A warehouse management method and system for chemical production

Abstract

The present application relates to the field of intelligent warehouse management, and specifically to a warehouse management method and system for chemical production. It uses computer vision technology based on deep learning to analyze and process the image data of chemical materials, extracts multi-level chemical material state feature representations, and uses attention mechanism and gated response fusion mechanism to enhance and interactively fuse the state features of chemical materials at different levels to achieve accurate characterization of the state of chemical materials, thereby performing type identification and hazard identification of chemical materials, so as to facilitate safety management according to their hazard levels. In this way, the efficiency and safety of chemical warehouse management can be greatly improved, the potential risks caused by human misjudgment can be reduced, and the intelligentization of warehouse management can be achieved.

Description

Technical Field

[0001] The present application relates to the field of intelligent warehouse management, and specifically to a warehouse management method and system for chemical production. Background Art

[0002] As an important part of the national economy, the efficiency and safety of the production and storage management of the chemical industry are directly related to the stable development of the entire industry. Due to the wide variety of chemical materials, their physical state, chemical properties and potential hazards are different. If they are improperly classified or stored, they can easily cause serious accidents such as fires and explosions, causing huge losses to enterprises and personnel.

[0003] Traditional chemical storage management methods mainly rely on manual material classification, storage and monitoring. However, this method has problems such as low efficiency, error-proneness, and difficulty in real-time monitoring. With the increase in the variety of chemical products and the expansion of production scale, the chemical industry has an increasing demand for intelligence and automation, and traditional management methods can no longer meet the needs of modern chemical production. Therefore, an optimized storage management method and system for chemical production is expected. Summary of the invention

[0004] This application is made in view of the above problems. One object of this application is to provide a storage management method and system for chemical production.

[0005] The embodiment of the present application provides a storage management method for chemical production, which includes:

[0006] Acquire chemical material images collected by a camera;

[0007] Perform multi-level material state feature extraction on chemical material images to obtain a shallow feature map of chemical materials and a deep feature map of chemical materials;

[0008] Performing feature selection and strengthening on the shallow characteristic map of chemical materials and the deep characteristic map of chemical materials respectively to obtain a strengthened shallow characteristic map of chemical materials and a strengthened deep characteristic map of chemical materials;

[0009] Dynamically interactively fuse the shallow feature map of enhanced chemical materials and the deep feature map of enhanced chemical materials to obtain a multi-scale gated fusion feature vector of chemical materials;

[0010] Based on the multi-scale gated fusion feature vector of chemical materials, the hazard level is assigned to chemical materials.

[0011] For example, according to the warehouse management method for chemical production of the embodiment of the present application, multi-level material state feature extraction is performed on the chemical material image to obtain a shallow feature map of the chemical material and a deep feature map of the chemical material, including:

[0012] The chemical material image is input into the material state feature extractor based on the pyramid network to obtain the chemical material shallow feature map and the chemical material deep feature map.

[0013] For example, according to the storage management method for chemical production of an embodiment of the present application, the shallow characteristic map of chemical materials and the deep characteristic map of chemical materials are respectively subjected to feature selection and enhancement to obtain an enhanced shallow characteristic map of chemical materials and an enhanced deep characteristic map of chemical materials, including:

[0014] The shallow feature map of chemical materials and the deep feature map of chemical materials are respectively input into the feature attention selection enhancement module based on the compression-suppression structure to obtain the enhanced shallow feature map of chemical materials and the enhanced deep feature map of chemical materials.

[0015] For example, according to the warehouse management method for chemical production of an embodiment of the present application, the shallow feature map of chemical materials and the deep feature map of chemical materials are respectively input into the feature attention selection enhancement module based on the compression-suppression structure to obtain the enhanced shallow feature map of chemical materials and the enhanced deep feature map of chemical materials, including:

[0016] Calculate the global mean of each feature matrix along the channel dimension of the shallow feature map of the chemical material to obtain a feature vector representing the shallow compressed information of the chemical material;

[0017] Performing one-dimensional convolution encoding on the characteristic vector representing the shallow compression information of chemical materials to obtain the characteristic vector representing the correlation between the shallow compression information of chemical materials;

[0018] The characteristic vector representing the shallow compression information of chemical materials and the characteristic vector representing the correlation between the shallow compression information of chemical materials are cascaded to obtain a multi-scale representation vector of the shallow feature compression information of chemical materials;

[0019] Inputting the multi-scale representation vector of the shallow feature compression information of the chemical material into the compression information feature extraction module including the multi-layer perceptron and the SiLU activation function to obtain the multi-scale correlation feature vector of the shallow feature compression information of the chemical material;

[0020] The Sigmoid function is used to normalize the multi-scale correlation feature vector of the shallow feature compression information of chemical materials to obtain the shallow feature weight vector of chemical materials;

[0021] Based on the weight vector of the shallow features of chemical materials, the shallow feature map of chemical materials is amplified and suppressed to obtain an enhanced shallow feature map of chemical materials.

[0022] For example, according to the storage management method for chemical production of an embodiment of the present application, based on the shallow feature weight vector of the chemical material, the shallow feature map of the chemical material is subjected to feature amplification and suppression operations to obtain an enhanced shallow feature map of the chemical material, including:

[0023] The Kronecker product of the chemical material shallow feature weight vector and each feature matrix of the chemical material shallow feature map along the channel dimension is calculated to obtain an enhanced chemical material shallow feature map.

[0024] For example, according to the storage management method for chemical production of the embodiment of the present application, the shallow feature map of enhanced chemical materials and the deep feature map of enhanced chemical materials are dynamically interactively fused to obtain a multi-scale gated fusion feature vector of the chemical materials, including:

[0025] Performing global mean pooling on the shallow feature map of the enhanced chemical material and the deep feature map of the enhanced chemical material to obtain a shallow compression feature vector of the enhanced chemical material and a deep compression feature vector of the enhanced chemical material;

[0026] The shallow compression feature vector of enhanced chemical materials and the deep compression feature vector of enhanced chemical materials are input into the feature vector dynamic interactive fusion module based on gated response to obtain the multi-scale gated fusion feature vector of chemical materials.

[0027] For example, according to the storage management method for chemical production of an embodiment of the present application, the shallow compression feature vector of the enhanced chemical material and the deep compression feature vector of the enhanced chemical material are input into the feature vector dynamic interactive fusion module based on the gated response to obtain the multi-scale gated fusion feature vector of the chemical material, including:

[0028] Perform feature concatenation on the shallow compression feature vector of the enhanced chemical material and the deep compression feature vector of the enhanced chemical material to obtain a multi-scale feature joint representation vector of the chemical material;

[0029] The multi-scale feature joint representation vector of chemical materials is input into the gated response function to obtain the information fusion response gate;

[0030] The difference between the response gate and the information fusion is calculated, and the response gate and the difference of the information fusion are used as weights to calculate the position-weighted sum of the enhanced chemical material shallow compression feature vector and the enhanced chemical material deep compression feature vector to obtain the chemical material multi-scale gated fusion feature vector.

[0031] For example, according to the warehouse management method for chemical production of an embodiment of the present application, the multi-scale feature joint representation vector of the chemical material is input into the gated response function to obtain the response gate of the information fusion, including:

[0032] Multiplying the multi-scale feature joint representation vector of chemical materials by a predetermined weight vector to obtain an information interaction fusion correlation coefficient;

[0033] The information interaction fusion correlation coefficient and the predetermined bias parameter are added together and then activated through a sigmoid function to obtain the response gate of the information fusion.

[0034] For example, according to the storage management method for chemical production of an embodiment of the present application, a hazard level is assigned to chemical materials based on a multi-scale gated fusion feature vector of the chemical materials, including:

[0035] Inputting the multi-scale gated fusion feature vector of the chemical material into the classifier to obtain a classification result, and the classification result is used to represent the type label of the chemical material;

[0036] Based on the classification results, a hazard level is assigned to the chemical material.

[0037] The embodiment of the present application also provides a warehouse management system for chemical production, which includes:

[0038] An image acquisition module, used to acquire chemical material images acquired by a camera;

[0039] A state feature extraction module is used to perform multi-level material state feature extraction on chemical material images to obtain a shallow feature map of chemical materials and a deep feature map of chemical materials;

[0040] A feature selection and strengthening module is used to perform feature selection and strengthening on the shallow characteristic map of chemical materials and the deep characteristic map of chemical materials to obtain a strengthened shallow characteristic map of chemical materials and a strengthened deep characteristic map of chemical materials;

[0041] A fusion module is used to dynamically and interactively fuse the shallow feature map of enhanced chemical materials and the deep feature map of enhanced chemical materials to obtain a multi-scale gated fusion feature vector of the chemical materials;

[0042] The hazard level assignment module is used to assign hazard levels to chemical materials based on the multi-scale gated fusion feature vectors of chemical materials.

[0043] According to the warehousing management method and system for chemical production of the embodiments of the present application, it can greatly improve the efficiency and safety of chemical warehousing management, reduce potential risks caused by human misjudgment, and realize intelligent warehousing management. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0045] Figure 1 A schematic diagram of the application architecture of the warehouse management method for chemical production in an embodiment of the present application is shown;

[0046] Figure 2 A flow chart of a storage management method for chemical production in an embodiment of the present application is shown;

[0047] Figure 3 A flowchart of sub-step S530 of the warehouse management method for chemical production in an embodiment of the present application is shown;

[0048] Figure 4 A flowchart of sub-step S540 of the warehouse management method for chemical production in an embodiment of the present application is shown;

[0049] Figure 5 A schematic diagram of the structure of a warehouse management system for chemical production in an embodiment of the present application is shown;

[0050] Figure 6 An application scenario diagram of a storage management method for chemical production in an embodiment of the present application is shown;

[0051] Figure 7 A schematic diagram of a storage medium according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0053] The terms used in this specification are those common terms currently widely used in the art in consideration of the functions of the present application, but these terms may vary according to the intention of a person of ordinary skill in the art, precedents, or new technologies in the art. In addition, specific terms may be selected by the applicant, and in this case, their detailed meanings will be described in the detailed description of the present application. Therefore, the terms used in the specification should not be understood as simple names, but rather as a general description based on the meaning of the terms and the present application.

[0054] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0055] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. At the same time, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0056] Figure 1 A schematic diagram of the application architecture of a warehouse management method for chemical production in an embodiment of the present application is shown, including a server 100 and a terminal device 200.

[0057] The terminal device 200 and the server 100 can be connected via the Internet to achieve mutual communication. Optionally, the above-mentioned Internet uses standard communication technology and / or protocol. The Internet is usually the Internet, but it can also be any network, including but not limited to any combination of a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a private network or a virtual private network. In some embodiments, the data exchanged through the network can be represented by technologies and / or formats including Hyper Text Markup Language (HTML), Extensible Markup Language (XML), etc. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec), etc. can be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above-mentioned data communication technologies.

[0058] The server 100 can provide various network services for the terminal device 200, wherein the server 100 can be a single server, a server cluster consisting of several servers, or a cloud computing center. Specifically, the server 100 may include a processor 110 (Center Processing Unit, CPU), a memory 120, an input device 130, and an output device 140, etc. The input device 130 may include a keyboard, a mouse, a touch screen, etc., and the output device 140 may include a display device, such as a liquid crystal display (Liquid Crystal Display, LCD), a cathode ray tube (Cathode Ray Tube, CRT), etc.

[0059] The memory 120 may include a read-only memory (ROM) and a random access memory (RAM), and provides the processor 110 with program instructions and data stored in the memory 120. In the embodiment of the present application, the memory 120 may be used to store the program of the warehouse management method for chemical production in the embodiment of the present application.

[0060] The processor 110 calls the program instructions stored in the memory 120, and the processor 110 is used to execute the steps of any one of the warehouse management methods for chemical production in the embodiments of the present application according to the obtained program instructions.

[0061] In addition, the application architecture diagram in the embodiment of the present application is intended to more clearly illustrate the technical solution in the embodiment of the present application, and does not constitute a limitation on the technical solution provided in the embodiment of the present application. Of course, for other application architectures and business applications, the technical solution provided in the embodiment of the present application is also applicable to similar problems.

[0062] The following is a non-restrictive description of the warehouse management method for chemical production provided according to at least one embodiment of the present application through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without conflicting with each other to obtain new examples or embodiments, and these new examples or embodiments also fall within the scope of protection of the present application.

[0063] In view of the above technical problems, the technical concept of this application is to use computer vision technology based on deep learning to analyze and process the image data of chemical materials, extract multi-level chemical material state feature representation, and use attention mechanism and gated response fusion mechanism to enhance and interactively fuse the state features of chemical materials at different levels to achieve accurate characterization of the state of chemical materials, so as to identify the type and hazard of chemical materials, so as to facilitate safety management according to their hazard level. In this way, the efficiency and safety of chemical storage management can be greatly improved, the potential risks caused by human misjudgment can be reduced, and intelligent storage management can be realized.

[0064] Based on this, Figure 2 The flowchart of the warehouse management method for chemical production in the embodiment of the present application is shown. For example, the warehouse management method for chemical production can be executed by a server, which can be Figure 1 The server 100 shown in FIG. Figure 2 As shown, the warehouse management method for chemical production according to the embodiment of the present application includes the following steps: S510, acquiring a chemical material image collected by a camera; S520, performing multi-level material state feature extraction on the chemical material image to obtain a shallow feature map of the chemical material and a deep feature map of the chemical material; S530, performing feature selection enhancement on the shallow feature map of the chemical material and the deep feature map of the chemical material, respectively, to obtain an enhanced shallow feature map of the chemical material and an enhanced deep feature map of the chemical material; S540, performing dynamic interactive fusion on the enhanced shallow feature map of the chemical material and the enhanced deep feature map of the chemical material to obtain a multi-scale gated fusion feature vector of the chemical material; S550, assigning a hazard level to the chemical material based on the multi-scale gated fusion feature vector of the chemical material.

[0065] Specifically, in the technical solution of the present application, first, an image of a chemical material captured by a camera is obtained. It should be understood that since many chemical materials have properties such as corrosiveness, flammability, explosion, toxicity and harm, direct contact may bring potential safety hazards or damage the detection equipment. The detection method of capturing images through a camera is a non-contact detection method, which does not require physical contact with the material and can ensure the safety of the detection process.

[0066] Next, due to the wide variety of chemical materials, the morphology of the materials may present different scales and levels in the image, and a single feature extraction method is often difficult to fully and accurately characterize its physical state. Therefore, in the technical solution of the present application, a material state feature extractor based on a pyramid network is used to perform multi-level feature extraction on chemical material images to obtain a shallow feature map of chemical materials and a deep feature map of chemical materials. Specifically, the pyramid network constructs feature pyramids of different scales to capture shallow local detail information such as color and texture of chemical materials, as well as deep contextual information such as the overall structure and spatial distribution of the materials, thereby achieving a comprehensive and detailed characterization of the state characteristics of chemical materials.

[0067] Accordingly, in step S520, multi-level material state feature extraction is performed on the chemical material image to obtain a chemical material shallow feature map and a chemical material deep feature map, including: inputting the chemical material image into a material state feature extractor based on a pyramid network to obtain a chemical material shallow feature map and a chemical material deep feature map.

[0068] Then, considering that there may be redundant information or noise interference in the feature extraction process, directly using the shallow features and deep features of chemical materials for subsequent material type identification may lead to inaccurate identification results. Therefore, in the technical solution of the present application, a feature attention selection and enhancement module based on a compression-suppression structure is further introduced to perform feature selection and enhancement on the shallow feature map of chemical materials and the deep feature map of chemical materials respectively. Specifically, the feature attention selection and enhancement module first compresses the information of the feature map through a pooling operation, and captures the feature association between the channel compression information through one-dimensional convolutional coding, so as to generate attention weights corresponding to each channel in the feature map, and based on the generated attention weights, the original feature map is subjected to feature amplification and suppression operations to achieve the enhancement of key features, as well as the suppression of redundant information and noise interference, thereby obtaining an enhanced shallow feature map of chemical materials and an enhanced deep feature map of chemical materials.

[0069] Furthermore, in order to streamline the feature dimensions and improve the effectiveness of feature representation, the global mean pooling operation is further used to compress the shallow feature maps and deep feature maps of enhanced chemical materials, retaining the global statistical information in the feature maps while reducing the dimensions of the feature maps to obtain the shallow compressed feature vectors and deep compressed feature vectors of enhanced chemical materials, thereby effectively reducing the computational complexity and preventing overfitting, providing more effective feature representation for subsequent chemical material type identification.

[0070] Accordingly, in step S530, if Figure 3As shown, the shallow feature map of chemical materials and the deep feature map of chemical materials are respectively input into the feature attention selection and enhancement module based on the compression-suppression structure to obtain the enhanced shallow feature map of chemical materials and the enhanced deep feature map of chemical materials, including: S531, calculating the global mean of each feature matrix along the channel dimension of the shallow feature map of chemical materials to obtain the characteristic vector representing the shallow compressed information of chemical materials; S532, performing one-dimensional convolution encoding on the characteristic vector representing the shallow compressed information of chemical materials to obtain the characteristic vector representing the correlation between the shallow compressed information of chemical materials; S533, converting the characteristic vector representing the shallow compressed information of chemical materials and the characteristic vector representing the correlation between the shallow compressed information of chemical materials into the characteristic vector representing the correlation between the shallow compressed information of chemical materials. The vectors are cascaded to obtain a multi-scale representation vector of the compressed information of the shallow features of chemical materials; S534, the multi-scale representation vector of the compressed information of the shallow features of chemical materials is input into a compression information feature extraction module including a multi-layer perceptron and a SiLU activation function to obtain a multi-scale correlation feature vector of the compressed information of the shallow features of chemical materials; S535, the multi-scale correlation feature vector of the compressed information of the shallow features of chemical materials is normalized using a Sigmoid function to obtain a weight vector of the shallow features of chemical materials; S536, based on the weight vector of the shallow features of chemical materials, a feature amplification and suppression operation is performed on the shallow feature map of chemical materials to obtain an enhanced shallow feature map of chemical materials.

[0071] Among them, in step S536, based on the shallow feature weight vector of the chemical material, the shallow feature map of the chemical material is subjected to feature amplification and suppression operations to obtain an enhanced shallow feature map of the chemical material, including: calculating the Kronecker product of the shallow feature weight vector of the chemical material and each feature matrix of the shallow feature map of the chemical material along the channel dimension to obtain an enhanced shallow feature map of the chemical material.

[0072] In a specific example, the shallow feature map of chemical materials and the deep feature map of chemical materials are respectively input into the feature attention selection and enhancement module based on the compression-suppression structure to obtain the enhanced shallow feature map of chemical materials and the enhanced deep feature map of chemical materials, including: performing attention selection on the shallow feature map of chemical materials according to the following feature enhancement formula to obtain the enhanced shallow feature map of chemical materials, wherein the feature enhancement formula is;

[0073] ,

[0074] ,

[0075] ;

[0076] in, It is the shallow characteristic diagram of chemical materials The coordinates in the channel are The characteristic value of and are the height and width of the shallow characteristic map of chemical materials, It is the first feature vector representing the shallow compression information of chemical materials. eigenvalues, is the characteristic vector representing the shallow compression information of chemical materials, represents one-dimensional convolutional coding, Indicates cascade, It is the multi-scale representation vector of the shallow feature compression information of chemical materials. represents a multi-layer perceptron, represents the sigmoid function, It is the shallow characteristic diagram of chemical materials. Represents the Kronecker product of the calculation vector and each feature matrix along the channel dimension of the feature map. It is a shallow characteristic diagram of enhanced chemical materials.

[0077] Next, in order to make full use of the complementary information in the shallow features and deep features of chemical materials to improve the richness and accuracy of feature expression, the shallow compressed feature vectors of enhanced chemical materials and the deep compressed feature vectors of enhanced chemical materials are further feature fused. In the technical solution of the present application, a feature vector dynamic interactive fusion module based on gated response is introduced to perform the fusion task. Specifically, in the process of feature fusion, feature information at different levels has different contributions to the type identification of chemical materials. The feature vector dynamic interactive fusion module can use the gating mechanism to dynamically control the weight distribution of feature vectors at two different levels in the fusion process, thereby realizing the dynamic interactive fusion of shallow features and deep features of chemical materials, so as to obtain a multi-scale gated fusion feature vector of chemical materials that contains both local detail information and overall structural information, so as to more comprehensively and accurately characterize the physical state of chemical materials.

[0078] Accordingly, in step S540, if Figure 4 As shown, the shallow feature map of enhanced chemical materials and the deep feature map of enhanced chemical materials are dynamically interactively fused to obtain a multi-scale gated fusion feature vector of chemical materials, including: S541, global mean pooling is performed on the shallow feature map of enhanced chemical materials and the deep feature map of enhanced chemical materials to obtain a shallow compression feature vector of enhanced chemical materials and a deep compression feature vector of enhanced chemical materials; S542, the shallow compression feature vector of enhanced chemical materials and the deep compression feature vector of enhanced chemical materials are input into a feature vector dynamic interactive fusion module based on gated response to obtain a multi-scale gated fusion feature vector of chemical materials.

[0079] Among them, in step S542, the enhanced chemical material shallow compression feature vector and the enhanced chemical material deep compression feature vector are input into the feature vector dynamic interactive fusion module based on gated response to obtain the chemical material multi-scale gated fusion feature vector, including: feature cascading the enhanced chemical material shallow compression feature vector and the enhanced chemical material deep compression feature vector to obtain the chemical material multi-scale feature joint representation vector; inputting the chemical material multi-scale feature joint representation vector into the gated response function to obtain the information fusion response gate; calculating the difference with the information fusion response gate, and using the information fusion response gate and the difference as weights to calculate the position-weighted sum of the enhanced chemical material shallow compression feature vector and the enhanced chemical material deep compression feature vector to obtain the chemical material multi-scale gated fusion feature vector.

[0080] Specifically, the multi-scale feature joint representation vector of chemical materials is input into the gated response function to obtain a response gate for information fusion, including: multiplying the multi-scale feature joint representation vector of chemical materials by a predetermined weight vector to obtain an information interaction fusion correlation coefficient; adding the information interaction fusion correlation coefficient and a predetermined bias parameter and then activating the result through a sigmoid function to obtain a response gate for information fusion.

[0081] In a specific example, the shallow compression feature vector of the enhanced chemical material and the deep compression feature vector of the enhanced chemical material are input into the feature vector dynamic interactive fusion module based on the gated response to obtain the multi-scale gated fusion feature vector of the chemical material, including: fusing the shallow compression feature vector of the enhanced chemical material and the deep compression feature vector of the enhanced chemical material with the following interactive fusion formula to obtain the multi-scale gated fusion feature vector of the chemical material, wherein the interactive fusion formula is:

[0082] ,

[0083] ;

[0084] in, is the shallow compression eigenvector of enhanced chemical materials, is the deep compression feature vector of enhanced chemical materials, is the sigmoid function, is the predetermined weight vector, is the predetermined bias parameter, is the response gate of information fusion, It is the multi-scale gated fusion feature vector of chemical materials.

[0085] Secondly, the multi-scale gated fusion feature vector of chemical materials is input into the classifier to obtain the classification result, and the classification result is used to represent the type label of the chemical material. In the technical solution of the present application, the classifier performs linear transformation and nonlinear mapping on the multi-scale gated fusion feature vector of chemical materials based on the weight parameters learned by pre-training to generate the type probability distribution of chemical materials, so as to determine the type label of chemical materials, thereby realizing the automatic classification of chemical materials. Based on the classification result, the hazard level is assigned to the chemical materials, and the corresponding safety marking and storage area allocation are performed to ensure the safe distance of dangerous materials during the storage process, prevent the potential risks caused by the mixed storage of materials, and further improve the safety of storage management.

[0086] Accordingly, in step S550, based on the multi-scale gated fusion feature vector of the chemical material, a hazard level is assigned to the chemical material, including: inputting the multi-scale gated fusion feature vector of the chemical material into a classifier to obtain a classification result, the classification result is used to represent the type label of the chemical material; based on the classification result, a hazard level is assigned to the chemical material.

[0087] It should be understood that the role of the classifier is to use the given categories and known training data to learn classification rules and classifiers, and then classify (or predict) unknown data. Logistic regression, SVM, etc. are often used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are required to form a multi-classification, but this is prone to errors and inefficient. Commonly used multi-classification methods include the Softmax classification function.

[0088] In the technical solution of the present application, the shallow feature map of chemical materials and the deep feature map of chemical materials are semantic features of chemical material images with different depths and feature receptive fields obtained by a material state feature extractor based on a pyramid network. After feature attention selection enhancement and global mean pooling based on a compression-suppression structure, the feature receptive field deviation and feature expression heterogeneity of the enhanced shallow compressed feature vector of chemical materials and the enhanced deep compressed feature vector of chemical materials in terms of feature value granularity will be amplified, resulting in the process of inputting the enhanced shallow compressed feature vector of chemical materials and the enhanced deep compressed feature vector of chemical materials into the dynamic interactive fusion module of feature vectors based on gated response, and the obtained multi-scale gated fusion feature vector of chemical materials has feature aggregation class representation imbalance, which causes the loss of aggregation key information and affects the expression effect of the multi-scale gated fusion feature vector of chemical materials.

[0089] In a preferred example, inputting the multi-scale gated fusion feature vector of chemical materials into a classifier to obtain a classification result includes the following steps:

[0090] All eigenvalues ​​of the multi-scale gated fusion feature vector of chemical materials are analyzed based on the eigenvalue interval The clustering of distances is performed, and the clustering features are arranged into multi-scale gated fusion clustering vectors of chemical materials;

[0091] Determine a clustering ratio value of the number of eigenvalues ​​of the multi-scale gated fusion clustering vector of the chemical material and the number of eigenvalues ​​of the multi-scale gated fusion feature vector of the chemical material;

[0092] Dividing the binary norm of the multi-scale gated fusion clustering vector of the chemical material by the binary norm of the multi-scale gated fusion feature vector of the chemical material to obtain the multi-scale gated fusion conflict representation value of the chemical material;

[0093] Divide the first power value of the one-norm of the multi-scale gated fusion clustering vector of the chemical material with the clustering ratio value as the exponent by the second power value of the one-norm of the multi-scale gated fusion feature vector of the chemical material with the clustering ratio value as the exponent to obtain the multi-scale gated fusion adversarial representation value of the chemical material;

[0094] For each eigenvalue of the chemical material multi-scale gated fusion clustering vector, multiply it by the inverse of the difference between the chemical material multi-scale gated fusion conflict representation value and the chemical material multi-scale gated fusion confrontation representation value to obtain the optimized eigenvalue of the chemical material multi-scale gated fusion clustering vector;

[0095] For each eigenvalue outside the cluster in the multi-scale gated fusion feature vector of chemical materials, multiply it by the reciprocal of the sum of the multi-scale gated fusion conflict representation value of chemical materials and the multi-scale gated fusion confrontation representation value of chemical materials to obtain the optimized out-of-class eigenvalue of the multi-scale gated fusion feature vector of chemical materials;

[0096] Combining the optimized eigenvalues ​​of the chemical material multi-scale gated fusion clustering vector and the optimized out-of-class eigenvalues ​​of the chemical material multi-scale gated fusion feature vector into an optimized chemical material multi-scale gated fusion feature vector;

[0097] The optimized multi-scale gated fusion feature vector of chemical materials is input into the classifier to obtain the classification result.

[0098] The calculation formula of the optimized multi-scale gated fusion feature vector of chemical materials is expressed as:

[0099] ,

[0100] ;

[0101] in, is each eigenvalue of the multi-scale gated fusion clustering vector of chemical materials, is the multi-scale gated fusion feature vector of chemical materials, is the multi-scale gated fusion clustering vector of chemical materials, is the number of features of the multi-scale gated fusion feature vector of chemical materials, is the number of features of the multi-scale gated fusion clustering vector of chemical materials, is the ratio of the number of features of the multi-scale gated fusion clustering vector of chemical materials to the number of features of the multi-scale gated fusion feature vector of chemical materials, Represents the clustering feature set corresponding to the multi-scale gated fusion clustering vector of chemical materials, and Represents the binary norm and the uninorm of the vector respectively. Power, It is each eigenvalue of the optimized multi-scale gated fusion feature vector of chemical materials.

[0102] Accordingly, in order to avoid the loss of key suffix semantic information of the multi-scale gated fusion feature vector of chemical materials relative to the original feature set as a whole due to aggregation conflict, the clustering ratio of the number of eigenvalues ​​of the multi-scale gated fusion clustering vector of chemical materials relative to the number of eigenvalues ​​of the multi-scale gated fusion feature vector of chemical materials is used as the decision function to perform adversarial judgment on the set absolute representation of the one-norm of the multi-scale gated fusion clustering vector of chemical materials and the multi-scale gated fusion feature vector of chemical materials, and positive and negative interactions are performed with the clustering intrinsic conflict representation of the two-norm of the multi-scale gated fusion clustering vector of chemical materials and the multi-scale gated fusion feature vector of chemical materials respectively to construct a solid alignment guardrail between the optimized multi-scale gated fusion feature vector of chemical materials and the original feature set as a whole based on the aggregation features, thereby achieving the mitigation of the harmful information loss intention of the optimized multi-scale gated fusion feature vector of chemical materials based on the aggregation risk transferability, improving the expression effect of the optimized multi-scale gated fusion feature vector of chemical materials, and thus improving the accuracy of the classification result obtained by inputting the multi-scale gated fusion feature vector of chemical materials into the classifier.

[0103] Based on the above embodiments, see Figure 5As shown, it is a structural schematic diagram of a warehouse management system 800 for chemical production in an embodiment of the present application. The warehouse management system 800 for chemical production includes: an image acquisition module 810, which is used to obtain chemical material images acquired by a camera; a state feature extraction module 820, which is used to perform multi-level material state feature extraction on the chemical material image to obtain a shallow feature map of the chemical material and a deep feature map of the chemical material; a feature selection and enhancement module 830, which is used to perform feature selection and enhancement on the shallow feature map of the chemical material and the deep feature map of the chemical material to obtain an enhanced shallow feature map of the chemical material and an enhanced deep feature map of the chemical material; a fusion module 840, which is used to perform dynamic interactive fusion of the enhanced shallow feature map of the chemical material and the enhanced deep feature map of the chemical material to obtain a multi-scale gated fusion feature vector of the chemical material; and a hazard level assignment module 850, which is used to assign a hazard level to the chemical material based on the multi-scale gated fusion feature vector of the chemical material.

[0104] Here, those skilled in the art will appreciate that the specific functions and operations of each module in the warehouse management system 800 for chemical production have been described in detail above. Figures 2 to 4 The description of the warehouse management method for chemical production has been introduced in detail, and therefore, its repeated description will be omitted.

[0105] Figure 6 FIG. 1 is an application scenario diagram of a storage management method for chemical production according to an embodiment of the present application. Figure 6 As shown, in this application scenario, first, a chemical material image captured by a camera is obtained (for example, Figure 6 D), and then, the chemical material image is input to a server deployed with a warehouse management algorithm for chemical production (for example, Figure 6 In S) as shown in , the server can use a warehouse management algorithm for chemical production to process the chemical material image to obtain a classification result for representing a type label of the chemical material.

[0106] Based on the above embodiments, another exemplary embodiment of an electronic device is also provided in the embodiments of the present application. In some possible implementations, the electronic device in the embodiments of the present application may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can implement the steps of the warehouse management method for chemical production in the above embodiments when executing the program.

[0107] For example, in the case of electronic equipment Figure 1 Taking the server 100 in the example as an example, the processor in the electronic device is the processor 110 in the server 100, and the memory in the electronic device is the memory 120 in the server 100.

[0108] An embodiment of the present application also provides a computer-readable storage medium. Figure 7 1 shows a schematic diagram of a computer-readable storage medium 1000 according to an embodiment of the present application. Figure 7 As shown, a computer executable instruction 1001 is stored on a computer readable storage medium 1000. When the computer executable instruction 1001 is executed by a processor, the warehouse management method for chemical production according to the embodiment of the present application described with reference to the above figures can be executed. The computer readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0109] The embodiment of the present application also provides a computer program product or a computer program, which includes computer executable instructions, and the computer executable instructions are stored in a computer readable storage medium. The processor of the computer device reads the computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the computer device executes the warehouse management method for chemical production according to the embodiment of the present application.

[0110] Those skilled in the art will appreciate that the contents disclosed in this application may be subject to various modifications and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.

[0111] In addition, although the present application makes various references to certain units in the system according to embodiments of the present application, any number of different units can be used and run on the client and / or server. The units are illustrative only, and different aspects of the system and method can use different units.

[0112] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of software function modules. The present application is not limited to any particular form of combination of hardware and software.

[0113] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined as such herein.

[0114] The above is an explanation of the present application and should not be considered as a limitation thereof. Although several exemplary embodiments of the present application have been described, it will be readily understood by those skilled in the art that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present application. Therefore, all of these modifications are intended to be included within the scope of the present application as defined by the claims. It should be understood that the above is an explanation of the present application and should not be considered to be limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present application is defined by the claims and their equivalents.

Claims

1. A storage management method for chemical production, characterized in that: include: Acquire chemical material images collected by a camera; Performing multi-level material state feature extraction on the chemical material image to obtain a chemical material shallow layer feature map and a chemical material deep layer feature map; Performing feature selection and strengthening on the shallow characteristic map of the chemical material and the deep characteristic map of the chemical material to obtain a strengthened shallow characteristic map of the chemical material and a strengthened deep characteristic map of the chemical material; Dynamically and interactively fusing the shallow feature map of the enhanced chemical material and the deep feature map of the enhanced chemical material to obtain a multi-scale gated fusion feature vector of the chemical material; Assigning a hazard level to the chemical material based on a multi-scale gated fusion feature vector of the chemical material; Based on the multi-scale gated fusion feature vector of the chemical material, a hazard level is assigned to the chemical material, including: Inputting the multi-scale gated fusion feature vector of the chemical material into a classifier to obtain a classification result, wherein the classification result is used to represent a type label of the chemical material; Based on the classification result, assign a hazard level to the chemical material; Inputting the multi-scale gated fusion feature vector of chemical materials into the classifier to obtain the classification result includes the following steps: All eigenvalues ​​of the multi-scale gated fusion feature vector of chemical materials are analyzed based on the eigenvalue interval The clustering of distances is performed, and the clustering features are arranged into multi-scale gated fusion clustering vectors of chemical materials; Determine a clustering ratio value of the number of eigenvalues ​​of the multi-scale gated fusion clustering vector of the chemical material and the number of eigenvalues ​​of the multi-scale gated fusion feature vector of the chemical material; Dividing the binary norm of the multi-scale gated fusion clustering vector of the chemical material by the binary norm of the multi-scale gated fusion feature vector of the chemical material to obtain the multi-scale gated fusion conflict representation value of the chemical material; Divide the first power value of the one-norm of the multi-scale gated fusion clustering vector of the chemical material with the clustering ratio value as the exponent by the second power value of the one-norm of the multi-scale gated fusion feature vector of the chemical material with the clustering ratio value as the exponent to obtain the multi-scale gated fusion adversarial representation value of the chemical material; For each eigenvalue of the chemical material multi-scale gated fusion clustering vector, multiply it by the inverse of the difference between the chemical material multi-scale gated fusion conflict representation value and the chemical material multi-scale gated fusion confrontation representation value to obtain the optimized eigenvalue of the chemical material multi-scale gated fusion clustering vector; For each eigenvalue outside the cluster in the multi-scale gated fusion feature vector of chemical materials, multiply it by the reciprocal of the sum of the multi-scale gated fusion conflict representation value of chemical materials and the multi-scale gated fusion confrontation representation value of chemical materials to obtain the optimized out-of-class eigenvalue of the multi-scale gated fusion feature vector of chemical materials; Combining the optimized eigenvalues ​​of the chemical material multi-scale gated fusion clustering vector and the optimized out-of-class eigenvalues ​​of the chemical material multi-scale gated fusion feature vector into an optimized chemical material multi-scale gated fusion feature vector; The optimized multi-scale gated fusion feature vector of chemical materials is input into the classifier to obtain the classification result.

2. The storage management method for chemical production according to claim 1, characterized in that: Performing multi-level material state feature extraction on the chemical material image to obtain a chemical material shallow feature map and a chemical material deep feature map, including: The chemical material image is input into a material state feature extractor based on a pyramid network to obtain a shallow feature map of the chemical material and a deep feature map of the chemical material.

3. The storage management method for chemical production according to claim 2, characterized in that: The shallow characteristic map of the chemical material and the deep characteristic map of the chemical material are respectively subjected to feature selection and enhancement to obtain an enhanced shallow characteristic map of the chemical material and an enhanced deep characteristic map of the chemical material, including: The shallow feature map of chemical materials and the deep feature map of chemical materials are respectively input into a feature attention selection enhancement module based on a compression-suppression structure to obtain the enhanced shallow feature map of chemical materials and the enhanced deep feature map of chemical materials.

4. The storage management method for chemical production according to claim 3, characterized in that: The shallow feature map of the chemical material and the deep feature map of the chemical material are respectively input into a feature attention selection and enhancement module based on a compression-suppression structure to obtain the enhanced shallow feature map of the chemical material and the enhanced deep feature map of the chemical material, including: Calculating the global mean of each feature matrix of the shallow feature map of the chemical material along the channel dimension to obtain a feature vector representing the shallow compression information of the chemical material; Performing one-dimensional convolution coding on the characteristic vector representing the shallow compression information of the chemical material to obtain a characteristic vector representing the correlation between the shallow compression information of the chemical material; Cascading the characteristic vector representing the shallow compression information of the chemical material and the characteristic vector representing the correlation between the shallow compression information of the chemical material to obtain a multi-scale representation vector of the shallow feature compression information of the chemical material; Inputting the multi-scale representation vector of the shallow feature compression information of the chemical material into a compression information feature extraction module including a multi-layer perceptron and a SiLU activation function to obtain a multi-scale correlation feature vector of the shallow feature compression information of the chemical material; Using a Sigmoid function to normalize the multi-scale correlation feature vector of the shallow feature compression information of the chemical material to obtain a shallow feature weight vector of the chemical material; Based on the chemical material shallow feature weight vector, the chemical material shallow feature map is subjected to feature amplification and suppression operations to obtain the enhanced chemical material shallow feature map.

5. The storage management method for chemical production according to claim 4, characterized in that: Based on the chemical material shallow feature weight vector, the chemical material shallow feature map is subjected to feature amplification and suppression operations to obtain the enhanced chemical material shallow feature map, including: The Kronecker product of the chemical material shallow characteristic weight vector and each characteristic matrix of the chemical material shallow characteristic map along the channel dimension is calculated to obtain the enhanced chemical material shallow characteristic map.

6. The storage management method for chemical production according to claim 5, characterized in that: Dynamically interactively fusing the shallow feature map of the enhanced chemical material and the deep feature map of the enhanced chemical material to obtain a multi-scale gated fusion feature vector of the chemical material, including: Performing global mean pooling on the shallow feature map of the enhanced chemical material and the deep feature map of the enhanced chemical material to obtain a shallow compression feature vector of the enhanced chemical material and a deep compression feature vector of the enhanced chemical material; The shallow compression feature vector of the enhanced chemical material and the deep compression feature vector of the enhanced chemical material are input into a feature vector dynamic interactive fusion module based on gated response to obtain the multi-scale gated fusion feature vector of the chemical material.

7. The storage management method for chemical production according to claim 6, characterized in that: Inputting the shallow compression feature vector of the enhanced chemical material and the deep compression feature vector of the enhanced chemical material into a feature vector dynamic interactive fusion module based on gated response to obtain the multi-scale gated fusion feature vector of the chemical material, including: Performing feature concatenation on the shallow compression feature vector of the enhanced chemical material and the deep compression feature vector of the enhanced chemical material to obtain a multi-scale feature joint representation vector of the chemical material; Inputting the multi-scale feature joint representation vector of the chemical material into a gated response function to obtain an information fusion response gate; A difference between a response gate and the information fusion is calculated, and the response gate of the information fusion and the difference are used as weights to calculate the position-weighted sum of the shallow compression feature vector of the enhanced chemical material and the deep compression feature vector of the enhanced chemical material to obtain the multi-scale gated fusion feature vector of the chemical material.

8. The storage management method for chemical production according to claim 7, characterized in that: The multi-scale feature joint representation vector of the chemical material is input into the gated response function to obtain the response gate of information fusion, including: Multiplying the multi-scale feature joint representation vector of the chemical material by a predetermined weight vector to obtain an information interaction fusion correlation coefficient; After the information interaction fusion correlation coefficient and the predetermined bias parameter are added, activation processing is performed through a sigmoid function to obtain a response gate for the information fusion.

9. A warehouse management system for chemical production, characterized in that: include: An image acquisition module, used to acquire chemical material images acquired by a camera; A state feature extraction module, used for performing multi-level material state feature extraction on the chemical material image to obtain a chemical material shallow layer feature map and a chemical material deep layer feature map; A feature selection and strengthening module is used to perform feature selection and strengthening on the shallow characteristic map of the chemical material and the deep characteristic map of the chemical material to obtain a strengthened shallow characteristic map of the chemical material and a strengthened deep characteristic map of the chemical material; A fusion module, used for dynamically and interactively fusing the shallow characteristic map of the enhanced chemical material and the deep characteristic map of the enhanced chemical material to obtain a multi-scale gated fusion characteristic vector of the chemical material; A hazard level designation module, used for designating a hazard level for the chemical material based on a multi-scale gated fusion feature vector of the chemical material; The hazard level designation module is used to: Inputting the multi-scale gated fusion feature vector of the chemical material into a classifier to obtain a classification result, wherein the classification result is used to represent a type label of the chemical material; Based on the classification result, assign a hazard level to the chemical material; The hazard level designation module is used to: All eigenvalues ​​of the multi-scale gated fusion feature vector of chemical materials are analyzed based on the eigenvalue interval The clustering of distances is performed, and the clustering features are arranged into multi-scale gated fusion clustering vectors of chemical materials; Determine a clustering ratio value of the number of eigenvalues ​​of the multi-scale gated fusion clustering vector of the chemical material and the number of eigenvalues ​​of the multi-scale gated fusion feature vector of the chemical material; Dividing the binary norm of the multi-scale gated fusion clustering vector of the chemical material by the binary norm of the multi-scale gated fusion feature vector of the chemical material to obtain the multi-scale gated fusion conflict representation value of the chemical material; Divide the first power value of the one-norm of the multi-scale gated fusion clustering vector of the chemical material with the clustering ratio value as the exponent by the second power value of the one-norm of the multi-scale gated fusion feature vector of the chemical material with the clustering ratio value as the exponent to obtain the multi-scale gated fusion adversarial representation value of the chemical material; For each eigenvalue of the chemical material multi-scale gated fusion clustering vector, multiply it by the inverse of the difference between the chemical material multi-scale gated fusion conflict representation value and the chemical material multi-scale gated fusion confrontation representation value to obtain the optimized eigenvalue of the chemical material multi-scale gated fusion clustering vector; For each eigenvalue outside the cluster in the multi-scale gated fusion feature vector of chemical materials, multiply it by the reciprocal of the sum of the multi-scale gated fusion conflict representation value of chemical materials and the multi-scale gated fusion confrontation representation value of chemical materials to obtain the optimized out-of-class eigenvalue of the multi-scale gated fusion feature vector of chemical materials; Combining the optimized eigenvalues ​​of the chemical material multi-scale gated fusion clustering vector and the optimized out-of-class eigenvalues ​​of the chemical material multi-scale gated fusion feature vector into an optimized chemical material multi-scale gated fusion feature vector; The optimized multi-scale gated fusion feature vector of chemical materials is input into the classifier to obtain the classification result.

Images