Automatic feeding systems for plastics processing

Abstract

The present invention discloses an automatic feeding system for plastic processing, which relates to the technical field of automatic plastic feeding. The system includes: obtaining a plastic raw material sensing data stream of a plastic storage bin through a storage sensing module; a plastic sorting module is used to identify the plastic raw material sensing data stream based on the plastic processing and molding standard to obtain plastic defect feature information, thereby dividing the raw materials and obtaining a set of available plastic raw materials; a control analysis module is used to perform loss optimization analysis on the traversed and retrieved matching feeding control parameters to obtain target feeding control parameters; a feeding conveying module is used to perform feeding and conveying monitoring on the available plastic raw material set based on the target feeding control parameters to obtain plastic feeding feedback parameters, and based on this, the target feeding control parameters are tuned for feeding control. The technical effect of realizing efficient automatic feeding for plastic processing, improving feeding control accuracy and feeding efficiency, and thus improving the quality of plastic processing products is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of automatic plastic feeding, in particular to an automatic feeding system for plastic processing. Background Art

[0002] In the plastic processing industry, the supply and delivery of raw materials is a crucial link, which involves delivering raw materials to processing equipment for subsequent processing and molding. However, existing feeding methods often rely on manual operation and have a low degree of automation, which is not only inefficient but also prone to errors, affecting product quality. Therefore, realizing the automatic supply and delivery of plastic raw materials is of great significance to improving production efficiency, reducing production costs, and reducing environmental pollution. Summary of the invention

[0003] The present application solves the technical problems in the prior art of low automation of plastic feeding, which is not only inefficient but also prone to errors and affects product quality, by providing an automatic feeding system for plastic processing. The application achieves the technical effect of realizing efficient automatic feeding for plastic processing, improving feeding control accuracy and feeding efficiency, and thereby improving the quality of plastic processing products.

[0004] In view of the above problems, the present invention provides an automatic feeding system for plastic processing.

[0005] The present application provides an automatic feeding system for plastic processing, the system comprising: a material storage sensing module, the material storage sensing module integrating a CMOS image device, a micro-thermal imaging lens and a material level sensor, the material storage bin is monitored and identified through the material storage sensing module, and a plastic raw material sensing data stream is obtained; a plastic sorting module, which is used to obtain a plastic processing and molding standard, perform defect identification and positioning on the plastic raw material sensing data stream based on the plastic processing and molding standard, obtain plastic defect feature information, divide the plastic storage bin into raw materials according to the plastic defect feature information, and obtain a set of available plastic raw materials; a control analysis module, which is used to obtain a target plastic molding demand, perform traversal retrieval based on the target plastic molding demand and a plastic feeding control space, determine matching feeding control parameters, perform loss optimization analysis on the matching feeding control parameters, and obtain target feeding control parameters; a feeding conveying module, which is used to perform feeding and conveying monitoring on the available plastic raw material set based on the target feeding control parameters, obtain plastic feeding feedback parameters, and optimize feeding control on the target feeding control parameters through the plastic feeding feedback parameters.

[0006] In a possible implementation, the method of obtaining the plastic raw material perception data stream also includes: obtaining the visual image data of the plastic storage bin through the CMOS image device, performing convolution feature extraction on the visual image data, and obtaining a plastic multi-dimensional feature set; dividing and marking the visual image data into regions of interest based on the plastic multi-dimensional feature set, and determining a plastic detection area of ​​interest; performing information imaging and acquisition on the plastic detection area of ​​interest using the micro-thermal imaging lens, generating plastic infrared thermal image data, and simultaneously obtaining the plastic storage location information of the plastic storage bin through the material level sensor; and obtaining the plastic raw material perception data stream based on the visual image data, the plastic infrared thermal image data, and the plastic storage location information.

[0007] In a possible implementation, the obtaining of plastic defect feature information further includes: determining plastic data acquisition attribute information according to the material storage perception module, configuring a processing program for the plastic data acquisition attribute information, and obtaining a multi-data attribute preprocessing program; mapping and preprocessing the plastic raw material perception data stream based on the multi-data attribute preprocessing program to obtain visual standard image data, standard infrared thermal image data, and standard material storage location information; marking defect feature identification marks for the visual standard image data and the standard infrared thermal image data respectively according to the plastic processing and molding standard to obtain a visual image defect feature set and an infrared image defect feature set; distributing and locating the visual image defect feature set and the infrared image defect feature set based on the standard material storage location information to determine the plastic defect feature information.

[0008] In a possible implementation, the determining of the plastic defect characteristic information further includes: respectively parsing the visual image defect characteristic set and the infrared image defect characteristic set by defect type and size quantity to obtain visual defect characteristic parameter information and infrared defect characteristic parameter information; distributing and locating the visual defect characteristic parameter information based on the standard material storage position information to generate a plastic surface defect distribution map; locating and marking the infrared defect characteristic parameter information based on the standard material storage position information to determine the infrared defect distribution characteristic parameter information; cross-validating the plastic surface defect distribution map through the infrared defect distribution characteristic parameter information to determine the plastic defect characteristic information.

[0009] In a possible implementation, the determining of matching feeding control parameters also includes: obtaining plastic molding attribute information, the plastic molding attribute information including plastic molding specification requirements, plastic physical and chemical properties and molding requirements, classifying the plastic feeding control space according to the plastic molding attribute information to obtain plastic molding attribute characteristic parameters; dividing and marking the plastic feeding control space based on the plastic molding attribute characteristic parameters to obtain a plastic feeding control characteristic database; parsing the target plastic molding requirements to obtain plastic molding attribute requirement parameters; traversing and matching the plastic feeding control characteristic database based on the plastic molding attribute requirement parameters to determine the matching feeding control parameters.

[0010] In a possible implementation, the method of obtaining the target feeding control parameters further includes: performing a feature loss analysis on the plastic molding property requirement parameters and the matching feeding control parameters through the plastic molding property information to obtain plastic property loss feature data; performing parameter optimization analysis based on the plastic property loss feature data to obtain a feeding control parameter optimization selection threshold; constructing a feeding control fitness function, and randomly generating multiple feeding control parameters within the feeding control parameter optimization selection threshold through the feeding control fitness function for evaluation calculation to obtain multiple feeding control effect fitnesses; and performing cluster learning optimization on the multiple feeding control parameters based on the multiple feeding control effect fitnesses to determine the target feeding control parameters.

[0011] In a possible implementation, the method of obtaining plastic property loss characteristic data further includes: determining attribute feature analysis dimension information based on the plastic molding attribute information, performing weight assignment on the attribute feature analysis dimension information, and obtaining feature dimension weight factor information; performing loss function fitting on the attribute feature analysis dimension information based on the feature dimension weight factor information, and constructing an attribute feature loss function; and performing feature loss calculation on the plastic molding attribute requirement parameters and the matching feeding control parameters using the attribute feature loss function, and obtaining the plastic property loss characteristic data.

[0012] In a possible implementation, the determining of the target feeding control parameters further includes: clustering the multiple feeding control parameters according to the fitness of the multiple feeding control effects to generate a feeding control parameter cluster; performing mutation crossover expansion on the remaining control parameters in the feeding control parameter cluster toward the center of the parameter cluster based on a preset variation step length to obtain an updated control parameter cluster, and performing global optimization on the updated control parameter cluster to determine the target feeding control parameters.

[0013] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0014] The plastic storage bin is monitored and identified by the storage sensing module to obtain the plastic raw material sensing data stream, the plastic sorting module is used to identify and locate the defects of the plastic raw material sensing data stream based on the plastic processing and molding standards, and obtain the plastic defect feature information, so as to divide the raw materials of the plastic storage bin and obtain the available plastic raw material set; the control analysis module is used to traverse and search based on the target plastic molding requirements and the plastic feeding control space, determine the matching feeding control parameters, perform loss optimization analysis on the matching feeding control parameters, and obtain the target feeding control parameters; the feeding conveying module is used to monitor the feeding and conveying of the available plastic raw material set based on the target feeding control parameters, obtain the plastic feeding feedback parameters, and adjust the feeding control of the target feeding control parameters based on the technical solution. Thus, the technical effect of realizing efficient and automated feeding of plastic processing, improving the feeding control accuracy and feeding efficiency, and thus improving the quality of plastic processing products is achieved.

[0015] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a schematic diagram of the structure of the automatic feeding system used for plastic processing in this application;

[0017] Figure 2 This is a schematic diagram of the process of obtaining plastic defect feature information in the automatic feeding system used in plastic processing in this application.

[0018] Explanation of the reference numerals: material storage sensing module 11 , plastic sorting module 12 , control analysis module 13 , material feeding and conveying module 14 . DETAILED DESCRIPTION

[0019] The present application solves the technical problems in the prior art of low automation of plastic feeding, which is not only inefficient but also prone to errors and affects product quality, by providing an automatic feeding system for plastic processing. The application achieves the technical effect of realizing efficient automatic feeding for plastic processing, improving feeding control accuracy and feeding efficiency, and thereby improving the quality of plastic processing products.

[0020] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0021] Embodiment 1, as Figure 1As shown, the present application provides an automatic feeding system for plastic processing, the system comprising:

[0022] The material storage sensing module 11 integrates a CMOS imaging device, a micro-thermal imaging lens and a material level sensor. The material storage sensing module is used to monitor and identify the plastic storage bin to obtain the plastic raw material sensing data stream.

[0023] Furthermore, the steps of obtaining the plastic raw material sensing data stream in this application also include:

[0024] The visual image data of the plastic storage bin is acquired through the CMOS image device, and convolution feature extraction is performed on the visual image data to obtain a plastic multi-dimensional feature set; based on the plastic multi-dimensional feature set, the visual image data is divided and marked with regions of interest to determine a plastic detection area of ​​interest; the micro-thermal imaging lens is used to perform information imaging and acquisition on the plastic detection area of ​​interest to generate plastic infrared thermal image data, and at the same time, the plastic storage position information of the plastic storage bin is acquired through the material level sensor; based on the visual image data, the plastic infrared thermal image data and the plastic storage position information, the plastic raw material perception data stream is acquired.

[0025] Specifically, in order to realize efficient and automated feeding of plastic processing, the plastic storage bin is monitored and identified by the storage sensing module 11, wherein the storage sensing module is a system function module for sensing raw materials of the plastic storage bin, and is provided with a variety of sensors, integrated CMOS image devices, micro thermal imaging lenses and material level sensors. First, the visual image data of the plastic raw materials in the plastic storage bin is obtained by the CMOS image device, which has the advantages of high integration, low power consumption and fast imaging capability. Convolution feature extraction is performed on the visual image data, which usually involves using a deep learning model (such as a convolutional neural network CNN) to automatically identify and extract key features in the image, and obtain a multi-dimensional feature set of the plastic in the storage bin, including the color, texture, shape, etc. of the plastic. Then, based on the multi-dimensional feature set of the plastic, the visual image data is divided and marked for regions of interest, and the plastic regions whose color, texture and shape are different from other raw materials are divided and marked, and the corresponding plastic detection regions of interest are determined, which is helpful to determine the plastic detection regions of interest that need to be inspected in more detail.

[0026] The micro thermal imaging lens is used to collect information imaging of the plastic detection area of ​​interest. The micro thermal imaging lens can capture the local temperature distribution of the plastic surface, which is very useful for detecting the storage status of the plastic (such as whether it is overheated, whether it is evenly distributed, etc.), thereby generating plastic infrared thermal image data. At the same time, the plastic storage position information of the plastic storage bin is obtained through the material level sensor. The material level sensor can directly measure the material level height or volume of the plastic in the storage bin, which is crucial for evaluating the storage status of the storage bin and planning subsequent operations. The visual image data, the plastic infrared thermal image data and the plastic storage position information are combined to obtain a plastic raw material perception data stream, which contains comprehensive information about the plastic storage bin, including the appearance distribution, temperature and material level of the plastic. Accurate and real-time perception of the state of the plastic storage bin is achieved, thereby improving the processing quality and efficiency of plastic raw materials.

[0027] The plastic sorting module 12 is used to obtain plastic processing and molding standards, identify and locate defects in the plastic raw material sensing data stream based on the plastic processing and molding standards, obtain plastic defect feature information, and classify the raw materials in the plastic storage bin according to the plastic defect feature information to obtain a set of available plastic raw materials.

[0028] like Figure 2 As shown, further, the steps of obtaining plastic defect characteristic information in this application also include:

[0029] According to the material storage perception module, the plastic data collection attribute information is determined, and a processing program is configured for the plastic data collection attribute information to obtain a multi-data attribute preprocessing program; based on the multi-data attribute preprocessing program, the plastic raw material perception data stream is mapped and preprocessed to obtain visual standard image data, standard infrared thermal image data and standard material storage location information; according to the plastic processing and molding standard, the visual standard image data and the standard infrared thermal image data are respectively marked for defect feature identification to obtain a visual image defect feature set and an infrared image defect feature set; based on the standard material storage location information, the visual image defect feature set and the infrared image defect feature set are distributed and located to determine the plastic defect feature information.

[0030] Furthermore, the step of determining the plastic defect characteristic information also includes:

[0031] The visual image defect feature set and the infrared image defect feature set are respectively analyzed for defect type and size quantity to obtain visual defect feature parameter information and infrared defect feature parameter information; the visual defect feature parameter information is distributed and located based on the standard material storage position information to generate a plastic surface defect distribution map; the infrared defect feature parameter information is located and marked based on the standard material storage position information to determine the infrared defect distribution feature parameter information; the plastic surface defect distribution map is cross-validated using the infrared defect distribution feature parameter information to determine the plastic defect feature information.

[0032] Specifically, the plastic sorting module 12 is a system functional module for sorting and dividing plastic raw materials that meet quality standards. First, the plastic processing and molding standards are obtained through industry specifications and standards. The plastic processing and molding standards are the quality requirements for plastic raw materials used for processing and molding plastics, usually including the material, size, surface defect requirements, etc. of the plastic raw materials. Based on the plastic processing and molding standards, the plastic raw material perception data stream is defect identified and located. First, according to the storage material perception module, the plastic data acquisition attribute information is determined. The plastic data acquisition attribute information is the perception data acquisition source attribute, including image data attributes, thermal imaging data attributes, and position data attributes. The processing program is configured for the plastic data acquisition attribute information. The preprocessing program steps for different data attributes are different. According to the collected plastic data attribute information, the corresponding processing program is configured. Exemplarily, the processing program steps for image data attributes include filtering, standardization, etc., so as to obtain the corresponding multi-data attribute preprocessing program for each data attribute.

[0033] Based on the multi-data attribute preprocessing program, the plastic raw material perception data stream is mapped, matched and preprocessed to obtain preprocessed visual standard image data, standard infrared thermal image data and standard storage location information. According to the plastic processing and molding standard, the visual standard image data and the standard infrared thermal image data are respectively marked for defect feature recognition, and the defect data that does not meet the plastic processing and molding standard is compared and identified using edge detection feature extraction, classification algorithm, etc., to mark the visual image defect feature set and infrared image defect feature set. Based on the standard storage location information, the visual image defect feature set and the infrared image defect feature set are distributed and located, specifically: the visual image defect feature set and the infrared image defect feature set are respectively analyzed for defect type and size quantity, and the corresponding infrared defect feature parameter information is determined according to the extracted visual image defect features, combined with the known defect type database or machine learning model, including the defect type (such as scratches, cracks, pits, etc.) and size quantity; similarly, according to the thermal features in the infrared image, the infrared defect feature parameter information related to the defect is analyzed and determined, including the thermal anomaly type (such as hot spots, cold spots) and size quantity.

[0034] The visual defect characteristic parameter information is distributed and located based on the standard storage location information, and the visual defect characteristic parameter information is mapped onto the plastic surface using the standard storage location information of the plastic (such as grid coordinates, area division, etc.), generating a plastic surface defect distribution map, which can intuitively show which positions have defects and the types and quantities of defects. The infrared defect characteristic parameter information is positioned and marked based on the standard storage location information, and the thermal anomaly distribution position parameters related to the infrared image are determined, which are related to the defects inside the plastic, and the corresponding infrared defect distribution characteristic parameter information is obtained. The plastic surface defect distribution map is cross-validated by the infrared defect distribution characteristic parameter information to more accurately determine the location, type and characteristics of the defect. Infrared thermal images and visual images may provide different aspects of information about the same defect, so more comprehensive defect characteristic information can be obtained by combining and validating. If both images detect defects at the same position, the possibility of this defect is higher, and the information of the two images can be combined to more accurately describe the characteristics of this defect, thereby determining the plastic defect characteristic information. The plastic storage bin is divided into raw materials according to the plastic defect feature information, defective and non-defective plastic raw materials are sorted, and a set of available plastic raw materials that meet the plastic processing and molding standards is obtained. The accuracy and comprehensiveness of plastic defect detection are improved, and a data basis is provided for plastic feeding control, thereby improving the quality of plastic processing products.

[0035] The control analysis module 13 is used to obtain the target plastic molding demand, perform traversal search based on the target plastic molding demand and the plastic feeding control space, determine the matching feeding control parameters, perform loss optimization analysis on the matching feeding control parameters, and obtain the target feeding control parameters.

[0036] Furthermore, the steps of determining and matching the feeding control parameters in this application also include:

[0037] Obtain plastic molding attribute information, the plastic molding attribute information including plastic molding specification requirements, plastic physical and chemical properties and molding requirements, classify the plastic feeding control space according to the plastic molding attribute information, and obtain plastic molding attribute characteristic parameters; divide and mark the plastic feeding control space based on the plastic molding attribute characteristic parameters to obtain a plastic feeding control characteristic database; parse the target plastic molding requirements to obtain plastic molding attribute requirement parameters; traverse and match the plastic feeding control characteristic database based on the plastic molding attribute requirement parameters to determine the matching feeding control parameters.

[0038] Furthermore, the step of obtaining the target feeding control parameter in this application also includes:

[0039] The plastic molding property information is used to perform feature loss analysis on the plastic molding property requirement parameters and the matching feeding control parameters to obtain plastic property loss feature data; parameter optimization analysis is performed based on the plastic property loss feature data to obtain a feeding control parameter optimization selection threshold; a feeding control fitness function is constructed, and a plurality of feeding control parameters are randomly generated within the feeding control parameter optimization selection threshold through the feeding control fitness function for evaluation calculation to obtain a plurality of feeding control effect fitnesses; cluster learning and optimization are performed on the plurality of feeding control parameters based on the plurality of feeding control effect fitnesses to determine the target feeding control parameters.

[0040] Furthermore, the steps of obtaining the characteristic data of plastic property loss in this application also include:

[0041] According to the plastic molding attribute information, the attribute feature analysis dimension information is determined, and weights are assigned to the attribute feature analysis dimension information to obtain feature dimension weight factor information; based on the feature dimension weight factor information, a loss function is fitted to the attribute feature analysis dimension information to build an attribute feature loss function; using the attribute feature loss function, the plastic molding attribute requirement parameters and the matching feeding control parameters are calculated for feature loss to obtain the plastic attribute loss feature data.

[0042] Specifically, the control analysis module 13 is a system function module for analyzing and optimizing the plastic feeding control parameters, and obtains the target plastic molding requirements through the system, and the target plastic molding requirements are the target plastic molding requirements, including the molding shape, size, precision, color, etc. Based on the target plastic molding requirements and the plastic feeding control space, a traversal search is performed to first obtain plastic molding attribute information, which is a material molding index associated with feeding control, including plastic molding specification requirements (such as density, melting point, hardness, etc.), plastic physical and chemical properties (such as temperature resistance, corrosion resistance, conductivity, etc.) and molding requirements (such as the required plastic weight size and total molding processing volume).

[0043] According to the plastic molding attribute information, the data in the plastic feeding control space are attribute-classified to obtain the corresponding plastic molding attribute characteristic parameters of each spatial data. Based on the plastic molding attribute characteristic parameters, the plastic feeding control space is divided and marked, and the spatial data with the same plastic molding attribute characteristic parameters are divided into the same attribute area, so as to obtain a plastic feeding control feature database after the characteristic parameter marking is integrated, so as to facilitate subsequent rapid retrieval and matching. The target plastic molding requirements are parsed through the plastic molding attribute information to obtain the corresponding specific plastic molding attribute requirement parameters. Then, based on the plastic molding attribute requirement parameters, the plastic feeding control feature database is traversed and matched and retrieved, and the one with the highest similarity to the plastic molding attribute requirement parameters is determined as the matching feeding control parameter. Rapid mapping and matching of feeding control parameters is achieved to adapt to plastic feeding control with different molding requirements, thereby improving plastic molding efficiency and control accuracy.

[0044] In order to improve the accuracy of the feeding control parameter analysis, the plastic molding attribute requirement parameters and the matching feeding control parameters are subjected to a feature loss analysis through the plastic molding attribute information. First, the attribute feature analysis dimension information is determined based on the plastic molding attribute information, and each plastic molding attribute is determined as an attribute feature analysis dimension. Based on the specific needs and process requirements of plastic molding, each attribute feature analysis dimension information needs to be assigned a weight. This process can be achieved through expert scoring, historical data analysis, or machine learning. The weight reflects the importance of each feature dimension in the entire molding process, and the feature dimension weight factor information assigned to each attribute feature analysis dimension information is obtained.

[0045] Based on the feature dimension weight factor information, the attribute feature analysis dimension information is fitted with a loss function to build an attribute feature loss function. The attribute feature loss function is a weighted multivariate loss function, so that the contribution of the difference in different feature dimensions to the loss function is proportional to its weight, and is used to measure the difference between the plastic molding attribute and the required standard. The attribute feature loss function is used to calculate the feature loss of the plastic molding attribute requirement parameters and the matching feeding control parameters to obtain the corresponding plastic attribute loss feature data, which reflects the degree of difference between the molding attribute feature parameters corresponding to the current matching feeding control parameters and the plastic molding attribute requirement parameters, and analyzes the attribute loss of plastic molding under given matching feeding control parameters, providing a basis for the subsequent adjustment and optimization of feeding control parameters.

[0046] Based on the characteristic data of the plastic property loss, parameter optimization analysis is performed, and the feed control parameters are optimized using an optimization algorithm (such as gradient descent method, genetic algorithm, etc.) to find the parameter selection range that can minimize the property loss, that is, the feed control parameter optimization selection threshold. A feed control fitness function is constructed. The feed control fitness function is a multivariate function used to evaluate the quality of the feed control parameters. It is obtained by empirical fitting based on the plastic processing demand effect data, and is usually constructed based on factors such as plastic molding property loss, production efficiency, and energy consumption. Multiple feed control parameters are randomly generated within the feed control parameter optimization selection threshold, and the multiple feed control parameters are evaluated and calculated by the feed control fitness function to obtain the corresponding multiple feed control effect fitness. Based on the multiple feed control effect fitness, the multiple feed control parameters are clustered and learned to optimize, improve the accuracy of global optimization of parameters, and determine the one with the largest fitness as the target feed control parameter. Optimization analysis of plastic molding feed control parameters is achieved, feeding control accuracy and feeding efficiency are improved, and the processing quality of plastic molding is ensured.

[0047] Furthermore, the step of determining the target feeding control parameter in this application further includes:

[0048] The multiple feeding control parameters are clustered and divided according to the fitness of the multiple feeding control effects to generate a feeding control parameter cluster; based on a preset variation step length, the remaining control parameters in the feeding control parameter cluster are mutated and cross-expanded toward the center of the parameter cluster to obtain an updated control parameter cluster, and the updated control parameter cluster is globally optimized to determine the target feeding control parameters.

[0049] Specifically, the specific process of clustering learning and optimizing the multiple feeding control parameters based on the multiple feeding control effect fitness is as follows: first, the multiple feeding control parameters are clustered and divided according to the multiple feeding control effect fitness, and the feeding control parameter with the largest fitness is used as the cluster center to generate a feeding control parameter cluster. Then, based on the preset variable step length, the remaining control parameters in the feeding control parameter cluster are mutated and fine-tuned toward the parameter cluster center, and then the parameters after the mutated fine-tuning are cross-expanded, wherein the preset variable step length determines the amplitude of the mutated cross-expanding in the feeding control parameter cluster, which can be set according to the demand experience. Then, some new feeding control parameters are generated to obtain an expanded updated control parameter cluster, which includes the parameters in the original parameter cluster and the new parameters generated by the mutated cross-expanding, thereby improving the diversity of the control parameter cluster population to ensure the accuracy of parameter optimization. The update control parameter cluster is globally optimized. The goal of this step is to find the feed control parameter with the best feed control effect fitness in the update control parameter cluster, which is achieved by comparing the fitness values ​​of the best parameters in each parameter cluster, thereby determining the target feed control parameter with the best feed control effect fitness in the parameter cluster. This realizes the global optimization analysis of the plastic molding feed control parameters, and improves the optimization accuracy of the feed control parameters and the feed analysis efficiency.

[0050] The feeding and conveying module 14 is used to monitor the feeding and conveying of the available plastic raw material set based on the target feeding control parameter, obtain plastic feeding feedback parameters, and adjust the target feeding control parameters through the plastic feeding feedback parameters.

[0051] Specifically, the feeding and conveying module 14 is a system function module for feeding and conveying the plastic raw material set and monitoring the process in real time. First, the feeding and conveying of the available plastic raw material set is monitored based on the target feeding control parameters, which may include feeding speed, feeding amount, feeding time, etc. It is transported to the processing equipment through the conveying system. During the feeding and conveying process, the feeding process will be monitored in real time to obtain plastic feeding feedback parameters, which may include actual feeding speed, actual feeding amount, raw material conveying stability, etc. The obtained plastic feeding feedback parameters are analyzed, and the difference between the actual parameters and the target parameters is compared to evaluate the accuracy and efficiency of feeding and conveying. And the target feeding control parameters are tuned for feeding control through the analysis results of the plastic feeding feedback parameters. If the actual feeding speed is too fast or too slow, the system may adjust the feeding speed; if there is a large difference between the actual feeding amount and the target amount, the system may adjust parameters such as feeding amount or feeding time. The optimized feeding control parameters will be used again for feeding and conveying, and will continue to be optimized through real-time monitoring and feedback parameter analysis until the ideal feeding control effect is achieved. This will achieve efficient and automated feeding of plastic processing, improve feeding control accuracy and feeding efficiency, and thus improve the quality of plastic processing products.

[0052] In summary, the automatic feeding system for plastic processing provided by this application has the following technical effects:

[0053] The plastic storage bin is monitored and identified by the storage sensing module to obtain the plastic raw material sensing data stream, the plastic sorting module is used to identify and locate the defects of the plastic raw material sensing data stream based on the plastic processing and molding standards, and obtain the plastic defect feature information, so as to divide the raw materials of the plastic storage bin and obtain the available plastic raw material set; the control analysis module is used to traverse and search based on the target plastic molding requirements and the plastic feeding control space, determine the matching feeding control parameters, perform loss optimization analysis on the matching feeding control parameters, and obtain the target feeding control parameters; the feeding conveying module is used to monitor the feeding and conveying of the available plastic raw material set based on the target feeding control parameters, obtain the plastic feeding feedback parameters, and adjust the feeding control of the target feeding control parameters based on the technical solution. Thus, the technical effect of realizing efficient and automated feeding of plastic processing, improving the feeding control accuracy and feeding efficiency, and thus improving the quality of plastic processing products is achieved.

[0054] This specification and drawings are merely exemplary illustrations of the present application, but the protection scope of the present application is not limited thereto. It should be noted that any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. Automatic feeding system for plastic processing, characterized in that: The system comprises: A material storage sensing module, which integrates a CMOS imaging device, a micro-thermal imaging lens and a material level sensor, monitors and identifies the plastic storage bin through the material storage sensing module to obtain a plastic raw material sensing data stream; A plastic sorting module is used to obtain plastic processing and molding standards, identify and locate defects on the plastic raw material sensing data stream based on the plastic processing and molding standards, obtain plastic defect feature information, and classify the plastic storage bin into raw materials according to the plastic defect feature information to obtain a set of available plastic raw materials, wherein the plastic processing and molding standards are obtained through industry specifications and standards, and the plastic processing and molding standards are quality requirements for plastic raw materials used for processing and molding plastics, including material, size, and surface defect requirements for plastic raw materials; A control analysis module is used to obtain target plastic molding requirements, perform traversal retrieval based on the target plastic molding requirements and the plastic feeding control space, determine matching feeding control parameters, perform loss optimization analysis on the matching feeding control parameters, and obtain target feeding control parameters; A feeding and conveying module, used for monitoring the feeding and conveying of the available plastic raw material set based on the target feeding control parameter, obtaining a plastic feeding feedback parameter, and optimizing the feeding control of the target feeding control parameter by using the plastic feeding feedback parameter; The obtaining of plastic defect characteristic information includes: Determine the plastic data collection attribute information according to the material storage sensing module, configure a processing program for the plastic data collection attribute information, and obtain a multi-data attribute preprocessing program; Based on the multi-data attribute preprocessing program, mapping preprocessing is performed on the plastic raw material perception data stream to obtain visual standard image data, standard infrared thermal image data and standard material storage location information; According to the plastic processing and molding standard, the visual standard image data and the standard infrared thermal image data are respectively marked with defect feature recognition to obtain a visual image defect feature set and an infrared image defect feature set; Based on the standard material storage position information, the visual image defect feature set and the infrared image defect feature set are distributed and located to determine the plastic defect feature information; The determining of the plastic defect characteristic information includes: Respectively analyzing the defect type and size quantity of the visual image defect feature set and the infrared image defect feature set to obtain visual defect feature parameter information and infrared defect feature parameter information; Based on the standard material storage position information, the visual defect characteristic parameter information is distributed and located to generate a plastic surface defect distribution map; Positioning and marking the infrared defect characteristic parameter information based on the standard material storage position information to determine the infrared defect distribution characteristic parameter information; The plastic surface defect distribution map is cross-validated using the infrared defect distribution characteristic parameter information to determine the plastic defect characteristic information.

2. The automatic feeding system for plastic processing according to claim 1, characterized in that: The step of obtaining the plastic raw material sensing data stream includes: Acquire visual image data of the plastic storage bin through the CMOS image device, perform convolution feature extraction on the visual image data, and obtain a multi-dimensional feature set of plastics; Based on the plastic multi-dimensional feature set, the visual image data is divided and marked with regions of interest to determine a plastic detection area of ​​interest; The micro-thermal imaging lens is used to collect information imaging of the plastic detection area of ​​interest to generate plastic infrared thermal image data, and at the same time, the plastic storage position information of the plastic storage bin is obtained through the material level sensor; The plastic raw material perception data stream is acquired based on the visual image data, the plastic infrared thermal image data and the plastic storage location information.

3. The automatic feeding system for plastic processing according to claim 1, characterized in that: The determining and matching feeding control parameters comprises: Acquire plastic molding attribute information, wherein the plastic molding attribute information includes plastic molding specification requirements, plastic physical and chemical properties, and molding demand, and classify the plastic feeding control space according to the plastic molding attribute information to obtain plastic molding attribute characteristic parameters; Dividing and marking the plastic feeding control space based on the plastic molding attribute characteristic parameters to obtain a plastic feeding control characteristic database; Analyzing the target plastic molding requirements to obtain plastic molding property requirement parameters; The plastic feeding control feature database is traversed and matched and retrieved based on the plastic molding property requirement parameters to determine the matching feeding control parameters.

4. The automatic feeding system for plastic processing according to claim 3, characterized in that: The obtaining of the target feeding control parameters comprises: Performing characteristic loss analysis on the plastic molding property requirement parameters and the matching feeding control parameters through the plastic molding property information to obtain plastic property loss characteristic data; Perform parameter optimization analysis based on the plastic property loss characteristic data to obtain a feed control parameter optimization selection threshold; Constructing a feeding control fitness function, and randomly generating a plurality of feeding control parameters within the feeding control parameter optimization selection threshold value through the feeding control fitness function for evaluation calculation, to obtain a plurality of feeding control effect fitnesses; Cluster learning and optimization are performed on the multiple feeding control parameters based on the multiple feeding control effect fitnesses to determine the target feeding control parameters.

5. The automatic feeding system for plastic processing according to claim 4, characterized in that: The step of obtaining plastic property loss characteristic data includes: Determine attribute feature analysis dimension information according to the plastic molding attribute information, perform weight assignment on the attribute feature analysis dimension information, and obtain feature dimension weight factor information; Based on the feature dimension weight factor information, a loss function is fitted to the attribute feature analysis dimension information to build an attribute feature loss function; The attribute characteristic loss function is used to perform characteristic loss calculation on the plastic molding attribute requirement parameter and the matching feeding control parameter to obtain the plastic attribute loss characteristic data.

6. The automatic feeding system for plastic processing according to claim 5, characterized in that: The determining of the target feeding control parameter comprises: Clustering the multiple feeding control parameters according to the multiple feeding control effect fitnesses to generate feeding control parameter clusters; Based on the preset variation step length, the remaining control parameters in the feeding control parameter cluster are mutated and cross-expanded toward the center of the parameter cluster to obtain an updated control parameter cluster, and the updated control parameter cluster is globally optimized to determine the target feeding control parameters.

Images