Industrial part detection method and detection system based on machine vision

Abstract

The invention is suitable for the technical field of machine vision, and provides an industrial part detection method and system based on machine vision. The method comprises the following steps: performing initialization shooting, and dividing a plurality of mounting structural surfaces; disassembly monitoring data are obtained, and screw parts in the disassembly process are received and collected; performing faceted positioning detection on the disassembly monitoring and shooting data; a plurality of corresponding surface structure contours are generated and displayed, and the multiple screw parts are positioned and placed in a face-dividing mode; and when the appearance is abnormal, replacement placement is carried out. Machine vision can be applied to industrial part detection of industrial product repair, faceted positioning data of a plurality of screw parts are detected and obtained, a plurality of surface structure contours are generated and displayed, and faceted positioning placement is performed on the plurality of screw parts, so that accurate reference is provided for manual assembly operation, and the situation that the screw parts are subjected to surface alignment during manual assembly is effectively avoided. For the situation that the positions and specifications of screw parts are mixed, the dismounting efficiency is improved, and the situation that industrial products are damaged due to dismounting can be reduced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of machine vision, and particularly relates to an industrial part detection method and system based on machine vision. BACKGROUND

[0002] Machine vision is a comprehensive technology that realizes the visual recognition, understanding and judgment ability of machines by using computer vision, image processing and artificial intelligence technologies. The core of machine vision is to obtain image data of a target object through optical imaging, input the image data into a computer or an embedded processing unit through an image acquisition card, and then analyze, recognize and measure the image by using an algorithm, so as to realize the detection, positioning, recognition, sorting, measurement and control of objects.

[0003] Compared with artificial vision, machine vision has the advantages of high speed, high precision, strong stability and adaptability to harsh environments, and is widely used in industrial automation, intelligent manufacturing, robot navigation, quality detection, security monitoring, medical imaging and other fields.

[0004] In the prior art, machine vision is not applied to the repair process of industrial products, and manual disassembly and assembly are usually relied on. For some industrial products with complex structures, there are a large number of screw parts of different types. When manual disassembly and assembly are performed, the positions and specifications of the screw parts may be confused, and errors may occur in the assembly process, which requires repeated trials, resulting in low disassembly and assembly efficiency and easy damage to the industrial products. SUMMARY

[0005] The purpose of the embodiments of the application is to provide an industrial part detection method and system based on machine vision, which aims to solve the technical problems existing in the prior art mentioned in the background.

[0006] The embodiments of the application are implemented as follows: An industrial part detection method based on machine vision, which specifically comprises the following steps: Initialization shooting is performed, a target industrial product is recognized, a three-dimensional structure model of the target industrial product is matched, structure installation recognition is performed, and a plurality of installation structure surfaces are divided; A manual disassembly process of the target industrial product is monitored and shot in real time, disassembly monitoring data is obtained, and screw parts in the disassembly process are collected; Based on a plurality of installation structure surfaces, the disassembly monitoring data is subjected to surface positioning detection, and surface positioning data of a plurality of screw parts is obtained; According to a plurality of installation structure surfaces, a plurality of corresponding surface structure contours are generated and displayed, and according to a plurality of surface positioning data, the plurality of screw parts are subjected to surface positioning and placement on the plurality of surface structure contours; The multiple screw parts placed in the sub-surface positioning are subjected to appearance detection to determine whether there is appearance abnormality, and corresponding replacement is performed when there is appearance abnormality.

[0007] As a further limitation of the technical scheme of the embodiment of the application, the initialization shooting is performed, the target industrial product is recognized, the three-dimensional structure model of the target industrial product is matched, the structure installation recognition is performed, and the multiple installation structure planes are divided, which specifically includes the following steps. The initialization shooting is performed to obtain initial shooting data; The target recognition is performed on the initial shooting data to determine the target industrial product; The three-dimensional structure model of the target industrial product is matched in a preset model database; The plane recognition is performed on the three-dimensional structure model to determine multiple model structure planes; The structure installation recognition is performed on the multiple model structure planes, and multiple installation structure planes with screw assembly are selected from the multiple model structure planes.

[0008] As a further limitation of the technical scheme of the embodiment of the application, the real-time monitoring and shooting of the manual disassembly process of the target industrial product is performed to obtain disassembly monitoring and shooting data, and the screw parts in the disassembly process are collected, which specifically includes the following steps. The real-time monitoring and shooting of the manual disassembly process of the target industrial product is performed to obtain disassembly monitoring and shooting data; The collection instruction is generated; According to the collection instruction, the collection disc is opened, and the screw parts in the disassembly process are collected through the collection disc.

[0009] As a further limitation of the technical scheme of the embodiment of the application, the sub-surface positioning detection of the multiple screw parts based on the multiple installation structure planes is performed on the disassembly monitoring and shooting data to obtain the sub-surface positioning data of the multiple screw parts, which specifically includes the following steps. The sub-surface detection is performed on the disassembly monitoring and shooting data to determine the sub-surface relationship between the multiple screw parts and the installation structure planes; The positioning detection is performed on the disassembly monitoring and shooting data to determine the plane positions of the multiple screw parts on the corresponding installation structure planes; The order recognition is performed on the disassembly monitoring and shooting data to determine the structure installation orders of the multiple installation structure planes; The sub-surface positioning data of the multiple screw parts are generated by comprehensively considering the multiple sub-surface relationships, the multiple plane positions, and the multiple structure installation orders.

[0010] As a further limitation of the technical scheme of the embodiment of the application, the generating and displaying of the plurality of corresponding surface structure contours according to the plurality of installation structure surfaces, and the sub-surface positioning and placing of the plurality of screw parts on the plurality of surface structure contours according to the plurality of sub-surface positioning data specifically comprises the following steps: projecting the three-dimensional structure model according to the plurality of installation structure surfaces to generate a plurality of surface structure contours corresponding to the plurality of installation structure surfaces; planning display positions of the plurality of surface structure contours according to the plurality of sub-surface positioning data; distinguishing and displaying the plurality of surface structure contours on the receiving disc according to the plurality of display positions; planning corresponding sub-surface placing positions on the plurality of surface structure contours according to the plurality of sub-surface positioning data; sub-surface positioning and placing the plurality of screw parts according to the plurality of sub-surface placing positions.

[0011] As a further limitation of the technical scheme of the embodiment of the application, the appearance detection of the sub-surface positioning and placing plurality of screw parts, the judgment of whether there is an appearance abnormality, and the corresponding replacement and placing when there is an appearance abnormality specifically comprises the following steps: detecting and photographing the sub-surface positioning and placing plurality of screw parts to obtain detection and photographing data; obtaining standard images of the plurality of screw parts; comparing and detecting the detection and photographing data based on the plurality of standard images to judge whether there is an appearance abnormality; selecting an abnormal screw from the plurality of screw parts when there is an appearance abnormality; matching a replacement screw corresponding to the abnormal screw; performing replacement and placing control to remove the abnormal screw and replace and place the replacement screw.

[0012] An industrial part detection system based on machine vision, the system comprising a structure installation identification module, a real-time monitoring and photographing processing module, a sub-surface positioning and detection module, a sub-surface positioning and placing module, and an abnormal replacement and placing module, wherein: The structure installation identification module is used for initial photographing, identifying a target industrial product, matching a three-dimensional structure model of the target industrial product, and performing structure installation identification to divide a plurality of installation structure surfaces. The real-time monitoring and photographing processing module is used for real-time monitoring and photographing of a manual disassembly process of the target industrial product, obtaining disassembly monitoring and photographing data, and receiving and collecting screw parts in the disassembly process. The sub-surface positioning and detection module is used for sub-surface positioning and detection of the disassembly monitoring and photographing data based on the plurality of installation structure surfaces to obtain sub-surface positioning data of the plurality of screw parts. The face positioning and placing module is configured to generate and display a plurality of corresponding face structure contours according to a plurality of installation structure faces, and to perform face positioning and placing of a plurality of screw parts on the plurality of face structure contours according to a plurality of face positioning data. The abnormal replacement and placing module is configured to perform appearance detection on the plurality of screw parts positioned and placed according to the faces, to determine whether the plurality of screw parts have appearance abnormalities, and to perform corresponding replacement and placing when the plurality of screw parts have appearance abnormalities.

[0013] As a further limitation of the technical scheme of the embodiment of the present application, the structure installation identification module specifically comprises: An initialization photographing unit is configured to perform initialization photographing to obtain initial photographing data. A target identification unit is configured to perform target identification on the initial photographing data to determine a target industrial product. A model matching unit is configured to match a three-dimensional structure model of the target industrial product in a preset model database. A plane identification unit is configured to perform plane identification on the three-dimensional structure model to determine a plurality of model structure planes. A structure face selection unit is configured to perform structure installation identification on the plurality of model structure planes, and to select a plurality of installation structure faces having screw assembly from the plurality of model structure planes.

[0014] As a further limitation of the technical scheme of the embodiment of the present application, the face positioning and detection module specifically comprises: A face detection unit is configured to perform face detection on the disassembly monitoring photographing data to determine a plurality of face relationships between screw parts and installation structure faces. A positioning detection unit is configured to perform positioning detection on the disassembly monitoring photographing data to determine a plurality of plane positions of the screw parts on corresponding installation structure faces. An order identification unit is configured to perform order identification on the disassembly monitoring photographing data to determine a structure installation order corresponding to the plurality of installation structure faces. A face positioning data generation unit is configured to generate face positioning data of the plurality of screw parts by comprehensively considering the plurality of face relationships, the plurality of plane positions, and the plurality of structure installation orders.

[0015] As a further limitation of the technical scheme of the embodiment of the present application, the abnormal replacement and placing module specifically comprises: A detection photographing unit is configured to perform detection photographing on the plurality of screw parts positioned and placed according to the faces to obtain detection photographing data. A standard image acquisition unit is configured to acquire standard images of the plurality of screw parts. A comparison detection unit is configured to perform comparison detection on the detection photographing data based on the plurality of standard images to determine whether there is an appearance abnormality. An abnormality selection unit is configured to select an abnormal screw from the plurality of screw parts when there is an appearance abnormality. A replacement matching unit is configured to match a replacement screw corresponding to the abnormal screw. A replacement placement control unit is configured to perform replacement placement control to remove the abnormal screw and place the replacement screw.

[0016] Compared with the prior art, the present application has the following advantages: (1) The present application can apply machine vision to the detection of industrial parts in the repair of industrial products, detect and obtain the surface positioning data of a plurality of screw parts, generate and display a plurality of surface structure contours, and position and place the plurality of screw parts by surface, thereby providing accurate reference for manual assembly operation, effectively avoiding the confusion of the positions and specifications of screw parts during manual assembly, improving the disassembly and assembly efficiency, and reducing the damage to industrial products caused by disassembly and assembly. (2) The present application can detect and photograph the plurality of screw parts positioned and placed by surface, compare with the standard images corresponding to the plurality of screw parts, determine whether there is an appearance abnormality, and then perform corresponding replacement placement when there is an appearance abnormality, thereby realizing automatic replacement of abnormal screws and further improving the disassembly and assembly efficiency of industrial product repair. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A flowchart of the machine vision-based industrial part detection method provided by the present application is shown; Figure 2 An application architecture diagram of the machine vision-based industrial part detection system provided by the present application is shown. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0019] It can be understood that in the prior art, machine vision is not applied to the repair process of industrial products, and manual disassembly and assembly is usually relied on. For some industrial products with complex structure, there are a large number of screw parts of different types. Manual disassembly and assembly may cause confusion of the positions and specifications of screw parts, and errors may occur during assembly, which requires repeated attempts, resulting in low disassembly and assembly efficiency and easy damage to industrial products.

[0020] To solve the above problems, the embodiment of the present application discloses an industrial part detection method and detection system based on machine vision, which identifies the target industrial product by initial shooting, matches the three-dimensional structure model of the target industrial product, and performs structure installation recognition and divides multiple installation structure surfaces; the manual disassembly process of the target industrial product is monitored in real time, disassembly monitoring data is obtained, and screw parts in the disassembly process are collected; based on multiple installation structure surfaces, the disassembly monitoring data is detected and positioned in multiple surfaces to obtain multiple surface positioning data of the screw parts; according to multiple installation structure surfaces, multiple corresponding surface structure contours are generated and displayed, and multiple screw parts are positioned and placed on multiple surface structure contours according to multiple surface positioning data; the multiple screw parts positioned and placed are subjected to appearance detection to determine whether they have appearance abnormalities, and are replaced and placed when they have appearance abnormalities. The machine vision can be applied to the industrial part detection of industrial product repair, the surface positioning data of multiple screw parts is detected and obtained, multiple surface structure contours are generated and displayed, and multiple screw parts are positioned and placed in multiple surfaces, thereby providing accurate reference for manual assembly operation, effectively avoiding the confusion of the positions and specifications of the screw parts during manual assembly, improving the disassembly efficiency, and reducing the damage of the industrial product caused by disassembly.

[0021] Specifically, Figure 1 The flowchart of the industrial part detection method based on machine vision provided by the embodiment of the present application is shown.

[0022] In a preferred embodiment provided by the present application, an industrial part detection method based on machine vision comprises the following steps: Step S101, initial shooting is performed to identify the target industrial product, match the three-dimensional structure model of the target industrial product, and perform structure installation recognition to divide multiple installation structure surfaces.

[0023] In the embodiment of the present application, before the manual disassembly operation of product repair, the product to be disassembled is placed on the disassembly workbench, the initial shooting data is obtained by initial shooting of the disassembly workbench, the target feature recognition and comparison analysis of the initial shooting data are performed based on the preset product feature data, the target industrial product is determined, the three-dimensional structure model of the target industrial product is matched in the preset model database, the plane recognition of the three-dimensional structure model at different rotation angles is performed by automatic rotation processing of the three-dimensional structure model, the multiple model structure planes of the surface of the three-dimensional structure model are determined, and the structure installation recognition of the multiple model structure planes is performed to select multiple installation structure surfaces with screw assembly from the multiple model structure planes.

[0024] Specifically, in another preferred embodiment provided by the present application, the initial shooting, the identification of the target industrial product, the matching of the three-dimensional structure model of the target industrial product, the structural installation identification, and the division of the plurality of installation structure surfaces specifically include the following steps: initial shooting is performed to obtain initial shooting data; target identification is performed on the initial shooting data to determine the target industrial product; a three-dimensional structure model of the target industrial product is matched in a preset model database; plane identification is performed on the three-dimensional structure model to determine a plurality of model structure planes; structural installation identification is performed on the plurality of model structure planes, and a plurality of installation structure surfaces with screw assembly are selected from the plurality of model structure planes.

[0025] Further, the industrial part detection method based on machine vision further includes the following steps: In step S102, the manual disassembly process of the target industrial product is monitored and shot in real time to obtain disassembly monitoring and shooting data, and screw parts in the disassembly process are collected.

[0026] In the embodiment of the present application, when the worker performs the manual disassembly operation on the target industrial product, the manual disassembly process of the target industrial product is monitored and shot in real time to obtain disassembly monitoring and shooting data, and a collection instruction is generated to open a collection disc and collect screw parts in the disassembly process of the target industrial product through the collection disc.

[0027] Specifically, in another preferred embodiment provided by the present application, the initial shooting, the identification of the target industrial product, the matching of the three-dimensional structure model of the target industrial product, the structural installation identification, and the division of the plurality of installation structure surfaces specifically include the following steps: initial shooting is performed to obtain initial shooting data; a collection instruction is generated; the collection disc is opened according to the collection instruction, and screw parts in the disassembly process are collected through the collection disc.

[0028] Further, the industrial part detection method based on machine vision further includes the following steps: In step S103, based on the plurality of installation structure surfaces, the disassembly monitoring and shooting data are subjected to surface positioning detection to obtain surface positioning data of the plurality of screw parts.

[0029] In the embodiment of the present application, by facet detection on the disassembly monitoring data, the corresponding facet relationship between the plurality of screw parts and the plurality of mounting structure surfaces is determined, and by positioning detection on the disassembly monitoring data, the planar positions of the plurality of screw parts on the corresponding mounting structure surfaces are determined, and by sequence recognition on the disassembly monitoring data, the structure disassembly sequence of the plurality of mounting structure surfaces in the process of disassembling the target industrial product is determined, and the structure disassembly sequence of the plurality of mounting structure surfaces is reversely sorted to obtain the structure mounting sequence of the plurality of mounting structure surfaces, and then the plurality of facet relationships, the plurality of planar positions and the plurality of structure mounting sequences are comprehensively arranged to generate the facet positioning data of the plurality of screw parts.

[0030] It can be understood that the facet positioning data of the plurality of screw parts records the corresponding facet relationship, planar position and structure mounting sequence of each screw part, wherein the facet relationship is the corresponding relationship between the screw part and the mounting structure surface, the planar position is the specific position of the screw part on the corresponding mounting structure surface, and the structure mounting sequence is the sequence of the mounting structure surface corresponding to the screw part in the plurality of mounting structure surfaces.

[0031] Specifically, in another preferred embodiment provided by the present application, the facet positioning detection on the disassembly monitoring data based on the plurality of mounting structure surfaces to obtain the facet positioning data of the plurality of screw parts specifically includes the following steps: facet detection on the disassembly monitoring data to determine the facet relationship between the plurality of screw parts and the mounting structure surface; positioning detection on the disassembly monitoring data to determine the planar positions of the plurality of screw parts on the corresponding mounting structure surfaces; sequence recognition on the disassembly monitoring data to determine the structure mounting sequence of the plurality of mounting structure surfaces; comprehensive arrangement of the plurality of facet relationships, the plurality of planar positions and the plurality of structure mounting sequences to generate the facet positioning data of the plurality of screw parts.

[0032] Further, the industrial part detection method based on machine vision further includes the following steps: Step S104, generating and displaying a plurality of corresponding surface structure contours according to the plurality of mounting structure surfaces, and placing a plurality of screw parts on the plurality of surface structure contours according to the plurality of facet positioning data.

[0033] In the embodiment of the present application, the plurality of installation structure surfaces are projected and planned, the projection normal directions of the plurality of installation structure surfaces are determined, the three-dimensional structure model is projected according to the plurality of projection normal directions, the surface structure contours corresponding to the plurality of installation structure surfaces are generated, the display positions of the plurality of surface structure contours are planned according to the plurality of structure installation sequences in the plurality of surface positioning data, then the plurality of surface structure contours are displayed and distinguished on the receiving disc according to the plurality of display positions, and the surface placement positions corresponding to the plurality of screw parts are planned on the plurality of surface structure contours according to the plurality of surface relationships and the plurality of plane positions in the plurality of surface positioning data, then the plurality of screw parts are positioned and placed in the surfaces according to the plurality of surface placement positions, so that each screw part can be placed at the plane position of the surface structure contour of the corresponding installation structure surface.

[0034] Specifically, in another preferred embodiment provided by the present application, the generating and displaying of the plurality of corresponding surface structure contours according to the plurality of installation structure surfaces, and the positioning and placing of the plurality of screw parts in the surfaces according to the plurality of surface positioning data specifically include the following steps: projecting the three-dimensional structure model according to the plurality of installation structure surfaces to generate the surface structure contours corresponding to the plurality of installation structure surfaces; planning the display positions of the plurality of surface structure contours according to the plurality of surface positioning data; displaying and distinguishing the plurality of surface structure contours on the receiving disc according to the plurality of display positions; planning the corresponding surface placement positions on the plurality of surface structure contours according to the plurality of surface positioning data; positioning and placing the plurality of screw parts in the surfaces according to the plurality of surface placement positions.

[0035] Further, the industrial part detection method based on machine vision further includes the following steps: Step S105, appearance detection is performed on the plurality of screw parts positioned and placed in the surfaces, it is judged whether there is appearance abnormality, and corresponding replacement and placement are performed when there is appearance abnormality.

[0036] In the embodiment of the present application, after the sub-surface positioning and placement of the plurality of screw parts are completed, the plurality of screw parts are detected and photographed to obtain detection and photographing data, and the detection and photographing data are subjected to model recognition to determine the screw models corresponding to the plurality of screw parts, and according to the plurality of screw models, the standard images corresponding to the plurality of screw parts are matched from a preset standard database, the detection and photographing data are subjected to comparative detection based on the plurality of standard images to determine whether there is an appearance abnormality, and in the case where it is determined that there is an appearance abnormality, the abnormal screw with the appearance abnormality is selected from the plurality of screw parts, and the corresponding replacement screw is matched according to the screw model corresponding to the abnormal screw to perform replacement placement control, remove the abnormal screw, and replace and place the replacement screw, thereby realizing automatic replacement of the abnormal screw.

[0037] It can be understood that the appearance abnormality of the screw part includes pits, scratches, cracks, rust marks, bending, etc.

[0038] Specifically, in another preferred embodiment provided by the present application, the appearance detection of the plurality of screw parts subjected to sub-surface positioning and placement, the determination of whether there is an appearance abnormality, and the corresponding replacement placement when there is an appearance abnormality specifically include the following steps: The plurality of screw parts subjected to sub-surface positioning and placement are detected and photographed to obtain detection and photographing data; The standard images of the plurality of screw parts are obtained; The detection and photographing data are subjected to comparative detection based on the plurality of standard images to determine whether there is an appearance abnormality; When there is an appearance abnormality, an abnormal screw is selected from the plurality of screw parts; The replacement screw corresponding to the abnormal screw is matched; Replacement placement control is performed to remove the abnormal screw and replace and place the replacement screw.

[0039] Further, Figure 2 An application architecture diagram of the industrial part detection system based on machine vision provided by the embodiment of the present application is shown.

[0040] Specifically, in another preferred embodiment provided by the present application, an industrial part detection system based on machine vision includes: A structure installation recognition module 101 is configured to perform initialization photographing, recognize a target industrial product, match a three-dimensional structure model of the target industrial product, and perform structure installation recognition to divide a plurality of installation structure surfaces.

[0041] In the embodiment of the present application, before the manual disassembly operation of product repair is performed, the product to be disassembled needs to be placed on the disassembly workbench, the structure installation recognition module 101 performs initial shooting on the disassembly workbench to obtain initial shooting data, based on the preset product feature data, the target feature recognition and comparative analysis of the initial shooting data are performed to determine the target industrial product, and then the three-dimensional structure model of the target industrial product is matched in the preset model database, the three-dimensional structure model is automatically rotated at multiple angles, the plane recognition of the three-dimensional structure model at different rotation angles is performed, the multiple model structure planes of the three-dimensional structure model are determined, and then the structure installation recognition of the multiple model structure planes is performed, and multiple installation structure planes with screw assembly are selected from the multiple model structure planes.

[0042] Specifically, in another preferred embodiment provided by the present application, the structure installation recognition module 101 specifically includes: An initial shooting unit for performing initial shooting to obtain initial shooting data; A target recognition unit for performing target recognition on the initial shooting data to determine a target industrial product; A model matching unit for matching a three-dimensional structure model of the target industrial product in a preset model database; A plane recognition unit for performing plane recognition on the three-dimensional structure model to determine multiple model structure planes; A structure plane selection unit for performing structure installation recognition on the multiple model structure planes, and selecting multiple installation structure planes with screw assembly from the multiple model structure planes.

[0043] Further, the industrial part detection system based on machine vision further includes: A real-time monitoring and shooting processing module 102 for performing real-time monitoring and shooting on the manual disassembly process of the target industrial product to obtain disassembly monitoring and shooting data, and collecting screws in the disassembly process.

[0044] In the embodiment of the present application, when the staff performs the manual disassembly operation of product repair on the target industrial product, the real-time monitoring and shooting processing module 102 performs real-time monitoring and shooting on the manual disassembly process of the target industrial product to obtain disassembly monitoring and shooting data, generates a collection instruction, opens the collection disc according to the collection instruction, and collects the screws in the disassembly process of the target industrial product through the collection disc.

[0045] A surface positioning detection module 103 for performing surface positioning detection on the disassembly monitoring and shooting data based on the multiple installation structure planes to obtain surface positioning data of the multiple screws.

[0046] In the embodiment of the present application, the face positioning detection module 103 determines the corresponding face relationship between the plurality of screw parts and the plurality of installation structure faces by performing face detection on the disassembly monitoring data, and determines the planar positions of the plurality of screw parts on the corresponding installation structure faces by performing positioning detection on the disassembly monitoring data. Meanwhile, the face positioning detection module 103 determines the structure disassembly sequence of the plurality of installation structure faces in the process of disassembling the target industrial product by performing sequence recognition on the disassembly monitoring data, and reversely sorts the structure disassembly sequence of the plurality of installation structure faces to obtain the structure installation sequence of the plurality of installation structure faces. Then, the face positioning detection module 103 comprehensively arranges the plurality of face relationships, the plurality of planar positions and the plurality of structure installation sequences to generate the face positioning data of the plurality of screw parts.

[0047] Specifically, in another preferred embodiment provided by the present application, the face positioning detection module 103 specifically includes: a face detection unit configured to perform face detection on the disassembly monitoring data to determine the face relationship between the plurality of screw parts and the installation structure face; a positioning detection unit configured to perform positioning detection on the disassembly monitoring data to determine the planar positions of the plurality of screw parts on the corresponding installation structure faces; a sequence recognition unit configured to perform sequence recognition on the disassembly monitoring data to determine the structure installation sequence of the plurality of installation structure faces; a face positioning data generation unit configured to comprehensively arrange the plurality of face relationships, the plurality of planar positions and the plurality of structure installation sequences to generate the face positioning data of the plurality of screw parts.

[0048] Further, the industrial part detection system based on machine vision further includes: a face positioning placement module 104 configured to generate and display a plurality of corresponding face structure contours according to the plurality of installation structure faces, and to perform face positioning placement of the plurality of screw parts on the plurality of face structure contours according to the plurality of face positioning data.

[0049] In the embodiment of the present application, the face positioning and placing module 104 projects and plans the plurality of installation structure faces, determines the projection normal directions of the plurality of installation structure faces, projects the three-dimensional structure model according to the plurality of projection normal directions, generates the face structure contours corresponding to the plurality of installation structure faces, plans the display positions of the plurality of face structure contours according to the plurality of structure installation sequences in the plurality of face positioning data, and then displays the plurality of face structure contours on the receiving disc according to the plurality of display positions, and plans the face placing positions of the plurality of screw parts on the plurality of face structure contours according to the plurality of face relationships and the plurality of plane positions in the plurality of face positioning data, and then positions and places the plurality of screw parts according to the plurality of face placing positions, so that each screw part can be placed at the plane position of the face structure contour of the corresponding installation structure face.

[0050] The abnormal replacement and placing module 105 is configured to detect the appearance of the plurality of screw parts positioned and placed in the faces, determine whether there is an appearance abnormality, and replace and place the corresponding screw part when there is an appearance abnormality.

[0051] In the embodiment of the present application, after the face positioning and placing of the plurality of screw parts is completed, the abnormal replacement and placing module 105 detects and photographs the plurality of screw parts to obtain detection and photographing data, identifies the screw models of the plurality of screw parts according to the detection and photographing data, matches the standard images corresponding to the plurality of screw parts from a pre-set standard database according to the plurality of screw models, compares and detects the detection and photographing data based on the plurality of standard images to determine whether there is an appearance abnormality, selects an abnormal screw part from the plurality of screw parts when there is an appearance abnormality, matches a replacement screw part corresponding to the abnormal screw part according to the screw model of the abnormal screw part, controls the replacement and placing of the replacement screw part, removes the abnormal screw part, and replaces and places the replacement screw part, thereby achieving automatic replacement of the abnormal screw part.

[0052] Specifically, in another preferred embodiment of the present application, the abnormal replacement and placing module 105 specifically includes: a detection and photographing unit configured to detect and photograph the plurality of screw parts positioned and placed in the faces to obtain detection and photographing data; a standard image acquisition unit configured to acquire standard images of the plurality of screw parts; a comparison and detection unit configured to compare and detect the detection and photographing data based on the plurality of standard images to determine whether there is an appearance abnormality; an abnormal selection unit configured to select an abnormal screw part from the plurality of screw parts when there is an appearance abnormality; a replacement matching unit configured to match a replacement screw part corresponding to the abnormal screw part; A replacement screw is placed in the hole of the workpiece, and a replacement screw placement control unit is used to perform replacement screw placement control to remove the abnormal screw and place the replacement screw.

[0053] The above embodiments only express several embodiments of the present application, and the description is more specific and detailed, but it cannot be understood as limiting the scope of the patent of the present application. It should be noted that for ordinary skilled persons in the art, without departing from the concept of the present application, several modifications and improvements can be made, which belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A machine vision-based industrial part inspection method, characterized by, The method specifically comprises the following steps: Performing initial shooting, identifying the target industrial product, matching the three-dimensional structure model of the target industrial product, and performing structure installation identification, dividing a plurality of installation structure surfaces; Real-time monitoring and shooting the manual disassembly process of the target industrial product, obtaining disassembly monitoring and shooting data, and receiving and collecting screw parts in the disassembly process; Based on a plurality of installation structure surfaces, the disassembly monitoring and shooting data are subjected to surface positioning detection to obtain surface positioning data of a plurality of screw parts; According to a plurality of installation structure surfaces, a plurality of corresponding surface structure contours are generated and displayed, and according to a plurality of surface positioning data, a plurality of screw parts are positioned and placed on a plurality of surface structure contours; The appearance of the plurality of screw parts positioned and placed is detected to determine whether there is an appearance abnormality, and when there is an appearance abnormality, corresponding replacement and placement are performed.

2. The machine vision-based industrial part inspection method of claim 1, wherein, The performing initial shooting, identifying the target industrial product, matching the three-dimensional structure model of the target industrial product, and performing structure installation identification, and dividing a plurality of installation structure surfaces specifically comprises the following steps: Performing initial shooting to obtain initial shooting data; Performing target identification on the initial shooting data to determine the target industrial product; Matching the three-dimensional structure model of the target industrial product in a preset model database; Performing plane identification on the three-dimensional structure model to determine a plurality of model structure planes; Performing structure installation identification on a plurality of model structure planes to select a plurality of installation structure surfaces with screw assembly from a plurality of model structure planes.

3. The machine vision-based industrial part inspection method of claim 1, wherein, The real-time monitoring and shooting the manual disassembly process of the target industrial product, obtaining disassembly monitoring and shooting data, and receiving and collecting screw parts in the disassembly process specifically comprises the following steps: Real-time monitoring and shooting the manual disassembly process of the target industrial product to obtain disassembly monitoring and shooting data; Generating a receiving and collecting instruction; According to the receiving and collecting instruction, opening the receiving and collecting disc, and receiving and collecting screw parts in the disassembly process through the receiving and collecting disc.

4. The machine vision-based industrial part inspection method of claim 1, wherein, The surface positioning detection based on a plurality of installation structure surfaces on the disassembly monitoring and shooting data to obtain surface positioning data of a plurality of screw parts specifically comprises the following steps: Performing surface detection on the disassembly monitoring and shooting data to determine the surface relationship between a plurality of screw parts and installation structure surfaces; Performing positioning detection on the disassembly monitoring and shooting data to determine the plane position of a plurality of screw parts on the corresponding installation structure surface; Performing sequence identification on the disassembly monitoring and shooting data to determine the structure installation sequence corresponding to a plurality of installation structure surfaces; Comprehensively generating surface positioning data of a plurality of screw parts based on a plurality of surface relationships, a plurality of plane positions, and a plurality of structure installation sequences.

5. The machine vision-based industrial part inspection method of claim 3, wherein, The generating and displaying a plurality of corresponding surface structure contours according to a plurality of installation structure surfaces, and positioning and placing a plurality of screw parts on a plurality of surface structure contours according to a plurality of surface positioning data specifically comprises the following steps: According to a plurality of installation structure surfaces, the three-dimensional structure model is projected to generate a plurality of surface structure contours corresponding to the installation structure surfaces; According to a plurality of surface positioning data, the display positions of a plurality of surface structure contours are planned; According to the plurality of display positions, the plurality of face structure profiles are displayed on the receiving disc; According to the plurality of face positioning data, the plurality of face structure profiles are planned on the plurality of face structure profiles; According to the plurality of face positioning data, the plurality of face structure profiles are planned on the plurality of face structure profiles; 6. The machine vision-based industrial part inspection method of claim 1, wherein, The appearance detection of the plurality of screw parts positioned and placed in the face is carried out to determine whether there is an appearance abnormality, and when there is an appearance abnormality, corresponding replacement and placement is carried out. Specifically includes the following steps: The detection and shooting of the plurality of screw parts positioned and placed in the face is carried out to obtain detection and shooting data; Obtain the standard image of the plurality of screw parts; Based on the plurality of standard images, the detection and shooting data are compared and detected to determine whether there is an appearance abnormality; When there is an appearance abnormality, select an abnormal screw from the plurality of screw parts; Match the replacement screw corresponding to the abnormal screw; Carry out replacement and placement control, remove the abnormal screw, and replace and place the replacement screw.

7. A machine vision-based industrial part inspection system, characterized by, The system comprises a structure installation identification module, a real-time monitoring and shooting processing module, a face positioning detection module, a face positioning and placing module, and an abnormal replacement and placing module, wherein: The structure installation identification module is used for initialization shooting, identifying a target industrial product, matching a three-dimensional structure model of the target industrial product, and carrying out structure installation identification to divide a plurality of installation structure faces; The real-time monitoring and shooting processing module is used for real-time monitoring and shooting of the manual disassembly process of the target industrial product, obtaining disassembly monitoring and shooting data, and collecting screw parts in the disassembly process; The face positioning detection module is used for face positioning detection of the disassembly monitoring and shooting data based on the plurality of installation structure faces to obtain face positioning data of the plurality of screw parts; The face positioning and placing module is used for generating and displaying a plurality of corresponding face structure profiles according to a plurality of installation structure faces, and positioning and placing a plurality of screw parts on a plurality of face structure profiles according to a plurality of face positioning data; The abnormal replacement and placing module is used for appearance detection of the plurality of screw parts positioned and placed in the face to determine whether there is an appearance abnormality, and when there is an appearance abnormality, corresponding replacement and placement is carried out.

8. The machine vision-based industrial part inspection system of claim 7, wherein, The structure installation identification module specifically comprises: An initialization shooting unit is used for initialization shooting to obtain initial shooting data; A target identification unit is used for target identification of the initial shooting data to determine a target industrial product; A model matching unit is used for matching a three-dimensional structure model of the target industrial product in a preset model database; A plane identification unit is used for plane identification of the three-dimensional structure model to determine a plurality of model structure planes; A structure face selection unit is used for structure installation identification of the plurality of model structure planes to select a plurality of installation structure faces with screw assembly from the plurality of model structure planes.

9. The machine vision-based industrial part inspection system of claim 7, wherein, The face positioning detection module specifically comprises: A face detection unit is used for face detection of the disassembly monitoring and shooting data to determine the face relationship between the plurality of screw parts and the installation structure face; A positioning detection unit is configured to perform positioning detection on the disassembly monitoring data to determine the planar positions of the plurality of screw parts on the corresponding mounting structure surfaces; An order identification unit is configured to perform order identification on the disassembly monitoring data to determine the structure mounting orders of the plurality of mounting structure surfaces; A sub-surface positioning data generation unit is configured to generate sub-surface positioning data of the plurality of screw parts by comprehensively considering the plurality of sub-surface relationships, the plurality of planar positions, and the plurality of structure mounting orders.

10. The machine vision-based industrial part inspection system of claim 7, wherein, The abnormal replacement placing module specifically includes: A detection photographing unit is configured to perform detection photographing on the plurality of screw parts placed in the sub-surface positions to obtain detection photographing data; A standard image acquisition unit is configured to acquire standard images of the plurality of screw parts; A comparison detection unit is configured to perform comparison detection on the detection photographing data based on the plurality of standard images to determine whether there is an appearance abnormality; An abnormal selection unit is configured to select an abnormal screw from the plurality of screw parts when there is an appearance abnormality; A replacement matching unit is configured to match a replacement screw corresponding to the abnormal screw; A replacement placing control unit is configured to perform replacement placing control, remove the abnormal screw, and replace the replacement screw.

Images