Substrate wrapper defect detection method and substrate wrapper defect detection mechanism

Abstract

The invention provides a substrate wrapper defect detection method and a detection mechanism, and the detection method comprises a wrapper defect layering detection step and a surface wrapper defect re-judgment step, the surface layer wrappage defect re-judgment step comprises the following sub-steps: S201, performing monochromatic laser irradiation on each possible wrappage defect coordinate point on the surface layer of the substrate which is preliminarily judged, and obtaining a spectrum of each coordinate point through Raman dispersion; and S202, comparing the characteristic peak position of the obtained spectrum with a prestored inclusion defect spectrum in a spectrum characteristic library so as to re-judge whether the inclusion defect exists or not, and marking the coordinate point which is confirmed to be the inclusion defect. The method for detecting the defects of the substrate wrappage has relatively high accuracy, and can provide quantitative and traceable analysis results, so that defect information can be known more comprehensively, and process analysis is also facilitated.

Description

Technical Field

[0001] The present application relates to the technical field of substrate defect detection, and in particular to a substrate packaging defect detection method and a substrate packaging defect detection mechanism. Background Art

[0002] ‌Substrate refers to the base or support used in the manufacturing of semiconductor devices, electronic components, optical devices, etc. It provides a stable platform for subsequent thin film growth, material deposition, device structure construction and other steps.

[0003] Inclusion defects may form during the production process of some substrates. For example, the current production of 4H crystal silicon carbide substrates uses the physical vapor transport (PVT) method. During the crystal growth process, due to the unstable partial pressure of multiple gas phases in the growth chamber, the Si / C ratio of the SiC crystal growth front gas phase deviates from the stoichiometric ratio of Si / C, thereby forming Si and C inclusion defects. In addition, impurity elements that may exist in raw materials and auxiliary materials such as crucibles are captured by the crystal front during high-temperature growth, thereby forming impurity inclusion defects. Inclusion defects will further induce other defects such as micropipes and triangles in the crystal, and extend them to the epitaxial layer, ultimately affecting the performance of the device such as breakdown strength, leading to device failure or even damage. Therefore, it is necessary to detect inclusion defects.

[0004] Currently, the method for detecting package defects is manual strong light observation, which has the following disadvantages:

[0005] 1. Low accuracy. Due to different observation angles, the brightness and appearance of packages under different lighting angles are different. In particular, shallow surface packages are easily confused with surface particles, which is prone to misjudgment. In addition, the human eye is prone to visual fatigue when working under high light intensity for a long time, which also affects observation accuracy.

[0006] 2. Unable to provide quantitative and traceable analysis results. Manual inspection can only make relatively qualitative judgments and cannot provide accurate quantitative analysis. For example, it is impossible to determine the size and type of inclusions, as well as the specific depth of the inclusions within the polished wafer. Moreover, manual visual inspection of carbon inclusions cannot currently be traced, making it difficult to conduct process analysis. Summary of the Invention

[0007] In order to more accurately detect defects in packaging, the present application provides a method for detecting defects in substrate packaging, including a step of re-judging defects in surface packaging; the step of re-judging defects in surface packaging includes the following sub-steps:

[0008] S201, irradiating each possible inclusion defect coordinate point on the substrate surface layer initially determined with a monochromatic laser and obtaining a spectrum of each coordinate point through Raman dispersion;

[0009] S202 , comparing the characteristic peak position of the acquired spectrum with the spectrum of inclusion defects in a pre-stored spectrum characteristic library to re-judge whether it is an inclusion defect, and marking the coordinate points confirmed to be inclusion defects.

[0010] An embodiment of a substrate wrapping defect detection method includes a wrapping defect layer detection step; the wrapping defect layer detection step includes the following sub-steps:

[0011] S101, collecting images of multiple layers at different depths of a substrate, including a surface layer image of the substrate;

[0012] S102, pre-processing the collected image;

[0013] S103, performing defect segmentation on the pre-processed image and marking coordinates on the surface image;

[0014] S104, preliminarily screening out package defects;

[0015] S105, obtaining the size of the inclusion defect and the depth of the inclusion defect;

[0016] After the step of detecting the layered defects of the package and the step of re-judging the defects of the surface package are completed, the detection results are output.

[0017] In one embodiment of a method for detecting defects in substrate inclusions, in S101, dual cameras synchronously capture images of two depth layers of the substrate. During the capture, the substrate moves in the x- and y-directions. After the capture, the dual cameras move a preset distance in the z-direction and then synchronously capture images of two other depth layers of the substrate until images of the entire depth of the substrate are captured. The x-, y-, and z-directions are perpendicular to each other, and the z-direction is the depth direction of the substrate.

[0018] In one embodiment of the method for detecting defects in substrate packaging, in S101, when collecting images of each depth layer, images of other areas on the depth layer except for the area blocked by the support part are first collected, and then the support position is changed to supplement the images of the area previously blocked by the support part.

[0019] In one embodiment of the substrate packaging defect detection method, in S102, the preprocessing of the image includes: eliminating background noise of the image, enhancing contrast of the image, flat field correction, and filtering noise reduction.

[0020] In one embodiment of the substrate packaging defect detection method, in S104, screening out packaging defects includes filtering out non-packaging defects based on morphological features of the defects and / or filtering out non-packaging defects based on texture features of the defects.

[0021] In an implementation manner of a method for detecting defects in substrate packaging, the morphological characteristics of the defects include the area of ​​the defects, the aspect ratio of the defects, and the roundness of the defects.

[0022] In one embodiment of the substrate inclusion defect detection method, in S105, obtaining the size of the inclusion defect includes: extracting the outer contour of the inclusion defect, obtaining the size of the inclusion defect, setting a large, medium, or small defect boundary value, and determining whether the inclusion defect belongs to a large, medium, or small defect.

[0023] In one embodiment of the substrate inclusion defect detection method, in S105, obtaining the depth of the inclusion defect includes: obtaining edge gradient values ​​of the same inclusion defect on multiple depth layer images and comparing them, and the depth of the depth layer with the largest edge gradient value is the depth of the inclusion defect.

[0024] An embodiment of a method for detecting defects in substrate inclusions, wherein the output of the detection results includes: splicing the inclusion defects on multiple depth layer images onto a MAP map based on the defect coordinates, and outputting the MAP map; and / or, outputting a report, wherein the report includes the size of the inclusion defects, the depth to which they belong, and the area yield of the substrate.

[0025] The present application also provides a substrate packaging defect detection mechanism, including a surface packaging defect re-judgment device and a control device, the surface packaging defect re-judgment device including: a laser emitter for irradiating each possible packaging defect coordinate point on the surface of the substrate that has been initially judged with a monochromatic laser, a Rayleigh filter for filtering Rayleigh scattering, a slit for controlling the light flux, a grating for Raman dispersion, and a receiver for receiving an optical signal and converting the optical signal into an electrical signal. The monochromatic laser irradiated onto the substrate is reflected by possible packaging defects on the surface of the substrate and then filtered by the Rayleigh filter, and then sequentially passed through the slit and the grating before being received by the receiver. The control device is communicatively connected to the receiver, and spectral characteristic peaks are pre-stored inside the control device. The control device can analyze and process the packaging defects and output the detection results.

[0026] An embodiment of a substrate packaging defect detection mechanism, the surface packaging defect re-judgment device includes an attenuation plate for adjusting laser power, a beam expander for controlling the laser spot size, an objective lens for focusing the light beam and magnifying the magnification, a lens for focusing the light beam, and three groups of reflectors for changing the direction of the light path. The monochromatic laser emitted by the laser emitter passes through the attenuation plate and the beam expander in sequence, then changes the light path through the Rayleigh filter and the first group of reflectors, and then is irradiated onto the substrate through the objective lens. The reflected light is then filtered by the Rayleigh filter and focused by the lens, then changes the light path through the second group of reflectors and then passes through the slit, and then changes the light path through the third group of reflectors and then passes through the grating.

[0027] An embodiment of a substrate packaging defect detection mechanism, the detection mechanism includes a stage, a drive device, a stage and a packaging defect layering detection device for detecting packaging defects at different depth layers of the substrate, the surface packaging defect re-judgment device is installed on the stage, and a re-judgment station is formed below the surface packaging defect re-judgment device, the packaging defect layering detection device is installed on the stage, and a detection station is formed below the packaging defect layering detection device, the stage is installed on the stage, and the control device can control the drive device to drive the stage to move along the x and y directions at the detection station and drive the stage to reciprocate between the re-judgment station and the detection station, wherein the x and y directions are perpendicular to each other and both perpendicular to the depth direction of the substrate.

[0028] An embodiment of a substrate packaging defect detection mechanism, the packaging defect layer detection device includes a light source and two cameras equipped with lenses, the two cameras can focus on different depth layers of the substrate respectively, the control device can control the drive system to drive the two cameras to move along the z direction to collect images of different depth layers of the substrate, wherein the x direction, y direction, and z direction are perpendicular to each other, the z direction is the depth direction of the substrate, and the control device is communicatively connected to the cameras to obtain the images collected by the cameras.

[0029] In one embodiment of a substrate wrapping defect detection mechanism, the stage includes a stage body and two groups of support parts, and the control device can control the drive device to drive the two groups of support parts to rise and fall along the z direction, so that the two groups of support parts alternately support the substrate.

[0030] The present application re-evaluates each possible inclusion defect on the surface of the substrate that has been initially identified. By obtaining the spectrum of each possible inclusion defect coordinate point and comparing the characteristic peak positions of these spectra with the spectra in the spectral feature library, it can be confirmed whether the defects at these coordinate points are inclusion defects, thereby avoiding mistaking surface particles or other defects for inclusion defects and ensuring the accuracy of inclusion defect detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a structural schematic diagram of a substrate wrapping defect detection mechanism according to the present invention;

[0032] Figure 2 yes Figure 1 Schematic diagram of the optical path of the device for re-judging defects in the middle and surface layers of the inclusions;

[0033] Figure 3 This is a schematic diagram of a device for detecting package defects and delamination;

[0034] Figure 4 is a schematic diagram of substrate layered imaging;

[0035] Figure 5 Area yield map of the substrate.

[0036] The following are the descriptions of the reference numerals:

[0037] 01 substrate;

[0038] A-Surface coating defect re-judgment device;

[0039] A1-laser transmitter, A2-Rayleigh filter, A3-slit, A4-grating, A5 receiver, A6-attenuator, A7-beam expander, A8-objective lens, A9-lens, A10-first group of reflectors, A11-second group of reflectors, A12-third group of reflectors;

[0040] B-package defect delamination detection device;

[0041] B1-light source, B2-camera, B3-lens, B4-MSG autofocus high-precision sensor;

[0042] C-frame, C1-crossbeam, C2-vibration-damping column, C3-platform;

[0043] D-control device; E-drive device;

[0044] F-stage, F1-stage body, F2-support part;

[0045] G Power plant. DETAILED DESCRIPTION

[0046] The present application provides a substrate wrapping defect detection method and a substrate wrapping defect detection mechanism. In order to enable those skilled in the art to better understand the technical solution of the present application, the technical solution of the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0047] The substrate wrapping defect detection method provided in the present application includes a surface wrapping defect re-judgment step.

[0048] The surface wrapping re-judgment step includes the following sub-steps:

[0049] S201 , irradiating each possible inclusion defect coordinate point on the surface of the substrate 01 that has been initially determined with a monochromatic laser and acquiring a spectrum of each coordinate point through Raman dispersion.

[0050] S202 , comparing the characteristic peak position of the acquired spectrum with the spectrum of inclusion defects in a pre-stored spectrum characteristic library to re-judge whether it is an inclusion defect, and marking the coordinate points confirmed to be inclusion defects.

[0051] Specifically, the spectral feature library is established based on confirmed samples, which includes characteristic peaks of various types of inclusion defects, such as carbon inclusion defect characteristic peaks, impurity inclusion defect characteristic peaks, etc., as well as substrate 01 background peaks, organic particle characteristic peaks, etc.

[0052] By comparing the characteristic peak position of the spectrum obtained at each coordinate point with the spectrum in the spectral feature library, it is possible to confirm whether the defects at these coordinate points are inclusion defects, thereby avoiding mistaking surface particles or other defects for inclusion defects and ensuring the accuracy of inclusion defect detection.

[0053] The substrate wrapping defect detection method provided in the present application also includes a wrapping defect layer detection step.

[0054] The package defect delamination detection step includes the following sub-steps:

[0055] S101 , collecting images of multiple layers at different depths of a substrate 01 , including a surface image of the substrate 01 .

[0056] like Figure 3 As shown, during acquisition, light source B1 illuminates substrate 01. If the illuminated area is free of defects and particles, the reflection angle of the reflected light is too large to be captured by the lens of camera B2, forming a dark field of view. If there are defects and particles in the illuminated area, the incident light will be segregated, causing the scattered light to be captured by the lens of camera B2, forming a bright light spot that is clearly distinguishable from the dark field of view.

[0057] Specifically, the dual camera B2 can synchronously capture images of two depth layers of the substrate 01. During the capture, the substrate 01 moves along the x-direction and the y-direction. After the capture, the dual camera B2 moves along the z-direction by a preset distance and then synchronously captures images of the other two depth layers of the substrate 01 until the full-depth image of the substrate 01 is captured. The x-direction, the y-direction, and the z-direction are perpendicular to each other, and the z-direction is parallel to the depth direction of the substrate 01. For example, Figure 4As shown, for a substrate 01 with a depth of 350 μm, the two cameras B2 can first focus on the 0 μm depth layer and the 50 μm depth layer, respectively, thereby capturing images of these two depth layers; then the two cameras B2 move down 100 μm and focus on the 100 μm depth layer and the 150 μm depth layer, respectively, thereby capturing images of these two depth layers; then the two cameras B2 move down another 100 μm and focus on the 200 μm depth layer and the 250 μm depth layer, respectively, thereby capturing images of these two depth layers; then the two cameras B2 move down another 100 μm and focus on the 300 μm depth layer and the 350 μm depth layer, respectively, thereby capturing images of these two depth layers. The two cameras B2 synchronously capture images of the two depth layers of substrate 01. In this way, during the process of capturing the full-depth image of substrate 01, the number of camera B2 movements is small, and the acquisition efficiency is high.

[0058] Specifically, the substrate 01 is supported by the support part F2. When collecting images of each depth layer, images of other areas on the depth layer except the area blocked by the support part F2 are collected first, and then the support position is changed to supplement the images of the areas previously blocked by the support part F2.

[0059] S102: Preprocess the collected image.

[0060] Specifically, preprocessing methods include removing background noise from the image, enhancing contrast to optimize subsequent gradient calculations, performing flat-field correction to eliminate uneven illumination, and filtering to reduce noise while preserving edge features. For example, a median filter algorithm can be used to remove background noise from the image, and contrast-limited adaptive histogram equalization (CLAHE) can be used to enhance local and edge contrast in the image.

[0061] S103 , performing defect segmentation on the pre-processed image and marking coordinates on the surface image.

[0062] Specifically, a global threshold segmentation algorithm may be used to output the preprocessed grayscale image as a binary image (one for background, the other for defects).

[0063] The coordinates of the surface image are marked to facilitate the acquisition of the coordinates of the surface inclusion defects. In this way, the surface inclusion defects can be accurately found when the surface inclusion defects are re-judged.

[0064] S104. Preliminary screening of package defects.

[0065] Specifically, non-inclusion defects can be filtered out based on their morphological characteristics. These include defect area, aspect ratio, and roundness. For example, inclusion defects are typically larger than 20 pixels² in area, less than or equal to 2 in aspect ratio, and within a range of 0.6-1.3 in roundness.

[0066] Specifically, non-encapsulated defects can be filtered out based on the texture characteristics of the defects. Specifically, the internal standard deviation of the defect area is calculated to find the grayscale eigenvalue of the encapsulated defect, and other defects that are not within the eigenvalue range are eliminated.

[0067] S105: Obtain the size of the inclusion defect and the depth of the inclusion defect.

[0068] Specifically, obtaining the size of the inclusion defect includes: extracting the outer contour of the defect, obtaining the size of the inclusion defect, setting the boundary value of large, medium and small defects, and defining whether the inclusion defect belongs to large, medium or small. For example, the Suzuki algorithm can be used to extract the outer contour of the defect, and inclusion defects of ≤10 um can be set as small inclusion defects, inclusion defects of 10 um-50 um can be set as medium inclusion defects, and inclusion defects of >50 um can be set as large inclusion defects.

[0069] Specifically, obtaining the depth of the inclusion defect includes obtaining and comparing edge gradient values ​​of the same inclusion defect on multiple depth layer images. The depth layer with the largest edge gradient value is the depth of the inclusion defect. The principle is that photographing the same inclusion defect at different focus layers will result in varying degrees of defocus. Defocus is characterized by image edge diffusion and reduced gradient values. The edge gradient values ​​of the images at each depth layer can be calculated using the Sobel operator. The depth layer with the largest edge gradient value is the optimal focus layer. For example, if the optimal focus layer is the 50μm depth layer, then the depth of the inclusion defect is 0-50μm.

[0070] The substrate packaging defect detection method provided herein also includes a result output step. Specifically, the surface packaging defect coordinate points can be initially determined through the packaging defect layer detection step. The surface packaging defect re-determination step then re-determines the surface packaging defect coordinate points, and after the re-determination is completed, the detection result is output.

[0071] Specifically, inclusion defects from multiple depth layer images can be stitched together into a single map based on the defect coordinates, and the map can be output. A report can also be generated, which can include the size of the inclusion defect, its depth, and the area yield of the substrate 01. It can also include the grade and type of the substrate 01 based on the size, depth, and usable area of ​​the inclusion defect.

[0072] The above-mentioned substrate inclusion defect detection method has high accuracy and can provide quantitative and traceable analysis results, which facilitates a more comprehensive understanding of defect information and also facilitates process analysis.

[0073] like Figure 1 As shown, the substrate 01 packaging defect detection mechanism provided in this application includes a surface packaging defect re-judgment device A and a control device.

[0074] like Figure 2 As shown, the surface coating defect re-judgment device A includes a laser transmitter A1, a Rayleigh filter A2, a slit A3, a grating A4, and a receiver A5.

[0075] Laser emitter A1 illuminates each initially identified possible inclusion defect coordinate point on the surface of substrate 01 with a monochromatic laser. For example, a 532nm monochromatic laser can be used. After reflection, the light passes through Rayleigh filter A2 to remove Rayleigh scattering, then passes through slit A3 to control the light flux. After Raman dispersion by grating A4, it is received by receiver A5. Receiver A5 converts the received optical signal into an electrical signal and transmits it to a control device, which processes it to generate a spectrum for each coordinate point.

[0076] A spectral feature library is pre-stored in the control device. The control device compares the characteristic peak position of the spectrum of each coordinate point with the spectral feature library to confirm whether the defects at these coordinate points are packaging defects.

[0077] Figure 3 In the figure, the surface inclusion defect re-inspection device A also includes an attenuator A6, a beam expander A7, an objective lens A8, a lens A9, and multiple groups of reflectors. The figure shows three groups: the first group of reflectors A10, the second group of reflectors A11, and the third group of reflectors A12. The attenuator A6 is used to adjust the laser power. The beam expander A7 controls the laser spot size to match the numerical aperture of the objective lens A8. The objective lens A8 is used to concentrate the light beam and magnify the magnification. The lens A9 focuses the light beam. The reflectors are used to change the optical path. In the figure, the monochromatic laser emitted by the laser transmitter A1 first passes through the attenuator A6 and the beam expander A7 in sequence, then changes the optical path through the Rayleigh filter A2 and the first group of reflectors, and then is irradiated onto the substrate 01 through the objective lens A8. The reflected light on the substrate 01 is filtered by the Rayleigh filter A2 and focused by the lens A9, then changes the optical path through the second group of reflectors A11, passes through the slit A3, and then changes the optical path through the third group of reflectors A12, and then passes through the grating A4.

[0078] like Figure 1As shown, the substrate wrapping defect detection mechanism provided in the present application includes a wrapping defect layer detection device B, and also includes a driving device E, a loading platform F, a stand C and an electric device for providing power.

[0079] The package defect layer detection device B includes a light source B1 and two cameras B2 equipped with lenses. For example, the light source B1 can be a circular parallel light source B1 with adjustable exposure time, and the lens can be a 5X optical lens, which can provide a larger imaging field of view and depth of field.

[0080] The incident angle of light source B1 is greater than the numerical aperture of the lens, ensuring that the reflected light irradiating the defect-free position is not captured by camera B2.

[0081] The two cameras B2 can respectively focus on different depth layers of the substrate 01 . The control device can control the drive system to drive the two cameras B2 to move along the z direction to capture images of different depth layers of the substrate 01 . The z direction is parallel to the depth direction of the substrate 01 .

[0082] In addition, the package defect layer detection device B can also be equipped with the MSG autofocus high-precision sensor B4 and a 660nm monochrome red laser real-time ranging for fast and accurate focusing.

[0083] The control device is in communication with the camera B2 of the package layer detection device and the receiver A5 of the surface package detection device. It can receive images captured by the camera B2 and electrical signals transmitted by the receiver A5, and can analyze and process package defects and output detection results.

[0084] The surface layer package defect re-judgment device A is mounted on the platform C, with a re-judgment station formed below the surface layer package defect re-judgment device A. The package defect delamination detection device B is mounted on the platform C, with an inspection station formed below the package defect delamination detection device B. The loading platform F is mounted on the platform C. Figure 1 In the figure, the platform C includes a beam C1 extending along the y direction, columns C2 supported at both ends of the beam C1, and a platform C3 located below the beam C1. The surface package defect re-judgment device A and the package defect delamination detection device B are respectively installed below the beam C1. The loading platform F is installed on the platform C3. The column C2 can be made of marble to reduce vibration and provide stable support.

[0085] The control device can control the drive device E to move the stage F along the x- and y-directions at the inspection station so that the camera B2 can capture a complete image of the substrate 01. The x- and y-directions are mutually perpendicular and both perpendicular to the depth direction of the substrate 01 (i.e., the z-direction). The control device can also control the drive device E to drive the stage F back and forth between the re-judgment station and the inspection station. After the delamination inspection is completed, the stage F moves from the inspection station to the re-judgment station for re-judgment. After the re-judgment is complete, the stage F moves from the re-judgment station to the inspection station to prepare for inspection of the next substrate 01. The drive device E can be a pneumatic or electric device.

[0086] In this embodiment, Figure 1 As shown, stage F comprises a stage body F1 and two sets of supports F2. A control device controls a drive device E to drive the two sets of supports F2 up and down relative to stage body F1 in the z-direction, causing the two sets of supports F2 to alternately support substrate 01. In the figure, four supports F2 are provided: two spaced apart in the x-direction and two spaced apart in the y-direction. When capturing images, substrate 01 can be first supported by the two spaced apart supports F2 in the x-direction, and then by the two spaced apart supports F2 in the y-direction. This allows for the capture of areas of substrate 01 previously obscured by the two spaced apart supports F2 in the x-direction.

[0087] The principles and implementation methods of the present application have been described above using specific examples. The description of the above embodiments is only intended to help understand the method and core concept of the present application. It should be noted that for those skilled in the art, without departing from the principles of the present application, various improvements and modifications may be made to the present application, and such improvements and modifications also fall within the scope of protection of the claims of the present application.

Claims

1. A method for detecting defects in substrate packaging, characterized in that: The method includes a step of re-judging defects of the surface wrapping; the step of re-judging defects of the surface wrapping includes the following sub-steps: S201, irradiating each possible inclusion defect coordinate point on the substrate surface layer initially determined with a monochromatic laser and obtaining a spectrum of each coordinate point through Raman dispersion; S202 , comparing the characteristic peak position of the acquired spectrum with the spectrum of inclusion defects in a pre-stored spectrum characteristic library to re-judge whether it is an inclusion defect, and marking the coordinate points confirmed to be inclusion defects.

2. The substrate packaging defect detection method according to claim 1, characterized in that: The method includes a step of detecting the delamination of the package defect; the step of detecting the delamination of the package defect includes the following sub-steps: S101, collecting images of multiple layers at different depths of a substrate, including a surface layer image of the substrate; S102, pre-processing the collected image; S103, performing defect segmentation on the pre-processed image and marking coordinates on the surface image; S104, preliminarily screening out package defects; S105, obtaining the size of the inclusion defect and the depth of the inclusion defect; After the step of detecting the layered defects of the package and the step of re-judging the defects of the surface package are completed, the detection results are output.

3. The substrate packaging defect detection method according to claim 2, characterized in that: In S101, the dual cameras synchronously capture images of two depth layers of the substrate. During the capture, the substrate moves along the x-direction and the y-direction. After the capture, the dual cameras move along the z-direction by a preset distance and then synchronously capture images of the other two depth layers of the substrate until the full-depth image of the substrate is captured. The x-direction, the y-direction, and the z-direction are perpendicular to each other, and the z-direction is the depth direction of the substrate.

4. The method for detecting defects in substrate packaging according to claim 2, wherein: In S101 , when collecting images of each depth layer, images of other areas on the depth layer except the area blocked by the support portion are first collected, and then the support position is changed to supplement the images of the area previously blocked by the support portion.

5. The method for detecting defects in substrate packaging according to claim 2, wherein: In S102, the pre-processing of the image includes: eliminating background noise of the image, enhancing contrast of the image, flat field correction, and filtering noise reduction. One or more of the following:

6. The method for detecting defects in substrate packaging according to claim 2, wherein: In S104, screening out inclusion defects includes filtering out non-inclusion defects based on morphological features of the defects and / or filtering out non-inclusion defects based on texture features of the defects.

7. The method for detecting defects in substrate packaging according to claim 6, wherein: The morphological characteristics of the defect include the area of ​​the defect, the aspect ratio of the defect, and the roundness of the defect.

8. The method for detecting defects in substrate packaging according to claim 2, wherein: In S105, obtaining the size of the inclusion defect includes: extracting the outer contour of the inclusion defect, obtaining the size of the inclusion defect, setting a boundary value of large, medium and small defects, and determining whether the inclusion defect belongs to large, medium or small defects.

9. The method for detecting defects in substrate packaging according to any one of claims 2 to 8, characterized in that: In S105, obtaining the depth of the inclusion defect includes obtaining edge gradient values ​​of the same inclusion defect on multiple depth layer images and comparing them, and the depth of the depth layer with the largest edge gradient value is the depth of the inclusion defect.

10. The method for detecting defects of substrate packaging according to claim 2, wherein: The output of the detection results includes: splicing the inclusion defects on multiple depth layer images into a MAP map according to the defect coordinates, and outputting the MAP map; and / or outputting a report, the report including the size of the inclusion defects, the depth to which they belong, and the area yield of the substrate.

11. A substrate packaging defect detection mechanism, characterized in that: The detection mechanism includes a surface inclusion defect re-judgment device and a control device. The surface inclusion defect re-judgment device includes: a laser emitter for irradiating each possible inclusion defect coordinate point on the surface of the substrate that has been initially judged with a monochromatic laser, a Rayleigh filter for filtering Rayleigh scattering, a slit for controlling the light flux, a grating for Raman dispersion, and a receiver for receiving an optical signal and converting the optical signal into an electrical signal. The monochromatic laser irradiated onto the substrate is reflected by possible inclusion defects on the surface of the substrate and then filtered by the Rayleigh filter. It is then received by the receiver after passing through the slit and the grating in sequence. The control device is communicatively connected to the receiver. The control device has spectral characteristic peaks pre-stored therein. The control device can analyze and process the inclusion defects and output the detection results.

12. The substrate wrapping defect detection mechanism according to claim 11, characterized in that: The surface package defect re-judgment device includes an attenuation plate for adjusting the laser power, a beam expander for controlling the laser spot size, an objective lens for focusing the light beam and magnifying the magnification, a lens for focusing the light beam, and three groups of reflectors for changing the direction of the light path. The monochromatic laser emitted by the laser emitter passes through the attenuation plate and the beam expander in sequence, then changes the light path through the Rayleigh filter and the first group of reflectors, and then is irradiated onto the substrate through the objective lens. The reflected light is then filtered by the Rayleigh filter and focused by the lens, then changes the light path through the second group of reflectors and passes through the slit, and then changes the light path through the third group of reflectors and passes through the grating.

13. The substrate wrapping defect detection mechanism according to any one of claims 11 to 12, characterized in that: The detection mechanism includes a stage, a driving device, a stage and a package defect layering detection device for detecting package defects on different depth layers of the substrate. The surface package defect re-judgment device is installed on the stage, and a re-judgment station is formed below the surface package defect re-judgment device. The package defect layering detection device is installed on the stage, and a detection station is formed below the package defect layering detection device. The stage is installed on the stage. The control device can control the driving device to drive the stage to move along the x and y directions on the detection station and drive the stage to reciprocate between the re-judgment station and the detection station, wherein the x and y directions are perpendicular to each other and both perpendicular to the depth direction of the substrate.

14. The substrate wrapping defect detection mechanism according to claim 13, characterized in that: The package defect layer detection device includes a light source and two cameras equipped with lenses. The two cameras can focus on different depth layers of the substrate respectively. The control device can control the drive system to drive the two cameras to move along the z direction to collect images of different depth layers of the substrate, wherein the x direction, y direction, and z direction are perpendicular to each other, and the z direction is the depth direction of the substrate. The control device is communicatively connected to the cameras to obtain the images collected by the cameras.

15. The substrate wrapping defect detection mechanism according to claim 13, characterized in that: The stage includes a stage body and two groups of supporting parts. The control device can control the driving device to drive the two groups of supporting parts to move up and down along the z direction, so that the two groups of supporting parts alternately support the substrate.

Images