CN-121982006-A - Film defect optimizing method and film defect optimizing system

CN121982006ACN 121982006 ACN121982006 ACN 121982006ACN-121982006-A

Abstract

The application relates to a film defect optimizing method and a film defect optimizing system, which are used for acquiring an image sequence of a film to be tested, wherein the image sequence comprises a plurality of images of the film to be tested, which are continuously acquired, and two adjacent images have a preset image overlapping rate, and splicing the image sequences according to the image overlapping rate to generate a panoramic image, so that the risk of defect omission caused by limited single shooting range is reduced, and the integrity of defect characteristic parameters of the film to be tested extracted based on the panoramic image is improved. Establishing a defect optimization model comprising a mapping relation among defect characteristic parameters of film samples, process parameters of the film samples and recipe parameters of the film samples, which are obtained based on a plurality of film samples, wherein precursor systems of the film samples are at least different, and at least one precursor system of the film samples is the same as that of a film to be tested. According to the application, through deep learning association, process-optimized data support is provided for reducing defects of the film to be tested, and the film preparation quality is improved.

Inventors

Request for anonymity
Request for anonymity
Request for anonymity
Request for anonymity
Request for anonymity
Request for anonymity

Assignees

现象创新(深圳)科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260129

Claims (10)

1. A method for optimizing defects in a thin film, comprising: The method comprises the steps of obtaining an image sequence of a film to be detected, wherein the image sequence comprises a plurality of images of the film to be detected, which are continuously obtained, and two adjacent images have a preset image overlapping rate; splicing the image sequences according to the image overlapping rate to generate panoramic images; extracting defect characteristic parameters of the film to be detected based on the panoramic image; Establishing a defect optimization model, wherein the defect optimization model comprises mapping relations among the defect characteristic parameters of the film samples, the process parameters of the film samples and the recipe parameters of the film samples, which are acquired based on various film samples, wherein precursor systems of the film samples are at least partially different, and at least one precursor system of the film samples is the same as the precursor system of the film to be tested; and outputting coating process adjustment parameters according to the defect characteristic parameters of the film to be tested and the defect optimization model.
2. The method for optimizing thin film defects according to claim 1, wherein in the process of acquiring the image sequence of the thin film to be tested, further comprising: Outputting alignment calibration instructions every time a preset number of images of the film to be tested are acquired, wherein the alignment calibration instructions are used for verifying the positions of the film to be tested in the movable objective table; and correcting the control parameter of the movable object table according to the offset when the offset of the position of the film to be detected in the movable object table and the preset position is larger than the target threshold.
3. The thin film defect optimization method according to claim 1, further comprising, before stitching the image sequence according to the image overlap ratio: performing filtering processing and contrast enhancement on each image in the image sequence; Extracting characteristic points of the image after filtering processing and contrast enhancement, wherein the splicing the image sequence according to the image overlapping rate comprises the following steps: And splicing two adjacent images with the image overlapping rate through feature point matching so as to generate the panoramic image.
4. The method for optimizing defects of a thin film according to claim 1, wherein the extracting defect characteristic parameters of the thin film to be tested based on the panoramic image comprises: identifying a defect area of the film to be detected based on the panoramic image of the film to be detected; identifying a target position in the defect area, wherein the gray value of the target position is in a first preset range; the gray value of the target position is reduced to a target range so as to fill the target position, wherein the target position represents the pore in each defect of the film to be tested; and extracting defect characteristic parameters in the panoramic image after filling processing.
5. The method of claim 1, wherein the establishing a defect optimization model comprises: Acquiring a plurality of related data sets of the film samples, wherein the related data sets comprise recipe parameters, process parameters and defect characteristic parameters of one film sample; And constructing the defect optimization model according to the associated data sets of the plurality of film samples.
6. The method of claim 5, wherein the correlation data set further comprises performance parameters of a target device provided with the film sample, wherein obtaining the correlation data set of one of the film samples comprises: identifying a defective area of the film sample based on the panoramic image of the film sample; extracting defect characteristic parameters of the film sample according to a defect region of the film sample, and determining defect types of the film sample according to a gray value range of a gray value of the defect region; Determining performance parameters of a target device according to the defect type and the defect characteristic parameters of the film sample; Wherein, the coating process adjustment parameters are determined according to the performance parameters, the defect characteristic parameters of the film to be tested and the process parameters of the film to be tested.
7. The method according to claim 6, wherein outputting coating process adjustment parameters according to the defect characteristic parameters and the defect optimization model of the thin film to be tested comprises: the different defect characteristic parameters are subjected to priority ranking according to the performance parameters of the target device; And outputting the coating process adjustment parameters according to the priority of the defect characteristic parameters and the process parameters.
8. The thin film defect optimization method of claim 5, wherein the defect optimization model comprises a feature extraction module and a full connection mapping module, wherein the defect feature parameters comprise at least two of defect area, defect perimeter, defect circularity, position coordinates of defects in the thin film samples, gray scale mean and standard deviation of the panoramic image, and wherein the constructing the defect optimization model from the associated data sets of the plurality of thin film samples comprises: Extracting the defect characteristic parameters through the characteristic extraction module, and mining deep association characteristics among at least two of the defect area, the defect perimeter, the defect circularity, the position coordinates of defects in the film sample, the gray average value and the standard deviation of the panoramic image; And establishing a mapping relation between the deep association characteristic and the formula parameter and the process parameter through the full-connection mapping module, and outputting the coating process adjustment parameter.
9. A thin film defect optimization system, comprising: The imaging equipment is used for bearing the film to be detected and shooting the image sequence of the film to be detected, wherein the image sequence comprises a plurality of images of the film to be detected, which are continuously acquired, and two adjacent images have a preset image overlapping rate; A master device, connected to the imaging device, for receiving the image sequence to perform the thin film defect optimization method according to any one of claims 1 to 8.
10. The thin film defect optimization system of claim 9, wherein the imaging device comprises: the backlight light source is used for providing light rays vertically incident to the film to be tested; The movable objective table is used for bearing and driving the film to be tested to move; the shooting component is used for shooting an image of the film to be detected; And the image transmission assembly is connected with the shooting assembly and the main control equipment, and is used for acquiring and transmitting the image of the film to be detected to the main control equipment.

Description

Film defect optimizing method and film defect optimizing system Technical Field The application relates to the technical field of film defect optimization, in particular to a film defect optimization method and a film defect optimization system. Background In the manufacturing fields of optical display, new energy, semiconductors and the like, the surface and internal quality of a film material serving as a core basic member directly determine the performance, reliability and service life of a terminal product, and the accurate detection of the film becomes an indispensable key link in the preparation process, so that the film material is an important premise for controlling the yield of the product and avoiding the subsequent production loss. In the current film detection technology, core research and development and application guidance are developed around defect identification, and the final target of the detection work is only to realize effective detection of the film defects and complete acquisition and output of basic information such as defect types, positions, sizes and the like. Although the detection mode can realize basic screening of the film quality, the detection result cannot be converted into an effective gripper for improving the film preparation quality, and the iterative upgrading of the film production process and the continuous improvement of the product yield are restricted. Disclosure of Invention In view of the foregoing, it is desirable to provide a thin film defect optimizing method and a thin film defect optimizing system capable of detecting thin film defects and optimizing the thin film defects. In a first aspect, the present application provides a method for optimizing defects of a thin film, including: The method comprises the steps of obtaining an image sequence of a film to be detected, wherein the image sequence comprises a plurality of images of the film to be detected, which are continuously obtained, and two adjacent images have a preset image overlapping rate; splicing the image sequences according to the image overlapping rate to generate panoramic images; extracting defect characteristic parameters of the film to be detected based on the panoramic image; Establishing a defect optimization model, wherein the defect optimization model comprises mapping relations among the defect characteristic parameters of the film samples, the process parameters of the film samples and the recipe parameters of the film samples, which are acquired based on various film samples, wherein precursor systems of the film samples are at least partially different, and at least one precursor system of the film samples is the same as the precursor system of the film to be tested; and outputting coating process adjustment parameters according to the defect characteristic parameters of the film to be tested and the defect optimization model. In one embodiment, in the process of acquiring the image sequence of the film to be measured, the method further includes: Outputting alignment calibration instructions every time a preset number of images of the film to be tested are acquired, wherein the alignment calibration instructions are used for verifying the positions of the film to be tested in the movable objective table; and correcting the control parameter of the movable object table according to the offset when the offset of the position of the film to be detected in the movable object table and the preset position is larger than the target threshold. In one embodiment, before the stitching the image sequence according to the image overlapping rate, the method further includes: performing filtering processing and contrast enhancement on each image in the image sequence; Extracting characteristic points of the image after filtering processing and contrast enhancement, wherein the splicing the image sequence according to the image overlapping rate comprises the following steps: And splicing two adjacent images with the image overlapping rate through feature point matching so as to generate the panoramic image. In one embodiment, the extracting the defect characteristic parameter of the film to be tested based on the panoramic image includes: identifying a defect area of the film to be detected based on the panoramic image of the film to be detected; identifying a target position in the defect area, wherein the gray value of the target position is in a first preset range; the gray value of the target position is reduced to a target range so as to fill the target position, wherein the target position represents the pore in each defect of the film to be tested; and extracting defect characteristic parameters in the panoramic image after filling processing. In one embodiment, the establishing the defect optimization model includes: Acquiring a plurality of related data sets of the film samples, wherein the related data sets comprise recipe parameters, process parameters and defect characteristic para