Device and method for automatically detecting appearance quality of solid wood table board of study table

Abstract

The invention belongs to the technical field of industrial automatic detection, and particularly relates to an automatic detection device and method for the appearance quality of a solid wood table board of a learning table, and the device comprises a conveying mechanism which is used for conveying a to-be-detected solid wood table board, and an image collection unit which is arranged above the conveying mechanism; the image processing and analyzing unit is electrically connected with the image acquisition unit and used for receiving and processing the acquired image, and the image processing and analyzing unit runs a software program; the executing mechanism is electrically connected with the image processing and analyzing unit and used for sorting the table plates according to the judgment result; and the man-machine interaction and control system is electrically connected with the conveying mechanism, the image acquisition unit, the image processing and analysis unit and the execution mechanism. The device is reasonable in design, simple in structure and capable of adapting to natural characteristics of solid wood, multiple surface defects are accurately recognized, the color difference and texture consistency is evaluated, and online, rapid and non-contact sorting of the appearance quality of the table board of the learning table is achieved.

Description

Technical Field

[0001] This invention belongs to the field of industrial automation testing technology, and in particular relates to an automatic testing device and method for the appearance quality of solid wood desks for study desks. Background Technology

[0002] Demand for study desks is growing: With increasing emphasis on children's learning environments, solid wood study desks are favored for their environmental friendliness, durability, and high-quality feel. Their appearance quality (surface smoothness, absence of knots and cracks, uniform color and texture) directly impacts product quality and user experience.

[0003] For the inspection of board materials during the production and processing of study desks: traditional inspection processes have certain drawbacks: Manual visual inspection relies on workers' experience and subjective judgment, resulting in low efficiency, high labor intensity, fatigue leading to missed or false inspections, poor consistency, high cost, and difficulty in standardization. Insufficient general-purpose equipment: Existing industrial vision inspection equipment is mostly designed for metals, plastics, or flat glass. It is poorly adapted to the complex textures, color changes, and diverse defect types (live knots, dead knots, cracks, wormholes, resin ducts, scratches, color spots, sanding marks, etc.) of natural wood, resulting in a high misjudgment rate.

[0004] Weak targeting: There is a lack of dedicated testing programs for specific dimensions, shapes (which may be curved or irregular), surface treatment processes (varnish, wood wax oil), and appearance quality standards of study desks; Limitations of existing technologies: Existing patents or equipment may focus on log grading, large board inspection, or single defect identification in wood inspection, lacking a solution for efficient, accurate, and integrated automatic detection of multi-dimensional appearance quality such as defects, color difference, and texture in the specific application scenario of finished solid wood desks. Summary of the Invention

[0005] This invention addresses the technical problems existing in the above-mentioned solid wood tabletop inspection process by proposing a reasonably designed, simple, and easy-to-process device that adapts to the natural characteristics of solid wood, accurately identifies various surface defects, assesses color difference and texture consistency, and achieves online, rapid, and non-contact sorting of the appearance quality of study desk tabletops. This provides an efficient, accurate, and highly automated automatic inspection device and method for the appearance quality of solid wood study desk tabletops.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is an automatic detection device and method for the appearance quality of a solid wood desk, comprising: A conveyor mechanism is used to transport the solid wood tabletops to be inspected. An image acquisition unit is disposed above the conveying mechanism and includes at least one industrial camera and a dedicated multi-light source lighting system used in conjunction with it. The dedicated multi-light source lighting system includes at least a coaxial light source, a low-angle light source, and a diffuse light source. An image processing and analysis unit is electrically connected to the image acquisition unit and is used to receive and process the acquired images. The image processing and analysis unit runs a software program, which includes a deep learning-based model module for identifying and locating various surface defects, and a color difference and texture evaluation module for calculating color difference values ​​and analyzing texture consistency. An actuator, electrically connected to the image processing and analysis unit, is used to sort the tabletops according to the judgment result; The human-computer interaction and control system is electrically connected to the transmission mechanism, image acquisition unit, image processing and analysis unit and execution mechanism respectively, and is used for parameter setting, process control, status monitoring and data management.

[0007] Preferably, the conveying mechanism includes a variable frequency speed control drive motor, a roller conveyor or belt conveyor, a precision positioning sensor, and a guiding device.

[0008] Preferably, the brightness and illumination angle of the light source in the dedicated multi-light source lighting system can be adjusted independently.

[0009] Preferably, the deep learning-based model module is a convolutional neural network model, which is trained on a sample image set labeled with various defect types of solid wood study table boards, and can identify defects including at least knots, cracks, wormholes, scratches, pits, and resin sacs.

[0010] Preferably, the color difference and texture evaluation module operates under a standard light source environment setting, calculates the color difference value (ΔE) between the tabletop area and the standard color swatch, and extracts texture feature parameters to evaluate uniformity.

[0011] Preferably, the software program of the image processing and analysis unit further includes an image preprocessing module and a feature fusion decision module. The feature fusion decision module integrates defect information, color difference value and texture consistency evaluation results, and makes a graded judgment based on preset quality standards.

[0012] Preferably, the actuator is a pneumatic push rod, a robotic arm, or a diversion conveyor belt.

[0013] Preferably, the human-computer interaction and control system includes a touch screen operating terminal and a data storage management system.

[0014] Preferably, an automatic inspection method for the appearance quality of a solid wood desk panel is provided, utilizing the automatic inspection device for the appearance quality of a solid wood desk panel as described in any one of claims 1-8, comprising the following inspection steps: S1: Transport the worktable to be inspected to the image acquisition station and position it; S2: Trigger the image acquisition unit to acquire an image of the tabletop surface under optimized lighting conditions; S3: Image processing and analysis unit processes images: segment target regions, use deep learning models to identify and locate defects, and evaluate color difference and texture consistency; S4: Based on the comprehensive information of defects, color difference, and texture, determine the quality grade of the tabletop according to the preset standards; S5: Send sorting instructions to the actuator to complete the automatic sorting of the table; S6: Store the detection data and generate reports.

[0015] Preferably, in step S3, the deep learning model is a target detection or image segmentation model based on a convolutional neural network.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: 1. High precision and high reliability: By utilizing deep learning and multi-algorithm fusion, it effectively overcomes wood texture interference, significantly improves the accuracy of defect identification (especially for live knots, dead knots, and fine cracks), and reduces the false positive and false negative rates. 2. Comprehensive testing: Simultaneously tests multiple appearance quality indicators (defects, color difference, texture) to meet the comprehensive appearance requirements of the study desk tabletop; 3. High efficiency in inspection: It achieves online automatic inspection, which is much faster than manual visual inspection and can meet the needs of mass production; 4. Standardization and objectivity in testing and processing: Judgments are made based on unified standards to eliminate the influence of human factors and ensure consistent product quality; 5. High degree of automation in the detection process: It integrates the entire process of automatic transmission, image acquisition, analysis, judgment, and sorting, reducing manual intervention; 6. High adaptability of the testing process: The light source and algorithm are adjustable, which can adapt to the testing of solid wood tables of different tree species, different surface treatments (paint / wood wax oil), and different specifications; 7. Data-driven management: Enables electronic recording, traceability, and analysis of quality data, facilitating quality control and process improvement; 8. Reduced production costs: In the long run, replacing manual inspection can significantly reduce labor costs and improve production efficiency; This device is rationally designed, has a simple structure, is easy to process, and can adapt to the natural characteristics of solid wood. It can accurately identify various surface defects, assess color differences and texture consistency, and realize online, fast, and non-contact sorting of the appearance quality of study desks, thus providing a highly efficient, accurate, and automated solution. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the detection device provided by the present invention; Figure 2 This is an internal block diagram of the image processing and analysis unit provided by the present invention. Figure 3 This is a software functional block diagram of the image processing and analysis unit provided by the present invention; Figure 4 A flowchart of the detection method provided by the present invention; Figure 5 This is a schematic diagram of the working principle of the detection device provided by the present invention; Figure 6 This is a schematic diagram of the detection method provided by the present invention regarding the detection of live joint types; Figure 7 This is a schematic diagram illustrating the detection method for dead junction type provided by the present invention; Figure 8 This is a schematic diagram illustrating the detection method for decay type provided by the present invention; Figure 9 This is a schematic diagram illustrating the detection method for detecting mineral line types provided by the present invention; Figure 10 This is a schematic diagram illustrating the detection method for mold type provided by the present invention; Figure 11 This is a schematic diagram illustrating the detection method for detecting insect eye types provided by the present invention; Figure 12 This is a schematic diagram illustrating the crack / crack type detection method provided by the present invention; Figure 13 This is a schematic diagram illustrating the detection method provided by the present invention regarding the detection of exposed adhesive lines. In the above figures, 1. Conveying mechanism; 11. Variable frequency speed control drive motor; 2. Image acquisition unit; 21. Industrial camera; 22. Dedicated multi-light source lighting system; 221. Coaxial light source; 222. Low-angle light source; 223. Diffuse light source; 3. Image processing and analysis unit; 4. Actuating mechanism; 41. Pneumatic push rod; 42. Robotic arm; 43. Diverting conveyor belt; 5. Human-machine interaction and control system. Detailed Implementation

[0019] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0020] Numerous specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways than those described herein, and therefore the invention is not limited to the specific embodiments disclosed in the following specification.

[0021] Examples, such as Figure 1 As shown, an automatic inspection device and method for the appearance quality of solid wood desks for study desks includes: a conveying mechanism 1 for conveying solid wood desks to be inspected; an image acquisition unit 2, disposed above the conveying mechanism 1, including at least one industrial camera 21 and a dedicated multi-light source lighting system 22 cooperating with it, the dedicated multi-light source lighting system 22 including at least a coaxial light source 221, a low-angle light source 222 and a diffuse light source 223; an image processing and analysis unit 3, electrically connected to the image acquisition unit 2, for receiving and processing the acquired images, the image processing and analysis unit 3 running a software program, the software program including a deep learning-based model module for identifying and locating various surface defects, and a color difference and texture evaluation module for calculating color difference values ​​and analyzing texture consistency; an execution mechanism 4, electrically connected to the image processing and analysis unit 3, for sorting the desks according to the judgment results; and a human-machine interaction and control system 5, electrically connected to the conveying mechanism 1, the image acquisition unit 2, the image processing and analysis unit 3 and the execution mechanism 4 respectively, for parameter setting, process control, status monitoring and data management.

[0022] To ensure convenient transport of workpieces to be processed and to guarantee the smooth implementation of inspection work, such as Figure 1 , 5 As shown, the conveying mechanism 1 includes a variable frequency speed control drive motor 11, a roller conveyor or belt conveyor, a precision positioning sensor, and a guiding device. Specifically, the conveying mechanism 1 is mainly used to carry and transport the solid wood table to be inspected along the inspection path. The roller conveyor or belt conveyor is preferably driven by a variable frequency speed control motor, and the speed is adjustable to adapt to different inspection accuracy requirements. Positioning sensors, such as photoelectric sensors or encoders, are selected to ensure that the table stops stably or passes through at a constant speed at the image acquisition position. In order to prevent the table from deviating during transport, guide bars can be designed above the conveying mechanism 1 to ensure the stability of the table posture.

[0023] To further improve the rationality of the setup of the equipment, such as Figure 1 , 5As shown in the embodiment, the brightness and illumination angle of the dedicated multi-source lighting system 22 can be independently adjusted. Specifically, the multi-source lighting system includes three types: a coaxial light source 221, a low-angle ring / strip light source, and a diffused dome light source. Each light source serves a different purpose. For example, the coaxial light source 221 is used for detecting high-reflectivity defects in flat areas; the low-angle ring / strip light source is used to highlight three-dimensional defects such as scratches and dents; and the diffused dome light source provides uniform illumination, reduces reflective interference, and facilitates color difference and texture evaluation. In addition to the light source, the spectrum of the lighting system is also appropriately selected. For instance, a multispectral camera or a specific wavelength light source such as infrared can be selected to detect internal defects such as internal decay or specific contaminants. Furthermore, the brightness and angle can be independently adjusted to meet the needs of detecting different parts, adapting to different wood species, paint gloss levels, and defect types.

[0024] To ensure the smooth progress of the testing work and to guarantee the comprehensiveness of the testing, such as Figure 1 , 5 As shown: The embodiment selects to set an industrial camera 21, including multiple high-resolution industrial cameras 21 such as CCD or CMOS area array cameras, equipped with fixed-focus lenses or zoom autofocus lenses.

[0025] For defects on the solid wood panels, including knots, cracks, wormholes, scratches, pits, and resin pockets, artificial intelligence technology was introduced, and a deep learning-based model module was set up. Specifically, a convolutional neural network (CNN) model was introduced for deep learning training. CNNs are existing technologies, as disclosed in the invention patent "Verification of Classification Decisions in Convolutional Neural Networks" (patent number: CN201980089516.9) published by the State Intellectual Property Office, and in the invention patent "Feature Extraction Method, System, Embedded Device and Medium of Convolutional Neural Networks" (patent number: CN202011391654.3). A deep learning model specifically designed for processing grid-structured data (such as images and speech) has the core advantage of efficiently extracting spatial features through local receptive fields, weight sharing, and pooling operations, and is widely used in image classification, object detection, and other fields. This embodiment mainly uses a convolutional neural network (CNN) for deep learning of images, training a sample image set labeled with various defect types on the solid wood tabletop, and combining this with the appearance quality inspection requirements of the beech straight-jointed board shown in Table 1 below. Figures 6-13 The schematic diagram of the detection results shown can identify defects including at least knots, cracks, wormholes, scratches, pits, and resin pockets.

[0026] Similarly, the color difference and texture evaluation module should operate under the standard light source environment settings set earlier, calculate the color difference value (ΔE) between the tabletop area and the standard color swatch, and extract texture feature parameters to evaluate uniformity.

[0027]

[0028] Table 1: Appearance Quality Requirements for Fagus Straight-Jointed Panels Implementation, for example Figure 3 As shown, the software program for setting up the image processing and analysis unit also includes an image preprocessing module, a feature fusion decision module, a background segmentation and target localization module, a defect detection module, and a color difference and texture consistency evaluation module.

[0029] The image preprocessing module is designed to perform functions such as noise reduction, filtering, enhancement, and distortion correction. The background segmentation and target localization module is used to accurately segment the tabletop area and eliminate background interference.

[0030] The defect detection module is designed to distinguish between natural and defective textures using multi-scale texture analysis algorithms such as Gabor filtering and Local Binary Pattern (LBP). It employs a deep learning-based defect recognition model, such as a Convolutional Neural Network (CNN), particularly object detection networks like YOLO, Faster R-CNN, or segmentation networks like U-Net. This model, trained on a large number of labeled solid wood table defect samples, can accurately identify and locate various defects such as knots, cracks, wormholes, scratches, dents, resin pockets, sand marks, and dirt. Traditional image processing algorithms, such as edge detection, thresholding, and blob analysis, are used as auxiliary methods or in specific scenarios.

[0031] The function of the color difference and texture consistency evaluation module is to extract color features such as Lab and RGB values ​​from different areas of the tabletop under a standard lighting environment, such as D65, and calculate the color difference (ΔE) with the standard sample or a set threshold. It also analyzes the overall or regional texture features of the tabletop, such as directionality, contrast, and roughness, to evaluate the uniformity and aesthetics of the texture.

[0032] The function of the feature fusion and decision module is to comprehensively detect information such as defect type, size, location, quantity, color difference, and texture consistency, and combine this information with preset standards for the appearance quality of study desks and grading standards for certain types of boards. Figure 5 The test results diagram shows the acceptance / non-acceptance / concession judgment for each tabletop, or the quality grade classification such as Grade A, Grade B, and Grade C.

[0033] Implementation, for example Figure 1As shown, the actuator is a roller conveyor or belt conveyor, equipped with pneumatic pushers 1-n and robotic arms 1-N. There are also 1-N branch conveyor belts representing different grades of workpieces. When the human-machine interface and control system delivers a workpiece of a certain grade to the exit of a branch conveyor belt, the pneumatic pusher and robotic arm at the corresponding exit receive instructions from the system. The pneumatic pusher, carrying a vacuum suction cup, picks up the workpiece, while the robotic arm moves up, down, and rotates, transferring the workpiece to the branch conveyor belt. The pneumatic pusher then releases the workpiece, and it slides down. The inspected workpieces are automatically sorted into the corresponding areas: qualified product area, unqualified product area, and different grade areas.

[0034] Implementation, for example Figure 1 As shown, the human-machine interface and control system is installed on the roller conveyor or belt conveyor, and specifically includes a touch screen operation terminal, a data storage management system, a PLC or motion controller. Figure 3 As shown, the touchscreen operating terminal allows for setting various parameters such as detection standards, speed, light source parameters, program start / stop, status monitoring, and alarm display. The touchscreen displays four operating interfaces: the conveyor mechanism interface, the image acquisition unit interface, the image processing and analysis unit interface, and the actuator interface. Each interface allows for parameter settings. Specific sub-interfaces include: the conveyor mechanism interface with main roller line operation settings; the actuator interface with pneumatic actuator settings, robotic arm settings, and diversion conveyor belt settings; the image acquisition unit interface with industrial camera settings and multi-light source lighting system settings; and the image processing and analysis unit interface with detection database and detection report interfaces.

[0035] The data storage management system of the human-computer interaction and control system is mainly used to store inspection results, such as images, defect information, judgment results, timestamps, and generate statistical reports. The report content should at least include the pass rate, defect type distribution, and support data traceability.

[0036] The PLC or motion controller in the human-machine interaction and control system mainly coordinates the timing of actions of the transmission mechanism, image acquisition trigger, and execution mechanism.

[0037] To ensure the smooth progress of the inspection work, an automatic inspection method for the appearance quality of solid wood desk panels is provided. The method utilizes an automatic inspection device for the appearance quality of solid wood desk panels to perform the inspection, which specifically includes the following inspection steps: S1: Transport the worktable to be inspected to the image acquisition station and position it; S2: Trigger the image acquisition unit to acquire an image of the tabletop surface under optimized lighting conditions; S3: Image processing and analysis unit processes images: segment target regions, use deep learning models to identify and locate defects, and evaluate color difference and texture consistency; S4: Based on the comprehensive information of defects, color difference, and texture, determine the quality grade of the tabletop according to the preset standards; S5: Send sorting instructions to the actuator to complete the automatic sorting of the table; S6: Store the detection data and generate reports.

[0038] In step S3 above, the deep learning model is a target detection or image segmentation model based on a convolutional neural network; In the above process: This invention features high precision and high reliability: Utilizing deep learning and multi-algorithm fusion, it effectively overcomes wood grain interference, significantly improving the accuracy of defect identification (especially for live knots, dead knots, and minor cracks), and reducing the false positive and false negative rates; Comprehensiveness: Simultaneously detecting multiple appearance quality indicators (defects, color difference, texture), meeting the comprehensive appearance requirements of study desks; High efficiency: Achieving online automatic detection, with a speed far exceeding manual visual inspection, adapting to the needs of mass production; Standardization and objectivity: Judgment based on unified standards, eliminating the influence of human factors and ensuring product quality consistency; High degree of automation: Integrating automatic transmission, image acquisition, analysis, judgment, and sorting throughout the entire process. The system offers several advantages: reduced manual intervention; high adaptability (adjustable light source and algorithm, adaptable to different tree species, surface treatments (paint / wood wax oil), and specifications of solid wood tables); data-driven management (electronic recording, traceability, and analysis of quality data, facilitating quality control and process improvement); cost reduction (in the long term, replacing manual inspection can significantly reduce labor costs and improve production efficiency); and a rational design, simple structure, and convenient processing. It adapts to the natural characteristics of solid wood, accurately identifies various surface defects, assesses color differences and texture consistency, and enables online, rapid, and non-contact sorting of the appearance quality of study tabletops, providing a highly efficient, accurate, and automated system.

[0039] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments for application in other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An automatic inspection device for the appearance quality of a solid wood desk, characterized in that, include: Conveying mechanism (1) is used to transport the solid wood table to be tested; Image acquisition unit (2) is located above the conveying mechanism (1) and includes at least one industrial camera (21) and a dedicated multi-source lighting system (22) used in conjunction with it. The dedicated multi-source lighting system (22) includes at least a coaxial light source (221), a low-angle light source (222) and a diffuse light source (223). The image processing and analysis unit (3) is electrically connected to the image acquisition unit (2) and is used to receive and process the acquired images. The image processing and analysis unit runs a software program, which includes a deep learning-based model module for identifying and locating various surface defects, and a color difference and texture evaluation module for calculating color difference values ​​and analyzing texture consistency. The actuator (4) is electrically connected to the image processing and analysis unit (3) and is used to sort the tabletops according to the judgment result; The human-computer interaction and control system (5) is electrically connected to the transmission mechanism (1), image acquisition unit (2), image processing and analysis unit (3) and execution mechanism (4) respectively, and is used for parameter setting, process control, status monitoring and data management.

2. The automatic inspection device for the appearance quality of a solid wood desk as described in claim 1, characterized in that, The conveying mechanism (1) includes a variable frequency speed control drive motor (11), a roller or belt conveyor, a precision positioning sensor, and a guiding device.

3. The automatic inspection device for the appearance quality of a solid wood desk as described in claim 1, characterized in that, The brightness and illumination angle of the light source of the dedicated multi-source lighting system (22) can be adjusted independently.

4. The automatic inspection device for the appearance quality of a solid wood desk as described in claim 1, characterized in that, The model module based on deep learning training is a convolutional neural network model. It is trained on a sample image set labeled with various defect types of solid wood study table boards, and can identify defects including at least knots, cracks, wormholes, scratches, pits, and resin sacs.

5. The automatic inspection device for the appearance quality of a solid wood desk as described in claim 4, characterized in that, The color difference and texture evaluation module operates under standard light source settings, calculates the color difference value (ΔE) between the tabletop area and the standard color swatch, and extracts texture feature parameters to evaluate uniformity.

6. The automatic inspection device for the appearance quality of a solid wood desk as described in claim 1, characterized in that, The software program of the image processing and analysis unit (3) also includes an image preprocessing module and a feature fusion decision module. The feature fusion decision module integrates defect information, color difference value and texture consistency evaluation results, and makes a graded judgment based on preset quality standards.

7. The automatic inspection device for the appearance quality of a solid wood desk as described in claim 1, characterized in that, The actuator (4) is a pneumatic push rod (41), a robotic arm (42), and a diversion conveyor belt (43).

8. The automatic inspection device for the appearance quality of a solid wood desk as described in claim 1, characterized in that, The human-computer interaction and control system (5) includes a touch screen operation terminal and a data storage management system.

9. An automatic inspection method for the appearance quality of a solid wood desk for a study table, characterized in that, The automatic inspection device for the appearance quality of solid wood desk panels as described in any one of claims 1-8 includes the following inspection steps: S1: Transport the worktable to be inspected to the image acquisition station and position it; S2: Trigger the image acquisition unit to acquire an image of the tabletop surface under optimized lighting conditions; S3: Image processing and analysis unit processes images: segment target regions, use deep learning models to identify and locate defects, and evaluate color difference and texture consistency; S4: Based on the comprehensive information of defects, color difference, and texture, determine the quality grade of the tabletop according to the preset standards; S5: Send sorting instructions to the actuator to complete the automatic sorting of the table; S6: Store the detection data and generate reports.

10. The automatic inspection method for the appearance quality of a solid wood desk as described in claim 9, characterized in that, In step S3, the deep learning model is a target detection or image segmentation model based on a convolutional neural network.

Images