Equal-part quantitative automatic packaging method and system for toy parts

Abstract

The invention relates to an equal-part quantitative automatic packaging method and system for toy parts. The equal-part quantitative automatic packaging method comprises the following steps that all the toy parts are conveyed into a conveying belt in a scattered and tiled mode through vibration of a vibration disc; a CCD camera is adopted to shoot the toy parts on the conveying belt; the obtained part image is preprocessed; a mechanical vision module is adopted to carry out part type identification on the preprocessed image; classifying the identified part types; generating a classification signal based on the identification classification result; a mechanical arm module is controlled to grab toy parts according to the target category, and the toy parts are put into a quantitative disc; after the weight reaches a set standard, the parts in the quantitative disc are emptied, packaged and sent out; according to the equal-part quantitative automatic packaging method and system for the toy parts, the accuracy, efficiency and flexibility of automatic packaging are effectively improved, and multiple technical defects in a traditional toy part packaging method are overcome.

Description

Technical Field

[0001] The invention belongs to the technical field of classification and packaging, and in particular relates to a method and a system for automatically packaging equal-portion quantitative toy parts. Background Art

[0002] In the toy manufacturing industry, the automatic packaging process of parts is crucial and directly affects production efficiency and product quality. The currently widely used toy parts packaging technology has the following major defects:

[0003] More manual intervention: Existing packaging systems often rely on manual sorting, measurement and packaging of parts, resulting in low production efficiency. This manual intervention not only increases labor intensity, but also easily introduces human errors, resulting in inconsistent packaging quantity and type, thus affecting the overall quality of the product.

[0004] Insufficient recognition accuracy: Many traditional automatic packaging systems use simple sensors or mechanical devices to identify and classify parts. Due to technical limitations, these systems are often not accurate enough in identifying part types and are prone to classification errors, resulting in confusion and rework in subsequent packaging links.

[0005] Lack of real-time monitoring and feedback: The existing system lacks real-time monitoring capabilities during the parts weighing and sorting process, and can often only perform quality checks after the entire packaging is completed. This lagging feedback mechanism makes it difficult to detect problems in a timely manner, affects the consistency and accuracy of packaging, and increases the risk of defective products.

[0006] Low production efficiency: The existing packaging process is often limited by the operating speed of the equipment and the efficiency of manual operation, and cannot achieve efficient and continuous production. This restricts the overall production efficiency and cannot meet the rapidly growing market demand.

[0007] In summary, the existing toy parts packaging technology has significant defects in efficiency, accuracy, flexibility and resource utilization. These defects not only affect the production efficiency and product quality of enterprises, but also hinder the development of enterprises in the fierce market competition. Therefore, a new automatic packaging method is urgently needed to solve these problems in order to improve the packaging efficiency and quality of toy parts. Summary of the invention

[0008] In order to overcome the deficiencies and shortcomings of the prior art, the primary purpose of the present invention is to provide a method and system for automatic packaging of toy parts in equal portions and in quantitative quantities. The method and system for automatic packaging of toy parts in equal portions and in quantitative quantities effectively improve the accuracy, efficiency and flexibility of automatic packaging, and solve many technical defects in traditional toy parts packaging methods.

[0009] The primary purpose of the present invention is achieved through the following technical solutions:

[0010] A method for automatically packaging equal-quantity toy parts comprises the following steps:

[0011] Pour the toy parts into the vibration plate, and send the toy parts to the conveyor belt in a scattered and flat manner through the vibration of the vibration plate;

[0012] Use a CCD camera to shoot toy parts on the conveyor belt to obtain part images;

[0013] Preprocessing the acquired part images;

[0014] Use machine vision module to identify part types from pre-processed images;

[0015] classifying the identified part type;

[0016] Generate a classification signal based on the recognition and classification results, indicating the category information of each part;

[0017] receiving the classification signal, controlling the robot arm module to grab toy parts according to the target category, and placing them into the quantitative tray;

[0018] The pressure sensor under the quantitative tray is used for real-time weighing. When the weight reaches the set standard, the parts in the quantitative tray are emptied, packaged and sent out.

[0019] Preferably, a CCD camera is used to photograph toy parts on a conveyor belt, and a method for obtaining images of the parts is as follows:

[0020] An LED lighting device with adjustable brightness is set on one side of the CCD camera to take photos of the toy parts on the conveyor belt;

[0021] A photoelectric sensor is set at the feeding port of the conveyor belt. When the parts pass through the photoelectric sensor, the CCD camera is triggered to collect images.

[0022] Preferably, the method for preprocessing the acquired part image is:

[0023] Resize the image to a uniform 224*224 pixels;

[0024] Normalize the image pixel values ​​to the range [0,1];

[0025] Data enhancement technology is used to enhance the image data.

[0026] Preferably, a machine vision module is used to identify the part type of the preprocessed image;

[0027] Input the preprocessed part image into the convolutional neural network to extract feature information;

[0028] The extracted features are transformed nonlinearly through the multi-layer structure of the convolutional neural network to generate the category probability distribution of each part;

[0029] According to the generated category probability distribution, the category with the highest probability is selected as the recognition result of the part;

[0030] The recognition result is converted into a classification signal for controlling the robot arm module to grab the corresponding toy parts.

[0031] Preferably, the convolutional neural network model is constructed as follows:

[0032] Construct at least one convolutional layer, wherein the convolutional layer uses a convolution kernel of 5x5, a stride of 1, and a ReLU activation function;

[0033] Construct at least one pooling layer, wherein the pooling layer is a maximum pooling layer, and its window size is 2x2 and the stride is 2;

[0034] Construct at least one fully connected layer with the number of output units equal to the number of categories, and use the softmax activation function to output the classification probability;

[0035] Select Adam optimizer as the training algorithm, and set the initial learning rate to 0.001;

[0036] Cross entropy is used as the loss function to evaluate the classification performance of the model;

[0037] Monitor validation loss using early stopping to prevent overfitting;

[0038] Evaluate model performance on an independent test set and calculate its accuracy, precision, recall, and F1-score;

[0039] Plot confusion matrix to analyze recognition performance of different part categories;

[0040] The model is adjusted based on the evaluation results and retrained to further improve the classification performance.

[0041] Preferably, the method for classifying the identified part types is:

[0042] Input the preprocessed part image into the convolutional neural network to extract feature information;

[0043] Through the multi-layer structure of CNN, the extracted features are nonlinearly transformed to generate the category probability distribution of each part;

[0044] According to the generated category probability distribution, the category with the highest probability is selected as the recognition result of the part;

[0045] Convert the recognition results into classification signals.

[0046] Preferably, the method of receiving the classification signal, controlling the robot arm module to grab toy parts according to the target category, and placing them into the quantitative tray is:

[0047] The robot arm control system receives a classification signal from the vision system through a communication interface, the signal including the identified part type and its corresponding position information;

[0048] Analyze the information in the signal and extract the category and location information of the target part;

[0049] According to the parsed position information, determine the specific position of each type of parts on the conveyor belt;

[0050] The robot control system calculates the optimal grasping path from the current position to the target part;

[0051] According to the path planning results, a motion command is issued to control the robot arm to move above the target part;

[0052] After reaching the target position, the robot arm opens the gripper to perform the grasping;

[0053] After grabbing the parts, the robot arm moves to the top of the quantitative plate according to the preset path;

[0054] Open the gripper and release the parts into the dosing tray;

[0055] The robot continues to receive classification signals for cyclic operation, and the robot arm continuously grabs parts of different categories and puts them into the quantitative tray until the quantitative requirements are met.

[0056] Another technical problem to be solved by the present invention is to provide a toy parts equal-portion quantitative automatic packaging system, comprising:

[0057] A vibration plate module is used to receive and disperse toy parts so that they can be fed into the conveyor belt in a flat manner;

[0058] An image acquisition module, including a CCD camera, configured to photograph toy parts on a conveyor belt and acquire images of the parts;

[0059] An image processing module, used for preprocessing the part image;

[0060] A machine vision module for part type recognition based on pre-processed part images;

[0061] A classification module, used to classify the identified part types and generate classification signals representing the category information of each part;

[0062] A control system module receives the classification signal and controls the robot arm module;

[0063] A robotic arm module, which grabs toy parts of a specified category according to the instructions of the control system module and puts them into a quantitative tray;

[0064] A quantitative tray module, used to store toy parts dropped by the robot arm;

[0065] A weighing sensor module is arranged below the quantitative plate to perform real-time weighing to monitor the weight of the parts;

[0066] The packaging module is used to empty the parts in the quantitative tray and package them after the weight reaches the set standard.

[0067] Another technical problem to be solved by the present invention is to provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a method for automatic packaging of equal portions of toy parts as described above is implemented.

[0068] Another technical problem to be solved by the present invention is to provide a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the method for automatic packaging of equal-sized toy parts as described above is implemented.

[0069] Compared with the prior art, the present invention has the following advantages:

[0070] Through the automated vibration plate and robotic arm, the solution can achieve continuous and efficient parts sorting and packaging, significantly improving overall production efficiency; manual operation often leads to errors, such as inaccurate part counts or misclassification. Using CCD cameras and mechanical vision systems for part identification and classification can effectively reduce interference from human factors and improve the accuracy of part classification.

[0071] The solution uses image preprocessing and machine vision modules to ensure accurate identification and classification of various types of parts, avoiding confusion and classification errors; and uses pressure sensors for real-time weighing to ensure that the parts in the quantitative tray meet the set standards before being emptied and packaged. This real-time monitoring mechanism enables the system to quickly feedback and adjust, ensuring the accuracy of each packaging and reducing the generation of defective products.

[0072] This automatic packaging method can generate classification signals based on the identified part categories and control the robotic arm to grasp them, which means that the system can adapt to different types and quantities of parts and effectively respond to the market demand for diversified products; by reducing manual intervention, improving production efficiency and reducing material waste, this method can effectively reduce production costs in the long run, thereby enhancing the competitiveness of the enterprise.

[0073] Through the optimization of the automated packaging system, the waste of materials and human resources is reduced, which meets the requirements of sustainable development; the automated quantitative packaging method ensures that the quantity and quality of parts in each package are consistent, thereby improving the overall quality of the final product and enhancing customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 The present invention is a flow chart of a method for automatically packaging equal portions of toy parts. DETAILED DESCRIPTION

[0075] The specific implementation modes of the present invention are further described in detail to make the technical solutions of the present invention easier to understand and grasp.

[0076] Example

[0077] See also Figure 1 As shown, a method for automatically packaging equal portions of toy parts comprises the following steps:

[0078] Pour the toy parts into the vibration plate, and send the toy parts to the conveyor belt in a scattered and flat manner through the vibration of the vibration plate;

[0079] Use a CCD camera to shoot toy parts on the conveyor belt to obtain part images;

[0080] Preprocessing the acquired part images;

[0081] Use machine vision module to identify part types from pre-processed images;

[0082] classifying the identified part type;

[0083] Generate a classification signal based on the recognition and classification results, indicating the category information of each part;

[0084] receiving the classification signal, controlling the robot arm module to grab toy parts according to the target category, and placing them into the quantitative tray;

[0085] The pressure sensor under the quantitative tray is used for real-time weighing. When the weight reaches the set standard, the parts in the quantitative tray are emptied, packaged and sent out.

[0086] Through the use of automated vibration plates, conveyor belts and robotic arms, the time for manual intervention is reduced, the overall production speed is increased, and 24-hour uninterrupted production can be achieved; real-time weighing is performed using pressure sensors to ensure that the number of parts in each package is accurate, reducing human errors and improving packaging accuracy; the automated system reduces dependence on manual labor and reduces labor costs, especially in large-scale production, which can effectively reduce labor expenses.

[0087] The automated parts identification and classification process reduces quality fluctuations caused by human factors, ensures consistent quality of parts in each package, and thus improves the overall quality of the final product; the system can generate classification signals based on the identified part type, flexibly respond to toy parts of different types and sizes, and adapt to the market demand for diversified products.

[0088] By real-time monitoring of part weight and classification status, the system can quickly provide feedback and adjust the workflow to ensure smooth operation of the production line and reduce downtime. By using robotic arms and automatic equipment, it reduces potential safety hazards in manual operations and improves the safety of the working environment. The automated quantitative control system ensures the accuracy of the number of parts, avoids material waste due to excess or insufficient quantities, and improves resource utilization efficiency.

[0089] An efficient packaging system enables companies to respond more quickly to market changes and customer needs, thereby gaining an advantage in competition; the system can be upgraded and expanded as needed to support new part types and packaging requirements, and has good sustainable development; through the automation system, real-time collection and analysis of production data can be achieved, providing data support for subsequent production optimization; it reduces manual intervention and material waste, helps reduce the impact of production on the environment, and meets the requirements of modern enterprises for sustainable development.

[0090] The CCD camera is used to shoot toy parts on the conveyor belt. The method of obtaining the part image is as follows:

[0091] An LED lighting device with adjustable brightness is set on one side of the CCD camera to take photos of the toy parts on the conveyor belt;

[0092] A photoelectric sensor is set at the feeding port of the conveyor belt. When the parts pass through the photoelectric sensor, the CCD camera is triggered to collect images.

[0093] The adjustable brightness of the LED lighting device can optimize the lighting effect according to different ambient lighting conditions and part materials, thereby improving image clarity and contrast and ensuring image quality; triggering the CCD camera for image acquisition through the photoelectric sensor can ensure that the picture is taken at the moment the part passes by, avoiding missed or wrong shots, and improving the efficiency and accuracy of image acquisition; the automated design of the solution reduces manual intervention, improves the intelligence level of the entire system, and reduces the workload of operators.

[0094] Photoelectric sensors can respond quickly to the passing of parts, and CCD cameras capture images almost instantly when parts pass by, reducing acquisition delays and making them suitable for high-speed production lines. Adjustable LED lighting can be adjusted according to the color, shape, and surface characteristics of different parts, ensuring that different types of parts can achieve the best shooting effects in various environments. High-quality images help subsequent machine vision systems identify part types more accurately, reduce recognition error rates, and improve overall packaging accuracy.

[0095] Compared with traditional light sources, LED lighting has lower energy consumption and longer service life, reducing energy costs and maintenance frequency in the production process; the brightness of LED lighting and the setting of photoelectric sensors can be adjusted according to production needs, increasing the flexibility of the system and adapting to different production needs; the combination of LED lighting and photoelectric sensors reduces the noise of mechanical equipment and improves the comfort of the working environment.

[0096] The method is simple in design, making subsequent maintenance and upgrades more convenient, and technicians can easily make adjustments and troubleshoot problems; the high-quality images taken by the CCD camera can be stored and analyzed, providing data support for subsequent quality control and production optimization; the system's automated design reduces direct contact between operators and machines, reducing potential safety risks.

[0097] The method for preprocessing the acquired part image is:

[0098] Resize the image to a uniform 224*224 pixels;

[0099] Normalize the image pixel values ​​to the range [0,1];

[0100] Data enhancement technology is used to enhance the image data.

[0101] Resize the image to a uniform 224×224 pixels to ensure consistent input size, making it easier to input into the neural network for training or reasoning. This is a standard input requirement for many deep learning models, reducing the complexity caused by different input sizes; uniform size can simplify the image processing process and avoid image alignment or padding problems caused by different sizes.

[0102] Normalizing the image pixel values ​​to the range of [0,1] can improve the convergence speed of the model and reduce the difficulty of training. Normalization can prevent gradient explosion or disappearance and help the model learn faster. After normalization, the distribution of input data is more consistent, which can reduce the sensitivity to the learning rate and improve the stability and performance of the model.

[0103] Generate a variety of training samples through data enhancement techniques (such as rotation, scaling, flipping, etc.) to increase the generalization ability of the model. This can effectively improve the performance of the model on new data and reduce the risk of overfitting; data enhancement can simulate different shooting conditions, angles and environmental changes, making the model more robust in the face of practical applications; data enhancement can create more samples on a limited data set, improve training efficiency, and help the model learn more comprehensive features.

[0104] By increasing the number and diversity of training samples, the overfitting phenomenon of the model on the training set can be effectively reduced, making it perform better on the validation set and test set; the preprocessed images can better reflect the characteristics of the parts, allowing the model to identify and classify more accurately, thereby improving the overall performance; unified size and normalization processing help convolutional neural networks better extract image features and improve the recognition ability of the model.

[0105] Use machine vision module to identify part types from pre-processed images;

[0106] Input the preprocessed part image into the convolutional neural network to extract feature information;

[0107] The extracted features are transformed nonlinearly through the multi-layer structure of the convolutional neural network to generate the category probability distribution of each part;

[0108] According to the generated category probability distribution, the category with the highest probability is selected as the recognition result of the part;

[0109] The recognition result is converted into a classification signal for controlling the robot arm module to grab the corresponding toy parts.

[0110] Using the machine vision module, parts can be identified quickly and automatically, reducing manual intervention and improving work efficiency; the system can realize real-time image processing and recognition, and is suitable for industrial applications that require quick response, such as automatic grasping on production lines; convolutional neural networks have powerful feature extraction capabilities and can automatically identify and extract important features in images, reducing the need for manual feature engineering; the nonlinear transformation of multi-layer structures can effectively capture complex patterns and improve classification accuracy, making it suitable for complex part recognition tasks.

[0111] By generating a probability distribution of categories, the classification results of each part can be evaluated more carefully, providing a more reliable basis for identification. Selecting the category with the highest probability as the identification result helps to improve the accuracy of identification and reduce misjudgments and missed identifications. The identification results can be quickly converted into control signals to drive the robotic arm module for precise grasping, thereby improving the intelligence level of the system. It can adapt to different part types and environmental changes to achieve flexible grasping strategies.

[0112] This solution uses machine vision modules and convolutional neural networks, which can be easily integrated with other systems to form a complete automated production line. The types of parts that can be identified can be expanded or other functions can be enhanced as needed, such as adding a variety of parts classifications. It reduces errors caused by human operation and ensures the stability and reliability of the recognition process. The combination of machine vision and CNN can provide consistent recognition standards and improve production quality.

[0113] The automated identification and grasping system reduces labor costs, improves production efficiency, and saves time. It can achieve higher production accuracy and consistency without affecting efficiency, thereby optimizing the entire production process. This solution is not only applicable to toy parts, but can also be extended to other fields, such as the identification and grasping of various industrial products such as automotive parts and electronic components. It can be adjusted and optimized according to different product characteristics and application scenarios.

[0114] The convolutional neural network model is constructed as follows:

[0115] Construct at least one convolutional layer, wherein the convolutional layer uses a convolution kernel of 5x5, a stride of 1, and a ReLU activation function;

[0116] Construct at least one pooling layer, wherein the pooling layer is a maximum pooling layer, and its window size is 2x2 and the stride is 2;

[0117] Construct at least one fully connected layer with the number of output units equal to the number of categories, and use the softmax activation function to output the classification probability;

[0118] Select Adam optimizer as the training algorithm, and set the initial learning rate to 0.001;

[0119] Cross entropy is used as the loss function to evaluate the classification performance of the model;

[0120] Monitor validation loss using early stopping to prevent overfitting;

[0121] Evaluate model performance on an independent test set and calculate its accuracy, precision, recall, and F1-score;

[0122] Plot confusion matrix to analyze recognition performance of different part categories;

[0123] The model is adjusted based on the evaluation results and retrained to further improve the classification performance.

[0124] Use the CI FAR-10 dataset, normalize and one-hot encode;

[0125] Use Conv2D to construct the convolution layer, and set the convolution kernel size to 5x5, the stride to 1, and the activation function to ReLU;

[0126] Use MaxPool ling2D to build a pooling layer with a window size of 2x2 and a stride of 2;

[0127] After flattening the feature map using Flatten, a fully connected layer is constructed, and finally the output layer uses the softmax activation function;

[0128] Select Adam optimizer, the initial learning rate is 0.001, and the loss function uses cross entropy;

[0129] Monitor validation loss to prevent model overfitting.

[0130] During training, 20% of the training data is used as a validation set;

[0131] Evaluate the model on the test set and output the test loss and accuracy;

[0132] Use class ifi cat i on_report to output indicators such as precision, recall, and F1-score;

[0133] Calculate and plot the confusion matrix to analyze the recognition performance of different part categories.

[0134] Using a 5x5 convolution kernel, local features in the image, such as edges, textures, etc., can be captured. This feature extraction capability enables the model to automatically learn and recognize complex patterns; ReLU activation function: The ReLU (Rectifi ed Linear Un it) activation function can introduce nonlinearity, allowing the model to learn more complex features, while avoiding the problem of gradient disappearance and accelerating the convergence of the model.

[0135] The use of a maximum pooling layer with a window size of 2x2 can effectively reduce the size of the feature map while retaining important features. This not only reduces the amount of calculation, but also improves the generalization ability of the model and reduces the risk of overfitting; Fully connected layer and softmax activation function: The design of the fully connected layer allows the network to integrate feature information from the convolutional layer, and the softmax activation function provides a multi-category probability distribution, which is suitable for the output of multi-classification problems.

[0136] Adam is selected as the optimizer, which combines the advantages of momentum method and RMSProp. It has strong adaptability, can accelerate the training process of the model, and is suitable for most deep learning tasks. The cross entropy loss function performs well in multi-classification tasks and can effectively evaluate the classification performance of the model. Early stopping is implemented by monitoring the validation loss, which can effectively prevent overfitting of the model and improve the generalization ability of the model on the test set.

[0137] Evaluating model performance on an independent test set can truly reflect the generalization ability of the model. By calculating accuracy, precision, recall, and F1-score, the performance of the model in different aspects can be comprehensively evaluated, which can provide a clear direction for subsequent optimization; the confusion matrix can intuitively show the recognition effect of the model on different categories, help analyze the weaknesses of the model, and guide further model improvement and training strategy adjustment.

[0138] The method to classify the identified part types is:

[0139] Input the preprocessed part image into the convolutional neural network to extract feature information;

[0140] Through the multi-layer structure of CNN, the extracted features are nonlinearly transformed to generate the category probability distribution of each part;

[0141] According to the generated category probability distribution, the category with the highest probability is selected as the recognition result of the part;

[0142] Convert the recognition results into classification signals.

[0143] Traditional image processing methods rely on manually designed features, while CNN can automatically extract local features (such as edges, corners, textures, etc.) from part images through convolutional layers. These features are critical to part recognition and can adapt to different scales and positions; through automated feature extraction, the step of manually designing features is eliminated, making the model more efficient and able to handle more complex situations.

[0144] Convolutional neural networks contain multiple convolutional layers, pooling layers, and fully connected layers. This multi-layer structure enables the model to gradually extract more abstract and complex features. As the number of network layers increases, CNN can gradually abstract from simple low-level features such as edges and shapes to higher-level part features; the ReLU activation function of each layer enables the model to learn nonlinear features, which is crucial for complex part classification tasks. Simple linear models may not be able to effectively distinguish complex part categories, while CNN can capture more diverse image features through multiple layers of nonlinear transformations.

[0145] Through the final output layer of CNN (such as the softmax layer), the probability distribution of each part belonging to each category can be generated. This probabilistic output method is more flexible than directly outputting category labels and can provide richer information, such as classification uncertainty. If necessary, different parts can be processed differently based on these probabilities (for example, setting a threshold to only accept high-probability classifications); the category with the highest probability is selected as the final recognition result. This simple and intuitive decision-making method can ensure high classification accuracy in most cases.

[0146] The recognition results of CNN are directly converted into classification signals for subsequent system processing. This signal can be directly used for applications such as decision-making in automated production lines and control of machine vision systems, without the need for complex subsequent processing steps, reducing latency and complexity. Since CNN can automatically learn useful features from data and its hierarchical structure has strong fault tolerance, it can adapt to different types of part images, including different shooting angles, lighting conditions, background noise, etc. This makes the model more robust and generalizable in practical applications. Even for parts of different shapes, sizes, and textures, CNN can effectively distinguish them through multi-layer feature extraction and classification probability distribution.

[0147] CNN can effectively learn on a large amount of labeled training data, thereby capturing complex part feature relationships. Through back propagation and optimization algorithms (such as Adam), model parameters can be efficiently adjusted to continuously improve classification accuracy; due to the hierarchical structure and efficient parameter sharing of CNN, model training is usually faster than traditional machine learning methods and has better performance.

[0148] The method of receiving the classification signal, controlling the robot arm module to grab toy parts according to the target category, and placing them into the quantitative plate is:

[0149] The robot arm control system receives a classification signal from the vision system through a communication interface, the signal including the identified part type and its corresponding position information;

[0150] Analyze the information in the signal and extract the category and location information of the target part;

[0151] According to the parsed position information, determine the specific position of each type of parts on the conveyor belt;

[0152] The robot control system calculates the optimal grasping path from the current position to the target part;

[0153] According to the path planning results, a motion command is issued to control the robot arm to move above the target part;

[0154] After reaching the target position, the robot arm opens the gripper to perform the grasping;

[0155] After grabbing the parts, the robot arm moves to the top of the quantitative plate according to the preset path;

[0156] Open the gripper and release the parts into the dosing tray;

[0157] The robot continues to receive classification signals for cyclic operation, and the robot arm continuously grabs parts of different categories and puts them into the quantitative tray until the quantitative requirements are met.

[0158] The entire grabbing and placing process is automated, reducing manual intervention and improving production efficiency. The robotic arm can accurately operate according to the classified signals, saving time for manual inspection and operation; the robotic arm can work continuously, automatically receive signals and perform repeated operations, which is suitable for large-scale production needs and improves production capacity.

[0159] Through the position information provided by the vision system, the robot arm can accurately find the position of the target part, reducing the possibility of grasping errors; through cyclic operation, the robot arm can ensure that each part is grasped and placed, reducing the risk of omissions and errors.

[0160] The system can handle multiple categories of toy parts and can adapt to different production needs by simply adjusting the classification signals, which has great flexibility. The system can receive and process new classification signals in real time and can promptly adapt to changes in parts on the production line to ensure a smooth production process.

[0161] The robotic arm can calculate the optimal grasping path from the current position to the target part, reducing movement time and energy consumption and improving work efficiency; through path planning, the system can avoid collisions with other objects and ensure the safe operation of the robotic arm.

[0162] The robotic arm can quickly receive classification signals and perform grabbing and delivery operations, improving the overall efficiency of the production line; the system can set quantitative requirements to ensure that parts are delivered on demand during the production process to avoid waste or overproduction.

[0163] Through the communication interface, the status and operation of the robot arm can be monitored in real time for easy management and maintenance; the system can set up an alarm mechanism to provide timely feedback on the working status of the robot arm, facilitating rapid diagnosis and repair of faults.

[0164] A toy parts equal-quantity automatic packaging system, comprising:

[0165] A vibration plate module is used to receive and disperse toy parts so that they can be fed into the conveyor belt in a flat manner;

[0166] An image acquisition module, including a CCD camera, configured to photograph toy parts on a conveyor belt and acquire images of the parts;

[0167] An image processing module, used for preprocessing the part image;

[0168] A machine vision module for part type recognition based on pre-processed part images;

[0169] A classification module, used to classify the identified part types and generate classification signals representing the category information of each part;

[0170] A control system module receives the classification signal and controls the robot arm module;

[0171] A robotic arm module, which grabs toy parts of a specified category according to the instructions of the control system module and puts them into a quantitative tray;

[0172] A quantitative tray module, used to store toy parts dropped by the robot arm;

[0173] A weighing sensor module is arranged below the quantitative plate to perform real-time weighing to monitor the weight of the parts;

[0174] The packaging module is used to empty the parts in the quantitative tray and package them after the weight reaches the set standard.

[0175] The system has a high degree of automation, and all modules work together to reduce manual intervention and improve production efficiency. The combination of the vibration plate and conveyor belt can quickly disperse and transport parts, reducing waiting and processing time. The CCD camera and image processing module can efficiently capture and process part images to ensure that parts are accurately identified. The machine vision module and classification module can intelligently classify parts in real time to ensure that each type of part can be correctly identified and processed.

[0176] The weighing sensor module monitors the weight of parts in real time to ensure that the quantity and quality of each package meet the set standards, thereby improving product consistency and reliability. The fully automated system reduces errors caused by human operation and ensures the stable quality of each batch of products. The system can handle a variety of toy parts and can flexibly adjust the production line configuration according to market demand. The design allows for the addition or replacement of modules according to production needs, such as introducing new recognition algorithms or replacing robotic arms to improve production capacity.

[0177] The weighing sensor provides real-time feedback and can monitor the weight of parts in the quantitative tray in time to ensure accurate packaging quantity; the coordinated work of various modules in the system can provide real-time monitoring, detect and handle faults in time, and improve system reliability; due to the high degree of automation of the system, it can significantly reduce dependence on manual operation and reduce labor costs; accurate weighing and classification reduce material waste and improve resource utilization.

[0178] This embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method for automatic packaging of equal-quantity toy parts when executing the program.

[0179] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for automatic packaging of equal-quantity toy parts as described above is implemented.

[0180] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0181] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0182] Of course, the above are only typical examples of the present invention. In addition to these, the present invention may also have many other specific implementations. Any technical solutions formed by equivalent replacement or equivalent transformation fall within the scope of protection required by the present invention. Of course, the above are only typical examples of the present invention. In addition to these, the present invention may also have many other specific implementations. Any technical solutions formed by equivalent replacement or equivalent transformation fall within the scope of protection required by the present invention.

Claims

1. A method for automatically packaging toy parts in equal quantities, characterized in that , including the following steps: Pour the toy parts into the vibration plate, and send the toy parts to the conveyor belt in a scattered and flat manner through the vibration of the vibration plate; Use a CCD camera to shoot toy parts on the conveyor belt to obtain part images; Preprocessing the acquired part images; Use machine vision module to identify part types from pre-processed images; classifying the identified part type; Generate a classification signal based on the recognition and classification results, indicating the category information of each part; receiving the classification signal, controlling the robot arm module to grab toy parts according to the target category, and placing them into the quantitative tray; The pressure sensor under the quantitative tray is used for real-time weighing. When the weight reaches the set standard, the parts in the quantitative tray are emptied, packaged and sent out.

2. The method for automatically packaging equal portions of toy parts according to claim 2, characterized in that: A CCD camera is used to photograph toy parts on a conveyor belt. A method for obtaining images of parts is to set an LED lighting device with adjustable brightness on one side of the CCD camera to photograph the toy parts on the conveyor belt; A photoelectric sensor is set at the feeding port of the conveyor belt. When the parts pass through the photoelectric sensor, the CCD camera is triggered to collect images.

3. The method for automatically packaging equal portions of toy parts according to claim 2, characterized in that: The method for preprocessing the acquired part image is: Resize the image to a uniform 224*224 pixels; Normalize the image pixel values ​​to the range [0,1]; Data enhancement technology is used to enhance the image data.

4. The method for automatically packaging equal portions of toy parts according to claim 2, characterized in that: Use machine vision module to identify part types from pre-processed images; Input the preprocessed part image into the convolutional neural network to extract feature information; The extracted features are transformed nonlinearly through the multi-layer structure of the convolutional neural network to generate the category probability distribution of each part; According to the generated category probability distribution, the category with the highest probability is selected as the recognition result of the part; The recognition result is converted into a classification signal for controlling the robot arm module to grab the corresponding toy parts.

5. The method for automatically packaging toy parts in equal quantities according to claim 4, characterized in that: The convolutional neural network model is constructed as follows: Construct at least one convolutional layer, wherein the convolutional layer uses a convolution kernel of 5x5, a stride of 1, and a ReLU activation function; Construct at least one pooling layer, wherein the pooling layer is a maximum pooling layer, and its window size is 2x2 and the stride is 2; Construct at least one fully connected layer with the number of output units equal to the number of categories, and use the softmax activation function to output the classification probability; Select Adam optimizer as the training algorithm, and set the initial learning rate to 0.001; Cross entropy is used as the loss function to evaluate the classification performance of the model; Monitor validation loss using early stopping to prevent overfitting; Evaluate model performance on an independent test set and calculate its accuracy, precision, recall, and F1-score; Plot confusion matrix to analyze recognition performance of different part categories; The model is adjusted based on the evaluation results and retrained to further improve the classification performance.

6. The method for automatic packaging of equal-quantity toy parts according to claim 5, characterized in that: The method to classify the identified part types is: Input the preprocessed part image into the convolutional neural network to extract feature information; Through the multi-layer structure of CNN, the extracted features are nonlinearly transformed to generate the category probability distribution of each part; According to the generated category probability distribution, the category with the highest probability is selected as the recognition result of the part; Convert the recognition results into classification signals.

7. The method for automatically packaging equal portions of toy parts according to claim 5, characterized in that: The method of receiving the classification signal, controlling the robot arm module to grab toy parts according to the target category, and placing them into the quantitative plate is: The robot arm control system receives a classification signal from the vision system through a communication interface, the signal including the identified part type and its corresponding position information; Analyze the information in the signal and extract the category and location information of the target part; According to the parsed position information, determine the specific position of each type of parts on the conveyor belt; The robot control system calculates the optimal grasping path from the current position to the target part; According to the path planning results, a motion command is issued to control the robot arm to move above the target part; After reaching the target position, the robot arm opens the gripper to perform the grasping; After grabbing the parts, the robot arm moves to the top of the quantitative plate according to the preset path; Open the gripper and release the parts into the dosing tray; The robot continues to receive classification signals for cyclic operation, and the robot arm continuously grabs parts of different categories and puts them into the quantitative tray until the quantitative requirements are met.

8. An automatic packaging system for equal-quantity toy parts, characterized in that: Included are: A vibration plate module is used to receive and disperse toy parts so that they can be fed into the conveyor belt in a flat manner; An image acquisition module, including a CCD camera, configured to photograph toy parts on a conveyor belt and acquire images of the parts; An image processing module, used for preprocessing the part image; A machine vision module for part type recognition based on pre-processed part images; A classification module, used to classify the identified part types and generate classification signals representing the category information of each part; A control system module receives the classification signal and controls the robot arm module; A robotic arm module, which grabs toy parts of a specified category according to the instructions of the control system module and puts them into a quantitative tray; A quantitative tray module, used to store toy parts dropped by the robot arm; A weighing sensor module is arranged below the quantitative plate to perform real-time weighing to monitor the weight of the parts; The packaging module is used to empty the parts in the quantitative tray and package them after the weight reaches the set standard.

9. An electronic device, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for automatic packaging of equal-portion quantitative toy parts as claimed in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for automatically packaging equal portions of toy parts as described in any one of claims 1 to 7 is implemented.

Images