An online detection method for toughened glass nickel sulfide inclusion

Abstract

The present application relates to the field of glass defect detection, and discloses an online detection method for nickel sulfide inclusions in tempered glass; in the continuous conveying process of the tempered glass, at least two detection lights with different incident directions are used to determine the plate coordinates, buried depth and equivalent radius of candidate inclusions, and the candidate inclusions are tracked according to the conveying displacement; the first modulation frequency and the second modulation frequency are determined according to the size, buried depth, glass thickness and thermal diffusivity of the candidate inclusions, the photothermal complex response and the dynamic birefringence complex response are obtained in the core detection area and the annular reference area; the differential photothermal complex response and the differential dynamic birefringence complex response are obtained through regional difference, and the dual-frequency thermal-stress transfer parameters are established; whether the candidate inclusions are nickel sulfide inclusions is judged by combining the transfer parameters and the change relationship thereof, the response position consistency, the phase causality consistency and the nickel sulfide calibration parameter range.

Description

Technical Field

[0001] This invention relates to the field of glass defect detection, and specifically to an online detection method for nickel sulfide inclusions in tempered glass. Background Technology

[0002] Tempered glass possesses high mechanical strength, impact resistance, and heat resistance, and is widely used in building curtain walls, doors and windows, vehicle glass, and photovoltaic modules. During the glass raw material melting process, nickel may react with sulfur to form nickel sulfide inclusions, which remain inside the glass. After tempering, these nickel sulfide inclusions may undergo crystal transformation and volume changes due to temperature variations or long-term use, causing localized stress concentration around the inclusions. This can lead to spontaneous breakage of the tempered glass without significant external force. Therefore, timely detection and location of nickel sulfide inclusions within the glass during the tempering process is crucial for reducing the risk of spontaneous breakage during subsequent use.

[0003] Current methods for detecting internal defects in glass typically include optical scattering, transmission imaging, infrared thermography, or polarization stress detection. Optical scattering and transmission imaging can detect scattering targets within the glass, but bubbles, glass inclusions, and other solid inclusions can all produce similar scattering characteristics, making it difficult to accurately determine the material of inclusions based solely on scattering brightness, shape, or size of the scattering area. Infrared thermography can detect temperature changes in locally heated areas, but the results are easily affected by glass thickness, surface condition, ambient temperature, and thermal diffusion boundaries. Polarization stress detection primarily reflects the static stress distribution within tempered glass. Background stress, edge stress, and local stress around other inclusions formed during the glass tempering process can all produce birefringence responses, making it difficult to use alone as a reliable basis for identifying nickel sulfide inclusions.

[0004] Furthermore, in the continuous conveying production environment of tempered glass, candidate inclusions move with the glass, resulting in conveying displacement and beam tracking errors between the initial inspection position and the subsequent confirmation inspection position. Using a fixed inspection area or fixed photothermal excitation parameters can easily lead to problems such as the excitation position deviating from the candidate inclusion, inclusions of different sizes and burial depths using the same thermal diffusion scale, and the ineffective subtraction of local background responses in the glass. Existing technologies typically focus on acquiring a single scattering signal, a single temperature response, or a static birefringence response, lacking a detection method that configures different modulation frequencies based on the spatial location, size, and burial depth of the candidate inclusion, and combines the amplitude and phase transfer relationship between the local photothermal response and the dynamic birefringence response to confirm the inclusion material. Therefore, an online detection method for nickel sulfide inclusions in tempered glass is needed to solve the difficulties in locating and tracking candidate inclusions under continuous conveying conditions, and the difficulty in accurately distinguishing nickel sulfide inclusions from other internal glass defects. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an online detection method for nickel sulfide inclusions in tempered glass, thereby solving the technical problems existing in the prior art.

[0006] The above-mentioned technical objective of the present invention is achieved through the following technical solution: An online detection method for nickel sulfide inclusions in tempered glass includes the following steps: S1: During the continuous conveying of tempered glass, the conveying displacement is obtained. The tempered glass is scanned with detection light from at least two different incident directions. The plate coordinates, burial depth and equivalent radius of the candidate inclusions are determined based on the scattering position of the same scattering target under different incident directions. S2: Determine the moment when the candidate inclusion arrives at the confirmation detection area based on the plate coordinates and conveying displacement, and control the excitation beam and probe beam to follow the movement of the candidate inclusion; S3: Determine the first modulation frequency and the second modulation frequency based on the equivalent radius, burial depth, thickness of tempered glass and thermal diffusivity, so that the first thermal diffusivity length corresponding to the first modulation frequency is not greater than the equivalent radius, and the second thermal diffusivity length corresponding to the second modulation frequency is greater than the equivalent radius and less than the distance from the center of the candidate inclusion to the nearest glass surface. S4: Set up a core detection area with the projection position of the candidate inclusion as the center, and set up an annular reference area without the candidate inclusion outside the core detection area. Perform periodic local photothermal excitation according to the first modulation frequency and the second modulation frequency in sequence, and simultaneously acquire the photothermal response and dynamic birefringence response of the two detection areas. Demodulate the photothermal response and dynamic birefringence response synchronously according to the output intensity change of the excitation beam to obtain the photothermal complex response and dynamic birefringence complex response. S5: For the first modulation frequency and the second modulation frequency, the difference between the photothermal complex response of the core detection area and the annular reference area is obtained to obtain the differential photothermal complex response. The difference between the dynamic birefringence complex response of the core detection area and the annular reference area is obtained to obtain the differential dynamic birefringence complex response. The heat-stress transfer parameters are determined based on the amplitude ratio and phase difference between the differential dynamic birefringence complex response and the differential photothermal complex response. S6: Compare the thermal-stress transfer parameters and their variation relationship between the first modulation frequency and the second modulation frequency with the parameter range formed by the nickel sulfide inclusion calibration sample confirmed by composition detection, and combine the response position consistency and phase causality consistency of the differential photothermal complex response and the differential dynamic birefringence complex response to determine whether the candidate inclusion is a nickel sulfide inclusion. S7: When determining that a candidate inclusion is a nickel sulfide inclusion, output the plate coordinates, burial depth and equivalent radius of the nickel sulfide inclusion.

[0007] Preferably, in step S1, the tempered glass is scanned with a detection light in the first incident direction to obtain the first scattering position of the scattering target in the first incident direction, and the tempered glass is scanned with a detection light in the second incident direction to obtain the second scattering position of the scattering target in the second incident direction. The first scattering position and the second scattering position are transformed into a plate coordinate system with the front edge and side edge of the tempered glass as the coordinate reference. The first scattering position is back-projected along the propagation path inside the tempered glass along the first incident direction to form a first projection path. The second scattering position is back-projected along the propagation path inside the tempered glass along the second incident direction to form a second projection path. When the minimum distance between the first projection path and the second projection path is not greater than the spatial positioning error of the detection light scan, and the positions of the two projection paths corresponding to the minimum distance are both located between the two surfaces of the tempered glass, the first scattering position and the second scattering position are determined as the scattering positions of the same scattering target under different incident directions. The midpoint of the line connecting the two projection path positions is determined as the spatial position of the scattering target. The projection coordinates of the spatial position on the glass surface on the side where the detection light is incident are determined as the plate coordinates. The normal distance from the spatial position to the glass surface on the side where the detection light is incident is determined as the burial depth. The equivalent radius is determined based on the size of the scattering region formed by the scattering target and the point diffusion calibration relationship formed by standard scatterers of different known sizes and burial depths. Scattering targets with a burial depth greater than the equivalent radius but less than the difference between the thickness of the tempered glass and the equivalent radius are identified as candidate inclusions.

[0008] Preferably, in step S2, the conveying speed of the tempered glass is determined based on the changes in conveying displacement obtained at adjacent sampling times; Based on the plate coordinates of the candidate inclusion, the distance between the detection light scanning position and the confirmation detection area along the tempered glass conveying direction, and the current conveying displacement, the remaining conveying distance between the candidate inclusion and the confirmation detection area is determined, and the time when the candidate inclusion arrives at the confirmation detection area is determined based on the remaining conveying distance and the conveying speed. After the candidate inclusion reaches the confirmation detection area, the position of the candidate inclusion in the confirmation detection area is updated according to the continuously acquired transport displacement, and the direction of the excitation beam and the probe beam is adjusted according to the updated position. When tempered glass stops being transported, the pause time is determined and the correspondence between candidate inclusions and the current transport displacement is retained. When tempered glass resumes transport or the transport direction is changed, the position of candidate inclusions and the time when they arrive at the confirmation detection area are re-determined based on the new transport displacement.

[0009] Preferably, in step S3, the thermal diffusion length corresponding to different modulation frequencies is determined based on the thermal diffusivity of tempered glass and the calibration correspondence between modulation frequency and thermal diffusion length. A first modulation frequency is determined within a modulation frequency range where the corresponding thermal diffusion length is not greater than the equivalent radius, and the thermal diffusion length corresponding to the first modulation frequency is determined as the first thermal diffusion length. Based on the thickness and burial depth of the tempered glass, the distances from the center of the candidate inclusion to the two surfaces of the tempered glass are determined, and the smaller of the two distances is determined as the nearest surface distance. A second modulation frequency is determined within a modulation frequency range where the corresponding thermal diffusion length is greater than the equivalent radius and less than the nearest surface distance, and the thermal diffusion length corresponding to the second modulation frequency is determined as the second thermal diffusion length, wherein the first modulation frequency is higher than the second modulation frequency; When there is no modulation frequency that satisfies the second thermal diffusion length condition, the candidate inclusions are identified as objects for re-inspection.

[0010] Preferably, in step S4, the projection position of the candidate inclusion in the confirmation detection area is determined based on the candidate inclusion position updated with the conveying displacement, and the position tracking error is determined based on the maximum deviation between the candidate inclusion position and the excitation beam illumination position. A core detection area is set with the projection position as the center, and the distance between the outer boundary of the core detection area and the projection position is not less than the sum of the second thermal diffusion length and the position tracking error; The detection spatial resolution is determined based on the spatial sampling interval of the probe beam in the plate coordinate system. An annular reference area is set around the core detection area, such that the distance between the inner boundary of the annular reference area and the outer boundary of the core detection area is not less than the sum of the detection spatial resolution and the position tracking error. The outer boundary of the annular reference area is determined based on the distance from the projection position to the edge of the tempered glass and the distance to the projection positions of other candidate inclusions, so that the annular reference area does not extend to the edge of the tempered glass and does not contain other candidate inclusions. When the effective sampling area of ​​the annular reference area is smaller than the effective sampling area of ​​the core detection area, the candidate inclusions are identified as objects for re-inspection.

[0011] Preferably, in step S4, the detection beam is expanded into a planar detection beam covering the core detection area and the annular reference area, and the planar detection beam after passing through the tempered glass is divided into a first detection beam and a second detection beam. The local wavefront deflection distribution of the first detection light at the corresponding positions in the core detection area and the annular reference area is obtained, and the periodic change of the local wavefront deflection distribution relative to the change in the output intensity of the excitation beam is determined as the photothermal response. The polarization intensity of the second detection light in at least four different polarization directions is obtained. The optical delay at each position in the core detection area and the annular reference area is determined based on the polarization intensity. The difference between the optical delay during the periodic local photothermal excitation and the optical delay before the periodic local photothermal excitation is determined as the dynamic birefringence response. The local wavefront deflection distribution and optical delay are mapped to the same plate coordinate system, and the photothermal response and dynamic birefringence response are demodulated synchronously with the output intensity change of the excitation beam to obtain the photothermal complex response and dynamic birefringence complex response at each position in the core detection area and the annular reference area.

[0012] Preferably, in step S5, for the first modulation frequency and the second modulation frequency, the in-phase component and the quadrature component of the photothermal complex response at each position in the core detection area are respectively normalized to form the core photothermal complex response, and the in-phase component and the quadrature component of the photothermal complex response at each position in the annular reference area are normalized to form the reference photothermal complex response, and the difference between the core photothermal complex response and the reference photothermal complex response is determined as the differential photothermal complex response. The core dynamic birefringence complex response and the reference dynamic birefringence complex response are formed in the same way, and the difference between the core dynamic birefringence complex response and the reference dynamic birefringence complex response is determined as the differential dynamic birefringence complex response. Without periodic local photothermal excitation, differential photothermal complex response and differential dynamic birefringence complex response are continuously acquired in at least three detection periods. The maximum value of the corresponding complex response amplitude and the sum of the corresponding detection repeatability error in each detection period are determined as the photothermal response detection boundary and the birefringence response detection boundary, respectively. The response at the corresponding modulation frequency is determined to be a valid response only when the amplitude of the differential photothermal complex response at the corresponding modulation frequency is higher than the photothermal response detection boundary, and the amplitude of the differential dynamic birefringence complex response is higher than the birefringence response detection boundary.

[0013] Preferably, before the start of the detection batch, a phase calibration device is used that can cause the first detection optical path and the second detection optical path to produce periodic detection changes according to the same period signal, so as to determine the system phase deviation between the two detection optical paths at the first modulation frequency and the second modulation frequency respectively. In step S5, the ratio of the amplitude of the differential dynamic birefringence complex response to the amplitude of the differential photothermal complex response at the first modulation frequency is determined as the first amplitude ratio, and the original phase difference of the two differential complex responses at the first modulation frequency is determined as the first phase difference after deducting the corresponding system phase deviation. The ratio of the amplitude of the differential dynamic birefringence complex response to the amplitude of the differential photothermal complex response at the second modulation frequency is determined as the second amplitude ratio, and the original phase difference of the two differential complex responses at the second modulation frequency is determined as the second phase difference after deducting the corresponding system phase deviation. The first amplitude ratio, the first phase difference, the second amplitude ratio, and the second phase difference are determined as heat-stress transfer parameters. Based on the ratio between the first amplitude ratio and the second amplitude ratio, and the difference between the first phase difference and the second phase difference, the variation relationship of the heat-stress transfer parameters between the two modulation frequencies is determined.

[0014] Preferably, in step S6, for the first modulation frequency and the second modulation frequency, the photothermal response energy centroid is determined according to the photothermal complex response amplitude and plate coordinates of each effective sampling position in the core detection area, and the birefringence response energy centroid is determined according to the dynamic birefringence complex response amplitude and plate coordinates of each effective sampling position in the core detection area. When the distance between the centroid of the photothermal response energy and the centroid of the birefringence response energy is no greater than the sum of the detection spatial resolution, the position tracking error, and the optical path spatial registration error between the two detection optical paths at both modulation frequencies, the consistency of the response position is determined. When both the first and second phase differences after system phase correction and phase expansion indicate that the differential dynamic birefringence complex response lags behind the differential photothermal complex response, it is determined that phase causality is valid. Tempered glass containing nickel sulfide inclusions, as confirmed by Raman spectroscopy and elemental analysis, was used as the nickel sulfide calibration sample. The nickel sulfide calibration samples were grouped according to the tempered glass thickness, the equivalent radius of the nickel sulfide inclusions, and the burial depth. Based on the repeated detection results of each group of calibration samples, the parameter ranges of the first amplitude ratio, the first phase difference, the second amplitude ratio, the second phase difference, and their variation relationships were formed. The parameter ranges were verified using a control sample without nickel sulfide inclusions, which was confirmed by detection. When the responses at both modulation frequencies are valid, the response position consistency and phase causality consistency are both valid, and the thermal-stress transfer parameters and their variation relationships are all within the parameter range corresponding to the thickness of the tempered glass to be tested, the equivalent radius of the candidate inclusion, and the burial depth, the candidate inclusion is identified as a nickel sulfide inclusion. If the response at any modulation frequency is not a valid response, or if the consistency of response position or phase causality is not established, the candidate inclusion is identified as a re-inspection object; if the response at both modulation frequencies is a valid response and both consistency conditions are met, but the thermal-stress transfer parameters or their variation relationship are not within the corresponding parameter range, the candidate inclusion is identified as a non-nickel sulfide inclusion.

[0015] Preferably, before performing periodic local photothermal excitation in step S4, a trial excitation is applied to the region where the candidate inclusion is located with a single-cycle excitation energy lower than that of the formal periodic local photothermal excitation, and the differential photothermal complex response corresponding to the trial excitation is obtained according to the photothermal response acquisition and regional difference method. The photothermal response sensitivity of the region where the candidate inclusion is located is determined based on the relationship between the single-cycle excitation energy of the trial excitation and the corresponding differential photothermal complex response amplitude. A glass calibration sample with the same thickness, glass composition and coating state as the tempered glass to be tested is used. The excitation energy is increased step by step. After each excitation, the optical delay when the glass calibration sample recovers to the temperature before excitation is compared with the optical delay before excitation. The differential photothermal complex response amplitude corresponding to the highest excitation level that does not cause irreversible optical delay change is determined as the safe response boundary after deducting the amplitude fluctuation margin caused by repeated excitation. The output power and duty cycle of the periodic local photothermal excitation are adjusted according to the photothermal response sensitivity to make the amplitude of the differential photothermal complex response higher than the photothermal response detection boundary and lower than the safety response boundary; when the amplitude of the differential photothermal complex response reaches the safety response boundary, the periodic local photothermal excitation is stopped, and the candidate inclusions are identified as the objects for re-inspection.

[0016] In summary, the present invention has the following main beneficial effects: By scanning continuously transported tempered glass with detection beams from at least two different incident directions, the plate coordinates, burial depth, and equivalent radius of candidate inclusions are determined based on the scattering position of the same scattering target under different incident directions. Combined with transport displacement, the position of candidate inclusions is updated and the beam is tracked. This achieves the effect of spatial positioning and online continuous detection of candidate inclusions inside the glass without interrupting the transport of tempered glass, reducing the impact of surface defects, changes in transport position, and detection beam offset on subsequent confirmation detection.

[0017] The first and second modulation frequencies are determined based on the equivalent radius, burial depth, tempered glass thickness, and thermal diffusivity of the candidate inclusions. These two modulation frequencies correspond to different thermal diffusivity scales in the neighborhood of the candidate inclusions and across the interface between the inclusions and the glass substrate, respectively. An annular reference area is set around the core detection area. The photothermal complex response and dynamic birefringence complex response of the core detection area and the annular reference area are differentially processed to reduce the influence of overall glass temperature changes, static tempering stress, surface condition, and fluctuations in the detection environment, thereby improving the signal-to-noise ratio of the local thermal response and thermoelastic stress response of the candidate inclusions.

[0018] By establishing thermal-stress transfer parameters based on the amplitude ratio and phase difference between the differential dynamic birefringence complex response and the differential photothermal complex response, and combining the parameter variation relationship between the two modulation frequencies, the consistency of response position, the consistency of phase causality, and the calibration range of nickel sulfide inclusions confirmed by component detection, a comprehensive judgment is made. This avoids judging the inclusion material solely based on scattering intensity, single temperature rise, or static birefringence characteristics, improves the ability to distinguish nickel sulfide inclusions from bubbles, glass stones, and other non-nickel sulfide inclusions, and reduces the risk of misjudgment caused by invalid responses and detection channel errors. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1 refer to Figure 1 An online detection method for nickel sulfide inclusions in tempered glass includes the following steps: S1: During the continuous conveying of tempered glass, the conveying displacement is obtained. The tempered glass is scanned with detection light from at least two different incident directions. The plate coordinates, burial depth and equivalent radius of the candidate inclusions are determined based on the scattering position of the same scattering target under different incident directions. S2: Determine the moment when the candidate inclusion arrives at the confirmation detection area based on the plate coordinates and conveying displacement, and control the excitation beam and probe beam to follow the movement of the candidate inclusion; S3: Determine the first modulation frequency and the second modulation frequency based on the equivalent radius, burial depth, thickness of tempered glass and thermal diffusivity, so that the first thermal diffusivity length corresponding to the first modulation frequency is not greater than the equivalent radius, and the second thermal diffusivity length corresponding to the second modulation frequency is greater than the equivalent radius and less than the distance from the center of the candidate inclusion to the nearest glass surface. S4: Set up a core detection area with the projection position of the candidate inclusion as the center, and set up an annular reference area without the candidate inclusion outside the core detection area. Perform periodic local photothermal excitation according to the first modulation frequency and the second modulation frequency in sequence, and simultaneously acquire the photothermal response and dynamic birefringence response of the two detection areas. Demodulate the photothermal response and dynamic birefringence response synchronously according to the output intensity change of the excitation beam to obtain the photothermal complex response and dynamic birefringence complex response. S5: For the first modulation frequency and the second modulation frequency, the difference between the photothermal complex response of the core detection area and the annular reference area is obtained to obtain the differential photothermal complex response. The difference between the dynamic birefringence complex response of the core detection area and the annular reference area is obtained to obtain the differential dynamic birefringence complex response. The heat-stress transfer parameters are determined based on the amplitude ratio and phase difference between the differential dynamic birefringence complex response and the differential photothermal complex response. S6: Compare the thermal-stress transfer parameters and their variation relationship between the first modulation frequency and the second modulation frequency with the parameter range formed by the nickel sulfide inclusion calibration sample confirmed by composition detection, and combine the response position consistency and phase causality consistency of the differential photothermal complex response and the differential dynamic birefringence complex response to determine whether the candidate inclusion is a nickel sulfide inclusion. S7: When determining that a candidate inclusion is a nickel sulfide inclusion, output the plate coordinates, burial depth and equivalent radius of the nickel sulfide inclusion.

[0022] It is used to initially locate the scattering target inside the glass during the continuous conveying of tempered glass, and to determine whether the candidate inclusion is nickel sulfide inclusion based on the photothermal response and dynamic birefringence response generated by the candidate inclusion at two different thermal diffusion scales.

[0023] In this embodiment, scattering detection is used to determine the plate coordinates, burial depth, and equivalent radius of the scattering target inside the glass, without directly determining the material of the scattering target based on scattering intensity, scattering color, or scattering region size. The determination of nickel sulfide inclusions is based on the differential photothermal complex response, differential dynamic birefringence complex response formed by the candidate inclusion under dual-modulation frequency photothermal excitation, and the thermal-stress transfer relationship between the two.

[0024] The periodic local photothermal excitation used in this embodiment is to induce detectable periodic temperature changes in the candidate inclusion and its adjacent glass substrate. It does not require heating the candidate inclusion to the nickel sulfide phase transition temperature, nor does it identify nickel sulfide inclusions by inducing glass fracture. Therefore, this embodiment differs from detection methods that only utilize static scattering images to identify inclusions, only detect local temperature rises, or only detect static stress stripes in tempered glass.

[0025] The online detection device implementing this method includes a conveying mechanism, a glass thickness detection mechanism, a conveying displacement detection mechanism, a preliminary inspection positioning mechanism, a confirmation detection mechanism, a beam tracking mechanism, a synchronous acquisition mechanism, and a processing control mechanism.

[0026] The conveying mechanism is used to continuously transport tempered glass along a set conveying direction. The conveying mechanism includes a set of conveying rollers to support the tempered glass. When the tempered glass passes through the initial inspection and positioning mechanism and the confirmation inspection mechanism, it maintains the same plate orientation and conveying reference.

[0027] The conveyor displacement detection mechanism is connected to the conveyor mechanism and is used to continuously acquire the conveyor displacement of the tempered glass. The conveyor displacement detection mechanism can be a rotary encoder coaxially connected to the conveyor rollers, or a linear displacement detector that directly detects the movement position of the tempered glass edge. The output of the conveyor displacement detection mechanism is associated with each tempered glass to avoid mixing displacement data between adjacent glass panes.

[0028] The glass thickness measurement mechanism is located upstream of the initial inspection positioning mechanism and is used to obtain the actual thickness of each piece of tempered glass to be inspected. The glass thickness measurement mechanism measures the normal position of the two surfaces of the tempered glass separately, and the normal distance between the two surfaces is determined as the thickness of the tempered glass.

[0029] When a glass thickness testing agency obtains multiple thickness measurements on the same piece of tempered glass, the thickness value corresponding to the thickness measurement position closest to the candidate inclusion is selected based on the plate coordinates of the candidate inclusion. If multiple thickness measurement positions are located around the candidate inclusion, the average of these multiple thickness measurements can also be used as the tempered glass thickness corresponding to the candidate inclusion. The method for selecting the thickness value is determined before the start of the same testing batch and remains unchanged during the testing process.

[0030] The measurement error of a glass thickness testing agency is determined using glass standard sheets with known thicknesses. Repeated measurements are taken on the same glass standard sheet, and the maximum deviation of the measured value from the standard thickness is taken as the thickness measurement error.

[0031] The initial inspection positioning mechanism includes a first detection light source, a second detection light source, and a scattering imaging device. The first detection light source projects detection light into the interior of the tempered glass in a first incident direction, and the second detection light source projects detection light into the interior of the tempered glass in a second incident direction different from the first incident direction.

[0032] The first and second detection light sources operate in a time-sharing manner. When the first detection light source is operating, the scattering imaging device acquires a first scattering image; when the second detection light source is operating, the scattering imaging device acquires a second scattering image, thereby preventing the scattered light from the two incident directions from superimposing on each other in the same image.

[0033] The testing facility is located downstream of the initial testing and positioning facility along the direction of tempered glass transport, and includes an excitation light source, a detection light source, a beam splitter, a wavefront detector, a polarization analysis component, and a polarization imaging device.

[0034] An excitation light source is used to form an excitation beam, and its output intensity is periodically changed according to a first modulation frequency and a second modulation frequency determined by a processing control mechanism. The excitation beam acts on the region where the candidate inclusion is located, causing the candidate inclusion and its adjacent glass substrate to form a periodic thermal response.

[0035] The wavelength of the excitation beam is determined based on the light absorption test of the tempered glass and nickel sulfide inclusions to be tested. The selected wavelength is not required to be absorbed only by nickel sulfide, but should meet the following condition: it should be able to induce a periodic thermal response above the photothermal response detection boundary in the region where the candidate inclusion is located without causing irreversible changes in the tempered glass.

[0036] The probe light source is used to form a planar probe beam covering the core detection area and the annular reference area. After passing through tempered glass, the planar probe beam is split into a first detection beam and a second detection beam by a beam splitter.

[0037] The first detection light enters the wavefront detector. The wavefront detector is used to acquire the local wavefront deflection distribution at various locations within the core detection region and the annular reference region. Local temperature changes cause changes in the glass refractive index and refractive index gradient, and the local wavefront deflection distribution is used to characterize the photothermal response of the region where the candidate inclusion is located.

[0038] The second detection light enters the polarization imaging device after passing through the polarization analysis component. The polarization imaging device is used to acquire the polarization intensity of light in different polarization directions, and determines the optical delay at each position in the core detection area and the annular reference area based on the pre-calibrated correspondence between polarization intensity and optical delay.

[0039] The beam tracking mechanism includes a first beam deflection assembly for adjusting the direction of the excitation beam and a second beam deflection assembly for adjusting the direction of the probe beam. The processing control mechanism updates the position of the candidate inclusions based on the transport displacement of the tempered glass and controls the first and second beam deflection assemblies to move the excitation and probe beams along with the candidate inclusions.

[0040] The synchronous acquisition mechanism is connected to the excitation light source, wavefront detector, polarization imaging device, and processing and control mechanism. The output intensity change signal of the excitation beam serves as the modulation reference signal, and the wavefront detector and polarization imaging device acquire detection data according to the same sampling clock.

[0041] The edge of the tempered glass entering the initial inspection positioning mechanism is taken as the leading edge of the tempered glass, and the edge parallel to the conveying direction of the tempered glass is taken as the side edge. The intersection of the leading edge and the side edge is taken as the origin of the plate surface coordinate system, the conveying direction of the tempered glass is taken as the first plate surface coordinate axis, and the direction parallel to the surface of the tempered glass and perpendicular to the conveying direction is taken as the second plate surface coordinate axis.

[0042] The plate coordinates refer to the coordinates of the spatial position of the scattering target in the plate coordinate system after it is projected along the thickness direction of the tempered glass onto the glass surface on the side where the detection light is incident.

[0043] Burial depth refers to the normal distance from the spatial location of the scattering target to the glass surface on the incident side of the detection light.

[0044] The equivalent radius refers to the radius of the circular region after the actual area of ​​the scattering region, after point diffusion correction, is equivalent to that region. For irregularly shaped scattering targets, the equivalent radius is only used to represent their equivalent size and does not indicate that the scattering target actually has a spherical structure.

[0045] The core detection area is the detection area centered on the projected position of the candidate inclusion on the plate surface and covering the candidate inclusion and its local thermal response range.

[0046] The annular reference area is a localized background detection region located outside the core detection area and excluding candidate inclusions and other potential inclusions. The annular reference area and the core detection area are situated in adjacent regions of the same tempered glass, providing a reference response under similar glass thickness, surface condition, ambient temperature, and static tempering stress conditions.

[0047] A calibration glass with multiple known plate position marks is placed on a conveying mechanism, and the pixel position of each mark in the first scattering image and the second scattering image are obtained respectively.

[0048] Based on the known plate position and corresponding pixel position, a transformation relationship from pixel coordinates of the scattering imaging device to plate coordinates is established. When the first scattering image and the second scattering image are acquired by different scattering imaging devices, a transformation relationship between the two scattering imaging devices and the same plate coordinate system is established respectively.

[0049] Both the first and second incident directions refer to the actual propagation direction of the detection light inside the tempered glass after it enters the glass, rather than the propagation direction of the detection light in the air.

[0050] In this embodiment, calibration glass with a standard scatterer at a known position inside is preferably used to directly calibrate the propagation paths of the first and second detection lights inside the glass. During calibration, the standard scatterer is placed at different plate positions and different burial depths, and the corresponding first and second scattering positions are obtained respectively. The propagation path of the detection light inside the glass is determined based on the known spatial position of the standard scatterer.

[0051] When the glass composition or refractive index grade of the tempered glass to be tested changes, the internal propagation path of the test light is recalibrated to avoid depth error caused by direct reverse projection along the direction of the test light in the air.

[0052] A calibration glass with a standard scatterer at a known position inside is used to repeatedly perform dual incident direction detection on the same standard scatterer.

[0053] Each detection determines the spatial position of the standard scatterer based on the first projection path and the second projection path, and compares the detected spatial position with the known spatial position of the standard scatterer.

[0054] The sum of the maximum spatial position deviation caused by repeated detections and the uncertainty of the standard scatterer position is determined as the spatial positioning error. This spatial positioning error is used to determine whether the first scattering position and the second scattering position originate from the same scattering target.

[0055] In the first scattering image and the second scattering image, the image region where there is no scattering target is selected as the background region, and the background gray level and the background gray level fluctuation range are determined according to the gray level distribution of the pixels in the background region.

[0056] Pixels whose gray level is higher than the sum of the background gray level and the upper limit of the background gray level fluctuation are identified as effective scattering pixels. Adjacent effective scattering pixels are connected to form a scattering region.

[0057] The grayscale determination rules used to form the scattering region are determined before the start of the same detection batch and are not adjusted individually based on the brightness of a single scattering target to be detected.

[0058] Calibration glass with standard scatterers of different known sizes and burial depths is used to obtain the scattering region size formed by each standard scatterer. Point diffusion calibration relationships are established based on the known size, burial depth, and corresponding scattering region size of the standard scatterers.

[0059] The point diffusion calibration relationship can be stored using a calibration data table, piecewise interpolation relationship, or fitted curve. During online detection, the corresponding point diffusion calibration relationship is selected according to the burial depth of the scattering target, the size of the scattering region is corrected, and then the equivalent radius is determined based on the corrected area of ​​the scattering region.

[0060] The detection results obtained from the first and second detection optical paths are converted to the plate coordinate system and resampled to the equally spaced plate grid.

[0061] When the sampling intervals of the first and second detection optical paths are different, the larger sampling interval is used as the detection spatial resolution.

[0062] A calibration target capable of being simultaneously identified by both the wavefront detector and the polarization imaging device is set in the detection area to confirm the detection position of the calibration target in the first and second detection optical paths, respectively.

[0063] The two detection positions are transformed to the plate coordinate system to establish a spatial mapping relationship between the two detection optical paths. Registration is performed repeatedly at multiple calibration positions, and the maximum residual between the two detection positions after spatial mapping is determined as the optical path spatial registration error.

[0064] Place a motion calibration glass with known position marks within the confirmation inspection area, and allow the motion calibration glass to pass through the confirmation inspection area in the same manner as the tempered glass to be inspected.

[0065] The processing control mechanism calculates the mark position based on the conveying displacement and controls the excitation beam to illuminate the mark position. The deviation between the actual illuminating position of the excitation beam and the actual mark position is detected.

[0066] The maximum deviation occurring during repeated tracking is defined as the position tracking error. This position tracking error includes the transport displacement detection error, the beam deflection execution error, and the position error caused by control delay.

[0067] Thermal diffusivity was determined using a calibration sample without inclusions that had the same glass composition and tempering process as the tempered glass to be tested.

[0068] Periodic local photothermal excitation is formed on the calibration sample, and photothermal responses are obtained at different known distances from the excitation center. The thermal diffusivity of the calibration sample is determined based on the change of photothermal response with distance and modulation frequency.

[0069] Tempered glass with the same glass composition and tempering process uses the same thermal diffusivity. When the glass composition, color, tempering process, or coating condition changes and affects the thermal conductivity, the thermal diffusivity is redefined.

[0070] The relationship between modulation frequency and thermal diffusion length is as follows: ; in, Indicates modulation frequency The corresponding thermal diffusion length; This indicates the thermal diffusivity of tempered glass; Indicates the modulation frequency of the excitation beam; It represents pi (π).

[0071] The above relationship forms the basis for selecting the first and second modulation frequencies in this embodiment. During actual testing, the thermal diffusion length obtained based on this relationship can be corrected against the measured thermal response attenuation range of the calibration sample, resulting in a table showing the correspondence between modulation frequencies and thermal diffusion lengths.

[0072] The wavefront detector, polarization imaging device, exposure circuit, and acquisition circuit each have their own response delays. Therefore, before comparing the phases of the photothermal complex response and the dynamic birefringence complex response, it is necessary to determine the system phase deviation between the first detection optical path and the second detection optical path.

[0073] Before the start of the testing batch, a phase calibration device that can synchronously change the output of the two detection optical paths according to a known periodic signal is used to obtain the response phase of the two detection optical paths at the proposed modulation frequency.

[0074] The phase difference detected by the first and second detection optical paths is determined as the system phase deviation at the corresponding modulation frequency. After obtaining the original phase difference between the dynamic birefringence complex response and the photothermal complex response through online detection, the system phase deviation at the corresponding modulation frequency is subtracted to obtain the phase difference used for subsequent judgment.

[0075] When the wavefront detector, polarization imaging device, exposure time, sampling clock, or detection optical path is adjusted, the system phase calibration should be performed again.

[0076] As tempered glass continuously enters the initial inspection and positioning mechanism, the conveying displacement detection mechanism continuously outputs the conveying displacement. The first and second detection light sources sequentially illuminate the tempered glass according to a set scanning rhythm, and the scattering imaging device obtains the first and second scattered images respectively.

[0077] Based on the background grayscale and its fluctuation range determined before detection, the first scattering region is extracted from the first scattering image, and the first scattering position is determined; the second scattering region is extracted from the second scattering image, and the second scattering position is determined.

[0078] The first and second scattering positions are both determined using the gray-level weighted center of the effective scattering pixels within the scattering region. The same position determination method is used for the same batch of tests.

[0079] The first scattering position and the second scattering position are transformed to the plate coordinate system. The first scattering position is projected backward along the propagation direction of the first detection light inside the tempered glass into the glass to form a first projection path; the second scattering position is projected backward along the propagation direction of the second detection light inside the tempered glass into the glass to form a second projection path.

[0080] The minimum distance between the first and second projection paths is determined within the thickness range of the tempered glass. When the minimum distance is greater than the spatial positioning error, the first and second scattering positions are not paired.

[0081] When the minimum distance is not greater than the spatial positioning error, and the first and second projection path positions corresponding to the minimum distance are both located between the two surfaces of the tempered glass, the first scattering position and the second scattering position are determined as the scattering positions of the same scattering target under different incident directions.

[0082] The midpoint of the line connecting the two projection path positions corresponding to the minimum distance is determined as the spatial position of the scattering target. The spatial position is projected along the glass thickness direction onto the surface of the detection light incident side to obtain the plate coordinates of the scattering target; the normal distance from the spatial position to the surface of the detection light incident side is determined as the burial depth.

[0083] The equivalent radius of the scattering target is determined based on the size of the scattering region formed by the scattering target and the point diffusion calibration relationship corresponding to the burial depth.

[0084] When the burial depth is greater than the equivalent radius and less than the difference between the tempered glass thickness and the equivalent radius, the equivalent boundary of the scattering target does not intersect with the two surfaces of the tempered glass, and the scattering target is identified as a candidate inclusion.

[0085] When the burial depth does not meet the above conditions, the scattering target will be included in the surface or near-surface defect detection results and will not be included in the subsequent nickel sulfide confirmation detection.

[0086] After the initial inspection and positioning mechanism identifies the candidate inclusions, the processing control mechanism records the plate coordinates of the candidate inclusions and the corresponding initial conveying displacement.

[0087] Based on the fixed distance along the conveying direction between the initial inspection location and the confirmation inspection area, and the current conveying displacement corresponding to the candidate inclusion, the remaining conveying distance between the candidate inclusion and the confirmation inspection area is determined.

[0088] The processing control mechanism determines the current conveying speed of the tempered glass based on the changes in conveying displacement between adjacent sampling times. To reduce the impact of minute speed fluctuations in the conveying rollers, the current conveying speed can be determined by using the changes in conveying displacement over multiple consecutive sampling times.

[0089] The estimated time when candidate inclusions will arrive at the confirmation detection area is determined based on the remaining conveying distance and the current conveying speed. Before the candidate inclusions arrive at the confirmation detection area, the remaining conveying distance and estimated time are continuously updated based on the latest conveying displacement.

[0090] When tempered glass stops being transported, the calculation of the estimated time is paused, and the correlation between the candidate inclusions and the transport displacement at the time of the stop is retained; after the tempered glass is transported again, the remaining transport distance is re-determined based on the new transport displacement.

[0091] When the tempered glass moves in the reverse direction, the position of the candidate inclusions is updated according to the direction of the conveying displacement, and the expected arrival time under the forward conveying state is not used.

[0092] After a candidate inclusion enters the confirmation detection area, the position of the candidate inclusion within the confirmation detection area is updated based on the continuously acquired transport displacement, and the excitation beam and probe beam are controlled to follow the movement of the candidate inclusion.

[0093] The processing control mechanism determines the first modulation frequency and the second modulation frequency based on the equivalent radius of the candidate inclusion, the burial depth, the thickness of the tempered glass, and the thermal diffusivity of the tempered glass.

[0094] Based on the correspondence between modulation frequency and thermal diffusion length, a first modulation frequency is determined within the modulation frequency range where the corresponding thermal diffusion length is not greater than the equivalent radius of the candidate inclusion, and the thermal diffusion length corresponding to the first modulation frequency is determined as the first thermal diffusion length.

[0095] The first modulation frequency is used to obtain a local response that is mainly concentrated in the neighborhood of the candidate inclusions, so as to reduce the influence of the glass substrate far away from the candidate inclusions on the detection results.

[0096] Based on the thickness of the tempered glass and the burial depth of the candidate inclusions, a first distance from the center of the candidate inclusion to the glass surface on the side where the detection light is incident is determined, and a second distance from the center of the candidate inclusion to the glass surface on the other side is determined. The smaller value between the first distance and the second distance is determined as the nearest surface distance.

[0097] A second modulation frequency is determined within a modulation frequency range where the corresponding thermal diffusion length is greater than the equivalent radius of the candidate inclusion and less than the nearest surface distance, and the thermal diffusion length corresponding to the second modulation frequency is determined as the second thermal diffusion length.

[0098] The second modulation frequency is used to make the thermal response cross the interface between the candidate inclusion and the glass substrate, while avoiding the main thermal diffusion range from extending to the tempered glass surface, thereby reducing the effects of surface convection heat transfer, surface coating and surface contamination.

[0099] The first modulation frequency is higher than the second modulation frequency, and the first thermal diffusion length is less than the second thermal diffusion length.

[0100] When the modulation frequencies that meet the conditions form a continuous interval, select the modulation frequency whose corresponding thermal diffusivity length is in the middle of the allowable range, so as to reserve a margin for the measurement error of thermal diffusivity, equivalent radius and burial depth.

[0101] If there is no modulation frequency within the allowable modulation frequency range of the excitation light source that satisfies the second thermal diffusion length condition, the candidate inclusion is identified as the object of re-inspection.

[0102] After a candidate inclusion enters the confirmation detection area, its projected position within the confirmation detection area is determined based on the candidate inclusion position updated with the conveying displacement.

[0103] A core detection area is set with the projected position as the center. When the core detection area is a circular region, its radius is not less than the sum of the second thermal diffusion length and the position tracking error.

[0104] When the cross section of the excitation beam or the thermal response distribution is non-circular, the core detection area can be set as an ellipse or other closed region. However, the boundaries of the core detection area in each direction should cover the thermal response range corresponding to the second thermal diffusion length and include the position tracking error.

[0105] A ring-shaped reference area is set up around the core detection area. The distance between the inner boundary of the ring-shaped reference area and the outer boundary of the core detection area is not less than the sum of the detection spatial resolution and the position tracking error, so as to reduce the impact of leakage of the core detection area response into the ring-shaped reference area.

[0106] The outer boundary of the annular reference area is determined based on the distance from the projection position of the candidate inclusion to the edge of the tempered glass, and the distance to the projection positions of other candidate inclusions. The annular reference area must not extend to the edge of the tempered glass, nor may it include the core detection area corresponding to other candidate inclusions.

[0107] Sampling points located outside the glass edge, at detector dead spots, or at locations where the detection signal is saturated are removed from the detection area. The area of ​​the plate corresponding to the remaining sampling points after removal is the effective sampling area.

[0108] When the effective sampling area of ​​the annular reference area is smaller than the effective sampling area of ​​the core detection area, the candidate inclusions are identified as objects for re-inspection.

[0109] When the effective sampling area of ​​the annular reference area is not less than the effective sampling area of ​​the core detection area, the positions of the core detection area and the annular reference area are updated synchronously according to the position of the candidate inclusions.

[0110] Before formal testing, the excitation beam is turned off, and the probe beam, wavefront detector, polarization imaging device, and synchronous acquisition mechanism are kept in formal testing state to obtain pre-test benchmark data for the core testing area and the annular reference area.

[0111] First, the output intensity of the excitation beam is periodically changed according to the first modulation frequency to perform periodic local photothermal excitation on the region where the candidate inclusions are located. Excitation is stopped after the detection period corresponding to the first modulation frequency ends.

[0112] The processing and control mechanism continuously monitors the photothermal response and dynamic birefringence response of the core detection area. Once both responses return to the baseline fluctuation range before detection, periodic local photothermal excitation is implemented according to the second modulation frequency to avoid the superposition of responses corresponding to the two modulation frequencies.

[0113] The planar probe beam covers the core detection area and the annular reference area, and moves with the candidate inclusions.

[0114] The wavefront detector acquires the local wavefront deflection distribution of the first detection light at corresponding positions in the core detection region and the annular reference region. The periodic changes in the local wavefront deflection distribution that are in sync with the modulation reference signal are defined as the photothermal response.

[0115] The polarization imaging device acquires the polarization intensity of the second detection light in at least four polarization directions, with an angular difference of 45 degrees between any two adjacent polarization directions. Based on the calibration relationship between polarization intensity and optical retardation established before detection, the optical retardation at each position within the core detection area and the annular reference area is determined.

[0116] The difference between the optical retardation during periodic local photothermal excitation and the optical retardation before periodic local photothermal excitation is defined as the dynamic birefringence response. This is then used to subtract the static optical retardation caused by the original static tempering stress of the tempered glass.

[0117] When a wavefront detector or polarization imaging device produces saturated pixels, the corresponding detection period becomes invalid. Reduce the probe light intensity or adjust the exposure time and re-detect.

[0118] After the first modulation frequency detection is completed, if neither of the two responses returns to the pre-detection reference fluctuation range, the second modulation frequency detection will not begin; if the response still fails to recover after the allowable recovery time, the candidate inclusion will be identified as a re-inspection target. The allowable recovery time is determined based on the longest time required for the inclusion-free calibration sample to recover to the pre-detection reference fluctuation range under the same excitation conditions.

[0119] For each sampling position on the plate surface, the photothermal response and dynamic birefringence response are synchronously demodulated with the modulation reference signal to extract the in-phase component and the quadrature component, and the corresponding complex response is formed by the in-phase component and the quadrature component.

[0120] The detection period includes an integer number of modulation cycles to reduce the impact of non-integer cycle truncation on the phase results.

[0121] Spatially register the detection results of the first and second detection optical paths and resample them to the same equally spaced plate grid.

[0122] For the first modulation frequency and the second modulation frequency, the photothermal complex response at each effective sampling position in the core detection area is averaged to form the core photothermal complex response; the photothermal complex response at each effective sampling position in the annular reference area is averaged to form the reference photothermal complex response.

[0123] The difference between the core photothermal complex response and the reference photothermal complex response is defined as the differential photothermal complex response.

[0124] The core dynamic birefringence complex response and the reference dynamic birefringence complex response are formed in the same way, and the difference between the core dynamic birefringence complex response and the reference dynamic birefringence complex response is determined as the differential dynamic birefringence complex response.

[0125] The regional averaging is performed directly on the in-phase and quadrature components of the complex response, without averaging the amplitude first and then the phase, in order to avoid errors when the phase crosses the period boundary.

[0126] When the excitation beam is off and the probe beam and synchronous acquisition mechanism are in the formal detection state, the same detection period and demodulation method as the formal detection are used to obtain the excitation-free differential photothermal complex response and the excitation-free differential dynamic birefringence complex response corresponding to the first modulation frequency and the second modulation frequency, respectively.

[0127] At least three unexcited detection periods are continuously acquired. The sum of the maximum value of the differential photothermal complex response amplitude in each unexcited detection period and the photothermal detection repeatability error is determined as the photothermal response detection boundary at the corresponding modulation frequency.

[0128] The sum of the maximum value of the differential dynamic birefringence complex response amplitude in each unexcited detection period and the polarization detection repeatability error is determined as the birefringence response detection boundary at the corresponding modulation frequency.

[0129] The repeatability error of photothermal detection and polarization detection were obtained by repeatedly detecting the same impurity-free calibration area, and the maximum absolute deviation of the repeated detection result relative to the average result was used.

[0130] The first modulation frequency and the second modulation frequency form their respective photothermal response detection boundaries and birefringence response detection boundaries, and the detection boundary under one modulation frequency is not directly used for another modulation frequency.

[0131] Only when the amplitude of the differential photothermal complex response at the corresponding modulation frequency is higher than the photothermal response detection boundary, and the amplitude of the differential dynamic birefringence complex response is higher than the birefringence response detection boundary, are the corresponding amplitude ratio, phase difference, and response position calculated.

[0132] If any response at any modulation frequency fails to reach the corresponding detection boundary, the candidate inclusion is identified as a re-inspection target.

[0133] For the first modulation frequency, the ratio of the amplitude of the differential dynamic birefringence complex response to the amplitude of the differential photothermal complex response is determined as the first amplitude ratio, and the difference between the phase of the dynamic birefringence complex response after system phase correction and the phase of the photothermal complex response is determined as the first phase difference.

[0134] For the second modulation frequency, the ratio of the amplitude of the differential dynamic birefringence complex response to the amplitude of the differential photothermal complex response is determined as the second amplitude ratio, and the difference between the phase of the dynamic birefringence complex response after system phase correction and the phase of the photothermal complex response is determined as the second phase difference.

[0135] The heat-stress transfer parameters are expressed as follows: ; in, Indicates the first Thermal-stress transfer parameters at each modulation frequency; Indicates the first Differential dynamic birefringence complex response at each modulation frequency; Indicates the first Differential photothermal complex response at each modulation frequency; This indicates the corresponding amplitude ratio; This represents the phase difference after deducting the system phase deviation; j represents the imaginary unit. When hour, and These are the first amplitude ratio and the first phase difference, respectively; when At that time, the second amplitude ratio and the second phase difference are respectively.

[0136] The amplitude ratio change is determined based on the first amplitude ratio and the second amplitude ratio, and the phase difference change is determined based on the first phase difference and the second phase difference. The amplitude ratio change can be expressed as the ratio of the first amplitude ratio to the second amplitude ratio, and the phase difference change can be expressed as the difference between the first phase difference and the second phase difference.

[0137] If the second amplitude ratio cannot form a valid value, the change in amplitude ratio is not calculated, and the candidate inclusion is identified as the object of re-inspection.

[0138] The thermal diffusion range corresponding to the first modulation frequency is mainly located in the neighborhood of the candidate inclusion, while the thermal diffusion range corresponding to the second modulation frequency spans the interface between the candidate inclusion and the glass substrate. The changes in amplitude ratio and phase difference at the two modulation frequencies reflect the frequency characteristics of how the candidate inclusion and its interface convert the local thermal response into a thermoelastic stress response.

[0139] For the first modulation frequency and the second modulation frequency, the centroid of the photothermal response energy is determined based on the amplitude of the photothermal complex response at each position in the core detection area and the corresponding plate coordinates; the centroid of the birefringence response energy is determined based on the amplitude of the dynamic birefringence complex response and the corresponding plate coordinates.

[0140] The energy centroid is determined by weighted averaging of the complex response amplitudes at each sampling location with respect to the corresponding plate coordinates. Only valid sampling locations above the corresponding response detection boundary are used in the energy centroid calculation.

[0141] When the distance between the centroid of the photothermal response energy and the centroid of the birefringence response energy is no greater than the sum of the detection spatial resolution, the position tracking error, and the optical path spatial registration error at both the first and second modulation frequencies, the consistency of the response position is determined.

[0142] After the first and second phase differences are corrected by the system phase, they are phase-expanded to represent the phase lag of the dynamic birefringence response relative to the photothermal response.

[0143] When both the first phase difference and the second phase difference indicate that the dynamic birefringence response lags behind the photothermal response within the corresponding detection period, the phase causality consistency is confirmed.

[0144] When the phase difference is near the boundary of the phase period, the phase expansion is performed using the continuous phase results of adjacent detection periods; when a continuous phase result cannot be formed, the candidate inclusion is identified as the object of re-inspection.

[0145] The calibration sample used to form the range of nickel sulfide determination parameters should be tempered glass that has been confirmed by independent component analysis to contain nickel sulfide inclusions.

[0146] Inclusions can be initially identified by Raman spectroscopy, and then inclusion samples can be obtained by slicing, local grinding or other methods that can expose inclusions. Elemental analysis can then confirm that the inclusions contain nickel and sulfur.

[0147] Raman spectroscopy and elemental composition analysis are used to establish nickel sulfide calibration samples and are not part of the online detection process.

[0148] Simultaneously, bubbles, glass inclusions, and non-nickel sulfide solid inclusions that had been confirmed by testing were used as control samples. The control samples were tested using the same spatial positioning, dual-frequency excitation, complex response extraction, and transfer parameter calculation process as the nickel sulfide calibration samples.

[0149] Record the tempered glass thickness, equivalent radius of candidate inclusions, burial depth, first modulation frequency, second modulation frequency, first amplitude ratio, first phase difference, second amplitude ratio, second phase difference, amplitude ratio change, and phase difference change for each calibration sample and control sample.

[0150] Nickel sulfide calibration samples were grouped according to tempered glass thickness, equivalent radius, and burial depth. The thickness range was determined based on the thickness grade of the tempered glass product, the width of the equivalent radius range was not less than twice the measurement uncertainty of the equivalent radius, and the width of the burial depth range was not less than twice the measurement uncertainty of the burial depth.

[0151] Each grouping interval is fixed before calibration begins. Except for the last interval, which includes both the upper and lower boundaries, the remaining intervals include the lower boundary but not the upper boundary, to ensure that candidate inclusions located at the grouping boundaries can be assigned to a unique group.

[0152] Repeated tests were performed on each nickel sulfide calibration sample in the same group. First, the average results of each parameter of each sample were determined, and then the minimum and maximum values ​​of the average results of each sample in the same group were determined.

[0153] The maximum absolute deviation of repeated test results relative to the corresponding average results is defined as the repeatability margin. The minimum value of the average results of parameters within the same group minus the repeatability margin is used as the lower limit of the parameter range, and the maximum value of the average results of parameters plus the repeatability margin is used as the upper limit of the parameter range.

[0154] The first amplitude ratio, the first phase difference, the second amplitude ratio, the second phase difference, the change in amplitude ratio, and the change in phase difference each form a corresponding parameter range.

[0155] After the parameter range is determined, verification is performed using nickel sulfide confirmation samples and non-nickel sulfide control samples that were not involved in the range formation. If all the determination parameters of a certain non-nickel sulfide control sample fall within the corresponding nickel sulfide parameter range, this parameter range should not be used for online determination. Instead, the parameters should be regrouped, the parameter range adjusted, or additional detection parameters that can distinguish the control sample should be added.

[0156] The instruction manual should not include parameter values ​​that have not been obtained through actual testing. Specific parameter ranges should be obtained by the applicant based on actual nickel sulfide calibration samples and non-nickel sulfide control samples, and should not be based on theoretical calculations or artificially created for writing purposes.

[0157] If there is no corresponding calibration group or the corresponding group has not yet completed independent sample verification, the candidate inclusions will be identified as the objects of re-examination, and the parameter range of the adjacent group will not be used as a substitute.

[0158] When the amplitudes of the differential photothermal complex response at both the first and second modulation frequencies are higher than the corresponding photothermal response detection boundaries, and the amplitudes of the differential dynamic birefringence complex response are both higher than the corresponding birefringence response detection boundaries, it is determined that the responses at both modulation frequencies are valid.

[0159] Based on the fact that the responses at both modulation frequencies are valid, the response positions are consistent, and the phase causality is consistent, the corresponding nickel sulfide calibration group is determined according to the thickness of the tempered glass to be tested, the equivalent radius of the candidate inclusions, and the burial depth.

[0160] When the first amplitude ratio, the first phase difference, the second amplitude ratio, the second phase difference, the amplitude ratio change, and the phase difference change are all within the parameter range of the corresponding nickel sulfide calibration group, the candidate inclusion is determined to be a nickel sulfide inclusion.

[0161] When the differential photothermal complex response or differential dynamic birefringence complex response at any modulation frequency fails to reach the corresponding detection boundary, or the response position consistency is not established, or the phase causal consistency is not established, the candidate inclusion is identified as the object of re-inspection.

[0162] When the responses at both modulation frequencies are valid, and the consistency of response position and phase causality are both established, but at least one thermal-stress transfer parameter or its variation relationship is not within the range of the corresponding nickel sulfide parameter, the candidate inclusion is determined to be a non-nickel sulfide inclusion.

[0163] The determination of non-nickel sulfide inclusions only indicates that the dual-frequency thermal-stress transfer characteristics of the candidate inclusions do not conform to the nickel sulfide calibration characteristics, and does not directly determine whether the inclusion is a bubble, glass stone or other specific material.

[0164] When a candidate inclusion is identified as a nickel sulfide inclusion, the processing control mechanism outputs the plate coordinates, burial depth, and equivalent radius of the nickel sulfide inclusion.

[0165] The test results may also include the tempered glass thickness, the first modulation frequency, the second modulation frequency, the first amplitude ratio, the first phase difference, the second amplitude ratio, the second phase difference, the amplitude ratio change, the phase difference change, the response position consistency result, and the phase causality consistency result.

[0166] The test results are linked to the production markings of the tempered glass to be tested, forming a test record that can be traced back to the corresponding tempered glass and the corresponding panel location.

[0167] Tempered glass of different colors, thicknesses, and coating conditions absorbs the excitation beam to varying degrees. Before performing formal periodic local photothermal excitation, a trial excitation is applied to the region containing the candidate inclusion.

[0168] The single-cycle excitation energy of the trial excitation is lower than that of the formal periodic local photothermal excitation. Following the same photothermal response acquisition and regional differential method as the formal detection, the differential photothermal complex response corresponding to the trial excitation is obtained.

[0169] Based on the correspondence between the excitation energy of the trial excitation and the amplitude of the corresponding differential photothermal complex response, the photothermal response sensitivity of the region where the candidate inclusion is located is determined.

[0170] The safety response boundary is determined using a glass calibration sample with the same thickness, glass composition, and coating condition as the tempered glass to be tested.

[0171] The excitation energy of the glass calibration sample is gradually increased. After each excitation, the glass calibration sample is allowed to return to its stable temperature before excitation, and then the optical retardation before and after excitation is compared.

[0172] If the difference in optical delay before and after the glass calibration sample returns to its stable temperature before excitation is still greater than the polarization detection repeatability error, it is determined that the excitation level caused an irreversible change in optical delay.

[0173] The differential photothermal complex response amplitude corresponding to the highest excitation level that does not cause an irreversible change in optical delay is determined as the safe response boundary after retaining the safety margin corresponding to the equipment control error.

[0174] The safety margin is obtained by repeatedly exciting the highest excitation level, using the maximum positive deviation of the differential photothermal complex response amplitude relative to the average amplitude during repeated excitation.

[0175] The output power and duty cycle of the formal periodic local photothermal excitation are adjusted according to the photothermal response sensitivity so that the amplitude of the differential photothermal complex response is higher than the photothermal response detection boundary but lower than the safety response boundary.

[0176] During the formal testing process, when the amplitude of the differential photothermal complex response reaches the safety response boundary, the periodic local photothermal excitation is immediately stopped, and the candidate inclusions are identified as the objects for re-inspection.

[0177] Candidate inclusions are not generated when the first projection path and the second projection path cannot meet the spatial positioning error requirements.

[0178] When the equivalent boundary of the scattering target intersects with any surface of the tempered glass, the scattering target is included in the surface or near-surface defect detection results.

[0179] If there is no second modulation frequency that meets the thermal diffusion length requirement, the candidate inclusion is identified as the object of re-inspection.

[0180] If an annular reference area with an effective sampling area not less than the effective sampling area of ​​the core detection area cannot be formed, the candidate inclusions will be identified as objects for re-inspection.

[0181] When the spatial registration error between the first and second detection optical paths exceeds the calibrated allowable range, detection is paused and spatial registration is performed again.

[0182] If any detection optical path has saturated pixels that cannot be eliminated by reducing the detection light intensity or adjusting the exposure time, the candidate inclusion will be identified as a re-inspection target.

[0183] If the response does not recover to the baseline fluctuation range before the first modulation frequency detection is completed, the second modulation frequency detection will not be performed.

[0184] If the differential photothermal complex response or differential dynamic birefringence complex response at any modulation frequency does not reach the corresponding detection boundary, the corresponding amplitude ratio, phase difference, and response position are not calculated.

[0185] If the first or second amplitude ratio cannot form a valid value, the change in amplitude ratio is not calculated.

[0186] When the phase difference cannot be converted into a continuous phase lag through system phase correction and phase expansion, phase causality consistency is not determined.

[0187] If a candidate inclusion does not have a corresponding nickel sulfide calibration group, or if the corresponding group has not yet completed independent sample verification, the candidate inclusion will be identified as a candidate for re-inspection.

[0188] When the trial excitation or formal periodic local photothermal excitation reaches the safety response boundary, the excitation is stopped and the candidate inclusions are identified as objects for re-inspection.

[0189] The plate coordinates obtained in step S1 are used for candidate inclusion position updates and beam tracking in step S2, and for setting the core detection area and annular reference area in step S4.

[0190] The burial depth and equivalent radius obtained in step S1, the tempered glass thickness obtained by the glass thickness detection mechanism, and the thermal diffusivity obtained by calibration are used together in step S3 to determine the first modulation frequency, the second modulation frequency, the first thermal diffusivity, and the second thermal diffusivity.

[0191] The thermal diffusion range corresponding to the first modulation frequency is mainly located in the neighborhood of the candidate inclusion; the thermal diffusion range corresponding to the second modulation frequency spans the interface between the candidate inclusion and the glass substrate, but does not extend to the nearest tempered glass surface. Therefore, the first and second modulation frequencies are not arbitrarily selected detection frequencies.

[0192] The core detection area is used to acquire detection data including the response of candidate inclusions, while the annular reference area is used to acquire local reference data under similar glass thickness, surface condition, ambient temperature, and static tempering stress background.

[0193] The differential photothermal complex response represents the effective thermal response after subtracting the local background, while the differential dynamic birefringence complex response represents the effective thermoelastic stress response after subtracting the static stress background.

[0194] The thermal-stress transfer parameters are characterized by the differential photothermal complex response as the actual thermal input and the differential dynamic birefringence complex response as the thermoelastic stress output. They are formed by the amplitude ratio between the two and the phase difference after system phase correction, rather than a simple weighting of two independent detection results.

[0195] Response location consistency is used to confirm that the photothermal response and dynamic birefringence response originate from the same candidate inclusion region; phase causality consistency is used to confirm that the dynamic birefringence response has a hysteresis relationship with the photothermal response that conforms to the heat transfer and thermoelastic stress formation process.

[0196] Candidate inclusions are only identified as nickel sulfide inclusions when both frequency responses are valid, the response positions are consistent, the phase causality is consistent, and the dual-frequency thermal-stress transfer parameters and their variation relationships are all within the nickel sulfide calibration range confirmed by independent component testing.

[0197] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An online detection method for nickel sulfide inclusions in tempered glass, characterized in that, Includes the following steps: S1: During the continuous conveying of tempered glass, the conveying displacement is obtained. The tempered glass is scanned with detection light from at least two different incident directions. The plate coordinates, burial depth and equivalent radius of the candidate inclusions are determined based on the scattering position of the same scattering target under different incident directions. S2: Determine the moment when the candidate inclusion arrives at the confirmation detection area based on the plate coordinates and conveying displacement, and control the excitation beam and probe beam to follow the movement of the candidate inclusion; S3: Determine the first modulation frequency and the second modulation frequency based on the equivalent radius, burial depth, thickness of tempered glass and thermal diffusivity, so that the first thermal diffusivity length corresponding to the first modulation frequency is not greater than the equivalent radius, and the second thermal diffusivity length corresponding to the second modulation frequency is greater than the equivalent radius and less than the distance from the center of the candidate inclusion to the nearest glass surface. S4: Set up a core detection area with the projection position of the candidate inclusion as the center, and set up an annular reference area without the candidate inclusion outside the core detection area. Perform periodic local photothermal excitation according to the first modulation frequency and the second modulation frequency in sequence, and simultaneously acquire the photothermal response and dynamic birefringence response of the two detection areas. Demodulate the photothermal response and dynamic birefringence response synchronously according to the output intensity change of the excitation beam to obtain the photothermal complex response and dynamic birefringence complex response. S5: For the first modulation frequency and the second modulation frequency, the difference between the photothermal complex response of the core detection area and the annular reference area is obtained to obtain the differential photothermal complex response. The difference between the dynamic birefringence complex response of the core detection area and the annular reference area is obtained to obtain the differential dynamic birefringence complex response. The heat-stress transfer parameters are determined based on the amplitude ratio and phase difference between the differential dynamic birefringence complex response and the differential photothermal complex response. S6: Compare the thermal-stress transfer parameters and their variation relationship between the first modulation frequency and the second modulation frequency with the parameter range formed by the nickel sulfide inclusion calibration sample confirmed by composition detection, and combine the response position consistency and phase causality consistency of the differential photothermal complex response and the differential dynamic birefringence complex response to determine whether the candidate inclusion is a nickel sulfide inclusion. S7: When determining that a candidate inclusion is a nickel sulfide inclusion, output the plate coordinates, burial depth and equivalent radius of the nickel sulfide inclusion.

2. The online detection method for nickel sulfide inclusions in tempered glass according to claim 1, characterized in that, In step S1, the tempered glass is scanned with detection light in the first incident direction to obtain the first scattering position of the scattering target in the first incident direction, and the tempered glass is scanned with detection light in the second incident direction to obtain the second scattering position of the scattering target in the second incident direction. The first scattering position and the second scattering position are transformed into a plate coordinate system with the front edge and side edge of the tempered glass as the coordinate reference. The first scattering position is back-projected along the propagation path inside the tempered glass along the first incident direction to form a first projection path. The second scattering position is back-projected along the propagation path inside the tempered glass along the second incident direction to form a second projection path. When the minimum distance between the first projection path and the second projection path is not greater than the spatial positioning error of the detection light scan, and the positions of the two projection paths corresponding to the minimum distance are both located between the two surfaces of the tempered glass, the first scattering position and the second scattering position are determined as the scattering positions of the same scattering target under different incident directions. The midpoint of the line connecting the two projection path positions is determined as the spatial position of the scattering target. The projection coordinates of the spatial position on the glass surface on the side where the detection light is incident are determined as the plate coordinates. The normal distance from the spatial position to the glass surface on the side where the detection light is incident is determined as the burial depth. The equivalent radius is determined based on the size of the scattering region formed by the scattering target and the point diffusion calibration relationship formed by standard scatterers of different known sizes and burial depths. Scattering targets with a burial depth greater than the equivalent radius but less than the difference between the thickness of the tempered glass and the equivalent radius are identified as candidate inclusions.

3. The online detection method for nickel sulfide inclusions in tempered glass according to claim 2, characterized in that, In step S2, the conveying speed of the tempered glass is determined based on the changes in conveying displacement obtained at adjacent sampling times; Based on the plate coordinates of the candidate inclusion, the distance between the detection light scanning position and the confirmation detection area along the tempered glass conveying direction, and the current conveying displacement, the remaining conveying distance between the candidate inclusion and the confirmation detection area is determined, and the time when the candidate inclusion arrives at the confirmation detection area is determined based on the remaining conveying distance and the conveying speed. After the candidate inclusion reaches the confirmation detection area, the position of the candidate inclusion in the confirmation detection area is updated according to the continuously acquired transport displacement, and the direction of the excitation beam and the probe beam is adjusted according to the updated position. When tempered glass stops being transported, the pause time is determined and the correspondence between candidate inclusions and the current transport displacement is retained. When tempered glass resumes transport or the transport direction is changed, the position of candidate inclusions and the time when they arrive at the confirmation detection area are re-determined based on the new transport displacement.

4. The online detection method for nickel sulfide inclusions in tempered glass according to claim 3, characterized in that, In step S3, the thermal diffusion length corresponding to different modulation frequencies is determined based on the thermal diffusivity of tempered glass and the calibration correspondence between modulation frequency and thermal diffusion length. A first modulation frequency is determined within a modulation frequency range where the corresponding thermal diffusion length is not greater than the equivalent radius, and the thermal diffusion length corresponding to the first modulation frequency is determined as the first thermal diffusion length. Based on the thickness and burial depth of the tempered glass, the distances from the center of the candidate inclusion to the two surfaces of the tempered glass are determined, and the smaller of the two distances is determined as the nearest surface distance. A second modulation frequency is determined within the modulation frequency range where the corresponding thermal diffusion length is greater than the equivalent radius and less than the nearest surface distance, and the thermal diffusion length corresponding to the second modulation frequency is determined as the second thermal diffusion length, wherein the first modulation frequency is higher than the second modulation frequency; When there is no modulation frequency that satisfies the second thermal diffusion length condition, the candidate inclusions are identified as objects for re-inspection.

5. The online detection method for nickel sulfide inclusions in tempered glass according to claim 4, characterized in that, In step S4, the projection position of the candidate inclusion in the confirmed detection area is determined based on the candidate inclusion position updated with the conveying displacement, and the position tracking error is determined based on the maximum deviation between the candidate inclusion position and the excitation beam illumination position. A core detection area is set with the projection position as the center, and the distance between the outer boundary of the core detection area and the projection position is not less than the sum of the second thermal diffusion length and the position tracking error; The detection spatial resolution is determined based on the spatial sampling interval of the probe beam in the plate coordinate system. An annular reference area is set around the core detection area, such that the distance between the inner boundary of the annular reference area and the outer boundary of the core detection area is not less than the sum of the detection spatial resolution and the position tracking error. The outer boundary of the annular reference area is determined based on the distance from the projection position to the edge of the tempered glass and the distance to the projection positions of other candidate inclusions, so that the annular reference area does not extend to the edge of the tempered glass and does not contain other candidate inclusions. When the effective sampling area of ​​the annular reference area is smaller than the effective sampling area of ​​the core detection area, the candidate inclusions are identified as objects for re-inspection.

6. The online detection method for nickel sulfide inclusions in tempered glass according to claim 5, characterized in that, In step S4, the detection beam is expanded into a planar detection beam covering the core detection area and the annular reference area, and the planar detection beam after passing through the tempered glass is divided into a first detection beam and a second detection beam. The local wavefront deflection distribution of the first detection light at the corresponding positions in the core detection area and the annular reference area is obtained, and the periodic change of the local wavefront deflection distribution relative to the change in the output intensity of the excitation beam is determined as the photothermal response. The polarization intensity of the second detection light in at least four different polarization directions is obtained. The optical delay at each position in the core detection area and the annular reference area is determined based on the polarization intensity. The difference between the optical delay during the periodic local photothermal excitation and the optical delay before the periodic local photothermal excitation is determined as the dynamic birefringence response. The local wavefront deflection distribution and optical delay are mapped to the same plate coordinate system, and the photothermal response and dynamic birefringence response are demodulated synchronously with the output intensity change of the excitation beam to obtain the photothermal complex response and dynamic birefringence complex response at each position in the core detection area and the annular reference area.

7. The online detection method for nickel sulfide inclusions in tempered glass according to claim 6, characterized in that, In step S5, for the first modulation frequency and the second modulation frequency, the in-phase component and the quadrature component of the photothermal complex response at each position in the core detection area are normalized by area to form the core photothermal complex response. The in-phase component and the quadrature component of the photothermal complex response at each position in the annular reference area are normalized by area to form the reference photothermal complex response. The difference between the core photothermal complex response and the reference photothermal complex response is determined as the differential photothermal complex response. The core dynamic birefringence complex response and the reference dynamic birefringence complex response are formed in the same way, and the difference between the core dynamic birefringence complex response and the reference dynamic birefringence complex response is determined as the differential dynamic birefringence complex response. Without periodic local photothermal excitation, differential photothermal complex response and differential dynamic birefringence complex response are continuously acquired in at least three detection periods. The maximum value of the corresponding complex response amplitude and the sum of the corresponding detection repeatability error in each detection period are determined as the photothermal response detection boundary and the birefringence response detection boundary, respectively. The response at the corresponding modulation frequency is determined to be a valid response only when the amplitude of the differential photothermal complex response at the corresponding modulation frequency is higher than the photothermal response detection boundary, and the amplitude of the differential dynamic birefringence complex response is higher than the birefringence response detection boundary.

8. The online detection method for nickel sulfide inclusions in tempered glass according to claim 7, characterized in that, Before the start of the testing batch, a phase calibration device that can cause the first and second detection optical paths to produce periodic detection changes according to the same period signal is used to determine the system phase deviation between the two detection optical paths at the first modulation frequency and the second modulation frequency, respectively. In step S5, the ratio of the amplitude of the differential dynamic birefringence complex response to the amplitude of the differential photothermal complex response at the first modulation frequency is determined as the first amplitude ratio, and the original phase difference of the two differential complex responses at the first modulation frequency is determined as the first phase difference after deducting the corresponding system phase deviation. The ratio of the amplitude of the differential dynamic birefringence complex response to the amplitude of the differential photothermal complex response at the second modulation frequency is determined as the second amplitude ratio, and the original phase difference of the two differential complex responses at the second modulation frequency is determined as the second phase difference after deducting the corresponding system phase deviation. The first amplitude ratio, the first phase difference, the second amplitude ratio, and the second phase difference are determined as heat-stress transfer parameters. Based on the ratio between the first amplitude ratio and the second amplitude ratio, and the difference between the first phase difference and the second phase difference, the variation relationship of the heat-stress transfer parameters between the two modulation frequencies is determined.

9. The online detection method for nickel sulfide inclusions in tempered glass according to claim 8, characterized in that, In step S6, for the first modulation frequency and the second modulation frequency, the photothermal response energy centroid is determined according to the photothermal complex response amplitude and plate coordinates of each effective sampling position in the core detection area, and the birefringence response energy centroid is determined according to the dynamic birefringence complex response amplitude and plate coordinates of each effective sampling position in the core detection area. When the distance between the centroid of the photothermal response energy and the centroid of the birefringence response energy is no greater than the sum of the detection spatial resolution, the position tracking error, and the optical path spatial registration error between the two detection optical paths at both modulation frequencies, the consistency of the response position is determined. When both the first and second phase differences after system phase correction and phase expansion indicate that the differential dynamic birefringence complex response lags behind the differential photothermal complex response, it is determined that the phase causality consistency holds. Tempered glass containing nickel sulfide inclusions, as confirmed by Raman spectroscopy and elemental analysis, was used as the nickel sulfide calibration sample. The nickel sulfide calibration samples were grouped according to the tempered glass thickness, the equivalent radius of the nickel sulfide inclusions, and the burial depth. Based on the repeated detection results of each group of calibration samples, the parameter ranges of the first amplitude ratio, the first phase difference, the second amplitude ratio, the second phase difference, and their variation relationships were formed. The parameter ranges were verified using a control sample without nickel sulfide inclusions, which was confirmed by detection. When the responses at both modulation frequencies are valid, the response position consistency and phase causality consistency are both valid, and the thermal-stress transfer parameters and their variation relationships are all within the parameter range corresponding to the thickness of the tempered glass to be tested, the equivalent radius of the candidate inclusion, and the burial depth, the candidate inclusion is identified as a nickel sulfide inclusion. If the response at any modulation frequency is not a valid response, or if the consistency of response position or phase causality is not established, the candidate inclusion is identified as a re-inspection object; if the response at both modulation frequencies is a valid response and both consistency conditions are met, but the thermal-stress transfer parameters or their variation relationship are not within the corresponding parameter range, the candidate inclusion is identified as a non-nickel sulfide inclusion.

10. The online detection method for nickel sulfide inclusions in tempered glass according to claim 9, characterized in that, Before implementing periodic local photothermal excitation in step S4, a trial excitation is applied to the region where the candidate inclusion is located with a single-cycle excitation energy lower than that of the formal periodic local photothermal excitation. The differential photothermal complex response corresponding to the trial excitation is obtained according to the photothermal response acquisition and regional difference method. The photothermal response sensitivity of the region where the candidate inclusion is located is determined based on the relationship between the single-cycle excitation energy of the trial excitation and the corresponding differential photothermal complex response amplitude. A glass calibration sample with the same thickness, glass composition and coating state as the tempered glass to be tested is used. The excitation energy is increased step by step. After each excitation, the optical delay when the glass calibration sample recovers to the temperature before excitation is compared with the optical delay before excitation. The differential photothermal complex response amplitude corresponding to the highest excitation level that does not cause irreversible optical delay change is determined as the safe response boundary after deducting the amplitude fluctuation margin caused by repeated excitation. The output power and duty cycle of the periodic local photothermal excitation are adjusted according to the photothermal response sensitivity to make the amplitude of the differential photothermal complex response higher than the photothermal response detection boundary and lower than the safety response boundary; when the amplitude of the differential photothermal complex response reaches the safety response boundary, the periodic local photothermal excitation is stopped, and the candidate inclusions are identified as the objects for re-inspection.

Images