CN112235500B - Visual detection system and method based on deep learning - Google Patents
Visual detection system and method based on deep learning Download PDFInfo
- Publication number
- CN112235500B CN112235500B CN202011425136.9A CN202011425136A CN112235500B CN 112235500 B CN112235500 B CN 112235500B CN 202011425136 A CN202011425136 A CN 202011425136A CN 112235500 B CN112235500 B CN 112235500B
- Authority
- CN
- China
- Prior art keywords
- product
- image information
- camera
- deep learning
- fixing
- 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.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/50—Constructional details
-
- G—PHYSICS
- G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
- G03B—APPARATUS OR ARRANGEMENTS FOR TAKING PHOTOGRAPHS OR FOR PROJECTING OR VIEWING THEM; APPARATUS OR ARRANGEMENTS EMPLOYING ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ACCESSORIES THEREFOR
- G03B17/00—Details of cameras or camera bodies; Accessories therefor
- G03B17/56—Accessories
- G03B17/561—Support related camera accessories
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The application relates to the technical field of visual detection equipment, in particular to a visual detection system and a visual detection method based on deep learning, which comprises a rack, wherein a fixing module used for placing a product, an image acquisition module used for acquiring watchband image information, a control center and a display module used for displaying the watchband image information are arranged on the rack; the fixed module comprises a supporting seat and a product carrier arranged on the supporting seat, and a positioning piece used for fixing the watchband is arranged at the top end of the product carrier. When the product is required to be detected, the product is placed on the product carrier, the product is fixed through the positioning piece, then the image information of the product buckle is obtained through the image acquisition module, the obtained image information is sent to the control center, the control center is used for sending the received image information to the display module to be displayed, and whether the serial number of the product is qualified or not is judged by a worker according to the image information displayed on the display module, so that the detection precision is improved.
Description
Technical Field
The application relates to the technical field of visual detection equipment, in particular to a visual detection system and method based on deep learning.
Background
A wristwatch, also called a wristwatch, is an instrument worn on a wrist for timing or displaying time, and is usually made of leather, rubber, nylon cloth, stainless steel, etc. to form a watch band, and a watch head for displaying time is tied on the wrist through the watch band.
The related watchband 1 includes a first watchband 11 and a second watchband 12 as shown in fig. 1, wherein the first watchband 11 and the second watchband 12 are fixed together by a clasp, and watchbuckles 13 are disposed at opposite ends of the first watchband 11 and the second watchband 12. And each meter is buckled 13 and is all corresponded a serial number, and the consumer can be according to the serial number through mobile client, little letter public number or website to carry out the inquiry of traceing back information to the product, be convenient for know the specific production information of product.
In order to ensure that a consumer can know the production information of a product through the serial number on the watch buckle 13, the serial number on the watch buckle 13 needs to be manually detected and identified in the production process of the watchband 1, and whether the serial number on the watch buckle 13 is qualified or not is judged.
With respect to the related art among the above, the inventors consider that the following drawbacks exist: the serial number on the buckle 13 is difficult to detect by naked eyes, so that the detection precision is low.
Disclosure of Invention
In order to improve the detection precision of the serial number, the application provides a visual detection system and method based on deep learning.
In a first aspect, the present application provides a visual inspection system based on deep learning, which adopts the following technical scheme:
a visual detection system based on deep learning comprises a rack, wherein a fixing module used for placing a product, an image acquisition module used for acquiring watchband image information, a control center connected with the image acquisition module and a display module used for displaying the watchband image information are arranged on the rack;
the fixed module comprises a supporting seat and a product carrier arranged on the supporting seat, and a positioning piece used for fixing the watchband is arranged at the top end of the product carrier.
By adopting the technical scheme, when the product is required to be detected, the product is placed on the product carrier, the product is fixed through the positioning piece, then the image information of the product buckle is acquired through the image acquisition module, the acquired image information is sent to the control center, the control center is used for sending the received image information to the display module to be displayed, and the staff judges whether the serial number of the product is qualified or not according to the image information displayed on the display module, so that the detection precision is improved.
Optionally, the positioning element includes a limiting block and a positioning pin disposed at two ends of the product carrier, and a positioning portion for fixing the watch buckle is formed between the limiting block and the corresponding positioning pin.
Through adopting above-mentioned technical scheme, place the product back on the product carrier, the table at product both ends is detained and is located respectively between stopper and the locating pin that corresponds, is convenient for play preliminary fixed effect to the product, effectively reduces the probability that the product takes place the skew, and then is convenient for improve the precision that the product detected.
Optionally, the fixing module further includes a cover plate disposed on the product carrier and used for pressing the product, and the cover plate is provided with an accommodating cavity for accommodating the product;
the bottom of apron is provided with the locating hole, be provided with magnet in the locating hole.
Through adopting above-mentioned technical scheme, the setting of apron further plays fixed effect to the product.
Optionally, two sides of the supporting seat are respectively provided with a clamping assembly for fixing a product;
the clamping assembly comprises an upper pressure head arranged above the product carrier and a lower pressure head arranged below the product carrier, and a first transmission piece used for driving the upper pressure head to move up and down and a second transmission piece used for driving the lower pressure head to move up and down are arranged on the rack.
Through adopting above-mentioned technical scheme, after the supporting seat that has put the product removed to image acquisition module department, go up the pressure head and remove under the driving action of first driving medium, lower pressure head removes under the driving action of second driving medium for go up the pressure head and fix the product with lower pressure head, further reduce the probability that the product takes place the skew, thereby be convenient for improve the detection precision of product.
Optionally, the clamping assembly further comprises pressing heads arranged on two sides of the upper pressing head, each pressing head comprises a pressing block and a positioning block arranged at the bottom end of the corresponding pressing block, the positioning blocks are abutted to the product carriers, and the two pressing blocks are abutted to two ends of the corresponding meter buckle.
Through adopting above-mentioned technical scheme, go up the pressure head when the lapse, the drive compresses tightly the pressure head lapse to make compact heap offset with the both ends that the correspondence table was buckled, thereby be convenient for fix the product, and then be convenient for improve the detection precision of product.
Optionally, a driving assembly for driving the supporting seat to move is arranged on the rack;
the driving assembly comprises a sliding rail, a sliding block arranged in the sliding rail and a driving piece used for driving the sliding block to slide, a sliding seat is arranged on the sliding block, and the sliding seat is connected with the supporting seat.
Through adopting above-mentioned technical scheme, the supporting seat removes on the slide rail, and the getting of the product of being convenient for is put.
Optionally, the both ends of slide rail all are provided with the backup pad, two the top of backup pad is provided with the protecting cover, be provided with the spout that is used for supplying the protecting cover to pass on the slide.
Through adopting above-mentioned technical scheme, the setting of protecting cover is convenient for play the effect of protection when not influencing the supporting seat and remove.
Optionally, the image acquisition module comprises a first camera component and a second camera component which are arranged on two sides of the sliding rail;
the first camera shooting assembly comprises a first fixed seat, a first camera arranged on the first fixed seat and a first lens connected with the first camera, and the first camera is connected with the control center;
the second camera shooting assembly comprises a second fixed seat, a second camera arranged on the second fixed seat and a second lens connected with the second camera, the second camera is connected with the control center, and a transmission mechanism used for driving the second fixed seat to move is arranged on the rack.
By adopting the technical scheme, the image information of the surface buttons at the two ends of the product can be simultaneously acquired through the arrangement of the first camera shooting assembly and the second camera shooting assembly, so that the detection efficiency is improved;
in addition, when the product moves to the image acquisition module, the position of the second camera shooting assembly is adjusted, so that the angle and the distance between the camera lens and the product can be conveniently adjusted, and the optimal image acquisition effect is achieved.
Optionally, a housing is arranged on the frame, and an opening is arranged on the housing;
an infrared grating is arranged on the shell and positioned at the opening, and a light receiver and a light projector of the infrared grating are arranged in a correlation mode and are respectively arranged on two sides of the opening.
Through adopting above-mentioned technical scheme, when the supporting seat withdrawed to the opening, infrared correlation took place to block, and the photic ware can send the signal to control center, is convenient for remind the staff in time to handle.
In a second aspect, the application provides a visual inspection method based on deep learning, which adopts the following technical scheme:
a deep learning visual inspection method based on the deep learning visual inspection system, comprising:
fixing the product to be detected through a clamping assembly;
sequentially acquiring image information of surface buttons at two ends of a product through an image acquisition module, and displaying the acquired image information on a display module;
and judging whether the serial number deducted by the table is qualified or not according to the displayed image information.
Through adopting above-mentioned technical scheme, when needs examine the product, fix the product through clamping component, then acquire the image information that the product table was buckled through the image acquisition module to image information who will acquire sends control center, and control center is used for sending received image information to display module and shows, and whether the staff judges the serial number of product qualified according to the image information that shows on the display module, thereby is convenient for improve and detects the precision.
In summary, the present application includes at least one of the following beneficial technical effects:
1. the product is fixed through the fixing module, then the image information of the product buckle is obtained through the image acquisition module, the obtained image information is sent to the control center, the control center is used for sending the received image information to the display module to be displayed, and a worker judges whether the serial number of the product is qualified or not according to the image information displayed on the display module, so that the detection precision is improved conveniently;
2. through the setting of last pressure head and lower pressure head, after the supporting seat of putting the product removed to image acquisition module department, go up the pressure head and remove under the driving action of first driving medium, lower pressure head removes under the driving action of second driving medium for go up the pressure head and fix the product with lower pressure head, further reduce the probability that the product takes place the skew, thereby be convenient for improve the detection precision of product.
Drawings
Fig. 1 is a schematic view of a band of the related art.
Fig. 2 is a schematic structural diagram of an overall visual inspection system according to an embodiment of the present application.
Fig. 3 is a schematic structural diagram of a visual inspection system according to an embodiment of the present application.
Fig. 4 is a schematic structural diagram of a fixing module and a driving assembly according to an embodiment of the present disclosure.
Fig. 5 is a schematic structural diagram of a product carrier according to an embodiment of the present application.
Fig. 6 is a schematic structural diagram of a first imaging assembly according to an embodiment of the present application.
Fig. 7 is a schematic structural diagram of a second imaging assembly according to an embodiment of the present application.
Fig. 8 is an enlarged view of a portion a in fig. 7.
Fig. 9 is a block flow diagram illustrating a visual inspection method according to an embodiment of the present application.
Description of reference numerals: 1. a watchband; 11. a first band; 12. a second band of watchbands; 13. a watch buckle; 2. a frame; 21. a housing; 22. an opening; 3. fixing the module; 31. a supporting seat; 32. a product carrier; 33. a positioning member; 331. a limiting block; 332. positioning pins; 34. a cover plate; 35. an accommodating chamber; 36. a magnet; 4. a drive assembly; 41. a slide rail; 42. a slider; 43. a motor; 44. fixing the screw rod; 45. a slide base; 46. a support plate; 47. a protective cover; 48. a chute; 5. an image acquisition module; 51. a first camera assembly; 511. a first fixed seat; 512. a first camera; 513. a first lens; 52. a second camera assembly; 521. a second fixed seat; 522. a second camera; 523. a second lens; 524. a transmission mechanism; 525. a drive motor; 526. a transmission screw rod; 527. connecting blocks; 528. a sliding seat; 529. a sliding block; 6. a clamping assembly; 61. an upper pressure head; 611. a first cylinder; 612. a first fixing plate; 613. a connecting seat; 614. a first LVDT displacement sensor; 615. pressing a pressure head; 616. a compression block; 617. positioning blocks; 62. a lower pressure head; 621. a second cylinder; 622. a second fixing plate; 623. a second LVDT displacement sensor; 7. a display module; 71. a display screen; 8. an infrared grating.
Detailed Description
The present application is described in further detail below with reference to figures 1-9.
Referring to fig. 1, the watchband 1 includes a first watchband 11 and a second watchband 12, the first watchband 11 and the second watchband 12 are fixed together by a hasp, a watch buckle 13 is disposed at one end of the first watchband 11, which is opposite to the end of the second watchband 12, and a serial number is disposed on an end surface of the watch buckle 13, which is far away from the watchband 1.
The embodiment of the application discloses a visual inspection system based on deep learning. Referring to fig. 2 and 3, the device comprises a frame 2, wherein a housing 21 is arranged on the frame 2, and an opening 22 for taking and placing a product is arranged on the housing 21. The frame 2 is provided with a fixed module 3 for placing a product, an image acquisition module 5 for acquiring image information of the watch buckle 13, a control center connected with the image acquisition module 5 and a display module 7 for displaying the image information of the watch buckle 13, wherein the fixed module 3, the image acquisition module 5 and the display module 7 are arranged in the shell 21.
When the product needs to be detected, the product is fixed on the fixing module 3, the image information of the product buckle 13 is obtained through the image acquisition module 5, the obtained image information is sent to the control center, the control center is used for sending the received image information to the display module 7 to be displayed, and whether the serial number of the product is qualified or not is judged manually according to the image information displayed on the display module 7.
Referring to fig. 4 and 5, the fixing module 3 includes a supporting seat 31 and a product carrier 32 disposed on the supporting seat 31, and a plurality of stations for placing products are disposed on the product carrier 32 side by side, so that a plurality of products can be placed at the same time. The top end of the product carrier 32 and the position corresponding to the station are provided with positioning elements 33 for fixing the product, specifically, the positioning elements 33 include limiting blocks 331 and positioning pins 332 arranged at two ends of the product carrier 32, and positioning portions for fixing the surface fastener 13 are formed between the limiting blocks 331 and the corresponding positioning pins 332.
Referring to fig. 4, the fixing module 3 further includes a cover plate 34 for fixing a plurality of products, the cover plate 34 is provided with a containing cavity 35 for containing the products, the bottom end of the cover plate 34 is provided with positioning holes at four corners, and magnets 36 are disposed in the positioning holes. Place a plurality of products respectively in the station department that corresponds and the joint is behind the location portion that stopper 331 and locating pin 332 formed, and the pressfitting apron 34 on a plurality of products to make the product be located the corresponding chamber 35 that holds. Because the product carrier 32 is made of metal material, the cover plate 34 is fixed on the product carrier 32 under the adsorption action of the magnet 36, and the product is fixed.
Referring to fig. 3 and 4, a driving assembly 4 for moving the supporting seat 31 is disposed on the frame 2, specifically, the driving assembly 4 includes a sliding rail 41, a sliding block 42 slidably disposed in the sliding rail 41, and a driving member for driving the sliding block 42 to slide, wherein the driving member is a motor 43, an output shaft of the motor 43 is connected with a fixed screw 44, the fixed screw 44 is threadedly connected with a ball, and the ball is fixedly connected with the sliding block 42. The slide block 42 is provided with a slide seat 45, and the slide seat 45 is connected with the support seat 31. The both ends of slide rail 41 all are provided with backup pad 46, and the one end that motor 43 was kept away from to fixed screw 44 rotates with one of them backup pad 46 to be connected, and the top of two backup pad 46 is provided with protecting cover 47, is provided with the spout 48 that is used for supplying protecting cover 47 to pass on the slide 45.
Referring to fig. 3 and 4, when a product needs to be detected, the product is sequentially placed on the product carrier 32 at the opening 22, and the product is fixed by the cover plate 34, the motor 43 is started, the motor 43 drives the fixing screw 44 to rotate, and since the sliding block 42 is slidably disposed in the sliding rail 41, the sliding block 42 slides along the length direction of the fixing screw 44, and drives the sliding seat 45, the supporting seat 31 and the product carrier 32 to move, so that the product is conveniently moved to the image acquisition module 5.
Referring to fig. 3, the image capturing module 5 includes a first camera module 51 and a second camera module 52 disposed on two sides of the sliding rail 41, so as to obtain image information of a product at the same time.
Referring to fig. 6, the first camera module 51 includes a first fixing base 511, a first camera 512 disposed on the first fixing base 511, and a first lens 513 connected to the first camera 512, wherein the first camera 512 is connected to the control center.
Referring to fig. 7, the second camera module 52 includes a second fixing base 521, a second camera 522 disposed on the second fixing base 521, and a second lens 523 connected to the second camera 522, wherein the second camera 522 is connected to the control center. The frame 2 is provided with a transmission mechanism 524 for driving the second fixing seat 521 to move, specifically, the transmission mechanism 524 includes a transmission motor 525 and a transmission screw 526 connected with an output shaft of the transmission motor 525, the transmission screw 526 is connected with a transmission ball through a thread, the transmission ball is fixedly connected with a connection block 527, and the connection block 527 is connected with the second fixing seat 521. The frame 2 is provided with sliding seats 528 at two sides of the driving screw 526, and the bottom end of the second fixed seat 521 is connected with a sliding block 529 which is slidably arranged on the sliding seats 528.
Referring to fig. 3, when the product moves to the image capturing module 5, the position of the second camera module 52 is adjusted, so that the angle and distance between the lens of the camera and the product can be adjusted conveniently, and the best image capturing effect can be achieved.
Referring to fig. 6 and 7, clamping assemblies 6 for fixing the product are disposed on the side walls of the first fixing seat 511 and the second fixing seat 521, which are disposed opposite to each other, and when the second camera module 52 moves to a proper position, the product is clamped by the clamping assemblies 6.
Referring to fig. 6 and 7, the clamping assembly 6 includes an upper pressing head 61 disposed above the product carrier 32 (see fig. 5), a first transmission member for driving the upper pressing head 61 to move up and down, a lower pressing head 62 disposed below the product carrier 32, and a second transmission member for driving the lower pressing head 62 to move up and down, specifically, the first transmission member includes a first air cylinder 611 disposed on the fixing base, a piston rod of the first air cylinder 611 is connected with a first fixing plate 612, a connecting base 613 is disposed on the first fixing plate 612, and the upper pressing head 61 is disposed at a bottom end of the connecting base 613. A first LVDT displacement sensor 614 is disposed on the fixing seat and above the connecting seat 613, and the connecting seat 613 is connected to an iron core of the first LVDT displacement sensor 614.
Referring to fig. 6 and 7, the second transmission member includes a second cylinder 621 disposed on the fixing base, a piston rod of the second cylinder 621 is connected to a second fixing plate 622, the second fixing plate 622 is fixedly connected to the lower pressing head 62, the fixing base is provided with a second LVDT displacement sensor 623, the lower pressing head 62 is connected to an iron core of the second LVDT displacement sensor 623, and the upper pressing head 61 and the lower pressing head 62 are disposed oppositely.
When the second camera module 52 moves to a proper position, the first cylinder 611 and the second cylinder 621 are activated, the first cylinder 611 drives the connecting seat 613 to move downward, and the upper pressing head 61 moves downward; the second cylinder 621 drives the lower pressing head 62 to move upwards, and the product is clamped by the upper pressing head 61 and the lower pressing head 62. In addition, when the connecting seat 613 and the lower pressing head 62 move, the iron core in the LVDT displacement sensor is driven to move, the displacement of the iron core can be determined according to induced electromotive forces generated by two secondary coils in the LVDT displacement sensor, and the displacement of the iron core, that is, the moving distance between the upper pressing head 61 and the lower pressing head 62, is convenient for determining whether the upper pressing head 61 and the lower pressing head 62 clamp a product.
Referring to fig. 7 and 8, the clamping assembly 6 further includes a pressing ram 615 disposed on two sides of the upper ram 61, the pressing ram 615 includes a pressing block 616 and a positioning block 617 disposed at a bottom end of the pressing block 616, when the connecting seat 613 moves downward, the pressing ram 615 is driven to move downward, and the positioning block 617 abuts against the product carrier 32, and the two pressing blocks 616 abut against two ends of the corresponding meter fastener 13, respectively, so as to further fix the product.
Referring to fig. 6 and 7, after the product is fixed, the first camera module 51 and the second camera module 52 acquire image information of the corresponding product, and send the acquired image information to the control center, and the control center sends the received image information to the display module 7 (see fig. 1) for display.
Referring to fig. 2, the display module 7 includes a display screen 71 disposed on the housing 21, and the display screen 71 is connected to the control center. The display screen 71 displays image information sent by the control center, so that the staff can conveniently judge whether the serial number on the product is qualified.
Referring to fig. 1, when the product image information is displayed on the display screen 71, the product completion detection is described.
Referring to fig. 4 and 7, at this time, the first cylinder 611 and the second cylinder 621 are activated to drive the connecting seat 613 to move upwards, and the lower pressing head 62 moves downwards to release the fixing of the product; the support base 31 and the product carrier 32 are then moved forward by the driving member to facilitate the detection of the next product on the product carrier 32.
Referring to fig. 2 and 4, after all the products on the product carrier 32 are detected, the supporting seat 31 is retracted to the opening 22 under the driving action of the driving member, and simultaneously, the image information of all the products on the product carrier 32 is displayed on the display screen 71, so that the worker can conveniently and accurately judge whether the products are qualified.
Referring to fig. 2 and 4, an infrared grating 8 is disposed on the housing 21 and at the opening 22, the infrared grating 8 is connected to the control center, the light receiver and the light projector of the infrared grating 8 are arranged in opposite directions and are respectively disposed at two sides of the opening 22, and an alarm connected to the control center is disposed on the housing 21. When the supporting seat 31 is retracted to the opening 22, the infrared correlation is blocked, the light receiver sends a signal to the control center, and the control center controls the alarm to give an alarm, so that the staff is reminded to process in time.
The implementation principle of the visual inspection system based on deep learning in the embodiment of the application is as follows:
when the product needs to be detected, the product is sequentially placed on the product carrier 32 at the opening 22, the product is fixed through the cover plate 34, the motor 43 is started, the motor 43 drives the fixing screw rod 44 to rotate, and the sliding block 42 slides along the length direction of the fixing screw rod 44 and drives the sliding seat 45, the supporting seat 31 and the product carrier 32 to move due to the sliding of the sliding block 42 in the sliding rail 41, so that the product is conveniently moved to the image acquisition module 5.
When the product moves to the image capturing module 5, the position of the second camera module 52 is adjusted according to the size of the product carrier 32 and the watch band 1, and after the second camera module 52 moves to a proper position, the first cylinder 611 and the second cylinder 621 are started, the first cylinder 611 drives the connecting seat 613 to move downwards, and the upper pressing head 61 moves downwards; the second cylinder 621 drives the lower pressing head 62 to move upwards, and the product is clamped by the upper pressing head 61 and the lower pressing head 62.
After the product is fixed, the first camera component 51 and the second camera component 52 acquire the image information of the corresponding buckle 13, and send the acquired image information to the control center, and the control center sends the received image information to the display module 7 for display.
When all the products on the product carrier 32 are detected, the supporting seat 31 is driven by the driving member to retract to the opening 22, and at this time, the image information of all the products on the product carrier 32 is displayed on the display screen 71. When the supporting seat 31 returns to the opening 22, the infrared correlation is blocked, the light receiver sends a signal to the control center, the control center controls the alarm to give an alarm, the worker is reminded to process the signal in time, the worker judges whether the serial number of the corresponding product is clear or complete through the image information displayed on the display screen 71, and the detection precision is improved.
The embodiment of the application also discloses a visual detection method based on deep learning, and with reference to fig. 9, the method comprises the following steps:
s1, fixing the product to be detected through the clamping assembly 6;
s2, sequentially acquiring image information of surface buckles 13 at two ends of the product through the image acquisition module 5, and displaying the acquired image information on the display module 7;
and S3, judging whether the serial number on the watch dog 13 is qualified or not according to the displayed image information.
The above embodiments are preferred embodiments of the present application, and the protection scope of the present application is not limited by the above embodiments, so: all equivalent changes made according to the structure, shape and principle of the present application shall be covered by the protection scope of the present application.
Claims (8)
1. A visual inspection system based on deep learning is characterized in that: the device comprises a rack (2), wherein a fixing module (3) for placing a product, an image acquisition module (5) for acquiring image information of a watch buckle (13), a control center connected with the image acquisition module (5) and a display module (7) for displaying the image information of the watch buckle (13) are arranged on the rack (2);
the fixing module (3) comprises a supporting seat (31) and a product carrier (32) arranged on the supporting seat (31), and a positioning piece (33) used for fixing the watchband (1) is arranged at the top end of the product carrier (32);
the fixing module (3) further comprises a cover plate (34) which is arranged on the product carrier (32) and used for pressing the product, and an accommodating cavity (35) used for accommodating the product is arranged on the cover plate (34);
the bottom end of the cover plate (34) is provided with a positioning hole, and a magnet (36) is arranged in the positioning hole;
clamping components (6) for fixing products are respectively arranged on two sides of the supporting seat (31);
the clamping assembly (6) comprises an upper pressure head (61) arranged above the product carrier (32) and a lower pressure head (62) arranged below the product carrier (32), and a first transmission piece used for driving the upper pressure head (61) to move up and down and a second transmission piece used for driving the lower pressure head (62) to move up and down are arranged on the rack (2).
2. The deep learning based visual inspection system of claim 1, wherein: the positioning piece (33) comprises a limiting block (331) and a positioning pin (332) which are arranged at two ends of the product carrier (32), and a positioning part for fixing the watch buckle (13) is formed between the limiting block (331) and the corresponding positioning pin (332).
3. The deep learning based visual inspection system of claim 1, wherein: the clamping assembly (6) further comprises pressing heads (615) arranged on two sides of the upper pressing head (61), each pressing head (615) comprises a pressing block (616) and a positioning block (617) arranged at the bottom end of the pressing block (616), the positioning blocks (617) are abutted to the product carriers (32), and the two pressing blocks (616) are abutted to two ends of the corresponding meter buckles (13).
4. The deep learning based visual inspection system of claim 1, wherein: the rack (2) is provided with a driving component (4) for driving the supporting seat (31) to move;
the driving assembly (4) comprises a sliding rail (41), a sliding block (42) arranged in the sliding rail (41) and a driving piece used for driving the sliding block (42) to slide, a sliding seat (45) is arranged on the sliding block (42), and the sliding seat (45) is connected with the supporting seat (31).
5. The deep learning based visual inspection system of claim 4, wherein: the both ends of slide rail (41) all are provided with backup pad (46), two the top of backup pad (46) is provided with protecting cover (47), be provided with spout (48) that are used for supplying protecting cover (47) to pass on slide (45).
6. The deep learning based visual inspection system of claim 1, wherein: the image acquisition module (5) comprises a first camera component (51) and a second camera component (52) which are arranged on two sides of the sliding rail (41);
the first camera shooting assembly (51) comprises a first fixed seat (511), a first camera (512) arranged on the first fixed seat (511) and a first lens (513) connected with the first camera (512), and the first camera (512) is connected with a control center;
the second camera shooting assembly (52) comprises a second fixed seat (521), a second camera (522) arranged on the second fixed seat (521) and a second lens (523) connected with the second camera (522), the second camera (522) is connected with a control center, and a transmission mechanism (524) used for driving the second fixed seat (521) to move is arranged on the rack (2).
7. The deep learning based visual inspection system of claim 1, wherein: a shell (21) is arranged on the frame (2), and an opening (22) is formed in the shell (21);
the utility model discloses a control system, including casing (21) just lie in opening (22) and be provided with on the casing (21) and be connected with control center's infrared grating (8), the photic ware and the light projector of infrared grating (8) are the correlation setting and set up respectively in the both sides of opening (22).
8. A deep learning visual inspection method based on the deep learning based visual inspection system according to any one of claims 1 to 7, comprising:
fixing the product to be detected through a clamping assembly (6);
sequentially acquiring image information of surface buttons (13) at two ends of a product through an image acquisition module (5), and displaying the acquired image information on a display module (7);
whether the serial number on the watchband (13) is qualified is judged through the displayed image information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011425136.9A CN112235500B (en) | 2020-12-09 | 2020-12-09 | Visual detection system and method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011425136.9A CN112235500B (en) | 2020-12-09 | 2020-12-09 | Visual detection system and method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112235500A CN112235500A (en) | 2021-01-15 |
| CN112235500B true CN112235500B (en) | 2021-03-02 |
Family
ID=74124639
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011425136.9A Active CN112235500B (en) | 2020-12-09 | 2020-12-09 | Visual detection system and method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112235500B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107132230A (en) * | 2017-06-16 | 2017-09-05 | 苏州祥宝机械设备有限公司 | Pcb board glue surface covers automatic detection device and detection method |
| WO2018196734A1 (en) * | 2017-04-26 | 2018-11-01 | 宁波舜宇光电信息有限公司 | Test data processing device, test data processing method and test apparatus |
| CN111693530A (en) * | 2019-03-11 | 2020-09-22 | 深圳市联得自动化装备股份有限公司 | Detection device and method |
| CN212032163U (en) * | 2020-05-21 | 2020-11-27 | 捷普电子(广州)有限公司 | Image acquisition device |
-
2020
- 2020-12-09 CN CN202011425136.9A patent/CN112235500B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018196734A1 (en) * | 2017-04-26 | 2018-11-01 | 宁波舜宇光电信息有限公司 | Test data processing device, test data processing method and test apparatus |
| CN107132230A (en) * | 2017-06-16 | 2017-09-05 | 苏州祥宝机械设备有限公司 | Pcb board glue surface covers automatic detection device and detection method |
| CN111693530A (en) * | 2019-03-11 | 2020-09-22 | 深圳市联得自动化装备股份有限公司 | Detection device and method |
| CN212032163U (en) * | 2020-05-21 | 2020-11-27 | 捷普电子(广州)有限公司 | Image acquisition device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112235500A (en) | 2021-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106621284B (en) | A kind of intelligent physical examinations platform | |
| CN201575974U (en) | Circuit board image and laser detection equipment | |
| US9223151B2 (en) | Method for determining reading distance | |
| CN112235500B (en) | Visual detection system and method based on deep learning | |
| CN112595583B (en) | Device for detecting hardness of material | |
| CN208060142U (en) | A kind of testing tool of electronic product key | |
| CN111398034A (en) | Mask tension tester | |
| EP4151989A1 (en) | Welding bead inspection apparatus | |
| CN113960047A (en) | Chip packaging detection equipment and use method thereof | |
| CN217688591U (en) | Product scanning and detecting device | |
| CN107748428A (en) | Screen detects Atomatic focusing method and device | |
| CN207396368U (en) | A kind of detecting tool | |
| CN210221454U (en) | Lens whole-circumference appearance inspection device | |
| CN109907728B (en) | Vision testing device and method | |
| CN217408022U (en) | Watch buckle installed in sliding mode | |
| CN213748286U (en) | Measuring device adopting line laser | |
| CN212082287U (en) | Handheld electronic circumference measuring instrument | |
| CN215676901U (en) | Visual detection device | |
| CN214710901U (en) | Wearing equipment with monitoring function | |
| CN213643028U (en) | Swimming goggles | |
| CN207263647U (en) | Wrist strap screen detection device | |
| CN210181403U (en) | Watch with easily-detachable watchband | |
| CN114994416A (en) | Model machine placing device of SAR (synthetic aperture radar) testing system, use method of model machine placing device and SAR testing system | |
| CN214630454U (en) | Watch for health of old people | |
| CN219288911U (en) | Bracelet convenient to wear |
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 |