CN-121982031-A - Sealing ring detection method and system based on image processing

CN121982031ACN 121982031 ACN121982031 ACN 121982031ACN-121982031-A

Abstract

The application provides a sealing ring detection method and a sealing ring detection system based on image processing, which are applicable to the technical field of image processing, wherein the method comprises the following steps: according to the seal ring specification parameter information, the seal ring material information, the seal ring size detection image information to be detected currently, the seal ring end face detection image information to be detected currently, the seal ring circumferential detection image information to be detected currently, the historical seal ring size detection information, the historical seal ring end face detection information, the historical seal ring circumferential detection information and the seal ring detection model, the current seal ring size detection information, the current seal ring end face detection information and the current seal ring circumferential detection information are generated. The application effectively solves the problems of profile identification deviation, end face defect omission, incomplete detection data and the like in the traditional detection, and improves the accuracy and consistency of the detection of the sealing ring.

Inventors

ZHOU YI
AI GUOHUA

Assignees

长沙和捷实业有限公司

Dates

Publication Date: 20260505
Application Date: 20260408

Claims (10)

1. The sealing ring detection method based on image processing is characterized by comprising the following steps of: Acquiring a plurality of seal ring specification parameter information, a plurality of seal ring material information, a plurality of current seal ring size detection image information to be detected, a plurality of current seal ring end face detection image information to be detected, a plurality of current seal ring circumferential detection image information to be detected, a plurality of historical seal ring size detection information, a plurality of historical seal ring end face detection information and a plurality of historical seal ring circumferential detection information; Obtaining current seal ring contour representing information to be detected and current seal ring end face defect representing information to be detected according to the seal ring specification parameter information, the seal ring material information, the current seal ring size detecting image information to be detected, the current seal ring end face detecting image information to be detected, the historical seal ring size detecting information and the historical seal ring end face detecting information; And generating a plurality of pieces of current seal ring size detection information, a plurality of pieces of current seal ring end face detection information and a plurality of pieces of current seal ring circumferential detection information according to the current seal ring contour representation information to be detected, the current seal ring end face defect representation information to be detected, the plurality of pieces of current seal ring size detection image information to be detected, the plurality of pieces of current seal ring end face detection image information to be detected, the plurality of pieces of current seal ring circumferential detection image information to be detected, the plurality of pieces of historical seal ring circumferential detection information and the seal ring detection model.
2. The seal ring detection method based on image processing as claimed in claim 1, wherein, The current size detection image information of the seal ring to be detected comprises current inner diameter forward detection image information of the seal ring to be detected, current inner diameter reverse detection image information of the seal ring to be detected and current axial height detection image information of the seal ring to be detected; The current seal ring end face detection image information to be detected comprises current seal ring port detection image information to be detected, current seal ring port deformation detection image information to be detected and current seal ring port burr detection image information to be detected; The current circumferential detection image information of the seal ring to be detected comprises current circumferential texture detection image information of the seal ring body to be detected, current collision damage detection image information of the seal ring body to be detected and surface flaw detection image information of the seal ring body to be detected; The historical seal ring size detection information comprises historical seal ring inner diameter deviation detection information, historical seal ring height detection information and historical seal ring size discrete distribution detection information; The historical seal ring end face detection information comprises historical seal ring port defect type detection information, historical seal ring port defect degree detection information and historical seal ring port defect frequency detection information; The historical seal ring circumferential detection information comprises historical seal ring appearance flaw category detection information, historical seal ring appearance flaw distribution detection information and historical seal ring appearance shape detection information.
3. The method for detecting a seal ring based on image processing according to claim 1, wherein the step of generating a plurality of pieces of current seal ring size detection information, a plurality of pieces of current seal ring end face detection information, and a plurality of pieces of current seal ring circumferential detection information according to the current seal ring contour representation information to be detected, the current seal ring end face defect representation information to be detected, the plurality of pieces of current seal ring size detection image information to be detected, the plurality of pieces of current seal ring end face detection image information to be detected, the plurality of pieces of current seal ring circumferential detection image information to be detected, the plurality of pieces of history seal ring circumferential detection information, and the seal ring detection model specifically includes: Based on a preset seal ring geometric feature extraction rule, generating a plurality of seal ring geometric feature reference information according to a preset seal ring geometric feature reference sample set; Coding and vector conversion processing is carried out on the size detection image information of the plurality of current sealing rings to be detected, the end face detection image information of the plurality of current sealing rings to be detected and the circumferential detection image information of the plurality of current sealing rings to be detected, so that global characteristic information of the plurality of current sealing rings to be detected is generated; training the sealing ring detection model according to the global characteristic information of the plurality of current sealing rings to be detected and the geometric characteristic reference information of the plurality of sealing rings to obtain a trained sealing ring detection model; And generating a plurality of pieces of current seal ring size detection information, a plurality of pieces of current seal ring end face detection information and a plurality of pieces of current seal ring circumferential detection information according to the current seal ring contour representation information to be detected, the current seal ring end face defect representation information to be detected, a plurality of pieces of historical seal ring circumferential detection information and the trained seal ring detection model.
4. The method for detecting a seal ring based on image processing as claimed in claim 3, wherein the step of training the seal ring detection model according to the global feature information of the plurality of seal rings to be detected currently and the geometric feature reference information of the plurality of seal rings to obtain a trained seal ring detection model specifically comprises: Vector conversion and merging processing are carried out according to the global feature information of the plurality of current sealing rings to be detected and the geometric feature reference information of the plurality of sealing rings to be detected, so that a plurality of global feature representation vectors of the sealing rings to be mapped are generated; according to the global feature expression vectors of the sealing rings to be mapped and the preset global feature mapping vectors of the sealing rings, calculating to obtain global feature mapping information of the sealing rings and global feature mapping association degree information of the sealing rings; carrying out normalization processing according to the global feature mapping association degree information of the plurality of sealing rings to obtain global feature mapping weight information of the plurality of sealing rings; according to the global feature mapping weight information of the sealing rings, carrying out weighted calculation on the global feature mapping information of the sealing rings to obtain global feature weighted information of the sealing rings; And training the sealing ring detection model according to the global characteristic information of the plurality of current sealing rings to be detected, the geometric characteristic reference information of the plurality of sealing rings and the global characteristic weighting information of the plurality of sealing rings to obtain a trained sealing ring detection model.
5. The method for detecting a seal ring based on image processing according to claim 4, wherein, The plurality of preset global feature map vectors of the sealing ring comprise preset global feature search map vectors of the sealing ring, preset global feature matching map vectors of the sealing ring and preset global feature quantization map vectors of the sealing ring; the seal ring universe feature mapping information comprises seal ring universe feature retrieval mapping information, seal ring universe feature matching mapping information and seal ring universe feature quantification mapping information; the step of calculating to obtain a plurality of global feature mapping information of the sealing ring and a plurality of global feature mapping association degree information of the sealing ring according to the global feature representation vectors of the sealing ring to be mapped and a plurality of preset global feature mapping vectors of the sealing ring specifically includes: calculating to obtain a plurality of global feature retrieval mapping information of the sealing ring according to the global feature expression vectors of the sealing ring to be mapped and a preset global feature retrieval mapping vector of the sealing ring; calculating to obtain a plurality of seal ring global feature matching mapping information according to the plurality of seal ring global feature representation vectors to be mapped and a preset seal ring global feature matching mapping vector; Calculating to obtain a plurality of seal ring global feature quantized mapping information according to the plurality of seal ring global feature representation vectors to be mapped and a preset seal ring global feature quantized mapping vector; According to the global feature retrieval mapping information of the sealing rings and the global feature matching mapping information of the sealing rings, calculating to obtain global feature interaction mapping information of the sealing rings; calculating to obtain global feature interaction quantization information of the plurality of sealing rings according to the global feature interaction mapping information of the plurality of sealing rings and global feature quantization mapping information of the plurality of sealing rings; And calculating to obtain the global feature mapping association degree information of the plurality of sealing rings according to the global feature interaction quantization information of the plurality of sealing rings and the global feature interaction mapping information of the plurality of sealing rings.
6. The method for detecting a seal ring based on image processing according to claim 1, wherein after the step of generating a plurality of pieces of current seal ring size detection information, a plurality of pieces of current seal ring end face detection information, and a plurality of pieces of current seal ring circumferential detection information, a plurality of pieces of history seal ring circumferential detection information, and a seal ring detection model according to the current seal ring contour representation information to be detected, the current seal ring end face defect representation information to be detected, the plurality of pieces of current seal ring size detection image information to be detected, the plurality of pieces of current seal ring end face detection image information to be detected, the plurality of pieces of current seal ring circumferential detection information to be detected, further comprises: performing defect feature extraction on the plurality of pieces of current seal ring size detection information, the plurality of pieces of current seal ring end face detection information and the plurality of pieces of current seal ring circumferential detection information to obtain a plurality of pieces of seal ring defect feature information, wherein the seal ring defect feature information comprises seal ring inner diameter abnormal feature information, seal ring height deviation feature information, seal ring port defect feature information and seal ring appearance defect feature information; Based on a preset mapping relation between seal ring defect characteristics and sorting outlet instructions, generating a plurality of seal ring sorting outlet instruction information according to the plurality of seal ring inner diameter abnormal characteristic information, the plurality of seal ring height deviation characteristic information, the plurality of seal ring port defect characteristic information and the plurality of seal ring appearance flaw characteristic information.
7. The method for detecting a seal ring based on image processing according to claim 6, wherein after the step of generating a plurality of seal ring sorting outlet instruction information based on the preset seal ring defect feature and sorting outlet instruction mapping relationship according to the plurality of seal ring inside diameter anomaly feature information, the plurality of seal ring height deviation feature information, the plurality of seal ring port defect feature information, and the plurality of seal ring appearance defect feature information, the method further comprises: Sorting instruction classification processing is carried out on the plurality of seal ring sorting outlet instruction information to generate a plurality of seal ring sorting outlet instruction type information, wherein the seal ring sorting outlet instruction type information comprises seal ring inner diameter bad discharge execution information, seal ring height bad discharge execution information, seal ring port defect discharge execution information, seal ring appearance flaw discharge execution information, seal ring qualified piece collecting execution information and seal ring suspicious piece rechecking execution information; And performing split-flow execution processing on the sealing rings according to the inner diameter bad discharge execution information of the sealing rings, the height bad discharge execution information of the sealing rings, the port defect discharge execution information of the sealing rings, the appearance defect discharge execution information of the sealing rings, the collecting execution information of qualified sealing ring parts and the rechecking execution information of suspected sealing ring parts.
8. An image processing-based seal ring detection system, comprising: The information acquisition module is used for acquiring a plurality of seal ring specification parameter information, a plurality of seal ring material information, a plurality of current seal ring size detection image information to be detected, a plurality of current seal ring end face detection image information to be detected, a plurality of current seal ring circumferential detection image information to be detected, a plurality of historical seal ring size detection information, a plurality of historical seal ring end face detection information and a plurality of historical seal ring circumferential detection information; The current seal ring contour representation and end surface defect representation information generation module is used for obtaining current seal ring contour representation information to be detected and current seal ring end surface defect representation information to be detected according to the plurality of seal ring specification parameter information, the plurality of seal ring material information, the plurality of current seal ring size detection image information to be detected, the plurality of current seal ring end surface detection image information to be detected, the plurality of historical seal ring size detection information and the plurality of historical seal ring end surface detection information; The current seal ring detection information generation module is used for generating a plurality of current seal ring size detection information, a plurality of current seal ring end face detection information and a plurality of current seal ring circumferential detection information according to the current seal ring contour representation information to be detected, the current seal ring end face defect representation information to be detected, the plurality of current seal ring size detection image information to be detected, the plurality of current seal ring end face detection image information to be detected, the plurality of current seal ring circumferential detection image information to be detected, the plurality of historical seal ring circumferential detection information and the seal ring detection model.
9. A terminal device, characterized in that it comprises a memory, a processor, on which a computer program is stored which is executable on the processor, the processor executing the computer program to carry out the steps of the method according to any one of claims 1 to 7.
10. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the method according to any one of claims 1 to 7.

Description

Sealing ring detection method and system based on image processing Technical Field The application belongs to the technical field of image processing, and particularly relates to a sealing ring detection method and system based on image processing. Background With the advanced progress of intelligent manufacturing and industrial industry, workpiece surface detection has become a core link of quality control in the high-end manufacturing field, and technologies such as machine vision, deep learning, multi-mode fusion and the like gradually become mainstream detection means. In the prior art, a single or multiple groups of industrial cameras are deployed to obtain 2D/3D images of the surface of a workpiece, algorithms such as threshold segmentation, template matching, convolutional Neural Network (CNN), visionTransformer and the like are adopted to extract defect characteristics, then preprocessing such as filtering, enhancement and the like is carried out on the acquired images, then a defect area is positioned and defect types are classified through a target detection model, and finally a detection result and judgment information are output. However, in the prior art, image acquisition is easily affected by illumination fluctuation, environmental noise and workpiece material reflection characteristics, so that the image quality is unstable and defect feature extraction is interfered, a detection model depends on manual design features or large-scale labeling of defect samples, and the identification and segmentation accuracy of unknown defects, tiny defects and slender fuzzy defects are insufficient. Disclosure of Invention In view of the above, the embodiment of the application provides a sealing ring detection method and a sealing ring detection system based on image processing, which aim to solve the problems of unstable feature extraction and insufficient precision of complex and tiny defect recognition and segmentation of a model caused by environmental interference in image acquisition in the prior art. A first aspect of an embodiment of the present application provides a seal ring detection method based on image processing, including: Acquiring a plurality of seal ring specification parameter information, a plurality of seal ring material information, a plurality of current seal ring size detection image information to be detected, a plurality of current seal ring end face detection image information to be detected, a plurality of current seal ring circumferential detection image information to be detected, a plurality of historical seal ring size detection information, a plurality of historical seal ring end face detection information and a plurality of historical seal ring circumferential detection information; Obtaining current seal ring contour representing information to be detected and current seal ring end face defect representing information to be detected according to the seal ring specification parameter information, the seal ring material information, the current seal ring size detecting image information to be detected, the current seal ring end face detecting image information to be detected, the historical seal ring size detecting information and the historical seal ring end face detecting information; And generating a plurality of pieces of current seal ring size detection information, a plurality of pieces of current seal ring end face detection information and a plurality of pieces of current seal ring circumferential detection information according to the current seal ring contour representation information to be detected, the current seal ring end face defect representation information to be detected, the plurality of pieces of current seal ring size detection image information to be detected, the plurality of pieces of current seal ring end face detection image information to be detected, the plurality of pieces of current seal ring circumferential detection image information to be detected, the plurality of pieces of historical seal ring circumferential detection information and the seal ring detection model. A second aspect of an embodiment of the present application provides a seal ring detection system based on image processing, including: The information acquisition module is used for acquiring a plurality of seal ring specification parameter information, a plurality of seal ring material information, a plurality of current seal ring size detection image information to be detected, a plurality of current seal ring end face detection image information to be detected, a plurality of current seal ring circumferential detection image information to be detected, a plurality of historical seal ring size detection information, a plurality of historical seal ring end face detection information and a plurality of historical seal ring circumferential detection information; The current seal ring contour representation and end surface defect representation information generation module i