel search method, and the like. It should be understood that the above-described processing method is only for example and should not be construed as being particularly limited.
In a specific embodiment, the speckle images may include a flat speckle image and a curved speckle image, and are not particularly limited herein.
In a particular embodiment, the flat speckle image may be one or more. When the area of the straight part is small, the number of the straight speckle images can be one; when the area of the flat portion is small, the number of the flat speckle images may be multiple, and different flat speckle images correspond to different regions of the flat portion. In practical application, the number of the flat speckle images can be set according to the area of the flat part of the glass, and the number is not limited in detail.
In a particular embodiment, the curved speckle image may be one or more. When the bending parts are concentrated in the same area, the number of the bending speckle images can be one; when the curved portion is dispersed over a plurality of regions, the number of curved speckle images may be multiple, with different curved speckle images corresponding to different regions of the curved portion. In practical applications, the number of the curved speckle images may be set according to the area where the curved portion of the glass is dispersed, and this time is not particularly limited.
In a specific embodiment, the acquisition detector 140 is used to determine whether the glass is defective based on the speckle images and a deep-learned neural network.
In a particular embodiment, the input to the deep-learning neural network includes a first curved speckle image, a second curved speckle image, a first flat speckle image, and a second flat speckle image. It should be understood that the above example of the input of the neural network is only an example, and in practical applications, the speckle image input by the neural network may be more or less, and is not limited in particular.
In a specific embodiment, the output result of the deep learning neural network may include: no defects, scratches, bubbles, dirt, etc., although the output may be represented in fewer or more levels. It is to be understood that the above-described level division is merely used as an example, and the more the level division is, the more accurately the output result is represented.
In a particular embodiment, the deep-learning neural network includes a plurality of training models, for example, the training models may include a defect-free training model, a bubble model, a dirty model, and so on. It should be understood that the above-mentioned training models are only used as an example, and in practical applications, more or less training models may be included, and are not specifically limited herein. The Neural network may be a BP Neural network, a Hopfield network, an ART network, a Kohonen network, a Long Short-Term Memory network (LSTM), a residual network (ResNet), a Recurrent Neural Network (RNN), and the like, and is not limited herein.
In a specific embodiment, when the input includes the first curved speckle image, the second curved speckle image, the first flat speckle image, and the second flat speckle image, and the output includes no defect, scratch, bubble, and dirt, the deep-learning neural network may be as shown in fig. 4.
It should be appreciated that the deep-learning neural network can be trained and learned using large sample speckle images. For example, a large number of speckle images are respectively collected in advance according to different defect samples, and classification training is performed on the speckle images capable of indirectly reflecting the glass microstructure by using a deep learning neural network to obtain a correct neural network. During identification, the collected speckle images are input into the trained neural network to obtain an identification result.
In the scheme, coherent light is generated by a coherent light source, then the coherent light is divided into a plurality of sub-coherent light beams by a spectroscope, the irradiation angle of partial or full-partial coherent light beams is adjusted by a reflector, a collecting detector collects a speckle image formed by the plurality of sub-coherent light beams penetrating through the object to be detected or the plurality of sub-coherent light beams being reflected by the object to be detected, and finally, the collecting detector determines whether the object to be detected has defects according to the speckle image. It can be seen that the above solution allows for rapid inspection of curved glass.
Referring to fig. 5, fig. 5 is a schematic flow chart of a detection method provided in the present application. For simplicity, the following examples are all described with the object to be measured as glass. As shown in fig. 5, the detection method of the present embodiment includes the following steps:
s101: the coherent light source generates coherent light.
In one embodiment, the coherent light source generates linearly polarized light with the same frequency and the same vibration direction. Due to the fact that the non-imaging speckle images are used for monitoring, the problem of chromatic dispersion of the lens of the imaging system in different spectral bands is solved, and the value range of the spectrum of the coherent light can be large, for example, the value range of the spectrum of the coherent light can be 215-2000 nm. That is, the spectrum of coherent light may range from ultraviolet light to near-infrared light. It is understood that the above-mentioned range of values of the spectrum of the coherent light is merely an example, and should not be specifically limited. The defect detection of coherent light speckles is adopted to obtain more defect information of the surface of the detected object, such as intensity information, phase information, incident angle information and the like after defect reflection, so that more defects which cannot be detected by the traditional method, such as micro-scratches, edge breakage, internal bubbles and the like, can be identified.
S102: the spectroscope divides the coherent light into a plurality of beams of partial coherent light, wherein the plurality of beams of partial coherent light are used for illuminating an object to be measured.
In a particular embodiment, the beam splitter may include one or more beam splitters therein. When the optical splitter splits coherent light, a plurality of beams of the split coherent light are distributed according to the proportion corresponding to the optical power. Wherein the optical power of each beam of the partially coherent light can be determined by the area of the glass irradiated by each beam of the coherent light. For example, the ratio of the area of the glass irradiated with the first partial coherent light to the area of the glass irradiated with the second partial coherent light is 2: 1, the ratio of the optical power of the first incoherent light to the optical power of the second incoherent light is also 2: 1. it can be understood that the first and second partial coherent light beams are obtained by splitting the same coherent light beam, and therefore, it is strictly ensured that the frequencies and vibration directions of the first and second partial coherent light beams are consistent.
In one embodiment, the beam splitter includes a first beam splitter and a second beam splitter. It should be understood that the above-mentioned beam splitters are only used as an example, and in other embodiments, the number of beam splitters may be smaller or larger, and is not limited specifically herein.
S103: the mirror adjusts the illumination angle of the partially or fully partially coherent light.
In a specific embodiment, the mirrors include a first mirror and a second mirror. It should be understood that the above-mentioned mirrors are only an example, and in other embodiments, the number of mirrors may be smaller or larger, and is not limited specifically herein.
In a specific embodiment, the detection system further comprises a concave lens, wherein the concave lens is used for diffusing the incoherent light. The concave lens includes a first concave lens and a second concave lens. It should be understood that the above concave lenses are only used as an example, and in other embodiments, the number of concave lenses may be smaller or larger, and is not limited specifically herein.
In a specific embodiment, the detection system further includes a beam expander, wherein the beam expander is configured to diffuse a portion of the coherent light that illuminates the flat portion of the object to be measured. Specifically, the detection system further includes a prism for expanding an irradiation area of the partially coherent light that irradiates the straight portion of the object to be detected.
Coherent light generated by the coherent light source is incident on the first beam splitter. The first optical splitter splits the first split coherent light beam from the coherent light beam. The first concave lens is disposed on an optical path of the first partial coherent light beam, and the first partial coherent light beam passes through an axis of the first concave lens. The first concave lens diffuses the first partial coherent light to obtain diffused first partial coherent light. The first reflecting mirror is disposed on an optical path of the diffused first partial coherent light, and reflects the diffused first partial coherent light so that the reflected first partial coherent light irradiates a first curved portion (left wing portion of the 3D glass). After passing through the first beam splitter, the remaining coherent illumination is incident on the second beam splitter. The second beam splitter splits the second split coherent light beam from the remaining coherent light beam. The second concave lens is disposed on an optical path of the second partial coherent light beam, and the second partial coherent light beam passes through an axis of the second concave lens. And the second concave lens diffuses the second partial coherent light to obtain diffused second partial coherent light. The second reflecting mirror is disposed on an optical path of the diffused second incoherent light, and reflects the diffused second incoherent light such that the reflected second incoherent light irradiates a second curved part (a right wing part of the 3D glass). After passing through the second beam splitter, the remaining split coherent light is incident on a beam expander. The beam expander is used for diffusing the rest of the split coherent light to obtain diffused split coherent light, and irradiating the diffused split coherent light on a straight part (the middle part of the 3D glass). In the above embodiment, the first concave lens, the second concave lens and the beam expander respectively diffuse the first partial coherent light, the second partial coherent light and the remaining partial coherent light, so that the light can be more uniformly irradiated on the first curved portion, the second curved portion and the straight portion.
It is understood that the above-mentioned detection system is only a specific embodiment, and in other embodiments, more reflectors, concave lenses, beam expanders, etc. may be included, and it is only necessary that the emergent first partial coherent light, second partial coherent light, and the rest partial coherent light are used to illuminate the glass of the curved portion and the flat portion in a transmission manner as perpendicular as possible, and this is not limited in particular.
S104: and the collecting detector collects a speckle image formed by the plurality of beams of the partial coherent light penetrating through the object to be detected or the plurality of beams of the partial coherent light being reflected by the object to be detected, and determines whether the glass has defects according to the speckle image.
In a specific embodiment, the collecting detector comprises one or more photosensors, wherein the photosensors are configured to collect speckle images. In practical applications, the positions and the number of the photosensors can be set according to practical requirements, and are not particularly limited herein.
In a specific embodiment, the collecting detector collects a speckle image formed by the plurality of beams of the partial coherent light transmitting through the object to be measured or the plurality of beams of the partial coherent light being reflected by the object to be measured in a non-imaging manner. The non-imaging mode is that whether the defects exist can be determined without performing recovery calculation on the 3D glass according to the speckle images, and the defects exist can be determined directly according to the speckle image calculation. It can be understood that when the speckle images are formed in a non-imaging mode, the defect information is embodied in the speckle images, compared with the traditional optical imaging detection method, the method does not need to design very complicated illumination and imaging optics, does not need to carry out phase solving reconstruction and other real images of the reducing object on the speckle images, directly carries out distinguishing detection on the defects on the speckle images, can effectively reduce the calculated amount of data, and improves the identification speed.
In a particular embodiment, the speckle image is an image formed when light passes through or is reflected by an optically rough surface of the vibrating object. It can be understood that when light irradiates on optically rough surfaces (or transmission plates through which the optically rough surfaces pass) with average fluctuation larger than the order of wavelength, such as walls, paper, ground glass, etc., the wavelets scattered by the irregularly distributed surfaces on the surfaces are mutually superposed, so that the reflected light field (or the transmitted light field) has random spatial light intensity distribution and presents a granular structure, namely speckle. For example, as shown in FIG. 3, when a bubble occurs inside a glass curve, it causes the speckle distribution on the image collector to change. Clearly, the light intensity in the region where the bubbles are present in the glass will increase significantly. In addition to air bubbles, surface scratches, residues, pits, etc. cause variations in light intensity. Generally, these defects are generally in the range of 10 microns to 5 millimeters, and for coherent light having a wavelength of less than 1 micron, the accuracy of detection can be up to 5 microns or more, thus being sufficient to detect most defects.
Because the characteristics such as geometric shape and texture are only used for describing low-level edge information in the image, the shape of the glass defect is complex and variable, and the characteristics can not well represent the defect target. If the speckle image with the glass defect structure is directly subjected to phase inversion to obtain the restored image of the image, the restored image is greatly distorted, and the image can hardly be identified whether the glass has defects or not. The application may then employ inputting the speckle images into a neural network to determine whether a defect is present in the glass.
In a specific embodiment, the speckle image initially acquired by the acquisition detector is a noise-saturated speckle image. Useful information in the speckle images is buried in a large amount of noise, and therefore, the collecting detector needs to process the initially collected speckle images to obtain processed speckle images so as to remove speckle noise and improve fringe contrast. The image processing method includes a phase shift method, a fringe gray scale method, a fringe central line method, a fourier transform method, a sub-pixel search method, and the like. It should be understood that the above-described processing method is only for example and should not be construed as being particularly limited.
In a specific embodiment, the speckle images may include a flat speckle image and a curved speckle image, and are not particularly limited herein.
In a particular embodiment, the flat speckle image may be one or more. When the area of the straight part is small, the number of the straight speckle images can be one; when the area of the flat portion is small, the number of the flat speckle images may be multiple, and different flat speckle images correspond to different regions of the flat portion. In practical application, the number of the flat speckle images can be set according to the area of the flat part of the glass, and the number is not limited in detail.
In a particular embodiment, the curved speckle image may be one or more. When the bending parts are concentrated in the same area, the number of the bending speckle images can be one; when the curved portion is dispersed over a plurality of regions, the number of curved speckle images may be multiple, with different curved speckle images corresponding to different regions of the curved portion. In practical applications, the number of the curved speckle images may be set according to the area where the curved portion of the glass is dispersed, and this time is not particularly limited.
In a specific embodiment, the acquisition detector is used for determining whether the glass has defects according to the speckle images and a deep learning neural network.
In a particular embodiment, the input to the deep-learning neural network includes a first curved speckle image, a second curved speckle image, a first flat speckle image, and a second flat speckle image. It should be understood that the above example of the input of the neural network is only an example, and in practical applications, the speckle image input by the neural network may be more or less, and is not limited in particular.
In a specific embodiment, the output result of the deep learning neural network may include: no defects, scratches, bubbles, dirt, etc., although the output may be represented in fewer or more levels. It is to be understood that the above-described level division is merely used as an example, and the more the level division is, the more accurately the output result is represented.
In a particular embodiment, the deep-learning neural network includes a plurality of training models, for example, the training models may include a defect-free training model, a bubble model, a dirty model, and so on. It should be understood that the above-mentioned training models are only used as an example, and in practical applications, more or less training models may be included, and are not specifically limited herein. The Neural network may be a BP Neural network, a Hopfield network, an ART network, a Kohonen network, a Long Short-Term Memory network (LSTM), a residual network (ResNet), a Recurrent Neural Network (RNN), and the like, and is not limited herein.
In a specific embodiment, when the input includes the first curved speckle image, the second curved speckle image, the first flat speckle image, and the second flat speckle image, and the output includes no defect, scratch, bubble, and dirt, the deep-learning neural network may be as shown in fig. 4.
It should be appreciated that the deep-learning neural network can be trained and learned using large sample speckle images. For example, a large number of speckle images are respectively collected in advance according to different defect samples, and classification training is performed on the speckle images capable of indirectly reflecting the glass microstructure by using a deep learning neural network to obtain a correct neural network. During identification, the collected speckle images are input into the trained neural network to obtain an identification result.
In the scheme, coherent light is generated by a coherent light source, then the coherent light is divided into a plurality of sub-coherent light beams by a spectroscope, the angle of the sub-coherent light beam irradiating the curved part of the object to be detected is adjusted by a reflector, a collecting detector collects a speckle image formed by the plurality of sub-coherent light beams penetrating through the object to be detected or the plurality of sub-coherent light beams being reflected by the object to be detected, and finally, the collecting detector determines whether the object to be detected has defects according to the speckle image. It can be seen that the above solution allows for rapid inspection of curved glass.
Although the above embodiment is illustrated by taking 3D glass as an example, in practical applications, the above detection method may also be applied to curved glass with any curved shape, or translucent plastic, ground glass, or other objects that are even opaque, and the like, and is not limited herein. The following description is made in connection with several specific embodiments.
Example one
The detection system of the embodiment is used for realizing the defect detection of the appearance of the curved glass. The curved glass is placed in a detection system, and then the reflector is adjusted, so that the reflected coherent light vertically irradiates the bending part of the curved glass, and the corresponding speckle image is changed due to the defects of scratching and dirt of the curved glass, bubbles in the curved glass and the like, so that the curved glass is distinguished.
Example two
The detection system of the embodiment is used for realizing the defect detection of large-format flat glass, such as display screens of tablet computers and liquid crystal displays, television glass and the like. The large-format plane glass is placed in a detection system, and then the angle of the reflector is adjusted, so that the coherent light can cover the large-format plane glass, and the corresponding speckle image is changed due to the defects of scratching and dirt of the large-format glass, bubbles in the large-format glass and the like, so that the large-format plane glass is distinguished.
EXAMPLE III
The detection system of the embodiment is used for realizing the defect detection of the complex curved surface with any curved surface shape. The irradiation angle of a reflector of the detection system and the position and angle distribution of the photoelectric sensor are changed, so that the defect detection of the complex curved surface with any curved surface shape is realized.
Example four
The detection system of the embodiment is used for realizing defect detection of the opaque object. The position of a photoelectric sensor of the detection system is changed, so that the photoelectric sensor can collect coherent light reflected by the opaque object, and the defect detection of the opaque object is realized.
EXAMPLE five
The detection system of the embodiment is used for realizing continuous online detection. The method comprises the following steps of placing a plurality of objects to be detected on an electric platform, dragging the plurality of objects to be detected to pass through a detection system by the electric platform, and carrying out defect detection on the plurality of objects to be detected in turn by the detection system so as to realize continuous online detection.
In the several embodiments provided in the present application, it should be understood that the disclosed system, terminal and method can be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may also be an electric, mechanical or other form of connection.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment of the present invention.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention essentially or partially contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
- A detection system, comprising:a coherent light source for generating coherent light;a spectroscope for dividing the coherent light into a plurality of beams of partial coherent light for illuminating an object to be measured;the reflector is used for adjusting the irradiation angle of the partially or fully partially coherent light;and the collecting detector is used for collecting a speckle image formed by the plurality of beams of the partial coherent light penetrating through the object to be detected or the plurality of beams of the partial coherent light being reflected by the object to be detected, and determining whether the glass has defects according to the speckle image.
- The system of claim 1, wherein the collecting detector is specifically configured to collect a speckle image of the plurality of beams of partially coherent light transmitted through the object under test or the plurality of beams of partially coherent light reflected by the object under test and formed in a non-imaging manner.
- The system of claim 1, wherein the collecting detector is further configured to determine whether the object under test has a defect from the speckle image and through a deep learning neural network.
- A system according to any one of claims 1 to 3, further comprising a beam expander for diffusing a portion of the coherent light impinging on the flat portion of the object to be measured.
- The system of claim 4, further comprising a prism for expanding an irradiation area of the partially coherent light irradiating the flat portion of the object to be measured.
- A method of detection, comprising:the coherent light source generates coherent light;the spectroscope divides the coherent light into a plurality of beams of partial coherent light, wherein the plurality of beams of partial coherent light are used for irradiating an object to be measured;the reflector adjusts the irradiation angle of the partially or fully partially coherent light;and the collecting detector collects a speckle image formed by the plurality of beams of the partial coherent light penetrating through the object to be detected or the plurality of beams of the partial coherent light being reflected by the object to be detected, and determines whether the glass has defects according to the speckle image.
- The method according to claim 6, wherein the speckle image formed by the plurality of beams of the partially coherent light transmitted through the object to be measured or the plurality of beams of the partially coherent light reflected by the object to be measured is collected:and collecting a speckle image formed by the plurality of beams of the partial coherent light passing through the object to be detected or the plurality of beams of the partial coherent light being reflected by the object to be detected in a non-imaging mode.
- The method of claim 6, further comprising:the acquisition detector determines whether the object to be detected has defects or not according to the speckle images and through a deep learning neural network; wherein the neural network is trained and learned using speckle images of large samples.
- The method according to any one of claims 6 to 8, further comprising:and diffusing the phase of the dry light irradiating the straight part of the object to be detected through a beam expander.
- The method of claim 9, further comprising:and enlarging the irradiation area of the split coherent light irradiating the straight part of the object to be measured through the prism lens.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2018/099921 WO2020029237A1 (en) | 2018-08-10 | 2018-08-10 | Detection method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111226110A true CN111226110A (en) | 2020-06-02 |
Family
ID=69414404
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201880067184.XA Pending CN111226110A (en) | 2018-08-10 | 2018-08-10 | Detection method and system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111226110A (en) |
| WO (1) | WO2020029237A1 (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111650201A (en) * | 2020-07-15 | 2020-09-11 | Oppo（重庆）智能科技有限公司 | Detection device and detection method |
| CN114638467A (en) * | 2022-01-31 | 2022-06-17 | 南通鑫鑫医药药材有限公司 | Medical glass container production quality detection system |
| CN114998216A (en) * | 2022-05-06 | 2022-09-02 | 湖北文理学院 | Method and device for rapidly detecting surface defects of transparent part |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN85101640A (en) * | 1985-04-01 | 1987-01-17 | 欧文斯-伊利诺衣公司 | The detection system of refractive power defective in the transparent substance |
| CN101175986A (en) * | 2005-04-06 | 2008-05-07 | 康宁股份有限公司 | Glass inspection systems and methods for using same |
| CN101221133A (en) * | 2008-01-25 | 2008-07-16 | 天津商业大学 | Honeycomb paper plate bonding defect detecting method and detecting system |
| CN201184867Y (en) * | 2008-04-09 | 2009-01-21 | 苏州大学 | Optical non-linear apparatus base on 4f phase coherent imaging system measuring material |
| CN101460829A (en) * | 2006-04-19 | 2009-06-17 | 光学传感公司 | Measuring water vapor in hydrocarbons |
| CN201555808U (en) * | 2009-11-19 | 2010-08-18 | 西北工业大学 | Device capable of nondestructively testing surface of planar object |
| CN201876417U (en) * | 2010-11-05 | 2011-06-22 | 深圳清华大学研究院 | Immunological turbidimetry measuring device |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008170207A (en) * | 2007-01-10 | 2008-07-24 | Olympus Corp | Macro lighting system for visual inspection apparatus, visual inspection apparatus, and visual inspection method of glass substrate |
| CN104482875B (en) * | 2014-12-19 | 2017-07-21 | 合肥工业大学 | Single slit spatial carrier speckle-shearing interferometry measuring system and measuring method |
-
2018
- 2018-08-10 CN CN201880067184.XA patent/CN111226110A/en active Pending
- 2018-08-10 WO PCT/CN2018/099921 patent/WO2020029237A1/en active Application Filing
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN85101640A (en) * | 1985-04-01 | 1987-01-17 | 欧文斯-伊利诺衣公司 | The detection system of refractive power defective in the transparent substance |
| CN101175986A (en) * | 2005-04-06 | 2008-05-07 | 康宁股份有限公司 | Glass inspection systems and methods for using same |
| CN101460829A (en) * | 2006-04-19 | 2009-06-17 | 光学传感公司 | Measuring water vapor in hydrocarbons |
| CN101221133A (en) * | 2008-01-25 | 2008-07-16 | 天津商业大学 | Honeycomb paper plate bonding defect detecting method and detecting system |
| CN201184867Y (en) * | 2008-04-09 | 2009-01-21 | 苏州大学 | Optical non-linear apparatus base on 4f phase coherent imaging system measuring material |
| CN201555808U (en) * | 2009-11-19 | 2010-08-18 | 西北工业大学 | Device capable of nondestructively testing surface of planar object |
| CN201876417U (en) * | 2010-11-05 | 2011-06-22 | 深圳清华大学研究院 | Immunological turbidimetry measuring device |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111650201A (en) * | 2020-07-15 | 2020-09-11 | Oppo（重庆）智能科技有限公司 | Detection device and detection method |
| CN111650201B (en) * | 2020-07-15 | 2023-06-06 | Oppo（重庆）智能科技有限公司 | Detection device and detection method |
| CN114638467A (en) * | 2022-01-31 | 2022-06-17 | 南通鑫鑫医药药材有限公司 | Medical glass container production quality detection system |
| CN114998216A (en) * | 2022-05-06 | 2022-09-02 | 湖北文理学院 | Method and device for rapidly detecting surface defects of transparent part |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020029237A1 (en) | 2020-02-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU728407B2 (en) | Moire interferometry system and method with extended imaging depth | |
| TWI664586B (en) | Method for defect inspection of transparent substrate through complex defect diffractionwavefront | |
| US9891168B2 (en) | Device and method for sensing at least one partially specular surface with column-by-column analysis of the global or local intensity maximum | |
| US20140268105A1 (en) | Optical defect inspection system | |
| CN110389021B (en) | Lens image generating system, diopter power and thickness determining and defect detecting method | |
| CN104964982B (en) | Glass surface true and false defect identification method and system based on OCT complex signals | |
| CN111226110A (en) | Detection method and system | |
| US20210279858A1 (en) | Material testing of optical test pieces | |
| JPH0242307A (en) | Apparatus and method for displaying surface contour of convex object | |
| CN111579564B (en) | Transparent flat plate defect and diopter change detection system and method | |
| EP1605241A1 (en) | Device for the optical test of transparent or reflective samples | |
| JP6542906B2 (en) | Method and apparatus for determining surface data and / or measurement data associated with the surface of an at least partially transparent object | |
| CN111879789A (en) | Metal surface defect detection method and system | |
| CN111316086B (en) | Optical detection method for surface defects and related device | |
| US10598604B1 (en) | Normal incidence phase-shifted deflectometry sensor, system, and method for inspecting a surface of a specimen | |
| CN114459395A (en) | Plane flat crystal calibrating device and calibrating method | |
| EP0756152B1 (en) | Method and device for detecting the state of the surface of parts with a reflecting surface, applicable for controlling the roughness of polished parts | |
| CN113624459A (en) | Film uniformity detection system based on coherent tomography | |
| WO2013132072A1 (en) | Method and tool for measuring the geometric structure of an optical component | |
| JP6196841B2 (en) | Transmitted wavefront measuring apparatus and transmitted wavefront measuring method | |
| JP6791081B2 (en) | Refractive index measuring device and refractive index measuring method | |
| JP4286605B2 (en) | Object surface inspection apparatus and surface inspection method | |
| CN110749606A (en) | Laser damage detection method and system based on optical element | |
| Šmíd et al. | Detection of decorative phase defects on a polymeric non-planar object | |
| CN115950890B (en) | Spectral domain optical coherence tomography detection method and system for industrial detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200602 |
|
| RJ01 | Rejection of invention patent application after publication |