Carton printing defect visual identification device

Abstract

The invention provides a carton printing defect visual identification device, which relates to the technical field of image data processing, and comprises an outer box body, a conveying component and a visual detection component, an image acquisition module is arranged in the visual detection component, an encoder is arranged in the conveying positioning unit, and the encoder is linked with a driving mechanism of the conveying component and a trigger signal of the industrial linear array camera, so that the conveying speed of the conveying component, the conveying position of the carton and image acquisition are synchronized; a three-level image processing flow is executed through the algorithm processing module, interference such as corrugated paper textures is eliminated through self-adaptive median filtering, multiple defects such as overprinting deviation and through printing can be accurately recognized and graded in combination with multi-dimensional feature extraction and a support vector machine classification model, and the accuracy and the accuracy of detection are improved. The industrial pain points that in the prior art, the image processing flow is simple, the targeted interference filtering capacity is lacked, various defect types and grades cannot be accurately distinguished, and the overall intelligent level and collaboration are insufficient are solved.

Description

Technical Field

[0001] This invention relates to the field of image data processing technology, and in particular to a visual recognition device for printing defects in cardboard boxes. Background Technology

[0002] In the current era of rapid development in the packaging industry, cardboard boxes have become the preferred outer packaging carrier for many fields such as food, electronics, and daily chemicals due to their advantages of low cost, stable structure, and recyclability. The market demand for cardboard boxes continues to rise, and the requirements for their printing quality are becoming increasingly stringent. Cardboard box printing not only carries the functions of brand identification and product information transmission, but its printing accuracy and appearance integrity also directly affect the market acceptance of products. Therefore, printing defect detection has become a core link in the cardboard box production process to ensure product quality. With the deep penetration of intelligent manufacturing technology into the manufacturing industry, the cardboard box production industry is gradually transforming from the traditional manual-dominated mode to an automated and intelligent production mode. The operating efficiency and automation level of the production line have been greatly improved, and the corresponding automated detection technology has also become an important direction for industry development. This invention is a visual recognition device for cardboard box printing defects.

[0003] Among the existing technologies and equipment for detecting defects in carton printing, some automated detection equipment has a relatively simple image processing flow and lacks the ability to filter out specific interference. It cannot accurately distinguish between various defect types and levels, making it difficult to meet the high-efficiency and high-precision detection requirements of modern carton production lines. Summary of the Invention

[0004] This invention relates to a visual recognition device for printing defects in cardboard boxes, in order to solve the technical problems mentioned in the background art.

[0005] In a first aspect, this invention provides a visual recognition device for printing defects in cardboard boxes, specifically comprising: an outer box body, a conveying component, and a visual inspection component; the visual inspection component has a built-in image acquisition module, and is correspondingly installed above the inspection station of the conveying component; the image acquisition module includes an industrial line scan camera with adjustable focal length, a ring light source, and a conveying positioning unit; the conveying positioning unit has a built-in encoder, which is linked to the drive mechanism of the conveying component and the trigger signal of the industrial line scan camera, respectively, to synchronize the conveying speed of the conveying component, the conveying position of the cardboard box, and the image acquisition, for acquiring and outputting image data of the cardboard box printing area on the conveying component; the visual recognition device for printing defects in cardboard boxes further includes an algorithm processing module, an execution feedback module, and a human-machine interaction module; the execution feedback module consists of a PLC controller and an audible and visual alarm component; the algorithm processing module is integrated into the PLC controller, and the algorithm processing module is electrically connected to the image acquisition module; the PLC controller is respectively linked to... The algorithm processing module, the audible and visual alarm device, and the drive mechanism of the conveying component are electrically connected. The PLC controller receives the defect judgment signal output by the algorithm processing module, triggers the audible and visual alarm device to remind staff to troubleshoot the fault, and can control the conveying component to slow down or stop if multiple unqualified cartons appear consecutively to avoid the accumulation of unqualified products. The human-machine interaction module consists of a display screen, which is communicatively connected to the PLC controller. The PLC controller enables the retrieval of detection data, the issuance of operation commands, and the setting of parameters. The display screen is embedded in the surface of the outer casing and electrically connected to the PLC controller. The human-machine interaction module is used to realize user operation and data management, supporting the import of standard printed patterns, real-time detection data display, detection parameter setting, detection data storage, and production quality report generation. The conveying component is used to realize the continuous and stable conveying of cartons, and adopts a variable frequency speed regulation structure. The algorithm processing module receives the image data output by the image acquisition module and executes a three-level image processing flow.

[0006] Furthermore, the conveying component includes a partition, a front baffle, a rear baffle, supports, and a transmission rod. Four leveling feet are fixed to the lower end of the partition. The front baffle is fixedly mounted on the partition, and the rear baffle slides back and forth on the partition. Two supports are provided, both mounted on the partition, and a rotating rod is mounted between the two supports via a bearing. The rotating rod has a spindle-shaped cross-section and two pulleys are mounted on it. The pulley near the rear end of the rotating rod can slide back and forth on the rotating rod. A waist-shaped groove is provided at the right end of both the front and rear baffles, and movable seats are mounted on both the front and rear baffles. Pulleys are mounted on both movable seats. Both the front and rear baffles are equipped with fixed seats, and tensioning screws are threaded onto the fixed seats. The tensioning screws are pressed against the movable seats. A conveyor belt connects the pulley on the transmission rod to the pulley on the movable seat. A base frame is installed at the lower end of the partition, and a reducer is installed on the base frame. A drive motor is installed on the reducer, and pulleys are installed on the output shaft of the reducer and the front end of the rotating rod. A belt connects the pulley on the output shaft of the reducer to the pulley at the front end of the rotating rod. The base frame is provided with a waist-shaped groove and a dust cover is installed on the base frame. The transmission rod is mounted on the partition through bearings, and the transmission rod is threaded. A handwheel is installed at the front end of the transmission rod.

[0007] Furthermore, a mating plate is installed on the rear baffle, which is threadedly engaged with the transmission rod. An adjustable plate that can move back and forth is installed on the rear baffle by bolts. A telescopic cylinder is installed on the adjustable plate, and a rear clamping plate is installed on the piston rod of the telescopic cylinder. A front clamping plate is installed on the front baffle.

[0008] Furthermore, the visual inspection component includes a horizontal rail plate, a vertical rail plate, and a movable plate. A support frame is mounted on the horizontal rail plate, and the vertical rail plate slides back and forth on the horizontal rail plate. A first connecting seat and a second connecting seat are respectively mounted on the horizontal and vertical rail plates, and a drag chain is provided between the first and second connecting seats. A horizontal motor frame and a first lead screw are mounted on the horizontal rail plate, and a horizontal servo motor is mounted on the horizontal motor frame. The output shaft of the horizontal servo motor is connected to the first lead screw through a coupling, and the first lead screw is threadedly engaged with the vertical rail plate. The movable plate slides up and down on the vertical rail plate, and a second lead screw is mounted on the vertical rail plate through a bearing. The second lead screw is threadedly engaged with the movable plate. A vertical motor frame is mounted on the vertical rail plate, and a vertical servo motor is mounted on the vertical motor frame. The output shaft of the vertical servo motor is connected to the second lead screw through a coupling. U-shaped photoelectric sensors are mounted on both the horizontal and vertical rail plates, and sensing plates are mounted on both the vertical rail plate and the movable plate.

[0009] Furthermore, two motor mounts are installed on the movable plate, with a forward motor and a reverse motor mounted on each mount. Pulleys are mounted on the output shafts of both the forward and reverse motors. A mounting base is also installed on the movable plate, and a rotating shaft is mounted within the mounting base via bearings. Two pulleys are mounted on the rotating shaft, and connecting belts connect the two pulleys on the rotating shaft to the pulleys on the output shafts of the forward and reverse motors. A photoelectric sensor plate is mounted on the upper end of the rotating shaft, and a photoelectric switch is mounted on the movable plate. A mounting bracket is fixed to the lower end of the rotating shaft, and an industrial line scan camera and a ring light source are both mounted on the mounting bracket.

[0010] Furthermore, the three-level image processing flow is as follows: S1, Image preprocessing, sequentially performs grayscale and normalization processing, adaptive median filtering processing, and histogram equalization processing to eliminate the effects of corrugated paper texture interference, ink-wash printing color difference fluctuations, and uneven lighting on image quality. S2, Feature Extraction: Extracting texture features, geometric features, and edge features of the defect area. Texture features are extracted using the gray-level co-occurrence matrix, geometric features include the area, perimeter, and shape factor of the defect, and edge features are extracted using the Canny edge detection algorithm. S3, Defect Classification and Recognition: Construct a support vector machine classification model, input the extracted multi-dimensional features into the model, compare it with the preset normal sample feature library and defect sample feature library, and output the defect type, defect location and defect level; The algorithm processing module also has a built-in self-learning update module, which can automatically collect defect samples confirmed by manual re-inspection and optimize the classification model parameters.

[0011] Furthermore, the industrial line scan camera has a resolution of 2048 pixels, and the shooting frame rate can be adaptively adjusted according to the conveying speed of the conveying components. The color temperature adjustment range of the ring light source is 3000K-6500K, and the brightness is continuously adjustable. The supplementary lighting parameters can be adjusted according to the color depth and ink thickness of the printed pattern on the carton. The light source can switch between direct and diffused modes to adapt to the supplementary lighting needs of different carton materials.

[0012] Furthermore, in the adaptive median filtering process, the size of the filtering window can be adaptively adjusted according to the image texture complexity, with an adjustment range of 33-1111, so as to completely filter out the periodic texture interference on the corrugated paper surface while retaining the edge information of the printed pattern.

[0013] Furthermore, the defect types include misregistration, missing print, ink spots, scratches, incomplete text, and blurred barcodes. The defect levels are divided into minor defects, general defects, and severe defects, and the judgment criteria can be manually set through the human-computer interaction module.

[0014] Furthermore, the PLC controller supports interface with the factory's MES system, enabling real-time uploading of detection data to the MES system for digital traceability and management of production quality.

[0015] This invention provides a visual recognition device for printing defects in cardboard boxes, which has the following beneficial effects: The algorithm processing module of this invention executes a three-level image processing flow. It eliminates interference such as corrugated paper texture through adaptive median filtering. Combined with multi-dimensional feature extraction and support vector machine classification model, it can accurately identify and classify various defects such as registration deviation and missing printing. It solves the industry pain points of low efficiency, strong subjectivity and susceptibility to fatigue in manual inspection, which leads to missed or misjudged defects. It significantly improves the standardization level of carton printing inspection.

[0016] This invention utilizes the integrated industrial line scan camera, ring light source, and conveying positioning unit within the vision inspection component to work in tandem. The encoder, in conjunction with the conveying component drive mechanism and the industrial line scan camera trigger signal, ensures that the carton conveying speed, position, and image acquisition are synchronized, thus avoiding missed or false detections due to conveying deviation.

[0017] Furthermore, the execution feedback module in this invention consists of a PLC controller and an audible and visual alarm component. The PLC controller can receive the defect judgment signal from the algorithm processing module in real time and quickly trigger the audible and visual alarm to remind staff to troubleshoot the fault. At the same time, in the case of multiple consecutive defective cartons, it can automatically control the conveyor component to slow down or stop, effectively avoiding problems such as production line congestion and increased rework costs caused by the accumulation of defective products, realizing timely control of defective products, and improving the timeliness and effectiveness of production control.

[0018] Furthermore, the human-machine interface module of this invention consists of a display screen embedded in the surface of the outer casing. Through communication connection with the PLC controller, it can realize functions such as importing standard printed patterns, displaying real-time detection data, setting detection parameters, storing detection data, and generating production quality reports. Staff can intuitively grasp the detection status and conveniently adjust detection parameters. At the same time, the PLC controller supports docking with the factory's MES system, which can upload detection data to the MES system in real time, realize digital traceability and global control of production quality, and provide data support for production optimization and quality improvement. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments will be briefly described below.

[0020] The accompanying drawings described below are only related to some embodiments of the invention and are not intended to limit the invention.

[0021] In the attached diagram: Figure 1A schematic diagram of the overall structure of the present invention is shown.

[0022] Figure 2 A schematic diagram of the conveying component and vision inspection component inside the outer casing of the present invention is shown.

[0023] Figure 3 A schematic diagram of the conveying component portion of the present invention is shown.

[0024] Figure 4 A schematic diagram of the rear baffle portion of the present invention is shown.

[0025] Figure 5 The present invention is shown Figure 4 A magnified structural diagram of part A in the middle.

[0026] Figure 6 The present invention is shown Figure 4 A magnified structural diagram of part B.

[0027] Figure 7 A schematic diagram of the structure of the visual detection component of the present invention is shown.

[0028] Figure 8 The present invention is shown Figure 7 A magnified structural diagram of section C.

[0029] Figure 9 A block diagram of the electrical and control connections of the present invention is shown.

[0030] Figure 10 A flowchart of the three-level image processing and defect recognition process of the present invention is shown.

[0031] Figure 11 A timing diagram illustrating the working principle of the present invention is shown.

[0032] List of reference numerals 1. Outer casing; 11. Display screen; 12. PLC controller; 13. Audible and visual alarm components; 2. Conveying components; 21. Partition plate; 211. Leveling feet; 212. Base frame; 213. Reducer; 214. Drive motor; 215. Dust cover; 22. Front baffle; 221. Front clamping plate; 23. Rear baffle; 231. Movable seat; 232. Fixed seat; 233. Tensioning screw; 234. Conveyor belt; 235. Adjusting plate; 2351. Telescopic cylinder; 2352. Rear clamping plate; 236. Mating plate; 24. Support; 241. Rotating rod; 25. Transmission rod; 251. Handwheel 3. Vision inspection components; 31. Horizontal rail plate; 311. Support frame; 312. First connecting seat; 313. Cable chain; 314. Horizontal motor frame; 315. First lead screw; 316. Horizontal servo motor; 32. Vertical rail plate; 321. Second connecting seat; 322. Vertical motor frame; 323. Vertical servo motor; 324. Second lead screw; 33. Moving plate; 331. Motor base; 332. Forward rotation motor; 333. Reverse rotation motor; 334. Mounting base; 335. Rotating shaft; 336. Connecting belt; 337. Fixing frame; 338. Industrial line scan camera; 339. Ring light source. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the described embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Please refer to Figures 1 to 11 Example 1: This invention proposes a visual recognition device for cardboard box printing defects, comprising: an outer box body 1, a conveying component 2, and a visual inspection component 3; the visual inspection component 3 has a built-in image acquisition module and is installed above the inspection station of the conveying component 2. The image acquisition module includes an industrial line scan camera 338 with adjustable focal length, a ring light source 339, and a conveying positioning unit. The conveying positioning unit has a built-in encoder, which is linked with the drive mechanism of the conveying component 2 and the trigger signal of the industrial line scan camera 338 to synchronize the conveying speed, cardboard box conveying position, and image acquisition of the conveying component 2, and is used to acquire and output image data of the cardboard box printing area on the conveying component 2; the cardboard box printing defect visual recognition device also includes an algorithm processing module, an execution feedback module, and a human-machine interaction module; the execution feedback module consists of a PLC controller 12 and an audible and visual alarm component 13, the algorithm processing module is integrated into the PLC controller 12, and the algorithm processing module is electrically connected to the image acquisition module. The PLC controller 12 is connected to the algorithm processing module and the image acquisition module. The processing module, the audible and visual alarm device, and the drive mechanism of the conveying component 2 are electrically connected. The PLC controller 12 is used to receive the defect judgment signal output by the algorithm processing module, trigger the audible and visual alarm device to remind the staff to check for faults. If multiple unqualified cartons appear consecutively, the conveying component 2 can be controlled to slow down or stop to avoid the accumulation of unqualified products. The human-machine interaction module consists of a display screen 11, which is communicatively connected to the PLC controller 12. The PLC controller 12 enables the retrieval of detection data, the issuance of operation instructions, and the setting of parameters. The display screen 11 is embedded in the surface of the outer casing 1 and is electrically connected to the PLC controller 12. The human-machine interaction module is used to realize user operation and data management, and supports the import of standard printed patterns, real-time detection data display, detection parameter setting, detection data storage, and production quality report generation. The conveying component 2 is used to realize the continuous and stable conveying of cartons, and adopts a variable frequency speed regulation structure. The algorithm processing module is used to receive the image data output by the image acquisition module and execute a three-level image processing flow.

[0035] In this embodiment of the invention, the conveying component 2 includes a partition 21, a front baffle 22, a rear baffle 23, a support 24, and a transmission rod 25. Four leveling feet 211 are fixed to the lower end of the partition 21. The front baffle 22 is fixedly mounted on the partition 21, and the rear baffle 23 slides back and forth on the partition 21. Two supports 24 are provided, both mounted on the partition 21, and a rotating rod 241 is mounted between the two supports 24 via a bearing. The rotating rod 241 has a spindle-shaped cross-section and two pulleys are mounted on it. The pulley near the rear end of the rotating rod 241 can slide back and forth on the rotating rod 241. A waist-shaped groove is provided at the right end of both the front baffle 22 and the rear baffle 23. Furthermore, movable seats 231 are installed on both the front baffle 22 and the rear baffle 23, and pulleys are installed on both movable seats 231. Fixed seats 232 are installed on both the front baffle 22 and the rear baffle 23, and tension screws 233 are threaded onto the fixed seats 232. The tension screws 233 press against the movable seats 231. A conveyor belt 234 connects the pulley on the transmission rod 25 and the pulley on the movable seat 231. A base frame 212 is installed at the lower end of the partition 21, and a reducer 213 is installed on the base frame 212. A drive motor 214 is installed on the reducer 213. Pulleys are installed on the output shaft of the reducer 213 and the front end of the rotating rod 241. A belt connects the pulley on the output shaft to the pulley at the front end of the rotating rod 241. A slotted groove is provided on the base frame 212, and a dust cover 215 is installed on the base frame 212. The transmission rod 25 is mounted on the partition plate 21 via bearings. The transmission rod 25 has threads, and a handwheel 251 is installed at its front end. A mating plate 236 is installed on the rear baffle 23, and the mating plate 236 is threadedly engaged with the transmission rod 25. An adjustable plate 235 that can move back and forth is installed on the rear baffle 23 via bolts. A telescopic cylinder 2351 is installed on the adjustable plate 235, and a rear clamping plate 2352 is installed on the piston rod of the telescopic cylinder 2351. A front clamping plate 2352 is installed on the front baffle 22. 21, its function is as follows: the leveling support 211 ensures that the conveying component 2 is level and avoids conveying deviation; the shuttle-shaped rotating rod 241, the sliding pulley and the tensioning screw 233 cooperate to facilitate the adjustment of the tension and spacing of the conveyor belt 234 to adapt to cartons of different widths; the telescopic cylinder 2351 drives the rear clamping plate 2352 to cooperate with the front clamping plate 221 to clamp the cartons in both directions, preventing the cartons from shaking or shifting during the identification and detection process, and ensuring that the printing area of ​​the cartons is accurately aligned with the visual inspection station; the waist-shaped groove design on the base frame 212 facilitates the adjustment of the position of the drive motor 214, realizes the tension of the belt between the reducer 213 and the rotating rod 241, and improves the adaptability and maintenance convenience of the device.

[0036] In this embodiment of the invention, the visual inspection component 3 includes a horizontal rail 31, a vertical rail 32, and a moving plate 33. A support frame 311 is mounted on the horizontal rail 31, and the vertical rail 32 slides back and forth on the horizontal rail 31. A first connecting seat 312 and a second connecting seat 321 are respectively mounted on the horizontal rail 31 and the vertical rail 32. A drag chain 313 is provided between the first connecting seat 312 and the second connecting seat 321. A horizontal motor frame 314 and a first lead screw 315 are mounted on the horizontal rail 31. A horizontal servo motor 316 is mounted on the horizontal motor frame 314. The output shaft of the horizontal servo motor 316 is connected to the first lead screw 315 through a coupling. The first lead screw 315 is threaded onto the vertical rail 32. In coordination, the movable plate 33 slides up and down on the vertical rail plate 32. A second lead screw 324 is mounted on the vertical rail plate 32 via bearings, and the second lead screw 324 is threadedly engaged with the movable plate 33. A vertical motor frame 322 is mounted on the vertical rail plate 32, and a vertical servo motor 323 is mounted on the vertical motor frame 322. The output shaft of the vertical servo motor 323 is connected to the second lead screw 324 via a coupling. U-shaped photoelectric sensors are mounted on both the horizontal rail plate 31 and the vertical rail plate 32. Sensing plates are mounted on both the vertical rail plate 32 and the movable plate 33. Two motor mounts 331 are mounted on the movable plate 33, and a forward rotation motor 332 and a reverse rotation motor 333 are mounted on the two motor mounts 331 respectively. Both the output shafts of the forward and reverse motors 332 and 333 are equipped with pulleys, and the movable plate 33 is equipped with a mounting base 334. A rotating shaft 335 is mounted within the mounting base 334 via bearings. Two pulleys are mounted on the rotating shaft 335, and a connecting belt 336 connects the two pulleys on the rotating shaft 335 to the pulleys on the output shafts of the forward and reverse motors 332 and 333. A photoelectric sensor plate is mounted on the upper end of the rotating shaft 335, and a photoelectric switch is mounted on the movable plate 33. A mounting bracket 337 is fixed to the lower end of the rotating shaft 335. The industrial line scan camera 338 and the ring light source 339 are both mounted on the mounting bracket 337. Their function is to enable multi-dimensional scanning of the industrial line scan camera 338 and the ring light source 339. Precise adjustment is achieved through the horizontal servo motor 316 and the vertical servo motor 323, driven by the first lead screw 315 and the second lead screw 324. This allows for precise adjustment of the horizontal and vertical positions of the detection components, such as the industrial line scan camera 338 and the ring light source 339, adapting to the detection of cartons of different widths and heights. The forward rotation motor 332 and the reverse rotation motor 333 drive the rotating shaft 335 to adjust the detection angle of the industrial line scan camera 338 and the ring light source 339, ensuring accurate alignment with the printing area of ​​the carton. The U-shaped photoelectric sensor, in conjunction with the sensing sheet, provides limit protection for the adjustment stroke, preventing damage caused by excessive movement of components. The overall structure allows for flexible adaptation of the detection angle and position, ensuring the clarity and comprehensiveness of image acquisition.

[0037] In this embodiment of the invention, the three-level image processing flow is as follows: S1, Image preprocessing, sequentially performs grayscale and normalization processing, adaptive median filtering processing, and histogram equalization processing to eliminate the effects of corrugated paper texture interference, ink-wash printing color difference fluctuations, and uneven lighting on image quality. S2, Feature Extraction: Extracting texture features, geometric features, and edge features of the defect area. Texture features are extracted using the gray-level co-occurrence matrix, geometric features include the area, perimeter, and shape factor of the defect, and edge features are extracted using the Canny edge detection algorithm. S3, defect classification and identification, constructs a support vector machine classification model, inputs the extracted multi-dimensional features into the model, compares them with the preset normal sample feature library and defect sample feature library, and outputs the defect type, defect location and defect level; the algorithm processing module also has a built-in self-learning update module, which can automatically collect defect samples confirmed by manual re-inspection and optimize the classification model parameters; Its functions are as follows: to accurately process and analyze the cardboard printing images acquired by the image acquisition module, effectively eliminate various interference factors through a three-level processing flow, improve image quality, and ensure accurate extraction of defect features; the support vector machine classification model realizes accurate identification and classification of various defects, providing a clear basis for defect control; the self-learning update module enables the model to adapt to the detection needs of new and subtle defects, continuously improves the recognition accuracy as samples accumulate, reduces the workload of manual re-inspection, and improves the intelligence level and reliability of defect detection.

[0038] In this embodiment of the invention, the industrial line scan camera 338 has a resolution of 2048 pixels, and its shooting frame rate can be adaptively adjusted according to the conveying speed of the conveying component 2. The ring light source 339 has a color temperature adjustment range of 3000K-6500K and continuously adjustable brightness. The supplementary lighting parameters can be adjusted according to the color depth and ink thickness of the printed pattern on the carton. The light source can switch between direct and diffused modes to adapt to the supplementary lighting needs of different carton materials. Its functions are as follows: With a resolution of 2048 pixels, the industrial line scan camera 338 can clearly capture the subtle textures and defects in the printed area of ​​the carton. The adaptive frame rate adjustment ensures that there are no omissions or ghosting during the conveying process, ensuring the integrity and accuracy of the image data. The ring light source 339 can adapt to different colors, ink thicknesses, and materials of cartons through color temperature, brightness adjustment, and illumination mode switching. It can effectively compensate for the image blurring problem caused by uneven lighting, highlight the contrast between printing defects and normal patterns, provide a high-quality image data source for subsequent algorithm processing, and improve the accuracy of defect detection.

[0039] In this embodiment of the invention, the size of the filtering window can be adaptively adjusted according to the complexity of the image texture, with an adjustment range of 33-1111. While preserving the edge information of the printed pattern, it thoroughly filters out the periodic texture interference on the surface of the corrugated paper. Its function is: by adaptively adjusting the size of the filtering window, it can not only thoroughly filter out irrelevant interference textures, but also preserve the edge details and defect features of the printed pattern to the greatest extent, avoid the blurring or loss of defects caused by filtering, ensure the accuracy of subsequent feature extraction and defect identification, and improve the adaptability of the algorithm processing module to complex working conditions.

[0040] Example 2, based on Example 1, includes defect types such as misregistration, missing print, ink spots, scratches, incomplete text, and blurred barcodes. Defect levels are categorized as minor, general, and severe. Judgment criteria can be manually set via the human-machine interface module. Its function is to allow staff to flexibly adjust defect judgment criteria according to production quality requirements, adapting to the quality needs of different customers and products; to accurately distinguish between different types and levels of defects, facilitating targeted troubleshooting of faults, subsequent repair of minor defects, and elimination of severe defects, thereby improving product qualification rates and providing clear direction for production process optimization.

[0041] In Example 3, based on Examples 1 and 2, the PLC controller 12 supports interface with the factory MES system, and can upload the test data to the MES system in real time to realize digital traceability and management of production quality. Its function is to upload various data in the testing process to the MES system in real time, realize full digital traceability of production quality, and facilitate managers to fully grasp the production quality status.

[0042] The working principle of this invention is as follows: First, by rotating the handwheel 251 at the front end of the transmission rod 25, the mating plate 236 on the rear baffle 23 moves along the thread of the transmission rod 25, adjusting the distance between the rear baffle 23 and the front baffle 22 to match the width of the carton to be inspected. After the distance is adjusted, the carton to be inspected is placed on the conveyor belt 234 of the conveying component 2. Then, the drive motor 214 on the base frame 212 is started. After being reduced in speed by the reducer 213, the drive motor 214 drives the rotating rod 241 between the supports 24 to rotate via the belt, thereby driving the conveyor belt 234 to rotate, realizing the conveying of the carton. The conveying component 2... The frequency-controlled speed adjustment structure allows for speed adjustment according to inspection requirements, while the leveling feet 211 ensure smooth and unbiased conveying. When the carton reaches the inspection position, the telescopic cylinder 2351 on the adjusting plate 235 extends and retracts, driving the rear clamping plate 2352 to extend and cooperate with the front clamping plate 221 on the front baffle 22 to clamp and fix the carton. At this time, the vision inspection component 3 begins operation. The horizontal servo motor 316 drives the first lead screw 315 to rotate, causing the vertical rail plate 32 to move back and forth along the horizontal rail plate 31. The vertical servo motor 323 drives the second lead screw 324 to rotate, causing the moving plate 33 to move vertically... The track plate 32 moves up and down. The forward rotation motor 332 or the reverse rotation motor 333 drives the rotating shaft 335 to rotate via the connecting belt 336, adjusting the position and angle of the industrial line scan camera 338 and the ring light source 339 on the fixed frame 337. The ring light source 339 adjusts the color temperature (3000K-6500K), brightness, and illumination mode according to the carton printing situation to provide supplementary lighting for image acquisition. The encoder built into the conveying positioning unit is linked with the drive mechanism of the conveying component 2 and the trigger signal of the industrial line scan camera 338 to achieve synchronization of conveying speed, carton position, and image acquisition. 338 acquires image data of the printing area of ​​the carton and outputs it to the algorithm processing module integrated in the PLC controller 12. After receiving the image data, the algorithm processing module performs a three-level preprocessing process of grayscale conversion and normalization, adaptive median filtering, and histogram equalization to filter out interference such as corrugated paper texture. Then, it extracts the texture, geometry (area, perimeter, etc.) and edge (Canny algorithm) features of the defect area, inputs them into the support vector machine classification model and compares them with the sample library, and outputs the defect type, location and level. The self-learning update module synchronously collects the optimized model parameters of the defect samples confirmed by manual review.After receiving the defect judgment signal from the algorithm processing module, the PLC controller 12 triggers the audible and visual alarm component 13 if a defective carton is detected, alerting staff to investigate. If multiple defective products are detected consecutively, the conveyor component 2 can be slowed down or stopped to prevent accumulation. Simultaneously, the PLC controller 12 communicates with the display screen 11 embedded in the surface of the outer casing 1 to retrieve detection data and set parameters. Furthermore, the PLC controller 12 supports integration with the factory's MES system, facilitating data flow between carton printing defect detection and overall factory production management. Various data during the detection process (including defect type, defect quantity, pass rate, equipment operating parameters, etc.) are uploaded to the MES system in real time, enabling full digital traceability of production quality and allowing managers to comprehensively understand the production quality status.

[0043] The following points should be noted in this article: 1. The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention; other structures can refer to general designs.

[0044] 2. Where there is no conflict, the embodiments of the present invention and the features thereof can be combined with each other to obtain new embodiments.

[0045] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A visual recognition device for printing defects in cardboard boxes, comprising: The outer casing (1), conveying component (2), and visual inspection component (3) are characterized in that the visual inspection component (3) has a built-in image acquisition module, and the visual inspection component (3) is installed above the inspection station of the conveying component (2). The image acquisition module includes an industrial line array camera (338) with adjustable focal length, a ring light source (339), and a conveying positioning unit. The conveying positioning unit has a built-in encoder, and the encoder is linked with the drive mechanism of the conveying component (2) and the trigger signal of the industrial line array camera (338) to realize the synchronization of the conveying speed, carton conveying position, and image acquisition of the conveying component (2). It is used to acquire and output image data of the carton printing area on the conveying component (2). The visual recognition device for carton printing defects also includes an algorithm processing module, an execution feedback module, and a human-machine interaction module. The execution feedback module is controlled by a PLC controller ( The system consists of a 12) and an audible and visual alarm component (13). The algorithm processing module is integrated into the PLC controller (12). The algorithm processing module is electrically connected to the image acquisition module. The PLC controller (12) is electrically connected to the algorithm processing module, the audible and visual alarm device, and the drive mechanism of the conveying component (2). The human-machine interaction module consists of a display screen (11). The display screen (11) is communicatively connected to the PLC controller (12). The PLC controller (12) enables the retrieval of detection data, the issuance of operation instructions, and the setting of parameters. The display screen (11) is embedded on the surface of the outer casing (1) and is electrically connected to the PLC controller (12). The conveying component (2) is used to realize the continuous and stable conveying of the carton. It adopts a variable frequency speed control structure. The algorithm processing module is used to receive the image data output by the image acquisition module and execute a three-level image processing flow.

2. The visual recognition device for printing defects in cardboard boxes according to claim 1, characterized in that, The conveying component (2) includes a partition (21), a front baffle (22), a rear baffle (23), a support (24), and a transmission rod (25). The lower end of the partition (21) is fixed with four leveling feet (211). The front baffle (22) is fixedly installed on the partition (21). The rear baffle (23) slides back and forth on the partition (21). There are two supports (24), both of which are installed on the partition (21) and are connected by a shaft. A rotating rod (241) is mounted on the bearing. The rotating rod (241) has a spindle-shaped cross-section. Two pulleys are mounted on the rotating rod (241). The pulley near the rear end of the rotating rod (241) can slide back and forth on the rotating rod (241). The right ends of the front baffle (22) and the rear baffle (23) are provided with waist-shaped grooves. Movable seats (231) are mounted on both the front baffle (22) and the rear baffle (23). Pulleys are mounted on both movable seats (231). (23) is equipped with a fixed seat (232), and a tensioning screw (233) is threaded onto the fixed seat (232). The tensioning screw (233) is pressed against the movable seat (231). A conveyor belt (234) is connected between the pulley on the transmission rod (25) and the pulley on the movable seat (231). A base frame (212) is installed at the lower end of the partition (21). A reducer (213) is installed on the base frame (212). A drive motor (214) is installed on the reducer (213). The output shaft of the machine (213) and the front end of the rotating rod (241) are both equipped with pulleys. A belt is connected between the pulley on the output shaft of the reducer (213) and the pulley at the front end of the rotating rod (241). The base frame (212) is provided with a waist-shaped groove and a dust cover plate (215) is installed on the base frame (212). The transmission rod (25) is installed on the partition plate (21) through a bearing. The transmission rod (25) is provided with a thread and a handwheel (251) is installed at the front end of the transmission rod (25).

3. The visual recognition device for printing defects in cardboard boxes according to claim 2, characterized in that, A mating plate (236) is installed on the rear baffle (23), and the mating plate (236) is threadedly engaged with the transmission rod (25). An adjusting plate (235) that can move back and forth is installed on the rear baffle (23) by bolts. A telescopic cylinder (2351) is installed on the adjusting plate (235), and a rear clamping plate (2352) is installed on the piston rod of the telescopic cylinder (2351). A front clamping plate (221) is installed on the front baffle (22).

4. The visual recognition device for printing defects in cardboard boxes according to claim 1, characterized in that, The visual inspection component (3) includes a horizontal rail (31), a vertical rail (32), and a moving plate (33). A support frame (311) is installed on the horizontal rail (31). The vertical rail (32) slides back and forth on the horizontal rail (31). A first connecting seat (312) and a second connecting seat (321) are respectively installed on the horizontal rail (31) and the vertical rail (32). A drag chain (313) is provided between the first connecting seat (312) and the second connecting seat (321). A horizontal motor frame (314) and a first lead screw (315) are installed on the horizontal rail (31). A horizontal servo motor (316) is installed on the horizontal motor frame (314). The output shaft of the horizontal servo motor (316) is connected to the first lead screw (315) via a coupling. A lead screw (315) is connected, and the first lead screw (315) is threadedly engaged with the vertical rail plate (32). The moving plate (33) slides up and down on the vertical rail plate (32). A second lead screw (324) is installed on the vertical rail plate (32) through a bearing. The second lead screw (324) is threadedly engaged with the moving plate (33). A vertical motor frame (322) is installed on the vertical rail plate (32). A vertical servo motor (323) is installed on the vertical motor frame (322). The output shaft of the vertical servo motor (323) is connected to the second lead screw (324) through a coupling. U-shaped photoelectric sensors are installed on both the horizontal rail plate (31) and the vertical rail plate (32). Sensing plates are installed on both the vertical rail plate (32) and the moving plate (33).

5. A visual recognition device for printing defects in cardboard boxes according to claim 4, characterized in that, Two motor mounts (331) are installed on the movable plate (33). A forward motor (332) and a reverse motor (333) are respectively installed on the two motor mounts (331). Pulleys are installed on the output shafts of the forward motor (332) and the reverse motor (333). A mounting base (334) is installed on the movable plate (33). A rotating shaft (335) is installed in the mounting base (334) through a bearing. Two pulleys are installed on the rotating shaft (335). A connecting belt (336) is connected between the two pulleys on the rotating shaft (335) and the pulleys on the output shafts of the forward motor (332) and the reverse motor (333). A photoelectric sensor plate is installed on the upper end of the rotating shaft (335). A photoelectric switch is installed on the movable plate (33). A fixing frame (337) is fixed on the lower end of the rotating shaft (335). An industrial line array camera (338) and a ring light source (339) are both installed on the fixing frame (337).

6. The visual recognition device for printing defects in cardboard boxes according to claim 1, characterized in that, The three-level image processing flow is as follows: S1, Image preprocessing, sequentially performs grayscale and normalization processing, adaptive median filtering processing, and histogram equalization processing to eliminate the effects of corrugated paper texture interference, ink-wash printing color difference fluctuations, and uneven lighting on image quality. S2, Feature Extraction: Extracting texture features, geometric features, and edge features of the defect area. Texture features are extracted using the gray-level co-occurrence matrix, geometric features include the area, perimeter, and shape factor of the defect, and edge features are extracted using the Canny edge detection algorithm. S3, Defect Classification and Recognition: Construct a support vector machine classification model, input the extracted multi-dimensional features into the model, compare it with the preset normal sample feature library and defect sample feature library, and output the defect type, defect location and defect level; The algorithm processing module also has a built-in self-learning update module, which can automatically collect defect samples confirmed by manual re-inspection and optimize the classification model parameters.

7. A visual recognition device for printing defects in cardboard boxes according to claim 5, characterized in that, The industrial line scan camera (338) has a resolution of 2048 pixels, and the shooting frame rate can be adaptively adjusted according to the conveying speed of the conveying component (2). The color temperature adjustment range of the ring light source (339) is 3000K-6500K, and the brightness is continuously adjustable. The supplementary lighting parameters can be adjusted according to the color depth and ink thickness of the printed pattern on the carton. The light source can switch between direct and diffused modes to adapt to the supplementary lighting needs of different carton materials.

8. A visual recognition device for printing defects in cardboard boxes according to claim 6, characterized in that, In the adaptive median filtering process, the size of the filtering window can be adaptively adjusted according to the image texture complexity, with an adjustment range of 33-1111. While preserving the edge information of the printed pattern, it completely filters out the periodic texture interference on the surface of the corrugated paper.

9. A visual recognition device for printing defects in cardboard boxes according to claim 6, characterized in that, The defect types include misregistration, missing print, ink spots, scratches, incomplete text, and blurred barcodes. The defect levels are divided into minor defects, general defects, and serious defects. The judgment criteria can be manually set through the human-computer interaction module.

10. A visual recognition device for printing defects in cardboard boxes according to claim 1, characterized in that, The PLC controller (12) supports interface with the factory MES system, and can upload the detection data to the MES system in real time to realize digital traceability and management of production quality.

Images