CN-122023963-A - Electronic paste screen printing defect data set expansion method, device, equipment and medium

CN122023963ACN 122023963 ACN122023963 ACN 122023963ACN-122023963-A

Abstract

The invention discloses a method, a device, equipment and a medium for expanding an electronic paste screen printing defect data set, which comprise the steps of collecting a printing image data set of an electronic paste screen; the method includes, for each print image in a print image dataset, extracting a print region image and a ring print image corresponding to the print region image from the print image, determining a defect mask image based on the ring print image and extracting a texture image corresponding to the print region image, determining a defect print image based on the defect mask image and the texture image, and constructing a defect dataset based on the defect print image. Based on the technical scheme, the defect printing image is obtained by processing the acquired printing image data, so that a comprehensive and diversified defect data set is obtained, the visual detection model can be trained based on the defect data set, the sample diversity in the defect data set can be improved, and the detection effect of the final visual detection model can be improved.

Inventors

ZENG WEIKANG
CAI HAOCONG
YANG CHENGHU
YUAN YONG
Lu Zhaocong
HUANG QINGJUN

Assignees

广东省科学院电子电器研究所
广东省电子技术研究所

Dates

Publication Date: 20260512
Application Date: 20260127

Claims (10)

1. An electronic paste screen printing defect data set expansion method is characterized by comprising the following steps: collecting a printed image dataset of the electronic paste screen; Extracting, for each print image in the print image dataset, a print area image and a ring-shaped print image corresponding to the print area image from the print image; Determining a defect mask image based on the annular printing image, and extracting a texture image corresponding to the printing region image; a defect print image is determined based on the defect mask image and the texture image, and a defect dataset is constructed based on the defect print image.
2. The method of claim 1, wherein the determining a defect mask image based on the ring-shaped print image comprises: Acquiring a preset random noise parameter, and generating a Berlin noise image based on the random noise parameter; The defect mask image is determined based on the berlin noise image and the loop-shaped print image.
3. The method of claim 2, wherein the determining the defect mask image based on the berlin noise image and the ring print image comprises: taking the Berlin noise image as a first mask image and taking the annular printing image as a second mask image; And performing AND operation on the first mask image and the second mask image to determine an overlapping region, and obtaining the defect mask image according to the overlapping region.
4. The method of claim 1, wherein the determining a defect print image based on the defect mask image and the texture image comprises: Determining an image overlapping region of the defect mask image and the printing image, and extracting texture pixels corresponding to the image overlapping region in the texture image; And replacing the image pixels in the image superposition area of the printing image with the texture pixels to obtain the defect printing image.
5. The method of claim 1, wherein the extracting the print area image from the print image comprises: performing color space conversion on the printed image to obtain a printed image in a target color space; Image extraction is carried out on the printing image in the target color space based on the color threshold value to obtain an electronic paste mask image; And cutting the printing image based on the electronic paste mask image to obtain the printing area image.
6. The method of claim 1, wherein extracting an annular print image corresponding to the print area image from the print image comprises: Performing expansion processing on the printing area image based on a preset expansion distance to obtain a printing area image after expansion processing; And performing exclusive OR operation on the print area image after the expansion treatment to obtain an annular print image corresponding to the print area image.
7. The method of claim 1, wherein the extracting a texture image corresponding to the print area image comprises: performing open operation processing on the printing area image to obtain a printing area image after the open operation processing; and performing expansion processing on the print area image subjected to the open operation processing to obtain the texture image.
8. An electronic paste screen printing defect data set expanding device, characterized by comprising: the image data collection module is used for collecting a printing image data set of the electronic paste silk screen; An image extraction module for extracting, for each print image in the print image dataset, a print region image and a ring-shaped print image corresponding to the print region image from the print images; an image processing module for determining a defect mask image based on the annular printing image and extracting a texture image corresponding to the printing region image; a defect dataset determination module for determining a defect print image based on the defect mask image and the texture image and constructing a defect dataset based on the defect print image.
9. An electronic device, the electronic device comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the electronic paste screen printing defect dataset expansion method of any of claims 1-7.
10. A computer readable storage medium storing computer instructions for causing a processor to perform the electronic paste screen printing defect dataset expansion method of any of claims 1-7.

Description

Electronic paste screen printing defect data set expansion method, device, equipment and medium Technical Field The invention relates to the technical field of visual inspection, in particular to a method, a device, equipment and a medium for expanding an electronic paste screen printing defect data set. Background The slurry is usually a paste or suspension formed by uniformly mixing solid powder and an organic solvent, and is generally composed of active materials, conductive agents, binders, solvents, and the like. According to the application, the electronic paste, the textile paste and the like can be classified. Electronic pastes are widely used in screen printing processes for forming conductive, resistive, insulating films. Common electronic pastes are silver paste, copper paste, carbon paste, molybdenum-manganese paste, and the like. The rheological properties of the paste, such as viscosity, thixotropic properties, etc., affect the printing accuracy. Too high viscosity, the paste is difficult to pass through meshes, which can lead to unclear lines and break points of a printed pattern, and too low viscosity, the paste is easy to spread in the printing process, which leads to the phenomena of line blurring, dot gain or ink seepage, and defective products appear. The paste with good thixotropy can keep a good shape in the printing process, and is beneficial to improving the printing precision. However, the existing method for detecting the screen printing defects of the electronic paste is also manual detection in the production process, so that eyes are easy to fatigue, missed detection is caused, and the spot check mode is often found that tens of bad products and hundreds of bad products are printed before the defects, so that the generation of bad products cannot be limited in time, a large amount of waste is caused, and further, the detection mode needs to be improved. Disclosure of Invention The invention provides a method, a device, equipment and a medium for expanding an electronic paste screen printing defect data set, which are used for processing acquired printing image data to obtain a defect printing image, further obtaining a comprehensive and diversified defect data set, further training a visual detection model based on the defect data set, further improving sample diversity in the defect data set, and further improving the detection effect of a final visual detection model. According to an aspect of the present invention, there is provided an electronic paste screen printing defect dataset expansion method, including: collecting a printed image dataset of the electronic paste screen; Extracting, for each print image in the print image dataset, a print area image and a ring-shaped print image corresponding to the print area image from the print image; Determining a defect mask image based on the annular printing image, and extracting a texture image corresponding to the printing region image; a defect print image is determined based on the defect mask image and the texture image, and a defect dataset is constructed based on the defect print image. According to another aspect of the present invention, there is provided an electronic paste screen printing defect data set expanding apparatus, comprising: an image data collection module for each printed image in the printed image dataset; An image extraction module for extracting, for each print image in the print image dataset, a print region image and a ring-shaped print image corresponding to the print region image from the print images; an image processing module for determining a defect mask image based on the annular printing image and extracting a texture image corresponding to the printing region image; a defect dataset determination module for determining a defect print image based on the defect mask image and the texture image and constructing a defect dataset based on the defect print image. According to another aspect of the present invention, there is provided an electronic apparatus including: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the electronic paste screen printing defect dataset expansion method according to any of the embodiments of the invention. According to another aspect of the present invention, there is provided a computer readable storage medium storing computer instructions for causing a processor to implement the electronic paste screen printing defect data set expansion method according to any of the embodiments of the present invention when executed. According to the technical scheme, a printing image data set of an electronic paste silk screen is collected, a printing area image and an annular printing image corresponding to the printing area image are extracted from each printing image in the printing image data set, a de