CN117092114A - Appearance detection system based on AI - Google Patents
Appearance detection system based on AI Download PDFInfo
- Publication number
- CN117092114A CN117092114A CN202311333946.5A CN202311333946A CN117092114A CN 117092114 A CN117092114 A CN 117092114A CN 202311333946 A CN202311333946 A CN 202311333946A CN 117092114 A CN117092114 A CN 117092114A
- Authority
- CN
- China
- Prior art keywords
- detection
- product
- image acquisition
- material taking
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 330
- 230000007306 turnover Effects 0.000 claims abstract description 48
- 238000012546 transfer Methods 0.000 claims abstract description 13
- 238000007599 discharging Methods 0.000 claims abstract description 12
- 230000000694 effects Effects 0.000 claims abstract description 11
- 239000000463 material Substances 0.000 claims description 92
- 238000009434 installation Methods 0.000 claims description 18
- 239000000969 carrier Substances 0.000 claims description 11
- 238000007689 inspection Methods 0.000 claims description 9
- 230000000007 visual effect Effects 0.000 claims description 7
- 238000010801 machine learning Methods 0.000 claims description 6
- 238000000034 method Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 241000252254 Catostomidae Species 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 claims description 3
- 230000001681 protective effect Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000001360 synchronised effect Effects 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 238000009960 carding Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/01—Arrangements or apparatus for facilitating the optical investigation
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Processing (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The application relates to an AI-based appearance detection system, which comprises a machine table, and a loading transfer module, a rear shell surface detection module, a product turn-over module, a discharging transplanting module and an inner cavity surface detection module which are sequentially arranged on the machine table. The machine bench is covered with a protective hood. The feeding transfer module, the rear shell surface detection module, the product turning module, the discharging transplanting module and the inner cavity surface detection module are connected with an appearance detection controller. The appearance detection controller generates a detection scheme to control the feeding transfer module, the rear shell surface detection module, the product turning module, the blanking transplanting module and the inner cavity surface detection module to operate based on detection requirement information of a product to be detected, automatically feeds and conveys the product to be detected, acquires rear shell surface image information of the product and inner cavity image information of the product, and performs appearance detection on the product. The application has the effect of effectively improving the detection efficiency and the detection precision.
Description
Technical Field
The application relates to the field of visual detection, in particular to an appearance detection system based on AI.
Background
In recent years, 3C products are newly developed with the development of halls and electronic industries, mainly refer to computer (computer), communication (communication) and consumer (consumer) electronic products, in this industry, the development of technologies is more and more advanced, the development of technologies is frequent, a large number of new products are required to be launched almost every month, and the development of the technologies is also a hotspot field of product design, and each design company establishes 3C departments in a dispute to specially design and develop the products. Among the most common are cell phones, tablet computers, notebooks, etc.
After the production of the existing 3C product rear shell is finished, a plurality of detection such as surface defect detection of BG surface (rear shell surface), flaw, glue overflow and the like, CG surface (inner cavity surface) laser carving detection, component (upright post, screw hole mounting seat, fixed mounting seat, electric component and the like) mounting detection, diaphragm pasting mounting detection, hole site detection and the like are needed. The detection items of different 3C products are different according to the structures, and the detection items of the rear shells of the products with complex part structures and strong functionality are as many as hundreds of items. In order to realize accurately detecting the back shell of the 3C product, the existing enterprises often need to integrate detection demands, a plurality of detection pipelines are arranged to detect the back shell of the 3C product through combination of manpower and machines, the detection process is complex, a large amount of manpower and material resources are required to be consumed, and the detection efficiency and the detection precision are low.
Aiming at the related technology, the inventor considers that the existing 3C product backshell detection has redundant and complex detection flow because of more detection items, so that a large amount of manpower and material resources are consumed, and the detection efficiency and the detection precision are low.
Disclosure of Invention
The application provides an appearance detection system based on AI (advanced technology) in order to solve the problems that the existing 3C product backshell detection has redundant and complex detection flow due to more detection items, a large amount of manpower and material resources are required to be consumed, and the detection efficiency and detection precision are low.
In a first aspect, the present application provides an AI-based appearance detection system, which adopts the following technical scheme:
an AI-based appearance detection system, comprising:
a machine table;
the feeding transfer module is arranged on the machine platform and comprises a feeding mechanical arm and a feeding transplanting assembly, wherein the feeding mechanical arm is used for conveying a product to be detected to the feeding transplanting assembly, and the product to be detected is in a rear shell surface upwards;
the back shell surface detection module is arranged on the machine platform and comprises a back shell detection installation frame and a back shell surface image acquisition assembly, wherein the back shell detection installation frame is arranged above the feeding transplanting assembly in a crossing mode, the back shell surface image acquisition assembly is arranged on the back shell detection installation frame, the feeding transplanting assembly is used for transplanting a product to be detected to a back shell surface image acquisition assembly, the back shell surface image acquisition assembly is used for acquiring images of the back shell surface of the product to be detected to obtain back shell surface image information of the product, and the back shell surface image information of the product comprises back shell surface plane image information and back shell 3D image information of the product;
the product overturning module is arranged on the machine and comprises a product overturning assembly and a material taking assembly, wherein the product overturning assembly is used for sucking and overturning a product to be detected on the feeding transplanting assembly, and the material taking assembly is used for taking out the overturned product;
The blanking transplanting module is arranged on the machine platform and comprises a blanking manipulator and a blanking transplanting assembly, the blanking transplanting assembly comprises a multi-shaft transplanting platform and a Y-shaft linear guide rail, the multi-shaft transplanting platform is arranged on a sliding table of the Y-shaft linear guide rail, and the material taking assembly places the overturned product on the multi-shaft transplanting platform;
the inner cavity surface detection module is arranged on the machine table and comprises an inner cavity image acquisition assembly and an inner cavity detection mounting frame which is arranged above the Y-axis linear guide rail in a crossing mode, the Y-axis linear guide rail conveys the multi-axis transplanting platform to the inner cavity image acquisition assembly, and the inner cavity image acquisition assembly acquires images of products with various placement angles displayed by the multi-axis transplanting platform to obtain inner cavity image information of the products; the method comprises the steps of,
the appearance detection controller is arranged in the machine table, based on detection requirement information of a product to be detected, a detection scheme is generated to control the feeding and transferring module, the rear shell surface detection module, the product turning module, the discharging and transplanting module and the inner cavity surface detection module to operate, automatic feeding and discharging conveying is carried out on the product to be detected, product rear shell surface image information and product inner cavity image information are collected, and appearance detection is carried out on the product.
Preferably, the feeding transplanting assembly comprises a feeding linear guide rail and a feeding transplanting platform, the feeding transplanting platform is arranged on a sliding block of the feeding linear guide rail, feeding auxiliary guide rails are arranged on two sides of the feeding linear guide rail in parallel, and the feeding transplanting platform is arranged on the two feeding auxiliary guide rails in a sliding manner; the feeding transplanting platform comprises a square frame body and a plurality of positioning carriers arranged at the top of the square frame body and used for positioning and bearing products to be detected, a plurality of code scanning devices used for scanning codes for identifying the products to be detected are arranged on the square frame body, the code scanning devices correspond to the positioning carriers one to one, and code scanning holes for scanning codes of the code scanning devices are formed in the positioning carriers.
Preferably, the rear shell detection mounting frame comprises a cross beam and two upright posts, the tops of the two upright posts are connected with the cross beam, a detection linear guide rail is arranged on the cross beam along the length direction of the cross beam, and the rear shell image acquisition assembly is arranged on a sliding table of the detection linear guide rail; the back shell face image acquisition assembly comprises a first mounting plate, and a high-definition camera, a coaxial light source and an annular light source which are sequentially mounted on the first mounting plate from top to bottom, wherein the first mounting plate is fixedly connected with a sliding table for detecting the linear guide rail, and one side of the first mounting plate is provided with a 3D camera.
Preferably, the product overturning assembly comprises an overturning installation frame and an overturning part, wherein the overturning installation frame comprises two support columns and a second installation plate provided with the tops of the two support columns, the overturning part comprises an overturning lifting driving piece, an overturning transverse plate and an overturning material taking plate, the overturning lifting driving piece is installed on the second installation plate, and the telescopic end of the overturning lifting driving piece is fixedly connected with the overturning transverse plate; the two ends of the bottom of the turnover transverse plate are respectively provided with a turnover connecting plate, the two ends of the turnover material taking plate are respectively connected with the two turnover connecting plates in a rotating way, and the turnover connecting plates are provided with turnover driving motors for driving the turnover material taking plate to rotate;
the material taking assembly comprises a material taking linear guide rail, a material taking lifting driving piece, a material taking transverse plate and a lifting material taking plate, wherein the material taking linear guide rail is arranged at the bottom of the second mounting plate, the material taking transverse plate is arranged on a sliding table of the material taking linear guide rail, the material taking lifting driving piece is arranged on the material taking transverse plate, and a telescopic end penetrates through the material taking transverse plate to be fixedly connected with the lifting material taking plate;
the turnover material taking plate and the lifting material taking plate are respectively provided with a plurality of vacuum suckers for sucking products, the second mounting plate is provided with a yielding hole for the material taking lifting driving piece to move, and the turnover transverse plate is provided with a yielding groove for accommodating the telescopic end of the material taking lifting driving piece, so that the lifting material taking plate moves to the upper side of the turnover material taking plate under the driving of the material taking linear guide rail to suck the turned products.
Preferably, both sides of the Y-axis linear guide rail are provided with Y-axis auxiliary guide rails, and the multi-axis transplanting platform is arranged on the two Y-axis auxiliary guide rails in a sliding manner; the multi-axis transplanting platform comprises a bearing bottom plate and a bearing seat, wherein connecting vertical plates are arranged at two ends of the bearing bottom plate in the width direction, two ends of the bearing seat are respectively and rotatably connected with the two connecting vertical plates, and a swinging driving piece for driving the bearing seat to rotate is arranged on the connecting vertical plates; the top of the bearing seat is embedded with a plurality of electric rotating platforms, and the output ends of the electric rotating platforms are connected with positioning carriers for bearing products.
Preferably, the positioning carrier comprises a bottom plate and a supporting middle plate, the supporting middle plate is fixedly arranged on the bottom plate, a positioning block for clamping a product is arranged on the bottom plate, a positioning column for positioning the product is arranged on the supporting middle plate, and a plurality of vacuum suction nozzles for sucking and fixing the product are further arranged on the supporting middle plate.
Preferably, the inner cavity detection mounting frame comprises a mounting cross beam and two mounting upright posts, wherein the tops of the two mounting upright posts are connected with the mounting cross beam, and an X-axis linear guide rail is arranged on the mounting cross beam; the inner cavity image acquisition assembly comprises a third mounting plate and a plurality of groups of CCD modules, wherein the third mounting plate is mounted on a sliding table of the X-axis linear guide rail, a Z-axis linear guide rail is arranged on the third mounting plate along the vertical direction, two sides of the Z-axis linear guide rail are provided with Z-axis auxiliary guide rails in the same direction, the Z-axis linear guide rail and the sliding tables of the two Z-axis auxiliary guide rails are jointly connected with a fourth mounting plate, and the CCD modules are sequentially arranged on the fourth mounting plate along the horizontal direction.
Preferably, the appearance detection controller performs appearance detection on the product, specifically including the following steps:
the appearance detection controller obtains detection requirement information of a product to be detected, wherein the detection requirement information comprises name information, detection standard information and image acquisition requirement information of a plurality of detection items, and the image acquisition requirement information of the detection items comprises a required image type, an image number, an image acquisition target and an image acquisition angle;
the appearance detection controller gathers and sorts the image acquisition demand information of each detection item based on the detection demand information of the product to be detected to generate a detection scheme, wherein the detection scheme comprises a detection control instruction and image acquisition flow information;
the appearance detection controller sends detection control instructions to the feeding transfer module, the rear shell surface detection module, the product turning module, the blanking transplanting module and the inner cavity surface detection module, so that the rear shell surface and the inner cavity surface of the product to be detected are automatically fed and discharged according to image acquisition flow information to acquire image information of the rear shell surface and the inner cavity surface of the product, and image information of the rear shell surface and image information of the inner cavity of the product are obtained;
the appearance detection controller performs appearance detection on each detection item through a preset AI visual detection model based on the product back shell surface image information and the product inner cavity image information, and outputs an appearance detection result; the AI visual detection model is obtained by performing deep learning iterative training on the basis of standard sample data of a product to be detected by a machine learning model.
Preferably, the appearance detection controller gathers and collates the image acquisition requirement information of each detection item based on the detection requirement information of the product to be detected to generate a detection scheme, and specifically comprises the following steps:
the appearance detection controller classifies each detection item based on the image acquisition target to obtain a rear shell surface detection set and an inner cavity surface detection set;
determining a placement driving angle of a product when each detection item acquires an image based on an image acquisition angle of each detection item in an inner cavity surface detection set, wherein the placement driving angle R= (R1, R2), R1 is a rotation angle of a swinging driving piece, R2 is a rotation angle of an electric rotating platform, and the value range of R1 and R2 is 0-360 degrees;
based on detection requirement information of products to be detected and the placement driving angles of the products when the detection items in the inner cavity surface detection set collect images, the placement driving angles of the detection items are summarized and ordered through a preset scheme generation model, a detection scheme is generated, and the scheme generation model is obtained through historical data training of a machine learning model.
Preferably, the method for ordering the placement driving angles of the detection items by using the scheme generation model specifically includes: the scheme generation model orders the images required by all detection items from small to large based on the rotation angle R1 of the swing driving piece, and orders the images required by all detection items from small to large based on the rotation angle of the electric rotating platform for the images required by the detection items with the same rotation angle R1 of the swing driving piece to generate an image acquisition flow of the inner cavity surface of the product; the scheme generation model also marks the placement driving angle with the repetition number being more than 2 as a key image acquisition angle, and when the multi-axis transplanting platform rotates to place products based on the key image acquisition angle, the inner cavity image acquisition assembly carries out time-delay shooting on the products so as to improve focusing effect.
In summary, the present application includes at least one of the following beneficial technical effects:
1. the appearance detection controller is used for realizing overall planning of various detection items of the product, carding detection circulation, orderly multi-angle image acquisition of the rear shell surface and the inner cavity surface of the product, synchronous and collaborative implementation of a plurality of detection items, and simplification of an integrated detection flow, saving of manpower and material resources and achievement of the effects of effectively improving detection efficiency and detection accuracy on the basis of meeting the requirements of various detection items of the product;
2. through the arrangement of the multi-axis transplanting platform, the appearance detection controller controls the multi-axis transplanting platform to rotate around the four axes to sequentially swing out the products according to the image acquisition requirements of all the inner cavity surface detection items of the products, so that the inner cavity surface detection module can conveniently and efficiently acquire images required by the detection of all the inner cavity surface detection items, a plurality of product detection items can be synchronously and cooperatively carried out, the integrated detection flow is simplified on the basis of meeting the requirements of a plurality of detection items of the products, manpower and material resources are saved, and the effects of effectively improving the detection efficiency and the detection precision are achieved;
3. the appearance detection controller determines the image acquisition requirements of each detection item based on the detection requirement information of the product, further gathers and sorts the image acquisition requirements of various acquisition angles of the inner cavity surface of the product, prefers a product placement time sequence, generates a detection scheme, further controls the appearance detection system to perform various back shell surface detection and inner cavity surface detection on the product to be detected, realizes the integration simplification of the product redundancy detection flow, realizes the synchronous and collaborative execution of a plurality of detection items, simplifies the integrated detection flow on the basis of meeting the requirements of a plurality of detection items of the product, saves manpower and material resources, and achieves the effect of effectively improving the detection efficiency and the detection precision;
4. The appearance detection controller uses the product placement angle as an anchor point, gathers and sorts the image acquisition angles of all detection items, breaks up the image acquisition requirement of all detection items by taking a single image acquisition as a unit, optimizes the product placement time sequence, is favorable for efficient cooperation of the CCD module and the multi-axis transplanting platform, realizes efficient and accurate image acquisition by accurately placing the products, and further improves the product appearance detection efficiency.
Drawings
FIG. 1 is a schematic diagram of an AI-based appearance detection system in accordance with an embodiment of the application;
FIG. 2 is a system block diagram of an AI-based appearance detection system in accordance with an embodiment of the application;
FIG. 3 is a schematic diagram of the overall structure of an AI-based appearance detection system in accordance with an embodiment of the present application;
FIG. 4 is a schematic diagram of the structure of a loading transfer module and a rear detection module according to an embodiment of the present application;
FIG. 5 is a schematic diagram of a product turn-over module according to an embodiment of the present application;
fig. 6 is a schematic structural diagram of a blanking transplanting assembly according to an embodiment of the present application;
FIG. 7 is a schematic diagram of a positioning carrier according to an embodiment of the application;
FIG. 8 is a schematic bottom view of a positioning carrier according to an embodiment of the application;
FIG. 9 is a schematic structural diagram of a blanking transplanting module and an inner cavity surface detection module in an embodiment of the present application;
FIG. 10 is a schematic view of another angle structure of the blanking transplanting module and the inner cavity surface detection module according to the embodiment of the present application;
FIG. 11 is a flowchart of a method for appearance inspection of a product by an appearance inspection controller in accordance with an embodiment of the present application;
fig. 12 is a flowchart of a method for generating a detection scheme by summarizing and sorting image acquisition requirement information of each detection item in an embodiment of the present application.
Reference numerals illustrate: 1. a machine table; 11. a hood; 2. a loading transfer module; 21. a feeding manipulator; 22. a feeding transplanting assembly; 221. feeding linear guide rails; 222. a feeding and transplanting platform; 223. a feeding auxiliary guide rail; 224. a square frame; 225. a code scanner; 3. a rear face detection module; 31. the rear shell detects the mounting bracket; 311. a cross beam; 312. a column; 313. detecting a linear guide rail; 32. a rear shell image acquisition assembly; 321. a first mounting plate; 322. a high definition camera; 323. a coaxial light source; 324. an annular light source; 325. a 3D camera; 4. a product turning module; 41. a product overturning assembly; 411. turning over the mounting frame; 412. a flipping member; 413. a support column; 414. a second mounting plate; 415. turning over the lifting driving piece; 416. turning the transverse plate; 417. turning over the material taking plate; 418. turning over the connecting plate; 419. a turnover driving motor; 42. a material taking assembly; 421. a material taking linear guide rail; 422. a material taking lifting driving piece; 423. a material taking transverse plate; 424. lifting the material taking plate; 425. a vacuum chuck; 426. a relief hole; 427. a relief groove; 5. a blanking transplanting module; 51. a blanking manipulator; 52. a blanking transplanting assembly; 53. a multi-axis transplanting platform; 531. a load-bearing bottom plate; 532. a bearing seat; 533. connecting the vertical plates; 534. a swing driving member; 535. an electric rotating platform; 54. a Y-axis linear guide rail; 55. y-axis auxiliary guide rail; 6. an inner cavity surface detection module; 61. an inner cavity image acquisition component; 611. a third mounting plate; 612. a CCD module; 613. a Z-axis linear guide rail; 614. a Z-axis auxiliary guide rail; 615. a fourth mounting plate; 62. an inner cavity detection mounting frame; 621. mounting a cross beam; 622. installing an upright post; 623. an X-axis linear guide rail; 7. an appearance detection controller; 8. positioning a carrier; 81. a code scanning hole site; 82. a bottom plate; 83. supporting a middle plate; 84. a positioning block; 85. positioning columns; 86. and (5) a vacuum suction nozzle.
Detailed Description
The application is described in further detail below with reference to fig. 1-12.
The embodiment of the application discloses an appearance detection system based on AI. Referring to fig. 1-3, an AI-based appearance detection system includes a machine 1, a loading transfer module 2, a rear shell surface detection module 3, a product turnover module 4, a unloading transplanting module 5, and an inner cavity surface detection module 6. The machine table 1 is covered with a protective hood 11. The feeding transfer module 2, the rear shell surface detection module 3, the product turnover module 4, the discharging transplanting module 5 and the inner cavity surface detection module 6 are connected with an appearance detection controller 7. The appearance detection controller 7 generates a detection scheme to control the feeding and transferring module 2, the rear shell surface detection module 3, the product turnover module 4, the discharging and transplanting module 5 and the inner cavity surface detection module 6 to operate based on detection requirement information of a product to be detected, automatically feeds and conveys the product to be detected, acquires rear shell surface image information of the product and inner cavity image information of the product, and performs appearance detection on the product. The appearance detection controller 7 is used for carrying out overall planning on various detection items of the product, carding, detecting and circulating, carrying out orderly multi-angle image acquisition on the rear shell surface and the inner cavity surface of the product, realizing synchronous and collaborative execution of a plurality of detection items, simplifying integrated detection flow, saving manpower and material resources and achieving the effect of effectively improving detection efficiency and detection precision on the basis of meeting the requirements of various detection items of the product.
Referring to fig. 2-4, the loading transfer module 2 includes a loading manipulator 21 and a loading transplanting assembly 22, the loading manipulator 21 is used for carrying a product to be detected onto the loading transplanting assembly 22, and the product to be detected is in a rear shell surface facing upwards. The feeding and transplanting assembly 22 comprises a feeding linear guide rail 221 and a feeding and transplanting platform 222, and the feeding and transplanting platform 222 is installed on a sliding block of the feeding linear guide rail 221. The two sides of the feeding linear guide rail 221 are provided with feeding auxiliary guide rails 223 in parallel, and the feeding transplanting platform 222 is arranged on the two feeding auxiliary guide rails 223 in a sliding manner. The loading and transplanting platform 222 comprises a square frame 224 and a plurality of positioning carriers 8 which are arranged at the top of the square frame 224 and used for positioning and bearing products to be detected. The square frame 224 is provided with a plurality of code scanner 225 for scanning and identifying the product to be detected, the code scanner 225 corresponds to the positioning carriers 8 one by one, and the positioning carriers 8 are provided with code scanning holes 81 for scanning the code scanner 225. The feeding manipulator 21 sucks the product to be detected and places the product on the positioning carrier 8 of the feeding and transplanting platform 222, and the code scanner 225 synchronously scans codes through the code scanning hole sites 81 to acquire product model information.
Referring to fig. 2-4, the rear face detection module 3 includes a rear face detection mount 31 straddling the upper side of the feed transplanting assembly 22 and a rear face image acquisition assembly 32 mounted on the rear face detection mount 31. The feeding transplanting assembly 22 is used for transplanting the product to be detected to the rear shell face image acquisition assembly 32, and the rear shell face image acquisition assembly 32 is used for acquiring images of the rear shell face of the product to be detected to obtain product rear shell face image information, wherein the product rear shell face image information comprises rear shell face plane image information and product rear shell 3D image information. The rear shell detection mounting frame 31 comprises a cross beam 311 and two upright posts 312, and the tops of the two upright posts 312 are connected with the cross beam 311. The beam 311 is provided with a detection linear guide 313 along its length direction, and the rear housing image acquisition assembly 32 is mounted on a slide table of the detection linear guide 313. The rear-shell image acquisition assembly 32 includes a first mounting plate 321, and a high-definition camera 322, a coaxial light source 323, and an annular light source 324 mounted on the first mounting plate 321 in this order from top to bottom. The first mounting plate 321 is fixedly connected with the sliding table of the detection linear guide 313, and a 3D camera 325 is arranged on one side of the first mounting plate 321. After the feeding transplanting platform 222 is stably conveyed to the rear shell surface image acquisition component 32 by the feeding linear guide rail 221 and the two feeding auxiliary guide rails 223, the detection linear guide rail 313 drives the high-definition camera 322 and the 3D camera 325 on the first mounting plate 321 to move, and rear shell surface plane image information and product rear shell 3D image information of each product to be detected are sequentially acquired, so that appearance detection is conveniently carried out on the rear shell surface of the product.
Referring to fig. 5, the product turn-over module 4 includes a product turn-over assembly 41 and a take-out assembly 42. The product overturning assembly 41 is used for sucking and overturning the product to be detected on the feeding and transplanting assembly 22, and the material taking assembly 42 is used for taking out the overturned product. The product turnover assembly 41 comprises a turnover installation frame 411 and a turnover part 412, wherein the turnover installation frame 411 comprises two support columns 413 and a second installation plate 414 arranged at the top of the two support columns 413, the turnover part 412 comprises a turnover lifting driving piece 415, a turnover transverse plate 416 and a turnover material taking plate 417, the turnover lifting driving piece 415 is installed on the second installation plate 414, and the telescopic end of the turnover lifting driving piece is fixedly connected with the turnover transverse plate 416; the two ends of the bottom of the turnover transverse plate 416 are provided with turnover connecting plates 418, two ends of the turnover material taking plate 417 are respectively connected with the two turnover connecting plates in a rotating mode, and the turnover connecting plates 418 are provided with turnover driving motors 419 used for driving the turnover material taking plate 417 to rotate.
Referring to fig. 5, the material taking assembly 42 includes a material taking linear guide 421, a material taking lifting driving member 422, a material taking transverse plate 423 and a lifting material taking plate 424, the material taking linear guide 421 is installed at the bottom of the second installation plate 414, the material taking transverse plate 423 is installed on a sliding table of the material taking linear guide 421, and the material taking lifting driving member 422 is installed on the material taking transverse plate 423 and the telescopic end penetrates through the material taking transverse plate 423 to be fixedly connected with the lifting material taking plate 424. The overturning reclaiming plate 417 and the lifting reclaiming plate 424 are respectively provided with a plurality of vacuum suckers 425 for sucking products, the second mounting plate 414 is provided with a yielding hole 426 for the reclaiming lifting driving member 422 to move, the overturning transverse plate 416 is provided with a yielding groove 427 for accommodating the telescopic end of the reclaiming lifting driving member 422, and the lifting reclaiming plate 424 moves to the upper part of the overturning reclaiming plate 417 under the driving of the reclaiming linear guide rail 421 to suck the overturned products. After the rear shell image acquisition assembly 32 completes image acquisition, the feeding linear guide rail 221 continuously conveys the feeding transplanting platform 222 to the lower part of the overturning and taking plate 417, the overturning lifting driving piece 415 drives the overturning transverse plate 416 to drive the overturning and taking plate 417 to descend, products are sucked through the vacuum chuck 425, and the overturning driving motor 419 drives the overturning and taking plate 417 to overturn with the products. The material taking linear guide rail 421 drives the material taking transverse plate 423 to move to the position above the overturning material taking plate 417, and the material taking lifting driving piece 422 on the material taking transverse plate 423 drives the lifting material taking plate 424 to descend so as to suck the turned product on the overturning material taking plate 417. The material taking linear guide rail 421 drives the material taking transverse plate 423 to move to the position of the blanking transplanting module 5 again, and the sucked product with the turning operation completed is placed at the position of the blanking transplanting module 5. The overturning lifting driving member 415 and the material taking lifting driving member 422 can be one or a combination of telescopic cylinders, telescopic electric cylinders or other devices capable of being linearly telescopic driven.
Referring to fig. 2 and 6, the discharging and transplanting module 5 includes a discharging manipulator 51 and a discharging and transplanting assembly 52, the discharging manipulator is used for taking down the product with appearance detection from the discharging and transplanting assembly 52, and placing the product in an OK tray and an NG tray based on the detection result. The blanking transplanting assembly 52 comprises a multi-shaft transplanting platform 53 and a Y-shaft linear guide rail 54, and the multi-shaft transplanting platform 53 is arranged on a sliding table of the Y-shaft linear guide rail 54. The two sides of the Y-axis linear guide rail 54 are provided with Y-axis auxiliary guide rails 55, and the multi-axis transplanting platform 53 is slidably arranged on the two Y-axis auxiliary guide rails 55. The material taking assembly 42 places the overturned product on the multi-axis transplanting platform 53, and the multi-axis transplanting platform 53 conveys the overturned product to the inner cavity surface detection module 6. The multi-axis transplanting platform 53 comprises a bearing bottom plate 531 and a bearing seat 532, wherein connecting risers 533 are arranged at two ends of the width direction of the bearing bottom plate 531, two ends of the bearing seat 532 are respectively and rotatably connected with the two connecting risers 533, and a swinging driving piece 534 for driving the bearing seat 532 to rotate is arranged on the connecting risers 533. The wobble drive 534 may be a stepper motor. A plurality of electric rotating platforms 535 are embedded at the top of the bearing seat 532, and the output ends of the electric rotating platforms 535 are connected with a positioning carrier 8 for bearing products. Through the setting of multiaxis transplanting platform 53, outward appearance detection controller 7 is based on the image acquisition demand of each inner chamber face detection item of product, control multiaxis transplanting platform 53 four-axis rotation and put out the gesture that satisfies each detection image acquisition demand with the product in proper order, the inner chamber face detection module 6 of being convenient for is high-efficient convenient to gather each inner chamber face detection item and detects required image, help going on in step with a plurality of product detection items in coordination, satisfy on the basis of the various detection item demands of product, simplify integrated detection flow, practice thrift manpower and materials, reach the effect that effectively improves detection efficiency and detection precision.
Referring to fig. 7 and 8, the positioning carrier 8 includes a bottom plate 82 and a supporting middle plate 83, the supporting middle plate 83 is fixedly disposed on the bottom plate 82, a positioning block 84 for clamping a product is disposed on the bottom plate 82, a positioning column 85 for positioning the product is disposed on the supporting middle plate 83, and a plurality of vacuum suction nozzles 86 for sucking the fixed product are further disposed on the supporting middle plate 83. Through the arrangement of the positioning block 84, the positioning column 85 and the vacuum suction nozzle 86, the product can be stably positioned and fixed. It should be noted that the structures and positions of the positioning blocks 84 and the positioning columns 85 can be customized according to the actual structure of the product to be detected and the reserved hole sites.
Referring to fig. 9 and 10, the inner cavity surface detection module 6 includes an inner cavity image acquisition assembly 61 and an inner cavity detection mounting frame 62 crossing over the Y-axis linear guide 54, the Y-axis linear guide 54 conveys the multi-axis transplanting platform 53 to the inner cavity image acquisition assembly 61, and the inner cavity image acquisition assembly 61 acquires image information of products of various placement angles displayed by the multi-axis transplanting platform 53. The inner cavity detection mounting frame 62 comprises a mounting cross beam 621 and two mounting upright posts 622, wherein the tops of the two mounting upright posts 622 are connected with the mounting cross beam 621, and an X-axis linear guide rail 623 is arranged on the mounting cross beam 621. The inner cavity image acquisition assembly 61 comprises a third mounting plate 611 and a plurality of groups of CCD modules 612, and the third mounting plate 611 is mounted on a sliding table of the X-axis linear guide rail 623. The third mounting plate 611 is provided with a Z-axis linear guide 613 along the vertical direction, and both sides of the Z-axis linear guide 613 are provided with a Z-axis auxiliary guide 614 along the same direction. The sliding tables of the Z-axis linear guide 613 and the two Z-axis auxiliary guide 614 are connected with a fourth mounting plate 615, and a plurality of groups of CCD modules 612 are sequentially arranged on the fourth mounting plate 615 along the horizontal direction. Through the setting of X axis linear guide 623 and Z axis linear guide 613, the relative position of multiunit CCD module 612 and product is adjusted steadily, ensures that CCD module 612 can be stable carry out clear, stable, the efficient image acquisition to the product that multiaxis transplanting platform 53 put.
Referring to fig. 11, the appearance detection controller 7 performs appearance detection of a product specifically including the steps of:
a1, acquiring detection requirement information of a product to be detected: the appearance detection controller 7 acquires detection requirement information of a product to be detected, wherein the detection requirement information comprises name information, detection standard information and image acquisition requirement information of a plurality of detection items, and the image acquisition requirement information of the detection items comprises a required image type, an image number, an image acquisition target and an image acquisition angle;
a2, generating a detection scheme: the appearance detection controller 7 gathers and sorts the image acquisition requirement information of each detection item based on the detection requirement information of the product to be detected to generate a detection scheme, wherein the detection scheme comprises a detection control instruction and image acquisition flow information;
a3, image acquisition is carried out on the rear shell surface and the inner cavity surface of the product: the appearance detection controller 7 sends detection control instructions to the positions of the feeding transfer module 2, the rear shell surface detection module 3, the product turnover module 4, the blanking transplanting module 5 and the inner cavity surface detection module 6, so that the rear shell surface and the inner cavity surface of a product to be detected are automatically fed and discharged according to image acquisition flow information to acquire image information of the rear shell surface and the inner cavity surface of the product, and image information of the rear shell surface and image information of the inner cavity of the product are obtained;
A4, carrying out appearance detection on each detection item: the appearance detection controller 7 performs appearance detection on each detection item through a preset AI visual detection model based on the image information of the rear shell surface of the product and the image information of the inner cavity of the product, and outputs an appearance detection result; the AI visual detection model is obtained by performing deep learning iterative training on the basis of standard sample data of a product to be detected by a machine learning model. The appearance detection controller 7 determines the image acquisition requirement of each detection item based on the detection requirement information of the product, and then gathers the arrangement to the image acquisition requirement of various acquisition angles of the inner cavity surface of the product, preferably the product placement time sequence, generates a detection scheme, and then controls the appearance detection system to carry out various back shell surface detection and inner cavity surface detection on the product to be detected, realizes the integration simplification of the product redundancy detection flow, realizes synchronous collaborative progress of a plurality of detection items, simplifies the integrated detection flow on the basis of meeting the requirement of a plurality of detection items of the product, saves manpower and material resources, and achieves the effect of effectively improving the detection efficiency and the detection precision.
Referring to fig. 12, the appearance detection controller 7 performs summary arrangement on the image acquisition requirement information of each detection item based on the detection requirement information of the product to be detected to generate a detection scheme specifically including the steps of:
B1, classifying each detection item: the appearance detection controller 7 classifies each detection item based on the image acquisition target to obtain a rear shell surface detection set and an inner cavity surface detection set;
b2, determining the placement driving angle of the product when each detection item collects images: determining a placement driving angle of a product when each detection item acquires an image based on an image acquisition angle of each detection item in the inner cavity surface detection set, wherein the placement driving angle R= (R1, R2), R1 is a rotation angle of a swinging driving piece 534, R2 is a rotation angle of an electric rotating platform 535, and the value range of R1 and R2 is 0-360 degrees;
b3, summarizing and sequencing the placement driving angles of all the detection items and generating a detection scheme: based on detection requirement information of products to be detected and the placement driving angles of the products when the detection items in the inner cavity surface detection set collect images, the placement driving angles of the detection items are summarized and ordered through a preset scheme generation model, a detection scheme is generated, and the scheme generation model is obtained through historical data training for a machine learning model. Because the inner cavity surface parts of the product are numerous, the derived detection items are more, the image acquisition requirements of all the detection items in the inner cavity surface detection set are faced, the appearance detection controller 7 gathers and sorts the image acquisition angles of all the detection items by taking the product placement angle as an anchor point (image acquisition angle), the image acquisition requirements of all the detection items are scattered by taking a single image acquisition as a unit, the product placement time sequence is optimized, the efficient cooperation of the CCD module 612 and the multi-axis transplanting platform 53 is facilitated, the efficient and accurate image acquisition is realized by accurately placing the product, and the product appearance detection efficiency is further improved.
In addition, the scheme generating model for sequencing the placement driving angles of the detection items specifically comprises the following steps: the scheme generation model orders the images required by all detection items from small to large based on the rotation angle R1 of the swinging driving piece 534, and orders the images required by all detection items with the same rotation angle R1 of the swinging driving piece 534 based on the rotation angle of the electric rotating platform 535 to generate an image acquisition flow of the inner cavity surface of the product; the scheme generating model also marks the placement driving angle with the repetition number being more than 2 as a key image acquisition angle, and when the multi-axis transplanting platform 53 rotates to place a product based on the key image acquisition angle, the inner cavity image acquisition assembly 61 performs time-delay shooting on the product to improve focusing effect. The scheme generating model is used for sorting based on the arrangement driving angles of all detection items, optimizing the arrangement time sequence of products, generating an image acquisition flow of the inner cavity surface of the products, marking the arrangement driving angles with high coincidence rate, carrying out time-delay shooting, improving the image acquisition quality and further improving the appearance detection precision.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the scope of the present invention. It will be apparent that the described embodiments are merely some, but not all, embodiments of the invention. Based on these embodiments, all other embodiments that may be obtained by one of ordinary skill in the art without inventive effort are within the scope of the invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art may still combine, add or delete features of the embodiments of the present invention or make other adjustments according to circumstances without any conflict, so as to obtain different technical solutions without substantially departing from the spirit of the present invention, which also falls within the scope of the present invention.
Claims (10)
1. An AI-based appearance detection system, comprising:
a machine table (1);
the feeding transfer module (2) is arranged on the machine table (1) and comprises a feeding manipulator (21) and a feeding transplanting assembly (22), wherein the feeding manipulator (21) is used for conveying a product to be detected to the feeding transplanting assembly (22), and the product to be detected is in a rear shell surface;
the rear shell surface detection module (3) is arranged on the machine table (1) and comprises a rear shell surface detection installation frame (31) arranged above the feeding transplanting assembly (22) in a crossing manner and a rear shell surface image acquisition assembly (32) arranged on the rear shell surface detection installation frame (31), the feeding transplanting assembly (22) is used for transplanting a product to be detected to the rear shell surface image acquisition assembly (32), the rear shell surface image acquisition assembly (32) is used for acquiring images of the rear shell surface of the product to be detected to obtain product rear shell surface image information, and the product rear shell surface image information comprises rear shell surface plane image information and product rear shell 3D image information;
the product turning module (4) is arranged on the machine table (1) and comprises a product turning assembly (41) and a material taking assembly (42), the product turning assembly (41) is used for sucking and turning over a product to be detected on the feeding transplanting assembly (22), and the material taking assembly (42) is used for taking out the turned product;
The blanking transplanting module (5) is arranged on the machine table (1) and comprises a blanking manipulator (51) and a blanking transplanting assembly (52), the blanking transplanting assembly (52) comprises a multi-shaft transplanting platform (53) and a Y-shaft linear guide rail (54), the multi-shaft transplanting platform (53) is arranged on a sliding table of the Y-shaft linear guide rail (54), and the material taking assembly (42) places a turned product on the multi-shaft transplanting platform (53);
the inner cavity surface detection module (6) is arranged on the machine table (1) and comprises an inner cavity image acquisition assembly (61) and an inner cavity detection mounting frame (62) which is arranged above the Y-axis linear guide rail (54) in a crossing mode, the Y-axis linear guide rail (54) conveys the multi-axis transplanting platform (53) to the inner cavity image acquisition assembly (61), and the inner cavity image acquisition assembly (61) acquires image information of products of various placing angles displayed by the multi-axis transplanting platform (53) through image acquisition; the method comprises the steps of,
the appearance detection controller (7) is arranged in the machine table (1), based on detection requirement information of a product to be detected, a detection scheme is generated to control the feeding and transferring module (2), the rear shell surface detection module (3), the product turn-over module (4), the blanking and transplanting module (5) and the inner cavity surface detection module (6) to operate, automatic feeding and discharging conveying is carried out on the product to be detected, product rear shell surface image information and product inner cavity image information are collected, and appearance detection is carried out on the product.
2. The AI-based appearance detection system of claim 1, wherein: the feeding transplanting assembly (22) comprises a feeding linear guide rail (221) and a feeding transplanting platform (222), the feeding transplanting platform (222) is arranged on a sliding block of the feeding linear guide rail (221), feeding auxiliary guide rails (223) are arranged on two sides of the feeding linear guide rail (221) in parallel, and the feeding transplanting platform (222) is arranged on the two feeding auxiliary guide rails (223) in a sliding mode; the feeding transplanting platform (222) comprises a square frame body (224) and a plurality of positioning carriers (8) which are arranged at the top of the square frame body (224) and used for positioning and bearing products to be detected, a plurality of code scanning devices (225) used for scanning the codes of the products to be detected are arranged on the square frame body (224), the code scanning devices (225) are in one-to-one correspondence with the positioning carriers (8), and code scanning hole sites (81) for scanning the codes of the code scanning devices (225) are formed in the positioning carriers (8).
3. The AI-based appearance detection system of claim 1, wherein: the rear shell detection mounting frame (31) comprises a cross beam (311) and two upright posts (312), the tops of the two upright posts (312) are connected with the cross beam (311), a detection linear guide rail (313) is arranged on the cross beam (311) along the length direction of the cross beam, and the rear shell image acquisition assembly (32) is arranged on a sliding table of the detection linear guide rail (313); the rear shell surface image acquisition assembly (32) comprises a first mounting plate (321), a high-definition camera (322), a coaxial light source (323) and an annular light source (324), wherein the high-definition camera (322), the coaxial light source (323) and the annular light source (324) are sequentially mounted on the first mounting plate (321) from top to bottom, the first mounting plate (321) is fixedly connected with a sliding table for detecting a linear guide rail (313), and a 3D camera (325) is arranged on one side of the first mounting plate (321).
4. The AI-based appearance detection system of claim 1, wherein: the product overturning assembly (41) comprises an overturning installation frame (411) and an overturning part (412), the overturning installation frame (411) comprises two supporting columns (413) and a second installation plate (414) arranged at the tops of the two supporting columns (413), the overturning part (412) comprises an overturning lifting driving piece (415), an overturning transverse plate (416) and an overturning material taking plate (417), the overturning lifting driving piece (415) is installed on the second installation plate (414), and the telescopic end of the overturning lifting driving piece is fixedly connected with the overturning transverse plate (416); two ends of the bottom of the turnover transverse plate (416) are provided with turnover connecting plates (418), two ends of the turnover material taking plate (417) are respectively and rotatably connected with the two turnover connecting plates (418), and the turnover connecting plates (418) are provided with turnover driving motors (419) for driving the turnover material taking plate (417) to rotate;
the material taking assembly (42) comprises a material taking linear guide rail (421), a material taking lifting driving piece (422), a material taking transverse plate (423) and a lifting material taking plate (424), the material taking linear guide rail (421) is arranged at the bottom of the second mounting plate (414), the material taking transverse plate (423) is arranged on a sliding table of the material taking linear guide rail (421), the material taking lifting driving piece (422) is arranged on the material taking transverse plate (423), and a telescopic end penetrates through the material taking transverse plate (423) to be fixedly connected with the lifting material taking plate (424);
The turnover material taking plate (417) and the lifting material taking plate (424) are respectively provided with a plurality of vacuum suckers (425) for absorbing products, the second mounting plate (414) is provided with a yielding hole (426) for the material taking lifting driving piece (422) to move, the turnover transverse plate (416) is provided with a yielding groove (427) for accommodating the telescopic end of the material taking lifting driving piece (422), and the lifting material taking plate (424) is driven by the material taking linear guide rail (421) to move to the upper side of the turnover material taking plate (417) to absorb the overturned products.
5. The AI-based appearance detection system of claim 1, wherein: two sides of the Y-axis linear guide rail (54) are provided with Y-axis auxiliary guide rails (55), and the multi-axis transplanting platform (53) is arranged on the two Y-axis auxiliary guide rails (55) in a sliding manner; the multi-axis transplanting platform (53) comprises a bearing bottom plate (531) and a bearing seat (532), connecting vertical plates (533) are arranged at two ends of the bearing bottom plate (531) in the width direction, two ends of the bearing seat (532) are respectively connected with the two connecting vertical plates (533) in a rotating mode, and a swinging driving piece (534) used for driving the bearing seat (532) to rotate is arranged on the connecting vertical plates (533); a plurality of electric rotating platforms (535) are embedded at the top of the bearing seat (532), and the output ends of the electric rotating platforms (535) are connected with positioning carriers (8) for bearing products.
6. The AI-based appearance detection system of claim 2 or 5, wherein: the positioning carrier (8) comprises a bottom plate (82) and a supporting middle plate (83), the supporting middle plate (83) is fixedly arranged on the bottom plate (82), a positioning block (84) for clamping a product is arranged on the bottom plate (82), a positioning column (85) for positioning the product is arranged on the supporting middle plate (83), and a plurality of vacuum suction nozzles (86) for sucking the fixed product are further arranged on the supporting middle plate (83).
7. The AI-based appearance detection system of claim 1, wherein: the inner cavity detection mounting frame (62) comprises a mounting cross beam (621) and two mounting upright posts (622), the tops of the two mounting upright posts (622) are connected with the mounting cross beam (621), and an X-axis linear guide rail (623) is arranged on the mounting cross beam (621); the inner cavity image acquisition assembly (61) comprises a third mounting plate (611) and a plurality of groups of CCD modules (612), the third mounting plate (611) is mounted on a sliding table of an X-axis linear guide rail (623), a Z-axis linear guide rail (613) is arranged on the third mounting plate (611) along the vertical direction, two sides of the Z-axis linear guide rail (613) are provided with Z-axis auxiliary guide rails (614) in the same direction, the Z-axis linear guide rails (613) and the sliding tables of the two Z-axis auxiliary guide rails (614) are jointly connected with a fourth mounting plate (615), and a plurality of groups of CCD modules (612) are sequentially arranged on the fourth mounting plate (615) along the horizontal direction.
8. The AI-based appearance detection system of claim 5, wherein the appearance detection controller (7) specifically comprises the steps of:
the appearance detection controller (7) acquires detection requirement information of a product to be detected, wherein the detection requirement information comprises name information, detection standard information and image acquisition requirement information of a plurality of detection items, and the image acquisition requirement information of the detection items comprises a required image type, an image number, an image acquisition target and an image acquisition angle;
the appearance detection controller (7) gathers and sorts the image acquisition demand information of each detection item based on the detection demand information of the product to be detected to generate a detection scheme, wherein the detection scheme comprises a detection control instruction and image acquisition flow information;
the appearance detection controller (7) sends detection control instructions to the positions of the feeding transfer module (2), the rear shell surface detection module (3), the product turning module (4), the blanking transplanting module (5) and the inner cavity surface detection module (6), so that the rear shell surface and the inner cavity surface of the product to be detected are automatically fed and discharged according to image acquisition flow information to acquire image information of the rear shell surface and the inner cavity surface of the product, and image information of the inner cavity surface of the product is obtained;
The appearance detection controller (7) carries out appearance detection on each detection item through a preset AI visual detection model based on the image information of the rear shell surface of the product and the image information of the inner cavity of the product, and outputs an appearance detection result; the AI visual detection model is obtained by performing deep learning iterative training on the basis of standard sample data of a product to be detected by a machine learning model.
9. The AI-based appearance inspection system according to claim 8, wherein the appearance inspection controller (7) performs a summary arrangement of image acquisition requirement information of each inspection item based on inspection requirement information of the product to be inspected to generate an inspection scheme, specifically comprising the steps of:
the appearance detection controller (7) classifies each detection item based on the image acquisition target to obtain a rear shell surface detection set and an inner cavity surface detection set;
determining a placement driving angle of a product when each detection item acquires an image based on an image acquisition angle of each detection item in an inner cavity surface detection set, wherein the placement driving angle R= (R1, R2), R1 is a rotation angle of a swinging driving piece (534), R2 is a rotation angle of an electric rotating platform (535), and the value range of R1 and R2 is 0-360 degrees;
based on detection requirement information of products to be detected and the placement driving angles of the products when the detection items in the inner cavity surface detection set collect images, the placement driving angles of the detection items are summarized and ordered through a preset scheme generation model, a detection scheme is generated, and the scheme generation model is obtained through historical data training of a machine learning model.
10. The AI-based appearance inspection system of claim 9, wherein the schema generation model orders the placement driving angles of the inspection items specifically comprises: the scheme generation model orders the images required by all detection items from small to large based on the rotation angle R1 of the swing driving piece (534), and orders the images required by all detection items from small to large based on the rotation angle of the electric rotation platform (535) for the images required by the detection items with the same rotation angle R1 of the swing driving piece (534) to generate an image acquisition flow of the inner cavity surface of the product; the scheme generating model is characterized in that the placement driving angle mark with the repetition frequency being more than 2 is a key image acquisition angle, and when the multi-axis transplanting platform (53) rotates to place products based on the key image acquisition angle, the inner cavity image acquisition assembly (61) carries out time-delay shooting on the products so as to improve focusing effect.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311333946.5A CN117092114B (en) | 2023-10-16 | 2023-10-16 | Appearance detection system based on AI |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311333946.5A CN117092114B (en) | 2023-10-16 | 2023-10-16 | Appearance detection system based on AI |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117092114A true CN117092114A (en) | 2023-11-21 |
| CN117092114B CN117092114B (en) | 2023-12-29 |
Family
ID=88775624
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311333946.5A Active CN117092114B (en) | 2023-10-16 | 2023-10-16 | Appearance detection system based on AI |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117092114B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117368200A (en) * | 2023-12-04 | 2024-01-09 | 荣旗工业科技(苏州)股份有限公司 | Product appearance detection equipment |
| CN118071743A (en) * | 2024-04-19 | 2024-05-24 | 苏州宣雄智能科技有限公司 | AI-based imaging lens defect detection method |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5096291A (en) * | 1990-05-16 | 1992-03-17 | Irvine Optical Corporation | Inspection systems having rotating motion |
| JPH07128256A (en) * | 1993-11-09 | 1995-05-19 | Murata Mach Ltd | Method for detecting untwilled thread of package |
| JP2007235490A (en) * | 2006-03-01 | 2007-09-13 | Shimatec:Kk | Automatic photographic apparatus |
| JP2010078355A (en) * | 2008-09-24 | 2010-04-08 | Sharp Corp | Inspection device |
| KR20100137065A (en) * | 2009-06-22 | 2010-12-30 | 오주환 | Surface inspection apparatus for electronic machine |
| CN104897053A (en) * | 2014-03-05 | 2015-09-09 | 泰科电子(上海)有限公司 | Programmable digital machine vision detection platform |
| CN105606627A (en) * | 2016-03-16 | 2016-05-25 | 武汉大学 | Remote appearance inspection and measurement method and system for nuclear power plant containment |
| CN110132166A (en) * | 2019-05-05 | 2019-08-16 | 广州佳帆计算机有限公司 | It is a kind of can automatic light distribution product image detection method and comparison device |
| WO2019185974A1 (en) * | 2018-03-28 | 2019-10-03 | Engemma Oy | Holder arrangement for gemological object recognition system |
| CN110596130A (en) * | 2018-05-25 | 2019-12-20 | 上海翌视信息技术有限公司 | Industrial detection device with auxiliary lighting |
| CN114719792A (en) * | 2022-03-25 | 2022-07-08 | 北京城建设计发展集团股份有限公司 | Intelligent scanning and automatic error identification system and method for prefabricated part assembling surface |
| CN115479895A (en) * | 2022-09-28 | 2022-12-16 | 苏州璐米智能科技有限公司 | Visual detection front sequencing and attitude control system |
| CN218331285U (en) * | 2022-07-22 | 2023-01-17 | 江苏润模汽车检测装备有限公司 | Rear cylinder appearance detection mechanism |
| US20230243758A1 (en) * | 2022-02-01 | 2023-08-03 | Gemological Institute Of America, Inc. (Gia) | Tilt and focusing adjustment for consistent gem imaging |
-
2023
- 2023-10-16 CN CN202311333946.5A patent/CN117092114B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5096291A (en) * | 1990-05-16 | 1992-03-17 | Irvine Optical Corporation | Inspection systems having rotating motion |
| JPH07128256A (en) * | 1993-11-09 | 1995-05-19 | Murata Mach Ltd | Method for detecting untwilled thread of package |
| JP2007235490A (en) * | 2006-03-01 | 2007-09-13 | Shimatec:Kk | Automatic photographic apparatus |
| JP2010078355A (en) * | 2008-09-24 | 2010-04-08 | Sharp Corp | Inspection device |
| KR20100137065A (en) * | 2009-06-22 | 2010-12-30 | 오주환 | Surface inspection apparatus for electronic machine |
| CN104897053A (en) * | 2014-03-05 | 2015-09-09 | 泰科电子(上海)有限公司 | Programmable digital machine vision detection platform |
| CN105606627A (en) * | 2016-03-16 | 2016-05-25 | 武汉大学 | Remote appearance inspection and measurement method and system for nuclear power plant containment |
| WO2019185974A1 (en) * | 2018-03-28 | 2019-10-03 | Engemma Oy | Holder arrangement for gemological object recognition system |
| CN110596130A (en) * | 2018-05-25 | 2019-12-20 | 上海翌视信息技术有限公司 | Industrial detection device with auxiliary lighting |
| CN110132166A (en) * | 2019-05-05 | 2019-08-16 | 广州佳帆计算机有限公司 | It is a kind of can automatic light distribution product image detection method and comparison device |
| US20230243758A1 (en) * | 2022-02-01 | 2023-08-03 | Gemological Institute Of America, Inc. (Gia) | Tilt and focusing adjustment for consistent gem imaging |
| CN114719792A (en) * | 2022-03-25 | 2022-07-08 | 北京城建设计发展集团股份有限公司 | Intelligent scanning and automatic error identification system and method for prefabricated part assembling surface |
| CN218331285U (en) * | 2022-07-22 | 2023-01-17 | 江苏润模汽车检测装备有限公司 | Rear cylinder appearance detection mechanism |
| CN115479895A (en) * | 2022-09-28 | 2022-12-16 | 苏州璐米智能科技有限公司 | Visual detection front sequencing and attitude control system |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117368200A (en) * | 2023-12-04 | 2024-01-09 | 荣旗工业科技(苏州)股份有限公司 | Product appearance detection equipment |
| CN117368200B (en) * | 2023-12-04 | 2024-04-19 | 荣旗工业科技(苏州)股份有限公司 | Product appearance detection equipment |
| CN118071743A (en) * | 2024-04-19 | 2024-05-24 | 苏州宣雄智能科技有限公司 | AI-based imaging lens defect detection method |
| CN118071743B (en) * | 2024-04-19 | 2024-06-25 | 苏州宣雄智能科技有限公司 | AI-based imaging lens defect detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117092114B (en) | 2023-12-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117092114B (en) | Appearance detection system based on AI | |
| CN107655897B (en) | Product appearance automatic checkout device | |
| CN107543827A (en) | Optical detection system | |
| CN216819897U (en) | Cell-phone center outward appearance detection mechanism | |
| CN107884702A (en) | A kind of FPC function automatic detections machine | |
| CN109702359A (en) | Multifunctional solar energy battery piece laser scribing device | |
| CN112548353A (en) | Double-station laser coding equipment and operation method thereof | |
| CN111010563A (en) | Multifunctional testing equipment for mobile phone camera module | |
| CN115646846B (en) | Detection equipment for 3C product | |
| CN112548350A (en) | Laser coding and code scanning equipment | |
| CN110842370A (en) | Full-automatic glass laser melting marking equipment and working method thereof | |
| CN112046157A (en) | Two-dimensional code marking system | |
| CN110949988B (en) | WIFI box marking device and method | |
| CN112548352A (en) | Laser coding equipment | |
| CN111495775A (en) | Optical device high accuracy just has equipment finished product detection machine of sorting mechanism | |
| CN217237794U (en) | Diversified outward appearance detection device | |
| CN108788474B (en) | Metal bottle cap laser marking device, preparation production line and preparation method | |
| CN214054071U (en) | Duplex position laser coding equipment | |
| CN213161952U (en) | AOI check out test set | |
| CN116008295B (en) | Copper-clad ceramic substrate detection equipment | |
| CN116300166A (en) | Multi-station detection mechanism and detection device for LCD screen | |
| CN116765615A (en) | Automatic coding equipment for battery | |
| CN215986793U (en) | Detection device | |
| CN217369288U (en) | Defect detection device for polyhedral material | |
| CN115682936A (en) | Turntable detection platform |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |