Defect detection system and method for special-shaped covering cap

Abstract

The invention discloses a special-shaped covering cap defect detection system and method, and relates to the field of image recognition. The system comprises a conveying belt and a detection mechanism, the conveying belt is used for conveying a to-be-detected special-shaped covering cap to the detection mechanism, and the detection mechanism comprises a control processing unit, a camera and a screening unit; the camera is used for shooting an appearance image of the special-shaped covering cap and sending the appearance image to the control processing unit; the control processing unit is used for identifying the appearance image through a special-shaped covering cap defect identification model so as to determine a defective covering cap and sending a screening instruction to the screening unit; and the screening unit is used for performing screening operation on the defective covering caps according to the screening instruction, so that the defective covering caps enter a defective area. The defect of the special-shaped covering cap is accurately detected through the machine learning technology, whether the special-shaped covering cap is qualified or not can be identified and distinguished under the condition that the structure of the special-shaped covering cap is complex or irregular, and the problem that the defect condition of the covering cap is difficult to intuitively and accurately judge due to the abnormal structure of the special-shaped covering cap is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a system and method for detecting defects of special-shaped mouth covers. Background Art

[0002] Flaps are components used to close container openings. They can be of various shapes and sizes to ensure the safety and hygiene of the container contents. In order to fit the shape of a specific container or meet special functional requirements, the closures also need to be adaptively designed into some irregular shapes. The design of special-shaped closures may include special edges, grooves, threads or other features that make them different from traditional round or square closures.

[0003] The design of the flap usually requires precise size and shape to ensure a good sealing performance of the container. At the same time, the flap needs to be inspected to determine whether there are any defects or substandard areas.

[0004] When inspecting the flap, it is convenient to directly obtain the top image of the flap with a regular appearance to determine whether the flap has defects. However, due to the abnormal structure of the special-shaped flap, it is difficult to intuitively and accurately determine the defects of the flap. Summary of the invention

[0005] In view of the above technical problems and defects, the purpose of the present invention is to provide a system and method for detecting defects in special-shaped flaps, which can alleviate the technical problem that it is difficult to intuitively and accurately judge the defects of the flaps due to the structural abnormalities of the special-shaped flaps.

[0006] To achieve the above-mentioned purpose, in a first aspect, the present invention provides a defect detection system for a special-shaped cover, comprising a conveyor belt and a detection mechanism, wherein the conveyor belt is used to transmit the special-shaped cover to be detected to the detection mechanism, and the detection mechanism comprises a control processing unit, a camera and a screening unit, wherein the special-shaped cover is formed with a normal special-shaped structure; the camera is used to capture an appearance image of the special-shaped cover, and send the appearance image to the control processing unit; the control processing unit is used to identify the appearance image through a preset special-shaped cover defect recognition model to determine whether the special-shaped cover has an abnormal structure other than the special-shaped structure, if the abnormal structure exists, the special-shaped cover is determined to be a defective inferior cover, and a screening instruction is sent to the screening unit, the abnormal structure is used to indicate that the special-shaped cover has defective features, and the special-shaped cover defect recognition model is obtained by training the defective and non-defective special-shaped cover images as sample data through a machine learning algorithm; the screening unit is used to perform a screening operation on the defective cover according to the screening instruction, so that the defective cover enters a defective area.

[0007] The present invention solves the problem of intuitively and accurately judging defects caused by abnormal structures of special-shaped flaps by using a high-precision camera to capture multi-angle appearance images of special-shaped flaps, and combines the image recognition and processing technology integrated in the control processing unit. The system can use machine learning technology to extract and identify features of captured images through a special-shaped flap defect recognition model, and can deeply understand the complex geometric shapes of special-shaped flaps, accurately distinguish between irregular shapes in design and actual production defect characteristics, such as cracks, deformations, missing parts, etc., to avoid misjudgment of defect detection. In this way, the system can automatically and accurately detect defects of special-shaped flaps, and even in the case of complex or irregular structures of special-shaped flaps, it can also identify and distinguish the special-shaped structures and defect characteristics of special-shaped flaps, and then judge whether the special-shaped flaps are qualified, thereby improving production efficiency and product quality.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the camera includes a top camera and a side camera, the appearance image includes a top image of the mouth cover and a side image of the mouth cover, the top camera is used to capture the top image of the mouth cover, and the side camera is used to capture the side image of the mouth cover.

[0009] By adopting the technical solution of the above embodiment, the system can fully capture the top and side images of the special-shaped cover by equipping the top camera and the side camera. This multi-angle detection method improves the comprehensiveness and accuracy of defect detection. The combined use of the top and side images enables the system to analyze the cover from different perspectives, increasing the depth and details of the detection, thereby more effectively identifying potential defects, no matter where the defects are located on the cover.

[0010] In combination with some embodiments of the first aspect, in some embodiments, the control processing unit is specifically used to identify the top image of the mouth cover through the special-shaped mouth cover defect recognition model to determine whether there is an abnormal structure other than the special-shaped structure on the top of the special-shaped mouth cover; if not, then identify the side image of the mouth cover through the special-shaped mouth cover defect recognition model to determine whether there is an abnormal structure other than the special-shaped structure on the side of the special-shaped mouth cover.

[0011] By adopting the technical solution of the above embodiment, the control processing unit can perform defect detection on the top and side of the special-shaped mouth cover in steps through machine learning technology. This step-by-step detection method ensures that each part is fully inspected. First check the top, and if there is no defect, check the side. This process improves the detection efficiency and ensures that only the mouth covers that are truly defect-free can enter the market, thereby improving product quality and consumer satisfaction.

[0012] In combination with some embodiments of the first aspect, in some embodiments, the screening unit includes a cylinder assembly, which is arranged at the side of the conveyor belt. The cylinder assembly is used to extend the cylinder in the direction of the conveyor belt when receiving a screening instruction, so as to push the defective product cover from the conveyor belt to the defective product area.

[0013] By adopting the technical solution of the above embodiment, the screening unit can accurately push the defective caps from the conveyor belt to the defective area through the setting of the cylinder assembly. This automated screening operation reduces manual intervention, improves screening speed and accuracy, ensures that the defective caps can be quickly identified and isolated, avoids them from being mixed with qualified products, and improves production efficiency and product quality.

[0014] In combination with some embodiments of the first aspect, in some embodiments, a feeder and a material sorting vibration plate are further included, the feeder is used to feed the special-shaped lids to the material sorting vibration plate, and the material sorting vibration plate is used to arrange the special-shaped lids into a directional arrangement through vibration.

[0015] By adopting the technical solution of the above embodiment, the system can arrange the special-shaped lids into a directional arrangement through the coordinated use of the feeder and the material arrangement vibration plate, which provides convenience for subsequent detection. The lids arranged in a directional manner are more easily captured by the camera with clear images, thereby improving the accuracy and efficiency of detection, and also reducing misjudgments caused by incorrect lid positions.

[0016] In combination with some embodiments of the first aspect, in some embodiments, a material handling elevator is further included to adjust the height of the material handling vibration plate so that the height of the material handling vibration plate matches the conveyor belt.

[0017] By adopting the technical solution of the above embodiment, the material handling elevator can adjust the height of the material handling vibration plate to match the conveyor belt. This height adjustment mechanism ensures a smooth transition of the flaps from the vibration plate to the conveyor belt. This design improves the conveying efficiency of the flaps, reduces flap damage or detection errors caused by height mismatch, and further improves the stability and reliability of the entire detection system.

[0018] In combination with some embodiments of the first aspect, in some embodiments, a mouth cover identification mark is provided on the surface of the special-shaped mouth cover, and the control processing unit is further used to identify the mouth cover identification mark from the appearance image through image processing and recognition technology, determine the mouth cover model of the special-shaped mouth cover according to the mouth cover identification mark, determine the mouth cover standard image according to the mouth cover model, and identify and compare the appearance image with the mouth cover standard image through a preset special-shaped mouth cover defect comparison model to determine whether the special-shaped mouth cover has defects.

[0019] By adopting the technical solution of the above embodiment, the system can determine the flap model by identifying the flap identification mark, and compare it with the standard image accordingly. This identification-based detection method improves the pertinence and accuracy of defect detection. Each flap model has a corresponding standard image of a special-shaped flap, making the comparison more accurate, thereby more effectively identifying flaps that do not meet the standards and ensuring product quality.

[0020] In a second aspect, the present invention provides a method for detecting defects in a special-shaped cover, which is applied to the special-shaped cover defect detection system provided by any embodiment of the first aspect above, and the method comprises: obtaining an appearance image of the special-shaped cover to be detected through a camera; recognizing the appearance image through a preset special-shaped cover defect recognition model to determine whether the special-shaped cover has an abnormal structure other than the special-shaped structure, the abnormal structure is used to indicate that the special-shaped cover has defective features, and the special-shaped cover defect recognition model is obtained by training the images of defective and non-defective special-shaped covers as sample data through a machine learning algorithm; if so, the special-shaped cover is determined to be a defective cover, and a screening instruction is sent to a screening unit, and the screening instruction is used to control the screening unit to perform a screening operation on the defective cover, so that the defective cover enters a defective area.

[0021] The method of the present invention improves the efficiency and accuracy of the detection process of special-shaped flaps by machine learning and image processing technology. The method uses a camera to obtain the appearance image of the flap, and the control processing unit analyzes and judges, thereby identifying defective defective flaps. The present invention can automatically and accurately detect the defects of special-shaped flaps, and can also identify and distinguish whether the special-shaped flaps are qualified when the structure of the special-shaped flaps is complex or irregular, thereby improving production efficiency and product quality.

[0022] In combination with some embodiments of the second aspect, in some embodiments, the camera includes a top camera and a side camera, and the appearance image includes a top image of the mouth cover and a side image of the mouth cover; the step of recognizing the appearance image through a preset special-shaped mouth cover defect recognition model to determine whether the special-shaped mouth cover has an abnormal structure other than the special-shaped structure includes: recognizing the top image of the mouth cover through the special-shaped mouth cover defect recognition model to determine whether the top of the special-shaped mouth cover has an abnormal structure other than the special-shaped structure; if not, recognizing the side image of the mouth cover through the special-shaped mouth cover defect recognition model to determine whether the side of the special-shaped mouth cover has an abnormal structure other than the special-shaped structure.

[0023] By adopting the technical solution of the above embodiment, the system can inspect each part of the special-shaped cover more carefully through the step-by-step analysis of the top and side images. This step-by-step inspection method improves the accuracy of defect detection. First check the top, and if there is no defect, check the side. This process not only improves the inspection efficiency, but also ensures that only covers without defects can enter the market, thereby improving the reliability of the product.

[0024] In combination with some embodiments of the second aspect, in some embodiments, the step of judging whether the special-shaped mouth cover has defects according to the appearance image includes: identifying a mouth cover identification mark of the special-shaped mouth cover from the appearance image through image processing and recognition technology; determining the mouth cover model of the special-shaped mouth cover according to the mouth cover identification mark; determining the mouth cover standard image according to the mouth cover model; and identifying and comparing the appearance image with the mouth cover standard image through a preset special-shaped mouth cover defect comparison model to judge whether the special-shaped mouth cover has defects.

[0025] By adopting the technical solution of the above embodiment, the system can realize accurate defect detection by identifying the flap identification mark and comparing the flap standard image. This method not only improves the accuracy of detection, but also makes the detection process more standardized and automated. By comparing with the standard image, the system can quickly identify any deviation, ensuring that only flaps that meet the quality standards can pass the inspection, thereby improving the market competitiveness of the product.

[0026] One or more technical solutions provided by the present invention have at least the following technical effects or advantages: 1. Using machine learning technology, the accuracy and efficiency of special-shaped flap defect detection are improved through the special-shaped flap defect recognition model. The system uses top and side cameras to capture all-round images of the flap, combined with the image processing technology of the control processing unit, it can accurately identify flap defects from different perspectives. This multi-angle, automated detection method not only reduces the subjectivity of manual detection, but also improves the detection speed, ensuring that each flap can be quickly and accurately evaluated, thereby improving production efficiency and product quality.

[0027] 2. An automated quality control process is implemented. By controlling the processing unit to analyze the appearance image, the system can automatically identify the identification mark on the flap, determine the flap model, and compare it with the preset standard image to determine whether there is a defect. This process does not require human intervention, reduces human errors, and improves the consistency and reliability of detection. The automated sending and execution of screening instructions further ensures that defective flaps can be quickly isolated, preventing unqualified products from entering the market and ensuring the final quality of the product.

[0028] 3. By introducing the feeder, the material sorting vibration plate and the material sorting elevator, the entire inspection system is made more flexible and adaptable. The material sorting vibration plate can arrange the lids into a directional arrangement, while the material sorting elevator can adjust the height of the vibration plate or the conveyor belt to match the height of the two, ensuring the smooth transition and directional transportation of the lids. This design enables the system to adapt to special-shaped lids of different shapes and sizes, improving the versatility and flexibility of the production line, while also facilitating the subsequent inspection and packaging processes, further improving the overall production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings herein are incorporated into and constitute a part of the specification, showing embodiments consistent with the present invention, and together with the specification, are used to explain the principles of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work. In the drawings: Figure 1 It is a top view of a defect detection system for a special-shaped cover according to an embodiment of the present invention; Figure 2 is a connection relationship diagram of the control processing unit in an embodiment of the present invention; Figure 3 is a schematic diagram of a special-shaped structure and an abnormal structure in an embodiment of the present invention; Figure 4 It is a schematic flow chart of a method for detecting defects of a special-shaped mouth cover according to an embodiment of the present invention; Figure 5 It is a schematic diagram of the architecture of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] The terms used in the following embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to be limiting of the present invention. As used in the specification of the present invention, the singular expressions "a", "a", "above", "the" and "this" are intended to also include plural expressions, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present invention refers to any or all possible combinations comprising one or more of the listed items.

[0031] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood as implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, unless otherwise specified, the meaning of "plurality" is two or more.

[0032] It should also be noted that, unless otherwise clearly specified and limited, in the embodiments of the present invention, the terms such as "setting" and "connection" should be understood in a broad sense. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two components; it can be a wired communication connection or a wireless communication connection. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. The embodiments of the present invention are described in detail below.

[0033] The embodiment of the present invention provides a defect detection system for special-shaped lids (hereinafter referred to as the system), such as Figure 1 and Figure 2 As shown, it includes a conveyor belt 101 and a detection mechanism 110. The conveyor belt 101 is used to transmit the special-shaped cover to be detected to the detection mechanism 110. The detection mechanism 110 includes a control processing unit 111, a camera 112 and a screening unit 113; the camera 112 is used to take the appearance image of the special-shaped cover, and send the appearance image to the control processing unit 111; the control processing unit 111 is used to identify the appearance image through a preset special-shaped cover defect recognition model to determine whether the special-shaped cover has an abnormal structure other than the special-shaped structure. If the abnormal structure exists, the special-shaped cover is determined to be a defective defective cover, and a screening instruction is sent to the screening unit. The abnormal structure is used to indicate that the special-shaped cover has defective features, and the screening instruction is sent to the screening unit 113; the screening unit 113 is used to perform a screening operation on the defective cover according to the screening instruction, so that the defective cover enters the defective area.

[0034] The workflow of the system of this embodiment is as follows: the special-shaped cover is first transported to the inspection area through the conveyor belt 101, and the camera 112 captures its appearance image and sends it to the control processing unit 111. The control processing unit 111 is used to identify the appearance image through the preset special-shaped cover defect recognition model to determine whether the special-shaped cover has an abnormal structure other than the special-shaped structure, and then determine whether the special-shaped cover has a defect. If a defect is detected, the control processing unit 111 will send a screening instruction to the screening unit 113. The screening unit 113 then performs a screening operation to separate the defective cover from the normal product and guide it to the defective area for further processing or recycling.

[0035] The system of this embodiment uses a high-precision camera 112 to capture multi-angle appearance images of special-shaped flaps, and combines the special-shaped flap defect recognition model integrated in the control processing unit 111 to solve the problem of intuitively and accurately judging defects caused by abnormal structures of special-shaped flaps. The system uses an automated conveyor belt and a high-precision camera to ensure that the flaps are captured at a stable speed and angle, providing stable input data for the special-shaped flap defect recognition model. The control processing unit 111 uses machine learning technology and the special-shaped flap defect recognition model to deeply understand the complex geometric shapes of special-shaped flaps, accurately distinguish between irregular shapes in design and actual production defect characteristics, such as cracks, deformations, missing parts, etc., avoids misjudgment of defect detection, and improves the accuracy and efficiency of special-shaped flap defect detection.

[0036] At the same time, this embodiment reduces labor costs, improves the safety of the working environment, and reduces worker fatigue and errors by reducing the need for manual inspection.

[0037] like Figure 3 As shown, Figure 3 Part (a) shows the normal special-shaped structure of the special-shaped mouth cover, and part (b) shows the defective features of the special-shaped mouth cover, that is, the abnormal structure. In this embodiment, the special-shaped mouth cover itself has some irregular special-shaped structures, and these irregular special-shaped structures are normal designs for the special-shaped mouth cover. If traditional image recognition technology is used, these irregular special-shaped structures will be regarded as defects. Therefore, this embodiment uses the special-shaped mouth cover defect recognition model to identify whether the special-shaped mouth cover has defects, and will not misjudge the special-shaped structure as an abnormal structure representing defective features.

[0038] In view of the special case of irregular mouth covers, the irregular mouth cover defect recognition model adopted in this embodiment is specially designed to distinguish the irregular structure of the inherent irregular shape of the mouth cover from the actual defects, and can accurately distinguish the normal irregular structure shape from the actual defect characteristics.

[0039] In this embodiment, the defect recognition model for the special-shaped mouth cover is obtained by training the defective and non-defective special-shaped mouth cover images as sample data through a machine learning algorithm. The specific training process is as follows: 1) Data collection: Collect a large number of images of special-shaped mouth covers as sample data, including normal images of special-shaped mouth covers without defects and images of special-shaped mouth covers with various defects (such as cracks, deformation, missing, etc.). These sample data images should contain the normal special-shaped structure of the special-shaped mouth covers and possible defect features.

[0040] 2) Data preprocessing: The collected images (sample data) are preprocessed, including adjusting the image size to match the model input requirements, normalizing pixel values, enhancing contrast, removing noise, etc., to improve the training efficiency and accuracy of the model.

[0041] 3) Data enhancement: In order to increase the generalization ability of the model, the images are enhanced by performing data enhancement such as rotation, scaling, translation, flipping, adding noise, etc. to simulate different shooting conditions and angles.

[0042] 4) Annotation: Annotate the images of the irregular palate, clearly indicating which are normal palates, which irregular shapes are normal irregular structures, which irregular shapes are defective abnormal structures, and the specific types and locations of the defects. This can be achieved by professional annotators or by using semi-automatic annotation tools.

[0043] 5) Model selection: Choose a suitable machine learning model architecture, such as convolutional neural network (CNN), ResNet (Residual Network), VGG (Very Deep Convolutional Networks) and other network structures. It is preferred to use some open source machine learning model architectures.

[0044] 6) Feature extraction: Taking CNN as an example, the convolutional neural network automatically learns features from the image. The convolution layer of the model can capture the local features of the palate, while the pooling layer helps to extract more abstract features.

[0045] 7) Model training: The preprocessed and enhanced images are input into the model for training. During the training process, the model learns features through forward propagation and adjusts weights through back propagation to minimize the loss function.

[0046] 8) Hyperparameter tuning: Adjust the model’s hyperparameters, such as learning rate, batch size, number of iterations, etc., to optimize model performance.

[0047] 9) Validation and testing: Use a portion of the sample data as a validation set to monitor the performance of the model during training to avoid overfitting. After training is completed, use a portion of the sample data as an independent test set to evaluate the final performance of the model.

[0048] 10) Model optimization: Based on the results of verification and testing, the model is further optimized, which may include adjusting the network structure, adding regularization terms, using more complex optimization algorithms, etc.

[0049] 11) Deployment and monitoring: Deploy the trained model to the production environment and continuously monitor its performance to ensure its stability and accuracy in practical applications.

[0050] Through the above training method, the special-shaped cover defect recognition model can be effectively trained to accurately identify and distinguish the special-shaped structure and actual defects (abnormal structure) of the special-shaped cover, and identify which special-shaped covers are qualified and which are defective substandard covers.

[0051] In this embodiment, the control processing unit 111, as a key hardware component of the defect detection system for the special-shaped cover, may include a series of high-performance computing hardware, such as industrial-grade processors, high-speed memory modules, large-capacity storage devices, image processing acceleration cards, and interface cards and communication modules for data exchange with peripheral devices such as cameras, sensors, cylinder components, and robotic arms. In addition, it may also include a power management module, a heat dissipation system, and an industrial control motherboard to ensure stable operation. These hardware together constitute a powerful computing platform, which enables the control processing unit 111 to quickly and accurately process large amounts of data, perform complex image analysis and control logic, and thus achieve efficient and reliable defect detection and screening operations.

[0052] In some embodiments, after receiving the appearance image of the special-shaped mouth cover transmitted by the camera 112, the control processing unit 111 uses the built-in image recognition and analysis algorithms to perform in-depth processing on the image. These algorithms may include machine learning models that can identify the expected shape, size and characteristics of the mouth cover and compare them with the actual captured image. Through the set defect standards, such as irregular edges, insufficient groove depth, incomplete threads, etc., the control processing unit 111 can identify the mouth cover that deviates from the standard, and thus determine that it is a defective inferior mouth cover.

[0053] Once a defect is detected, the control processing unit 111 will immediately generate screening instructions and send these instructions to the screening unit 113 through the communication interface with the screening unit 113. After receiving the instructions, the screening unit 113 will perform corresponding selection operations to separate the defective caps from the production line and guide them to the defective area, thereby realizing automated quality control and rapid elimination of defective products.

[0054] In some embodiments, the camera 112 includes a top camera 1121 and a side camera 1122, and the appearance image includes a top image of the mouth cover and a side image of the mouth cover. The top camera 1121 is used to capture the top image of the mouth cover, and the side camera 1122 is used to capture the side image of the mouth cover.

[0055] Specifically, in the defect detection system for special-shaped flaps, the camera 112 includes a top camera 1121 and a side camera 1122, which are responsible for capturing images of the flap from different perspectives. The top camera 1121 is specifically used to obtain a top view of the flap, which is essential for detecting the center area, inner edge, and top features of the flap. The side camera 1122 is responsible for capturing a side view of the flap, which helps evaluate the side features of the flap, such as grooves, threads, and other side structures.

[0056] By combining the top and side images, the system is able to fully analyze the appearance of the flap, ensuring that potential defects are detected from multiple angles, thereby improving the accuracy and reliability of inspection. This dual-camera 112 configuration enables the system to more accurately identify and evaluate the complex shape and structure of the flap, ensuring that only defect-free flaps pass quality inspection.

[0057] Furthermore, the control processing unit 111 is specifically used to identify the top image of the special-shaped mouth cover through the special-shaped mouth cover defect recognition model to determine whether there is an abnormal structure other than the special-shaped structure on the top of the special-shaped mouth cover; if not, then identify the side image of the mouth cover through the special-shaped mouth cover defect recognition model to determine whether there is an abnormal structure other than the special-shaped structure on the side of the special-shaped mouth cover.

[0058] Specifically, the control processing unit 111 first receives and processes the top image of the flap captured by the top camera 1121, and first uses the special-shaped flap defect recognition model to analyze the top image of the flap on the conveyor belt. The model is trained to recognize and distinguish between the normal special-shaped structure of the flap top and possible abnormal structures, such as depressions, cracks or other irregularities on the top. The model extracts features and identifies defects on the top image through machine learning algorithms, especially convolutional neural networks.

[0059] If the top image analysis shows that there is no abnormal structure, that is, the top of the flap meets the design requirements, the system will guide the flap to continue moving along the conveyor belt so that the camera can capture the side image of the flap. Then, the defect recognition model is called again, this time to analyze the side image of the flap to determine whether there are abnormal structures on the side other than the normal special-shaped structure, such as scratches, deformations or other defects on the side.

[0060] Through this step-by-step inspection approach, the system is able to fully evaluate the top and sides of the flap, ensuring that potential defects are identified from multiple angles. If no abnormal structure is found on the top and sides of the flap, the model will determine that the flap is qualified and allow it to continue along the conveyor belt. On the contrary, if an abnormal structure is identified at any stage, the model will determine that the flap is defective and trigger the screening unit to separate it from the production line, thereby ensuring that only defect-free flaps enter the market. This automated inspection process not only improves the accuracy of inspection, but also significantly improves production efficiency and product quality.

[0061] In some embodiments, the screening unit 113 includes a cylinder assembly, which is arranged at the side of the conveyor belt 101. When receiving a screening instruction, the cylinder assembly is used to extend the cylinder in the direction of the conveyor belt 101 to push the defective mouth cover from the conveyor belt 101 to the defective area.

[0062] Specifically, after completing the image analysis of the irregular-shaped lid, if a defect is detected, the control processing unit 111 will generate a specific screening instruction according to the preset logic and algorithm. This instruction contains all the information required to push the defective lid from the conveyor belt 101 to the defective area, such as the type of action, force and execution time that the cylinder assembly needs to perform. The control processing unit 111 transmits this instruction to the control circuit of the cylinder assembly in real time through a digital communication interface or an industrial field bus.

[0063] After the cylinder assembly receives the command, the solenoid valve inside it will be activated to control the compressed air or liquid to flow into the cylinder, pushing the piston rod to move in the specified direction, thereby lifting the defective cover from the conveyor belt 101 and pushing it to the defective area. The entire process is precisely controlled by the control processing unit 111 to ensure that the cylinder assembly can accurately perform the screening operation and achieve efficient and accurate quality control.

[0064] In some embodiments, the screening unit 113 may also include a robotic arm that can perform precise operations in more complex scenarios. For example, when a defective flap is detected, the robotic arm can flexibly adjust its position and posture, accurately grab or push the defective flap, and transfer it from the conveyor belt 101 to the defective area. This screening unit 113 with an integrated robotic arm not only improves the accuracy and efficiency of screening, but also can adapt to flaps of different shapes and sizes, and handle more diverse defect types, thereby improving the performance and adaptability of the entire detection system.

[0065] In some embodiments, the system further includes a feeder 102 and a material sorting vibration plate 103, wherein the feeder 102 is used to feed the special-shaped lids to the material sorting vibration plate 103, and the material sorting vibration plate 103 is used to arrange the special-shaped lids into a directional arrangement through vibration.

[0066] Specifically, the feeder 102 is responsible for continuously and stably feeding the special-shaped lids to be inspected from the packaging or storage area into the system, ensuring that the lids can smoothly enter the material sorting vibration plate 103 .

[0067] The material sorting vibration plate 103 uses vibration to sort the special-shaped caps, and by precisely controlling the vibration frequency and amplitude, the special-shaped caps are arranged in a predetermined direction and order, thus achieving directional arrangement. This directional arrangement is crucial for the subsequent inspection process, because it ensures that the key features of the special-shaped caps can be accurately captured by the camera 112 when they pass through the inspection mechanism 110, thereby improving the accuracy and efficiency of the inspection.

[0068] Through this automated feeding and sorting process, the system can achieve efficient and continuous special-shaped lid inspection operations, while reducing manual intervention and improving overall production efficiency and quality control levels.

[0069] In this implementation, the reason why the material handling vibration disk 103 can achieve the directional arrangement of special-shaped flaps by precisely controlling the vibration frequency and amplitude is because it utilizes the influence of vibration force on the movement of objects. When the material handling vibration disk 103 is started, the special-shaped flaps on the disk surface will roll or slide with the vibration. By adjusting the vibration frequency, the rate at which the flaps roll can be controlled, while adjusting the vibration amplitude can affect the distance and force of the flaps' movement. By accurately setting these parameters, the flaps can move along a specific trajectory on the material handling vibration disk 103, thereby achieving orderly arrangement of the flaps. For example, a lower frequency may cause the flaps to roll slowly, while an appropriate amplitude ensures that the flaps can move a certain distance in a predetermined direction after each roll, ultimately allowing all flaps to be arranged on the material handling vibration disk 103 in a uniform direction and order, providing convenience for subsequent automated detection and processing procedures. This method of precisely controlling the vibration frequency and amplitude is a key technology for achieving directional arrangement of special-shaped flaps.

[0070] In some embodiments, the system further includes a material handling elevator 104 for adjusting the height of the material handling vibration plate 103 so that the height of the material handling vibration plate 103 matches the height of the conveyor belt 101 .

[0071] Among them, the main function of the material handling elevator 104 is to adjust the height of the material handling vibration plate 103 or the conveyor belt 101 to ensure that the relative positions of the material handling vibration plate 103 and the conveyor belt 101 are accurately matched. Through this height adjustment mechanism, the system can adapt to conveyor belts 101 of different heights. Whether in the installation, maintenance or production process, the position of the material handling vibration plate 103 or the conveyor belt 101 can be quickly adjusted to achieve smooth transition and transportation of the flaps. This design not only improves the adaptability and flexibility of the system, but also ensures the stability and consistency of the flaps during the transfer process from the material handling vibration plate 103 to the conveyor belt 101, thereby optimizing the efficiency and accuracy of the entire detection process.

[0072] In some embodiments, a cover identification mark is provided on the surface of the special-shaped cover, and the control processing unit 111 is specifically used to identify the cover identification mark from the appearance image, determine the cover model of the special-shaped cover according to the cover identification mark, determine the cover standard image according to the cover model, and compare the appearance image with the cover standard image to determine whether the special-shaped cover has defects.

[0073] The flap identification mark is a specific mark set on the surface of the special-shaped flap, which is used to uniquely identify the model, batch or other key information of each flap. The flap identification mark can be a one-dimensional barcode, a two-dimensional code, a digital sequence, a letter combination or a specific pattern. The flap identification mark is designed to be quickly identified during high-speed automated inspection.

[0074] The purpose of the flap identification mark is to enable the control processing unit 111 to accurately read and parse the specific information of the outlet cover from the appearance image, so as to determine the model of the flap and perform defect detection against the corresponding standard image. This identification system is crucial to ensure product quality, realize automated quality control, and track product batches. It enables special-shaped flaps of different structures to be accurately compared with their preset standards to identify and exclude any defective products that do not meet the quality standards.

[0075] Specifically, the control processing unit 111 identifies the outlet cover identification mark from the appearance image captured by the camera 112 through image processing and recognition technology. First, the appearance image is preprocessed, such as adjusting the brightness, contrast and denoising, to improve the image quality; then, the image segmentation technology is applied to separate the mouth cover identification mark from the background. Then, the feature extraction algorithm is used to identify the key information in the mark, such as a barcode, a two-dimensional code or a specific pattern. Finally, the text or digital information in the mark is parsed and interpreted through pattern recognition technology, such as OCR (optical character recognition) or template matching, so as to accurately identify the key data of the outlet cover, such as the model and batch. This process enables the control processing unit 111 to automatically and accurately extract the mouth cover identification mark from the image, providing important information for subsequent quality control and defect detection.

[0076] Subsequently, the control processing unit 111 retrieves a standard image of a corresponding mouth flap model from the database according to the determined mouth flap model. The standard mouth flap image is a pre-set, defect-free mouth flap image for identification and comparison.

[0077] Finally, the control processing unit 111 accurately identifies and compares the actually captured appearance image with the retrieved standard image of the mouth cover, and determines whether the special-shaped mouth cover has defects in shape, size or other aspects by analyzing and identifying the difference between the two.

[0078] Based on this automated recognition and comparison process, the system can quickly identify special-shaped lids with different structures and their unique special-shaped structures, and conduct targeted and precise detection, which improves the accuracy of defect detection and ensures the efficiency and reliability of the detection process, thereby effectively screening out defective lids and ensuring product quality.

[0079] In some embodiments, the control processing unit 111 receives the actual appearance image of the mouth flap captured by the camera, and retrieves the corresponding standard image of the mouth flap from the database; then, the appearance image and the corresponding standard image of the mouth flap are input into a pre-trained special-shaped mouth flap defect comparison model, which can be based on machine learning or deep learning technology and can identify and learn the normal and abnormal features of the mouth flap. The special-shaped mouth flap defect comparison model extracts features and compares and analyzes the input image, and identifies any features that deviate from the standard by comparing the difference between the actual appearance image and the standard image of the mouth flap, such as shape deviation, size inconsistency or surface defects; finally, based on the analysis results of the model, the control processing unit determines whether the mouth flap has defects, and makes classification decisions accordingly, and uses the results in the subsequent quality control process.

[0080] Through this intelligent image recognition processing method, the control processing unit 111 can accurately determine whether the special-shaped cover has defects in shape, size or other aspects, which greatly improves the speed and accuracy of defect detection.

[0081] Among them, the training method of the special-shaped mouth cover defect comparison model includes the following process: 1) Dataset construction: Collect images of special-shaped palpebral flaps, including images of standard palpebral flaps without defects and images of palpebral flaps with defects. Make sure the dataset contains various shapes, sizes, and defect types to improve the generalization ability of the model.

[0082] 2) Image pair preparation: Prepare one or more actually captured palate images for each standard palate image, which will be used to train the model to identify and compare normal and abnormal features.

[0083] 3) Image annotation and pairing: Annotate the image to indicate the type and location of the defect. Pair the standard palate image with the actual captured palate image to form a training data pair.

[0084] 4) Data preprocessing: Preprocess the images, including resizing, normalization, contrast enhancement, noise removal, etc., to improve the training efficiency and accuracy of the model.

[0085] 5) Feature Extraction: Design or select a feature extraction network, such as a convolutional neural network (CNN), to extract key features from the image.

[0086] 6) Defect contrast model design: Build a defect contrast model that can accept a pair of images (a standard palpebral image and an actual captured palpebral image) as input and output the difference between the two.

[0087] 7) Model training: The model is trained using paired image pairs to learn to identify differences by comparing the features of the actual captured palate images with the standard palate images. During training, the model optimizes parameters by minimizing the difference loss function (such as L1 or L2 loss).

[0088] 8) Transfer learning: If the available training data is limited, transfer learning can be used, that is, pre-training the model on a large-scale dataset and then fine-tuning it based on it.

[0089] 9) Hyperparameter tuning: Adjust the model’s hyperparameters, such as learning rate, batch size, number of iterations, etc., to optimize model performance.

[0090] 10) Model evaluation: Evaluate the performance of the model on an independent test set, using indicators such as accuracy, recall, and F1 score to measure the recognition ability of the model.

[0091] 11) Error analysis and feedback: Analyze samples with model prediction errors to understand the model’s weaknesses, further adjust the model, or collect more relevant data for retraining.

[0092] 12) Model optimization: Based on the evaluation results, the model is further optimized, which may include adjusting the network structure, adding regularization terms, using more complex optimization algorithms, etc.

[0093] 13) Model deployment: Deploy the trained model to the production environment and continuously monitor its performance to ensure its stability and accuracy in practical applications.

[0094] Through these steps, the defect comparison model for special-shaped flaps can learn how to accurately compare the actual captured flap images with the standard flap images, and by analyzing the differences between the two, determine whether the special-shaped flaps have defects in shape, size, or other aspects. This enables the model to achieve efficient and accurate quality control on the production line.

[0095] The defect detection system for special-shaped lids of this embodiment provides an efficient, automated, and intelligent quality control solution, which is designed specifically for detecting lids with complex shapes and structures. The system integrates a variety of high-precision components and image processing technology based on machine learning to achieve comprehensive detection of special-shaped lids, ensuring the sealing performance and hygienic safety of the products.

[0096] The system includes a conveyor belt 101, a detection mechanism 110, and a screening unit 113. The conveyor belt 101 is responsible for conveying the special-shaped mouth cover to be detected to the detection area, and the detection mechanism 110 is composed of a control processing unit 111, a camera 112, and a screening unit 113. The camera 112 captures the appearance images of the mouth cover from different angles, including top and side images, and provides detailed visual information to the control processing unit 111. The control processing unit 111 uses these images, combined with the mouth cover identification mark, to identify the outlet cover model, and compares it with the standard image stored in the database to determine whether the mouth cover has defects.

[0097] The flap identification mark is a specific mark set on the surface of the flap, which contains key information such as model and batch, allowing the system to accurately identify and match the standard image of each flap. This identification system not only improves the accuracy of detection, but also enhances the traceability of the product.

[0098] The screening unit 113 includes a cylinder assembly and a possible robot arm, which, after receiving the screening instruction from the control processing unit 111, will accurately push the defective lids from the conveyor belt 101 to the defective area, or remove them through the precise operation of the robot arm. In addition, the system may also include a material sorting vibration plate 103 and a material sorting elevator 104, which are used to sort the lids and adjust the height of the vibration plate to ensure that the lids match the height of the conveyor belt 101, so as to achieve smooth transportation and inspection.

[0099] The design of the entire system takes into account the irregular shape and special structure of special-shaped flaps, making it adaptable to various complex and changeable flap designs. Through the automated inspection process, enterprises can reduce the cost and error of manual inspection, while improving product quality and enhancing market competitiveness. The system can also be customized according to user needs to adapt to different inspection standards and environments, showing strong adaptability and scalability.

[0100] In summary, the defect detection system for special-shaped flaps provides an efficient, reliable and cost-effective solution for the flap manufacturing industry through technological innovation. It not only improves the accuracy and stability of detection, but also reduces the cost of enterprises and enhances the market competitiveness of products by reducing the production of defective products. With the continuous advancement and innovation of technology, the system is expected to achieve more intelligent and automated detection in the future, contributing greater value to the development of the flap manufacturing industry.

[0101] The present invention provides a method for detecting defects in a special-shaped mouth cover, which is applied to the special-shaped mouth cover defect detection system provided in the above embodiment. Figure 4 As shown, the method comprises the following steps: Step 201, obtaining an appearance image of a special-shaped mouth cover to be inspected through a camera.

[0102] Specifically, the control processing unit starts the image capture function of the camera through an interface connection with the camera, and the camera is then aimed at the special-shaped mouth cover to be inspected, and captures the detailed appearance image of the mouth cover through its high-resolution lens.

[0103] Step 202, identifying the appearance image by using a preset special-shaped mouth cover defect recognition model to determine whether the special-shaped mouth cover has any abnormal structure other than the special-shaped structure.

[0104] Among them, the abnormal structure is used to indicate the defective features of the special-shaped mouth cover. The special-shaped mouth cover defect recognition model is obtained by training the defective and non-defective special-shaped mouth cover images as sample data through a machine learning algorithm. The specific training process has been described above and will not be repeated here.

[0105] Specifically, after receiving the appearance image of the special-shaped mouth cover transmitted by the camera, the control processing unit uses the preset special-shaped mouth cover defect recognition model to perform in-depth recognition analysis on the image. It first pre-processes the image, including denoising, contrast enhancement and brightness adjustment, to improve image quality. Then, the defect recognition model analyzes the image and identifies the geometric shape and surface features of the outlet cover through the feature extraction and classification capabilities learned during its training process. The model checks whether the mouth cover has abnormal structures that are inconsistent with its normal special-shaped structure, such as additional protrusions, depressions, cracks or other irregularities.

[0106] If the model identifies these abnormal structures, it will determine that the flap is defective. These abnormal structures are indicator features of defects. They are different from normal design features and indicate that there may be problems with the flap during production. Once the model determines that the flap is defective, the control processing unit sends a screening instruction to the screening unit to separate the defective flaps from the production line, ensuring that only non-defective flaps continue to enter the subsequent production process. This automated inspection process not only improves the accuracy of inspection, but also significantly improves production efficiency.

[0107] Through the above-mentioned image recognition and processing technologies, the control processing unit can identify any abnormal features that deviate from the standard, such as inconsistent size, deformed shape or surface flaws, and thus determine whether the special-shaped mouth cover has defects.

[0108] If the special-shaped mouth cover has an abnormal structure, the process proceeds to step 203. If the special-shaped mouth cover does not have an abnormal structure, the special-shaped mouth cover is determined to be a qualified mouth cover.

[0109] Step 203: determine the special-shaped mouth cover as a defective mouth cover, and send a screening instruction to the screening unit.

[0110] The screening instructions are used to control the screening unit to perform screening operations on the defective caps, so that the defective caps enter the defective area. The screening instructions contain all the information required to perform the screening operation, such as the location of the defective caps, the actions to be performed, etc. The control processing unit sends the screening instructions to the screening unit, such as the cylinder assembly or the robotic arm, through the internal communication network, instructing it to perform specific physical operations, separate the defective caps from the production line and guide them to the designated defective area for subsequent processing and analysis.

[0111] This embodiment uses a high-precision camera to capture multi-angle appearance images of special-shaped lids, and combines the advanced image recognition and processing technology integrated in the control processing unit to solve the problem of intuitively and accurately judging defects caused by abnormal structures of special-shaped lids. By using machine learning algorithms and special-shaped lid defect recognition models, it is possible to deeply understand the complex geometric shapes of special-shaped lids, accurately distinguish between irregular shapes in design and actual production defect characteristics, such as cracks, deformations, missing parts, etc., and avoid misjudgment of defect detection. In this way, defects of special-shaped lids can be automatically and accurately detected, and even if the structure of the special-shaped lids is complex or irregular, it is possible to identify and distinguish whether the special-shaped lids are qualified, thereby improving production efficiency and product quality.

[0112] In some embodiments, the camera includes a top camera and a side camera, and the appearance image includes a mouth cover top image and a mouth cover side image. Step 202 specifically includes: The top image of the mouth cover is identified by the defect recognition model of the special-shaped mouth cover to determine whether there is an abnormal structure other than the special-shaped structure on the top of the special-shaped mouth cover; If not, the side image of the mouth cover is identified by the special-shaped mouth cover defect recognition model to determine whether there is any abnormal structure other than the special-shaped structure on the side of the special-shaped mouth cover.

[0113] Specifically, when the control processing unit detects the special-shaped flaps, it first analyzes the top image of the flaps on the conveyor belt using the special-shaped flap defect recognition model. The model is trained to recognize and distinguish between the normal special-shaped structure of the flap top and possible abnormal structures, such as depressions, cracks or other irregularities on the top. The model extracts features and identifies defects from the top image through deep learning algorithms, especially convolutional neural networks.

[0114] If the top image analysis shows that there is no abnormal structure, that is, the top of the flap meets the design requirements, the system will guide the flap to continue moving along the conveyor belt so that the camera can capture the side image of the flap. Then, the special-shaped flap defect recognition model is called again, this time to analyze the side image of the flap to determine whether there are abnormal structures other than normal special-shaped structures on the side, such as scratches, deformations or other defects on the side.

[0115] Through this step-by-step inspection approach, the system is able to fully evaluate the top and sides of the flap, ensuring that potential defects are identified from multiple angles. If no abnormal structure is found on the top and sides of the flap, the model will determine that the flap is qualified and allow it to continue along the conveyor belt. On the contrary, if an abnormal structure is identified at any stage, the model will determine that the flap is defective and trigger the screening unit to separate it from the production line, thereby ensuring that only defect-free flaps enter the market. This automated inspection process not only improves the accuracy of inspection, but also significantly improves production efficiency and product quality.

[0116] In some embodiments, step 202 may include the following steps: (1) Identify the cover identification mark of the special-shaped cover from the appearance image through image processing and recognition technology.

[0117] Specifically, the control processing unit first performs the image recognition task by analyzing the appearance image of the special-shaped cover captured by the camera and extracting the cover identification mark from the image using image processing technology. This step involves technologies such as image segmentation, feature extraction and pattern recognition to ensure accurate identification of the outlet cover identification mark from a complex background.

[0118] (2) Determine the cover model of the special-shaped cover according to the cover identification mark.

[0119] Specifically, once the identification mark is successfully extracted, the control processing unit will determine the specific model of the special-shaped outlet cover based on the information in the mark, such as barcode, QR code, digital code and graphic pattern. This step involves decoding the identification information and matching it with the model information in the database to identify the exact model of the outlet cover.

[0120] (3) Determine the standard image of the flap according to the flap model.

[0121] Specifically, after the flap model is determined, the control processing unit will access the built-in or externally connected database to retrieve the corresponding flap standard images according to the identified model. These standard images are pre-defined, defect-free flap images for subsequent comparison and analysis.

[0122] (4) The appearance image is identified and compared with the standard image of the mouth cover through the preset special-shaped mouth cover defect comparison model to determine whether the special-shaped mouth cover has defects.

[0123] Specifically, the control processing unit compares the actual captured appearance image with the retrieved standard image in detail. This step uses image comparison algorithms, such as feature matching or pixel-level comparison, to identify any differences between the two, thereby determining whether the special-shaped cover has defects. Any features that deviate from the standard image may indicate a defect, and the control processing unit makes a final quality judgment based on this.

[0124] The method of this embodiment achieves efficient and accurate detection of special-shaped flaps through systematic steps. First, the automated flap identification mark extraction and model determination process reduces manual intervention, reduces the risk of human error, and improves the speed and consistency of detection. Secondly, the image recognition technology based on machine learning is used to ensure the accuracy of detection through comparison and analysis of standard images, so that even special-shaped flaps with complex structures can be accurately evaluated, thereby effectively identifying potential defects. This method not only improves the quality and reliability of the product and prevents unqualified products from entering the market, but also reduces production costs and enhances the market competitiveness of the company by optimizing the detection process. In addition, real-time data feedback and analysis capabilities help to continuously improve the production process and promote the improvement of overall production efficiency.

[0125] The method of this embodiment, applying the system provided by the above embodiment, has the same technical effect as the above system, and will not be described in detail here.

[0126] The method provided in the above embodiment can be executed by a control processing unit, which is an electronic device. The following describes the electronic device in the embodiment of the present invention from the perspective of hardware processing. Figure 5 , is a schematic diagram of a physical device structure of an electronic device in an embodiment of the present invention.

[0127] It should be noted that Figure 5 The structure of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0128] like Figure 5 As shown, the electronic device includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 to the random access memory (RAM) 403, such as executing the method described in the above embodiment. In RAM 403, various programs and data required for system operation are also stored. CPU 401, ROM 402 and RAM 403 are connected to each other through bus 404. Input / output (I / O) interface 405 is also connected to bus 404.

[0129] The following components are connected to the I / O interface 405: an input section 406 including an audio input device, a button switch, etc.; an output section 407 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed so that a computer program read therefrom is installed into the storage section 408 as needed.

[0130] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 409, and / or installed from a removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, various functions defined in the present invention are performed.

[0131] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0132] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0133] Specifically, the electronic device of this embodiment includes a processor and a memory, the memory is coupled to one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the electronic device to execute the method provided by the above embodiment.

[0134] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the electronic device, the electronic device implements the method provided in the above embodiment.

[0135] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

[0136] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0137] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A defect detection system for special-shaped lids, characterized in that: It includes a conveyor belt and a detection mechanism, wherein the conveyor belt is used to transmit the special-shaped mouth cover to be detected to the detection mechanism, and the detection mechanism includes a control processing unit, a camera and a screening unit, and the special-shaped mouth cover is formed with a normal special-shaped structure; The camera is used to capture the appearance image of the special-shaped cover and send the appearance image to the control processing unit; The control processing unit is used to identify the appearance image through a preset special-shaped cover defect recognition model to determine whether the special-shaped cover has an abnormal structure other than the special-shaped structure. If the abnormal structure exists, the special-shaped cover is determined to be a defective inferior cover, and a screening instruction is sent to the screening unit. The abnormal structure is used to indicate that the special-shaped cover has defective features. The special-shaped cover defect recognition model is obtained by training the defective and non-defective special-shaped cover images as sample data through a machine learning algorithm. The screening unit is used to perform a screening operation on the defective flaps according to the screening instruction, so that the defective flaps enter the defective area.

2. The defect detection system for special-shaped lids according to claim 1, characterized in that: The camera comprises a top camera and a side camera, and the appearance image comprises a top image of the mouth cover and a side image of the mouth cover. The top camera is used to capture the top image of the mouth cover, and the side camera is used to capture the side image of the mouth cover.

3. The defect detection system for special-shaped lids according to claim 2, characterized in that: The control processing unit is specifically configured to identify the top image of the mouth cover through the special-shaped mouth cover defect recognition model to determine whether the top of the special-shaped mouth cover has the abnormal structure other than the special-shaped structure; if not, identify the side image of the mouth cover through the special-shaped mouth cover defect recognition model to determine whether the side of the special-shaped mouth cover has the abnormal structure other than the special-shaped structure.

4. The defect detection system for special-shaped lids according to claim 1, characterized in that: The screening unit includes a cylinder assembly, which is arranged at the side of the conveyor belt. When receiving the screening instruction, the cylinder assembly is used to extend the cylinder in the direction of the conveyor belt to push the defective product cover from the conveyor belt to the defective product area.

5. The defect detection system for special-shaped lids according to claim 1, characterized in that: It also includes a feeder and a material arrangement vibration plate. The feeder is used to feed the special-shaped mouth covers to the material arrangement vibration plate. The material arrangement vibration plate is used to arrange the special-shaped mouth covers into a directional arrangement state through vibration.

6. The defect detection system for special-shaped lids according to claim 5, characterized in that: It also includes a material handling elevator, which is used to adjust the height of the material handling vibration plate so that the height of the material handling vibration plate matches the conveyor belt.

7. The defect detection system for special-shaped lids according to any one of claims 1 to 6, characterized in that: A flap identification mark is provided on the surface of the special-shaped flap, and the control processing unit is further used to identify the flap identification mark from the appearance image through image processing and recognition technology, determine the flap model of the special-shaped flap according to the flap identification mark, determine the flap standard image according to the flap model, and identify and compare the appearance image with the flap standard image through a preset special-shaped flap defect comparison model to determine whether the special-shaped flap has defects.

8. A method for detecting defects of special-shaped lids, characterized in that: The method applied to the irregular-shaped cover defect detection system according to any one of claims 1 to 7 comprises: Acquire the appearance image of the special-shaped cover to be inspected through a camera; The appearance image is identified by a preset special-shaped mouth cover defect recognition model to determine whether the special-shaped mouth cover has an abnormal structure other than the special-shaped structure, wherein the abnormal structure is used to indicate that the special-shaped mouth cover has defective features, and the special-shaped mouth cover defect recognition model is obtained by training the defective and non-defective special-shaped mouth cover images as sample data through a machine learning algorithm; If so, the special-shaped cover is determined as a defective cover, and a screening instruction is sent to the screening unit, wherein the screening instruction is used to control the screening unit to perform a screening operation on the defective cover so that the defective cover enters a defective area.

9. The method for detecting defects of special-shaped lids according to claim 8, characterized in that: The camera includes a top camera and a side camera, and the appearance image includes a mouth cover top image and a mouth cover side image; The step of identifying the appearance image by using a preset special-shaped mouth cover defect recognition model to determine whether the special-shaped mouth cover has an abnormal structure other than the special-shaped structure, comprises: Identify the top image of the mouth cover by using the special-shaped mouth cover defect recognition model to determine whether the top of the special-shaped mouth cover has the abnormal structure other than the special-shaped structure; If not, the side image of the mouth cover is identified by the special-shaped mouth cover defect recognition model to determine whether the side of the special-shaped mouth cover has the abnormal structure other than the special-shaped structure.

10. The method for detecting defects of special-shaped lids according to claim 8, characterized in that: After the step of acquiring the appearance image of the special-shaped cover to be inspected by the camera, the method further includes: Identify the cover identification mark of the special-shaped cover from the appearance image through image processing and recognition technology; Determining the cover model of the special-shaped cover according to the cover identification mark; Determine a standard image of the mouth cover according to the mouth cover model; The appearance image is identified and compared with the mouth cover standard image through a preset special-shaped mouth cover defect comparison model to determine whether the special-shaped mouth cover has defects.

Images