Data modeling method and system for second-hand recycling products based on deep learning

Abstract

The present invention relates to the field of product recycling, and proposes a data modeling method and system for realizing second-hand recycling products based on deep learning. The method comprises: obtaining second-hand products to be recycled, mining their attributes, obtaining detailed attribute data, identifying product images, extracting appearance features and structural features, and calculating their corresponding weights, and constructing a primary recycling model according to their corresponding weights, and training it using product attribute mining data, deploying the trained model in a recycling environment, identifying recycling groups and extracting classification indicators, using the primary recycling model to parameterize second-hand products, determining the recycling process according to the recycling marking parameters, identifying the parameters in the process nodes and converting them, and finally obtaining the product recycling model. The present invention can improve the recycling efficiency of second-hand products.

Description

Technical Field

[0001] The present invention relates to the field of product recycling, and in particular to a data modeling method and system for second-hand recycled products based on deep learning. Background Art

[0002] As society pays more attention to sustainable development and resource recycling, the second-hand product market is gradually emerging, and second-hand recycled products have potential business opportunities and environmental significance.

[0003] At present, the traditional processing and management of second-hand recycled products often relies on manual classification and processing, which is inefficient and prone to errors. In addition, since the data of second-hand recycled products usually contains a large number of characteristic data information, it is necessary to extract effective information from the multiple types of characteristic data information and perform accurate classification. This process has great uncertainty, which affects the recycling efficiency of second-hand recycled products. Summary of the invention

[0004] The present invention provides a data modeling method and system for second-hand recycled products based on deep learning, the main purpose of which is to improve the recycling efficiency of second-hand products.

[0005] To achieve the above purpose, the present invention provides a data modeling method for second-hand recycling products based on deep learning, including:

[0006] Acquire the second-hand products to be recycled, perform product attribute mining on the second-hand products, obtain attribute mining data corresponding to the second-hand products, and obtain detailed attribute data after removing abnormal data from the attribute mining data;

[0007] Identify a product image corresponding to the second-hand product, perform feature extraction on the product image to obtain product appearance features and structural features, and calculate appearance feature weights and structural feature weights corresponding to the product appearance features and structural features, respectively;

[0008] Based on the appearance feature weights and the structure feature weights, a primary recycling model corresponding to the product image is constructed, and the primary recycling model is trained using the product attribute mining data to obtain a trained primary recycling model;

[0009] After deploying the trained product primary recycling model to a preset recycling environment, identifying the recycling groups in the preset recycling environment, extracting the classification indicators corresponding to the recycling groups, and based on the classification indicators, using the primary recycling model to mark the recycling parameters of the second-hand products to obtain product recycling marking parameters;

[0010] Based on the product recycling mark parameters, determine the recycling process corresponding to the second-hand product, query the recycling grouping in the preset recycling environment, identify the node parameters in the process nodes, convert the node parameters to obtain conversion parameters, and based on the conversion parameters, perform data modeling on the second-hand product to obtain the final product recycling model.

[0011] Optionally, the performing product attribute mining on the second-hand product to obtain attribute mining data corresponding to the second-hand product includes:

[0012] Collect product data of the second-hand product to obtain a product data source;

[0013] Performing text processing on the product description in the product data source to obtain text keywords;

[0014] Performing image recognition on the product images in the product data source to obtain image recognition features;

[0015] Calculating the product association between the text keyword and the image recognition feature;

[0016] Based on the product association, product attribute mining is performed on the second-hand product to obtain attribute mining data corresponding to the second-hand product.

[0017] Optionally, the calculating the product association between the text keyword and the image recognition feature includes:

[0018] The following formula is used to calculate the product association between the text keyword and the image recognition feature:

[0019]

[0020] Wherein, GD represents the product association between the text keyword and the image recognition feature, c represents the quantity index corresponding to the text keyword and the image feature, V represents the total quantity corresponding to the text keyword and the image feature, TF c represents the TF-IDF value corresponding to the cth keyword in the text, TZ c represents the feature value corresponding to the cth image recognition feature, WL(t) represents the text length at time t, TP(t) represents the image resolution at time t, and t represents the time range.

[0021] Optionally, after removing abnormal data from the attribute mining data, detailed attribute data is obtained, including:

[0022] Performing integrity verification on the attribute mining data to obtain verification data;

[0023] Performing outlier detection on the verification data to obtain outlier detection data;

[0024] Based on the outlier detection data, interpolating the verification data to obtain interpolated data;

[0025] Performing standardization processing on the interpolation data to obtain standardized data;

[0026] Performing redundant attribute screening on the standardized data to obtain redundant screening attributes;

[0027] Redundant screening attributes in the standardized data are removed to obtain detailed attribute data.

[0028] Optionally, the extracting features of the product image to obtain product appearance features and structural features includes:

[0029] Performing image preprocessing on the product image to obtain a preprocessed product image;

[0030] Performing edge detection on the preprocessed product image to obtain a binary image;

[0031] Performing contour extraction on the binary image to obtain contour features;

[0032] Performing shape recognition on the binary image to obtain shape features;

[0033] Using the outline feature and the shape feature as product appearance features corresponding to the product image;

[0034] Performing region segmentation on the preprocessed product image to obtain a region segmentation image;

[0035] Performing feature scanning on the region segmentation image to obtain scanning feature information;

[0036] Performing structural correlation analysis on the scan feature information to obtain structural features.

[0037] Optionally, respectively calculating the appearance feature weights and the structural feature weights corresponding to the product appearance feature and the structural feature includes:

[0038] The following formula is used to calculate the appearance feature weight and structural feature weight corresponding to the product appearance feature and the structural feature:

[0039]

[0040] Wherein, WG represents the appearance feature weight corresponding to the product appearance feature, i represents the quantity index corresponding to the product appearance feature, n represents the total number of product appearance features, f(P i) represents the contour feature function corresponding to the product appearance feature, P i represents the i-th contour feature in the product appearance feature, a and b represent the upper and lower limits of the integral corresponding to the product appearance feature, f(x) represents the shape feature function corresponding to the product appearance feature, x represents the shape feature in the product appearance feature, k represents the feature dimension corresponding to the product appearance feature, j represents the dimension index corresponding to the product appearance feature, g(Q j )

[0041] represents the appearance feature function corresponding to the appearance feature of the product, Q j Indicates the jth appearance feature of the product

[0042] appearance features, WJ represents the structural feature weight corresponding to the structural feature, p represents the number index corresponding to the segmentation feature in the structural feature, m represents the total number of segmentation features, h(R m ) represents the segmentation feature function corresponding to the structural feature, R m represents the mth segmentation feature in the structural feature, s represents the number index of the structural feature, t' represents the total number of the structural features, e(S s ) represents the scanning feature function corresponding to the structural feature, S s represents the sth scanning feature in the structural feature, n' represents the feature dimension corresponding to the structural feature, l represents the dimension index corresponding to the structural feature, l(T i ) represents the structural correlation function corresponding to the structural feature, T i Represents the i-th structure association value in the structural feature.

[0043] Optionally, based on the classification index, using the primary recycling model to mark the second-hand product with recycling parameters to obtain product recycling marking parameters includes:

[0044] Identifying classification indicator data corresponding to the classification indicator;

[0045] Performing data cleaning on the classification indicator data to obtain indicator cleaning data;

[0046] Performing feature extraction on the indicator cleansing data to obtain indicator features;

[0047] Based on the indicator characteristics, the primary recycling model is used to mark the recycling parameters of the second-hand product to obtain product recycling marking parameters.

[0048] Optionally, determining a recycling process corresponding to the second-hand product based on the product recycling mark parameter includes:

[0049] Classifying the product recycling label parameters to obtain a product label category;

[0050] Performing type separation on the product tag categories to obtain separated categories;

[0051] Performing material identification on the separation category to obtain separation materials;

[0052] Performing product recovery on the separated material to obtain a recovered product material;

[0053] After the recycled product materials are reused, a recycling process corresponding to the second-hand product is obtained.

[0054] Optionally, performing node parameter conversion on the node parameters to obtain conversion parameters includes:

[0055] Querying original parameter data corresponding to the node parameters;

[0056] Normalizing the original parameter data to obtain standardized data;

[0057] Performing dimensionality reduction processing on the standard data to obtain dimensionality reduction processing data;

[0058] Perform node parameter conversion on the dimension reduction processed data to obtain conversion parameters.

[0059] In order to solve the above problems, the present invention also provides a data modeling system for second-hand recycling products based on deep learning, the system comprising:

[0060] The attribute mining module is used to obtain the second-hand products to be recycled, perform product attribute mining on the second-hand products, obtain attribute mining data corresponding to the second-hand products, and obtain detailed attribute data after removing abnormal data in the attribute mining data;

[0061] A weight calculation module, used to identify a product image corresponding to the second-hand product, perform feature extraction on the product image to obtain product appearance features and structural features, and calculate appearance feature weights and structural feature weights corresponding to the product appearance features and structural features, respectively;

[0062] A model training module, used to construct a primary recycling model corresponding to the product image based on the appearance feature weights and the structure feature weights, and to train the primary recycling model using the product attribute mining data to obtain a trained primary recycling model;

[0063] A recycling marking module, which is used to deploy the trained product primary recycling model to a preset recycling environment, identify the recycling groups in the preset recycling environment, extract the classification indicators corresponding to the recycling groups, and use the primary recycling model to mark the recycling parameters of the second-hand products based on the classification indicators to obtain product recycling marking parameters;

[0064] A data modeling module is used to determine the recycling process corresponding to the second-hand product based on the product recycling mark parameters, query the recycling grouping in the preset recycling environment, identify the node parameters in the process nodes, perform node parameter conversion on the node parameters to obtain conversion parameters, and perform data modeling on the second-hand product based on the conversion parameters to obtain a final product recycling model.

[0065] The present invention obtains second-hand products to be recycled, performs product attribute mining on the second-hand products, and obtains attribute mining data corresponding to the second-hand products, which can more accurately grasp the characteristics and advantages of the products, thereby improving the competitiveness of the products and meeting the needs of consumers. At the same time, it is also helpful to environmental protection, extend the life cycle of products, promote resource recycling, and achieve sustainable development goals. The present invention helps to obtain more comprehensive product information by identifying the product image corresponding to the second-hand product, thereby improving the accuracy and completeness of the data and providing a more reliable data basis for subsequent modeling. The present invention constructs a primary recycling model corresponding to the product image based on the appearance feature weight and the structural feature weight, which can fully capture the visual information and structural properties of the product image, and can be analyzed in multiple aspects, so as to better preliminarily identify and classify different types of products. After the trained product primary recycling model is deployed to a preset recycling environment, the present invention can collect real-time data feedback for continuously optimizing the recycling model to adapt to new situations and improve performance. The present invention determines the recycling process corresponding to the second-hand product based on the product recycling mark parameter, and can accurately judge the recyclability and value of the second-hand product, thereby maximizing the recycling of resources, avoiding waste and improving the recycling rate of resources. Therefore, the data modeling method and system for second-hand recycling products based on deep learning proposed in the present invention can improve the efficiency of second-hand product recycling. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A schematic diagram of a flow chart of a method for implementing data modeling of second-hand recycled products based on deep learning provided by an embodiment of the present invention;

[0067] Figure 2 A schematic diagram of a module of a data modeling system for second-hand recycling products based on deep learning provided by an embodiment of the present invention;

[0068] Figure 3A schematic diagram of the internal structure of an electronic device that implements a data modeling method for second-hand recycled products based on deep learning provided in one embodiment of the present invention.

[0069] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0070] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0071] The embodiment of the present application provides a data modeling method for second-hand recycled products based on deep learning. The execution subject of the data modeling method for second-hand recycled products based on deep learning includes but is not limited to at least one of the electronic devices such as the server, the terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the data modeling method for second-hand recycled products based on deep learning can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

[0072] Reference Figure 1 FIG. 1 is a flow chart of a method for implementing data modeling of second-hand recycled products based on deep learning according to an embodiment of the present invention. In this embodiment, the method for implementing data modeling of second-hand recycled products based on deep learning includes:

[0073] S1. Obtain second-hand products to be recycled, perform product attribute mining on the second-hand products, obtain attribute mining data corresponding to the second-hand products, and obtain detailed attribute data after removing abnormal data in the attribute mining data.

[0074] The present invention obtains second-hand products to be recycled, mines product attributes of the second-hand products, and obtains attribute mining data corresponding to the second-hand products, which can more accurately grasp the characteristics and advantages of the products, thereby improving the competitiveness of the products and meeting the needs of consumers. At the same time, it also helps to protect the environment, extend the life cycle of products, promote resource recycling, and achieve sustainable development goals.

[0075] Among them, the second-hand products refer to goods that are released by the original users after a certain use cycle and are prepared for recycling and reuse, usually including: electronic equipment, furniture, clothing and other items; the attribute mining data refers to the information about the characteristics, quality, value and other aspects of second-hand products obtained through data analysis and processing, which may include: product specifications, brand, age, usage status, market demand and other aspects of data.

[0076] As an embodiment of the present invention, the method of performing product attribute mining on the second-hand product to obtain attribute mining data corresponding to the second-hand product includes: collecting product data on the second-hand product to obtain a product data source; performing text processing on the product description in the product data source to obtain text keywords; performing image recognition on the product image in the product data source to obtain image recognition features; calculating the product association degree corresponding to the text keywords and the image recognition features; and performing product attribute mining on the second-hand product based on the product association degree to obtain attribute mining data corresponding to the second-hand product.

[0077] Among them, the product data source refers to the data source of the second-hand product, including product descriptions, images and other information; the text keywords refer to the keyword information extracted from the product data source through text processing; the image recognition features refer to the feature information extracted after identifying the product images in the product data source; the product association degree refers to measuring the attribute association degree of second-hand products by calculating the correlation between text keywords and image recognition features.

[0078] Furthermore, the product data collection can be achieved through ETL tools, such as Talend, Apache NiFi, etc.; the text processing can be achieved through deep learning frameworks, such as TensorFlow, PyTorch and other tools; the image recognition can be achieved through convolutional neural networks, such as AlexNet, VGG, ResNet, etc.; the product association can be achieved through the following algorithms; the product attribute mining can be achieved through NLTK tools, such as word segmentation, part-of-speech tagging, syntactic analysis and other tools.

[0079] As an embodiment of the present invention, the calculating the product association between the text keyword and the image recognition feature includes:

[0080] The following formula is used to calculate the product association between the text keyword and the image recognition feature:

[0081]

[0082] Wherein, GD represents the product association between the text keyword and the image recognition feature, c represents the quantity index corresponding to the text keyword and the image feature, V represents the total quantity corresponding to the text keyword and the image feature, TF c represents the TF-IDF value corresponding to the cth keyword in the text, TZ c represents the feature value corresponding to the cth image recognition feature, WL(t) represents the text length at time t, TP(t) represents the image resolution at time t, and t represents the time range.

[0083] The present invention obtains detailed attribute data by removing abnormal data from the attribute mining data, which can better analyze data patterns, discover potential trends, and make more accurate predictions and decisions, thereby bringing more beneficial and practical effects to business and research.

[0084] The detailed attribute data refers to a cleaned and processed data set, which contains the accurate value or description of each attribute and eliminates outliers and erroneous information.

[0085] As an embodiment of the present invention, after removing the abnormal data in the attribute mining data, detailed attribute data is obtained, including: performing integrity verification on the attribute mining data to obtain verification data; performing outlier detection on the verification data to obtain outlier detection data; based on the outlier detection data, performing interpolation processing on the verification data to obtain interpolation data; performing standardization processing on the interpolation data to obtain standardized data; performing redundant attribute screening on the standardized data to obtain redundant screened attributes; removing redundant screened attributes in the standardized data to obtain detailed attribute data.

[0086] Among them, the verification data refers to the data after integrity verification; the outlier detection data refers to the outliers identified in the verification data; the interpolation data refers to the data obtained after the interpolation processing of the outlier detection data; the standardized data refers to the data obtained by standardizing the interpolation data; the redundant screening attributes refer to the attributes that are determined to be redundant after screening; the detailed attribute data refers to the final data obtained after removing the redundant screening attributes.

[0087] Furthermore, the integrity verification can be achieved through a verification algorithm, such as: a differential comparison algorithm, a CRC check algorithm, a BDS algorithm, etc.; the outlier detection can be achieved through the ELKI tool; the interpolation processing can be achieved through an interpolation model, such as: Kriging, radial basis function interpolation and other models; the standardization processing can be achieved through a range scaling method; the redundant attribute screening can be achieved through a recursive feature elimination algorithm.

[0088] S2. Identify a product image corresponding to the second-hand product, perform feature extraction on the product image to obtain product appearance features and structural features, and respectively calculate appearance feature weights and structural feature weights corresponding to the product appearance features and structural features.

[0089] The present invention helps to obtain more comprehensive product information by identifying the product image corresponding to the second-hand product, thereby improving the accuracy and completeness of the data and providing a more reliable data basis for subsequent modeling.

[0090] The product image refers to the visual representation of a second-hand product, that is, a photograph or image of the product that has been taken or collected. Optionally, the product image can be obtained through a deep learning framework, such as TensorFlow, PyTorch and other tools.

[0091] The present invention obtains product appearance features and structural features by performing feature extraction on the product image, and can effectively extract appearance and structural features from the product image, realize automatic identification of different types of products, and improve subsequent product classification efficiency.

[0092] The product appearance features include contour features and shape features, which are used to describe the appearance of the product; and the structural features refer to the internal structural features of the product image obtained through structural association analysis.

[0093] As an embodiment of the present invention, the feature extraction of the product image to obtain product appearance features and structural features includes: performing image preprocessing on the product image to obtain a preprocessed product image; performing edge detection on the preprocessed product image to obtain a binary image; performing contour extraction on the binary image to obtain contour features; performing shape recognition on the binary image to obtain shape features; using the contour features and the shape features as product appearance features corresponding to the product image; performing region segmentation on the preprocessed product image to obtain a region segmentation image; performing feature scanning on the region segmentation image to obtain scan feature information; and performing structural association analysis on the scan feature information to obtain structural features.

[0094] Among them, the preprocessed product image refers to the original product image that has been preprocessed; the binarized image refers to the image obtained by converting the original product image into black and white; the contour feature refers to the edge feature of the object extracted from the binarized image; the shape feature refers to the shape-based feature description used to characterize the shape of the object in the image; the region segmentation image refers to the segmented image segmented into different regions; and the scanning feature information refers to the feature information extracted from the region segmentation image.

[0095] Furthermore, the image preprocessing can be achieved through image processing tools, such as OpenCV tools; the edge detection can be achieved through the Canny edge detection algorithm; the contour extraction can be achieved through the findContours function in the OpenCV library; the shape recognition can be achieved through a shape matching algorithm; the region segmentation can be achieved through an image segmentation algorithm, such as region growing, watershed algorithm and other algorithms; the feature scanning can be achieved through a feature detection algorithm, such as SIFT, SURF, HOG and other algorithms; the structural association analysis can be achieved through a structural association algorithm.

[0096] As an embodiment of the present invention, the respectively calculating the appearance feature weights and the structural feature weights corresponding to the product appearance feature and the structural feature includes:

[0097] The appearance feature weight and the structural feature weight corresponding to the product appearance feature and the structural feature are calculated using the following formula:

[0098]

[0099] Wherein, WG represents the appearance feature weight corresponding to the product appearance feature, i represents the quantity index corresponding to the product appearance feature, n represents the total number of product appearance features, f(P i ) represents the contour feature function corresponding to the product appearance feature, P i represents the i-th contour feature in the product appearance feature, a and b represent the upper and lower limits of the integral corresponding to the product appearance feature, f(x) represents the shape feature function corresponding to the product appearance feature, x represents the shape feature in the product appearance feature, k represents the feature dimension corresponding to the product appearance feature, j represents the dimension index corresponding to the product appearance feature, g(Q j ) represents the appearance feature function corresponding to the appearance feature of the product, Q j represents the jth appearance feature in the product appearance feature, WJ represents the structural feature weight corresponding to the structural feature, p represents the number index corresponding to the segmentation feature in the structural feature, m represents the total number of segmentation features, h(R m ) represents the segmentation feature function corresponding to the structural feature, R m represents the mth segmentation feature in the structural feature, s represents the number index of the structural feature, t' represents the total number of the structural features, e(S s ) represents the scanning feature function corresponding to the structural feature, S s represents the sth scanning feature in the structural feature, n' represents the feature dimension corresponding to the structural feature, l represents the dimension index corresponding to the structural feature, l(T i) represents the structural correlation function corresponding to the structural feature, T i Represents the i-th structure association value in the structural feature.

[0100] In detail, the contour feature function is used to describe the edge or contour information of the product image, which usually involves edge detection, contour extraction and other operations in image processing; the shape feature function refers to a function that describes and analyzes the shape, including shape-related features such as area, perimeter, and compactness; the appearance feature value refers to a certain visual feature of the product image, which can be a numerical representation of information such as color, texture, brightness, etc., which helps to distinguish different appearance characteristics; the segmentation feature function is used to describe the features of the image or data after segmentation, which may include statistical features and boundary information of each segmented area; the feature scanning information describes the various feature information obtained when scanning the data or image, which may include scanning mode, scanning depth, scanning frequency, etc.; the structural association function is used to analyze the degree of association or structural relationship between different elements, which may include network connection strength, node interaction, etc., and is suitable for exploring the internal structure or association of the system.

[0101] S3. Based on the appearance feature weights and the structure feature weights, a primary recycling model corresponding to the product image is constructed, and the primary recycling model is trained using the product attribute mining data to obtain a trained primary recycling model.

[0102] The present invention constructs a primary recycling model corresponding to the product image based on the appearance feature weights and the structural feature weights, which can comprehensively capture the visual information and structural properties of the product image and can perform analysis in multiple aspects, thereby better preliminarily identifying and classifying different types of products.

[0103] The primary recycling model refers to a basic model for classifying, identifying or recycling product images based on given appearance feature weights and structural feature weights. Optionally, the primary recycling model can be implemented in Python programming language, such as PyTorch, TensorFlow, Scikit-learn and other languages.

[0104] In detail, the primary recycling model mainly uses appearance features and structural feature weights to analyze the visual information and structural properties of product images to achieve preliminary classification or identification of different types of products. In addition, the primary recycling model usually serves as a basic component in the system, providing a basis for subsequent more complex and refined recycling processing, and can be adjusted and optimized as needed in practice.

[0105] The present invention uses the product attribute mining data to train the primary recycling model to obtain a trained primary recycling model, which can learn more and richer product appearance and structural features, thereby improving the accuracy of the model in identifying and classifying product images.

[0106] The trained primary recycling model refers to a model with high accuracy and generalization ability after being trained using product attribute mining data. Optionally, the model training of the primary recycling model can be achieved through the TensorFlow tool.

[0107] In detail, the trained primary recycling model can be used as a basic model to achieve preliminary product classification and identification, and provide basic support for the execution of subsequent more complex tasks. In addition, the trained primary recycling model can quickly and accurately perform preliminary processing on different types of products in practice by learning the characteristics and rules in a large amount of product attribute mining data, thereby improving the efficiency and accuracy of the recycling process.

[0108] S4. After deploying the trained product primary recycling model to a preset recycling environment, identifying the recycling groups in the preset recycling environment, extracting the classification indicators corresponding to the recycling groups, and based on the classification indicators, using the primary recycling model to mark the recycling parameters of the second-hand products to obtain product recycling marking parameters.

[0109] After the trained product primary recycling model is deployed to a preset recycling environment, the present invention can collect real-time data feedback for continuously optimizing the recycling model to adapt to new situations and improve performance.

[0110] The preset recycling environment refers to an environment or system for product recycling that is planned and designed in advance, and often includes: physical facilities, technical equipment, data management systems and other environments. Optionally, the deployment of the trained product primary recycling model to the preset recycling environment can be achieved through the TensorFlow Serving tool.

[0111] The present invention can realize more intelligent, efficient and sustainable product recycling management by identifying recycling groups in the preset recycling environment and extracting classification indicators corresponding to the recycling groups, thereby making a positive contribution to environmental protection and resource recycling.

[0112] Among them, the recycling grouping refers to the process of classifying and grouping waste products or items according to their characteristics, types or materials and other attributes, and at the same time, it can help effectively manage and handle different types of waste for proper recycling, reuse or treatment; the classification index refers to the standard or parameter used to describe and distinguish the characteristics or attributes between different recycling groups, often including: material type, recyclability and treatment method and other indicators. Optionally, the identification of recycling groups in the preset recycling environment can be obtained through the OpenCV tool; the extraction of the classification index corresponding to the recycling group can be obtained through the NumPy, Pandas and other libraries in Python.

[0113] The present invention uses the primary recycling model to mark the recycling parameters of the second-hand products based on the classification indicators to obtain product recycling marking parameters, which can realize the intelligence and personalization of the recycling process, promote sustainable development and maximize the recycling and reuse of resources, thereby achieving positive impacts on both environmental protection and economic levels.

[0114] The product recycling marking parameters refer to the recycling attributes and marking information of the product described according to the output of the primary recycling model.

[0115] As an embodiment of the present invention, based on the classification index, the recycling parameter marking of the second-hand product is performed using the primary recycling model to obtain product recycling marking parameters, including: identifying classification index data corresponding to the classification index; performing data cleaning on the classification index data to obtain indicator cleaning data; performing feature extraction on the indicator cleaning data to obtain indicator features; based on the indicator features, the recycling parameter marking of the second-hand product is performed using the primary recycling model to obtain product recycling marking parameters.

[0116] Among them, the classification index data refers to the key information of second-hand products, such as product type, brand, newness, etc.; the index cleaning data refers to the classification index data that has been cleaned; and the index feature refers to the feature extracted from the cleaned data for training the model.

[0117] Furthermore, the identification of the classification indicator data corresponding to the classification indicator can be obtained through a named entity recognition algorithm, such as: conditional random field, recursive neural network and other algorithms; the data cleaning of the classification indicator data can be obtained through a missing value processing algorithm, such as: mean / median filling, K nearest neighbor filling, interpolation method and other algorithms; the feature extraction of the indicator cleaning data can be obtained through feature engineering tools, such as: scikit-learn, Featuretools, tsfresh and other tools; the use of the primary recycling model to label the recycling parameters of the second-hand product can be obtained through a machine learning framework, such as: TensorFlow, PyTorch, Scikit-learn and other tools.

[0118] S5. Based on the product recycling mark parameters, determine the recycling process corresponding to the second-hand product, query the recycling group in the preset recycling environment, identify the node parameters in the process nodes, perform node parameter conversion on the node parameters to obtain conversion parameters, and based on the conversion parameters, perform data modeling on the second-hand product to obtain a final product recycling model.

[0119] The present invention can accurately judge the recyclability and value of the second-hand product by determining the recycling process corresponding to the second-hand product based on the product recycling mark parameters, thereby maximizing the recycling of resources, avoiding waste and improving the recycling rate of resources.

[0120] The recycling process refers to a series of systematic operational steps or procedures for collecting, sorting, processing and reusing second-hand products or waste.

[0121] As an embodiment of the present invention, determining the recycling process corresponding to the second-hand product based on the product recycling marking parameters includes: performing product marking classification on the product recycling marking parameters to obtain product marking categories; performing type separation on the product marking categories to obtain separation categories; performing material identification on the separation categories to obtain separation materials; performing product recycling on the separation materials to obtain recycled product materials; and obtaining the recycling process corresponding to the second-hand product after reusing the recycled product materials.

[0122] Among them, the product marking category refers to the category for classifying products according to product recycling marking parameters; the separation category refers to the different categories obtained after type separation based on the product marking category; the separation material refers to the specific material type identified after separation category; the recycled product material refers to the specific material used for product recycling after material identification.

[0123] Furthermore, the product labeling classification of the product recycling labeling parameters can be achieved by using the SVM model in the Scikit-learn library or the CNN model in TensorFlow to classify the products; the type separation of the product labeling categories can be achieved by separating the product labeling categories using the K-means algorithm; the material identification of the separated categories can be achieved by using the YOLO algorithm; the product recycling of the separated materials can be achieved by using computer vision technology to guide automated robots to perform product recycling tasks, such as: using a robotic arm to remove or separate specific materials; the reuse of the recycled product materials can be achieved by applying a circular economy management system and reprocessing the recycled product materials through resource recovery and remanufacturing technology.

[0124] In detail, the specific steps of the recycling process may include: disassembling the second-hand product into components to obtain product components; extracting component attribute values ​​corresponding to the product components; identifying component characteristic values ​​corresponding to the product components; calculating the product evaluation value corresponding to the second-hand product based on the component attribute values ​​and the component characteristic values; constructing a price evaluation model corresponding to the second-hand product based on the product evaluation value and preset external second-hand product value data; and using the price evaluation model to achieve price correction and market forecasting of the second-hand product.

[0125] Among them, the product components refer to the various components obtained after disassembling a second-hand product; the component attribute values ​​refer to the extracted attribute information related to the product components; the component characteristic values ​​refer to the identified values ​​or characteristics used to describe the characteristics of the product components; the product evaluation value refers to the valuation of the second-hand product calculated based on the component attribute values ​​and component characteristic values; the preset external second-hand product value data refers to the external second-hand product price data extracted from external facilities for reference; the price evaluation model refers to a model constructed based on the product evaluation value and external second-hand product value data, which is used for second-hand product price evaluation, correction and market forecasting.

[0126] Furthermore, the extraction of component attribute values ​​corresponding to the product components can be obtained through a decision tree algorithm in machine learning, such as Random Forest, XGBoost, etc.; the identification of component feature values ​​corresponding to the product components can be obtained through a feature extraction and description algorithm, such as SIFT, SURF, ORB, etc.; the calculation of the product evaluation value corresponding to the second-hand product can be obtained through a regression algorithm, such as linear regression, polynomial regression, support vector regression, etc.; the construction of a price evaluation model corresponding to the second-hand product can be obtained through the Scikit-learn library in the Python programming language.

[0127] The present invention can quickly and accurately retrieve required recycling group information by querying the recycling group in the preset recycling environment and identifying the node parameters in the process node, which helps to optimize the recycling process and improve the recycling efficiency.

[0128] Among them, the recycling grouping refers to the process of classifying or grouping recycled items according to specific attributes or standards in the recycling process; the node parameters refer to the key parameters or variables in each stage or step in the recycling process. Optionally, the query of the recycling grouping in the preset recycling environment can be obtained through a database management system, such as MySQL, PostgreSQL, MongoDB, etc.; the identification of the node parameters in the process node can be obtained through data acquisition equipment, such as sensors, monitoring equipment and other equipment.

[0129] The present invention converts the node parameters into node parameters to obtain conversion parameters, which can improve data quality, simplify the data processing process, and enhance model performance, thereby better supporting decision-making and optimizing process management.

[0130] The conversion parameters refer to result data obtained by performing node parameter conversion on the dimension reduction processing data.

[0131] As an embodiment of the present invention, the node parameters are converted into node parameters to obtain conversion parameters, including: querying the original parameter data corresponding to the node parameters; normalizing the original parameter data to obtain normalized data; performing dimensionality reduction processing on the normalized data to obtain dimensionality reduction processing data; performing node parameter conversion on the dimensionality reduction processing data to obtain conversion parameters.

[0132] Among them, the original parameter data refers to the initial data directly extracted from the node; the standard data refers to the data after standardization processing, which is used to ensure the consistency of the data; the dimension reduction processing data refers to the data obtained after reducing the data dimension using dimension reduction technology.

[0133] Furthermore, the query of the original parameter data corresponding to the node parameters can be obtained through data collection tools, such as Scrapy, BeautifulSoup and other tools; the normalization of the original parameter data can be obtained through the Z-score standardization algorithm; the dimensionality reduction of the normalized data can be obtained through the principal component analysis model; the node parameter conversion of the dimensionality reduction data can be obtained through Scikit-learn tools, such as linear regression and decision tree.

[0134] The present invention performs data modeling on the second-hand products based on the conversion parameters to obtain a final product recycling model, which can provide data support to help enterprises more accurately evaluate which second-hand products are suitable for recycling, thereby optimizing the recycling process and resource allocation.

[0135] The final product recycling model refers to a trained and optimized data model for predicting, evaluating and deciding which second-hand products are suitable for recycling, and how to efficiently recycle and reuse products. Optionally, the data modeling of the second-hand products can be achieved through the Python programming language, such as tools such as NumPy, Pandas and Scikit-learn.

[0136] The present invention obtains second-hand products to be recycled, performs product attribute mining on the second-hand products, and obtains attribute mining data corresponding to the second-hand products, which can more accurately grasp the characteristics and advantages of the products, thereby improving the competitiveness of the products and meeting the needs of consumers. At the same time, it is also helpful to environmental protection, extend the life cycle of products, promote resource recycling, and achieve sustainable development goals. The present invention helps to obtain more comprehensive product information by identifying the product image corresponding to the second-hand product, thereby improving the accuracy and completeness of the data and providing a more reliable data basis for subsequent modeling. The present invention constructs a primary recycling model corresponding to the product image based on the appearance feature weight and the structural feature weight, which can fully capture the visual information and structural properties of the product image, and can be analyzed in multiple aspects, so as to better preliminarily identify and classify different types of products. After the trained product recycling model is deployed to a preset recycling environment, the present invention can collect real-time data feedback for continuously optimizing the recycling model to adapt to new situations and improve performance. The present invention determines the recycling process corresponding to the second-hand product based on the product recycling mark parameter, and can accurately judge the recyclability and value of the second-hand product, thereby maximizing the recycling of resources, avoiding waste and improving the recycling rate of resources. Therefore, the data modeling method and system for second-hand recycling products based on deep learning proposed in the present invention can improve the efficiency of second-hand product recycling.

[0137] like Figure 2 As shown, it is a functional module diagram of a data modeling system for second-hand recycling products based on deep learning provided by an embodiment of the present invention.

[0138] The data modeling system 200 for second-hand recycled products based on deep learning of the present invention can be installed in an electronic device. According to the functions implemented, the data modeling system 200 for second-hand recycled products based on deep learning can include an attribute mining module 201, a weight calculation module 202, a model training module 203, a recycling marking module 204 and a data modeling module 205. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0139] In this embodiment, the functions of each module / unit are as follows:

[0140] The attribute mining module 201 is used to obtain the second-hand products to be recycled, perform product attribute mining on the second-hand products, obtain attribute mining data corresponding to the second-hand products, and obtain detailed attribute data after removing abnormal data from the attribute mining data;

[0141] The weight calculation module 202 is used to identify the product image corresponding to the second-hand product, extract features from the product image, obtain product appearance features and structural features, and calculate the appearance feature weights and structural feature weights corresponding to the product appearance features and structural features, respectively;

[0142] The model training module 203 is used to construct a primary recycling model corresponding to the product image based on the appearance feature weights and the structure feature weights, and to train the primary recycling model using the product attribute mining data to obtain a trained primary recycling model;

[0143] The recycling marking module 204 is used to deploy the trained product primary recycling model to a preset recycling environment, identify the recycling groups in the preset recycling environment, extract the classification indicators corresponding to the recycling groups, and use the primary recycling model to mark the recycling parameters of the second-hand products based on the classification indicators to obtain product recycling marking parameters;

[0144] The data modeling module 205 is used to determine the recycling process corresponding to the second-hand product based on the product recycling mark parameters, query the recycling grouping in the preset recycling environment, identify the node parameters in the process nodes, convert the node parameters to obtain conversion parameters, and perform data modeling on the second-hand product based on the conversion parameters to obtain the final product initial recycling model.

[0145] In detail, each module described in the data modeling system 200 for second-hand recycled products based on deep learning described in the embodiment of the present invention adopts the same technical means as the data modeling method for second-hand recycled products based on deep learning described in the accompanying drawings when used, and can produce the same technical effects, which will not be repeated here.

[0146] like Figure 3 As shown, it is a structural schematic diagram of an electronic device that implements a data modeling method for second-hand recycling products based on deep learning according to the present invention.

[0147] The electronic device may include a processor 30, a memory 31, a communication bus 32 and a communication interface 33, and may also include a computer program stored in the memory 31 and executable on the processor 30, such as an engineering safety supervision program based on artificial intelligence.

[0148] In some embodiments, the processor 30 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 30 is the control core (ControlUnit) of the electronic device, and uses various interfaces and lines to connect the various components of the entire electronic device, and executes or executes programs or modules stored in the memory 31 (for example, executing engineering safety supervision programs based on artificial intelligence, etc.), and calls the data stored in the memory 31 to execute various functions of the electronic device and process data.

[0149] The memory 31 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory 31 may be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory 31 may also be an external storage device of an electronic device, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Further, the memory 31 may also include both an internal storage unit of the electronic device and an external storage device. The memory 31 may not only be used to store application software and various types of data installed in the electronic device, such as the code of the database configuration connection program, but also be used to temporarily store data that has been output or is to be output.

[0150] The communication bus 32 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 31 and at least one processor 30, etc.

[0151] The communication interface 33 is used for communication between the above-mentioned electronic device 3 and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally, the user interface may also be a standard wired interface, a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device and to display a visual user interface.

[0152] Figure 3 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 3The structure shown does not constitute a limitation on the electronic device, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0153] For example, although not shown, the electronic device may also include a power source (such as a battery) for supplying power to various components. Preferably, the power source may be logically connected to the at least one processor 30 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include any components such as one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, etc. The electronic device may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0154] It should be understood that the embodiment is for illustrative purposes only and the scope of the patent invention is not limited to this structure.

[0155] The database configuration connection program stored in the memory 31 in the electronic device is a combination of multiple computer programs. When running in the processor 30, it can achieve:

[0156] Acquire the second-hand products to be recycled, perform product attribute mining on the second-hand products, obtain attribute mining data corresponding to the second-hand products, and obtain detailed attribute data after removing abnormal data from the attribute mining data;

[0157] Identify a product image corresponding to the second-hand product, perform feature extraction on the product image to obtain product appearance features and structural features, and calculate appearance feature weights and structural feature weights corresponding to the product appearance features and structural features, respectively;

[0158] Based on the appearance feature weights and the structure feature weights, a primary recycling model corresponding to the product image is constructed, and the primary recycling model is trained using the product attribute mining data to obtain a trained primary recycling model;

[0159] After deploying the trained product primary recycling model to a preset recycling environment, identifying the recycling groups in the preset recycling environment, extracting the classification indicators corresponding to the recycling groups, and based on the classification indicators, using the primary recycling model to mark the recycling parameters of the second-hand products to obtain product recycling marking parameters;

[0160] Based on the product recycling mark parameters, determine the recycling process corresponding to the second-hand product, query the recycling grouping in the preset recycling environment, identify the node parameters in the process nodes, convert the node parameters to obtain conversion parameters, and based on the conversion parameters, perform data modeling on the second-hand product to obtain the final product recycling model.

[0161] Specifically, the specific implementation method of the processor 30 for the above computer program can refer to Figure 1 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0162] Furthermore, if the module / unit integrated in the electronic device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium. The storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0163] The present invention further provides a storage medium, wherein the readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:

[0164] Acquire the second-hand products to be recycled, perform product attribute mining on the second-hand products, obtain attribute mining data corresponding to the second-hand products, and obtain detailed attribute data after removing abnormal data from the attribute mining data;

[0165] Identify a product image corresponding to the second-hand product, perform feature extraction on the product image to obtain product appearance features and structural features, and calculate appearance feature weights and structural feature weights corresponding to the product appearance features and structural features, respectively;

[0166] Based on the appearance feature weights and the structure feature weights, a primary recycling model corresponding to the product image is constructed, and the primary recycling model is trained using the product attribute mining data to obtain a trained primary recycling model;

[0167] After deploying the trained product primary recycling model to a preset recycling environment, identifying the recycling groups in the preset recycling environment, extracting the classification indicators corresponding to the recycling groups, and based on the classification indicators, using the primary recycling model to mark the recycling parameters of the second-hand products to obtain product recycling marking parameters;

[0168] Based on the product recycling mark parameters, determine the recycling process corresponding to the second-hand product, query the recycling grouping in the preset recycling environment, identify the node parameters in the process nodes, convert the node parameters to obtain conversion parameters, and based on the conversion parameters, perform data modeling on the second-hand product to obtain the final product recycling model.

[0169] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0170] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0171] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0172] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0173] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0174] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0175] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A data modeling method for second-hand recycling products based on deep learning, characterized in that: The method comprises: Acquire the second-hand products to be recycled, perform product attribute mining on the second-hand products, obtain attribute mining data corresponding to the second-hand products, and obtain detailed attribute data after removing abnormal data from the attribute mining data; Identify a product image corresponding to the second-hand product, perform feature extraction on the product image to obtain product appearance features and structural features, and calculate appearance feature weights and structural feature weights corresponding to the product appearance features and structural features, respectively; Based on the appearance feature weights and the structure feature weights, a primary recycling model corresponding to the product image is constructed, and the primary recycling model is trained using the attribute mining data of the product to obtain a trained primary recycling model; After deploying the trained product primary recycling model to a preset recycling environment, identifying the recycling groups in the preset recycling environment, extracting the classification indicators corresponding to the recycling groups, and based on the classification indicators, using the primary recycling model to mark the recycling parameters of the second-hand products to obtain product recycling marking parameters; Based on the product recycling mark parameters, determine the recycling process corresponding to the second-hand product, query the recycling grouping in the preset recycling environment, identify the node parameters in the process nodes, convert the node parameters to obtain conversion parameters, and based on the conversion parameters, perform data modeling on the second-hand product to obtain the final product recycling model.

2. The method for realizing data modeling of second-hand recycled products based on deep learning according to claim 1, characterized in that: The performing product attribute mining on the second-hand product to obtain attribute mining data corresponding to the second-hand product includes: Collect product data of the second-hand product to obtain a product data source; Performing text processing on the product description in the product data source to obtain text keywords; Performing image recognition on the product images in the product data source to obtain image recognition features; Calculating the product association between the text keyword and the image recognition feature; Based on the product association, product attribute mining is performed on the second-hand product to obtain attribute mining data corresponding to the second-hand product.

3. The data modeling method for second-hand recycling products based on deep learning as claimed in claim 2 is characterized in that: The calculating the product association between the text keyword and the image recognition feature includes: The following formula is used to calculate the product association between the text keyword and the image recognition feature: Wherein, GD represents the product association between the text keyword and the image recognition feature, c represents the quantity index corresponding to the text keyword and the image recognition feature, V represents the total quantity corresponding to the text keyword and the image recognition feature, TF c represents the TF-IDF value corresponding to the cth keyword in the text, TZ c represents the feature value corresponding to the cth image recognition feature, WL(t) represents the text length at time t, TP(t) represents the image resolution at time t, and t represents the time range.

4. The method for realizing data modeling of second-hand recycled products based on deep learning according to claim 1, characterized in that: After removing abnormal data from the attribute mining data, detailed attribute data is obtained, including: Performing integrity verification on the attribute mining data to obtain verification data; Performing outlier detection on the verification data to obtain outlier detection data; Based on the outlier detection data, interpolating the verification data to obtain interpolated data; Performing standardization processing on the interpolation data to obtain standardized data; Performing redundant attribute screening on the standardized data to obtain redundant screening attributes; Redundant screening attributes in the standardized data are removed to obtain detailed attribute data.

5. The method for realizing data modeling of second-hand recycled products based on deep learning according to claim 1, characterized in that: The feature extraction of the product image to obtain product appearance features and structural features includes: Performing image preprocessing on the product image to obtain a preprocessed product image; Performing edge detection on the preprocessed product image to obtain a binary image; Performing contour extraction on the binary image to obtain contour features; Performing shape recognition on the binary image to obtain shape features; Using the outline feature and the shape feature as product appearance features corresponding to the product image; Performing region segmentation on the preprocessed product image to obtain a region segmentation image; Performing feature scanning on the region segmentation image to obtain scanning feature information; Performing structural correlation analysis on the scan feature information to obtain structural features.

6. The method for realizing data modeling of second-hand recycled products based on deep learning according to claim 1, characterized in that: The respectively calculating the appearance feature weights and the structural feature weights corresponding to the product appearance feature and the structural feature comprises: The appearance feature weight and the structural feature weight corresponding to the product appearance feature and the structural feature are calculated using the following formula: Wherein, WG represents the appearance feature weight corresponding to the product appearance feature, i represents the quantity index corresponding to the product appearance feature, n represents the total number of product appearance features, f(P i ) represents the contour feature function corresponding to the product appearance feature, P i represents the i-th contour feature in the product appearance feature, a and b represent the upper and lower limits of the integral corresponding to the product appearance feature, f(x) represents the shape feature function corresponding to the product appearance feature, x represents the shape feature in the product appearance feature, k represents the feature dimension corresponding to the product appearance feature, j represents the dimension index corresponding to the product appearance feature, g(Q j ) represents the appearance feature function corresponding to the appearance feature of the product, Q j represents the jth appearance feature in the product appearance feature, WJ represents the structural feature weight corresponding to the structural feature, p represents the number index corresponding to the segmentation feature in the structural feature, m represents the total number of segmentation features, h(R m ) represents the segmentation feature function corresponding to the structural feature, R m represents the mth segmentation feature in the structural feature, s represents the number index of the structural feature, t′ represents the total number of the structural features, e(S s ) represents the scanning feature function corresponding to the structural feature, S s represents the sth scanning feature in the structural feature, n′ represents the feature dimension corresponding to the structural feature, l represents the dimension index corresponding to the structural feature, l(T i ) represents the structural correlation function corresponding to the structural feature, T i Represents the i-th structure association value in the structural feature.

7. The method for realizing data modeling of second-hand recycled products based on deep learning according to claim 1, characterized in that: Based on the classification index, the primary recycling model is used to mark the second-hand product with recycling parameters to obtain product recycling marking parameters, including: Identifying classification indicator data corresponding to the classification indicator; Performing data cleaning on the classification indicator data to obtain indicator cleaning data; Performing feature extraction on the indicator cleansing data to obtain indicator features; Based on the indicator characteristics, the primary recycling model is used to mark the recycling parameters of the second-hand product to obtain product recycling marking parameters.

8. The method for realizing data modeling of second-hand recycled products based on deep learning according to claim 1, characterized in that: The determining, based on the product recycling mark parameter, a recycling process corresponding to the second-hand product includes: Classifying the product recycling label parameters to obtain a product label category; Performing type separation on the product tag categories to obtain separated categories; Performing material identification on the separation category to obtain separation materials; Performing product recovery on the separated material to obtain a recovered product material; After the recycled product materials are reused, a recycling process corresponding to the second-hand product is obtained.

9. The method for realizing data modeling of second-hand recycled products based on deep learning according to claim 1, characterized in that: The step of converting the node parameters to obtain conversion parameters includes: Querying original parameter data corresponding to the node parameters; Normalizing the original parameter data to obtain standardized data; Performing dimensionality reduction processing on the standard data to obtain dimensionality reduction processing data; Perform node parameter conversion on the dimension reduction processed data to obtain conversion parameters.

10. A data modeling system for second-hand recycling products based on deep learning, characterized in that: The system is used to implement the data modeling method for second-hand recycling products based on deep learning as described in any one of claims 1 to 9, and comprises: The attribute mining module is used to obtain the second-hand products to be recycled, perform product attribute mining on the second-hand products, obtain attribute mining data corresponding to the second-hand products, and obtain detailed attribute data after removing abnormal data in the attribute mining data; A weight calculation module, used to identify a product image corresponding to the second-hand product, perform feature extraction on the product image to obtain product appearance features and structural features, and calculate appearance feature weights and structural feature weights corresponding to the product appearance features and structural features, respectively; A model training module, used to construct a primary recycling model corresponding to the product image based on the appearance feature weights and the structural feature weights, and to train the primary recycling model using the product attribute mining data to obtain a trained primary recycling model; A recycling marking module, which is used to deploy the trained product primary recycling model to a preset recycling environment, identify the recycling groups in the preset recycling environment, extract the classification indicators corresponding to the recycling groups, and use the primary recycling model to mark the recycling parameters of the second-hand products based on the classification indicators to obtain product recycling marking parameters; A data modeling module is used to determine the recycling process corresponding to the second-hand product based on the product recycling mark parameters, query the recycling grouping in the preset recycling environment, identify the node parameters in the process nodes, perform node parameter conversion on the node parameters to obtain conversion parameters, and perform data modeling on the second-hand product based on the conversion parameters to obtain a final product recycling model.

Images