Warehousing logistics automatic distribution system and method

By using robots and artificial intelligence technologies in the warehousing and logistics distribution system, the problem of inefficient manual inspection and verification in the existing technology is solved, and the efficiency and accuracy of warehousing and logistics distribution is improved.

CN118823686BActive Publication Date: 2024-12-03HANGZHOU OUROBOROS TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411302322.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-18
Publication Date
2024-12-03
Estimated Expiration
2044-09-18

AI Technical Summary

Technical Problem

The existing warehousing and logistics distribution methods rely on manual labor in detection and verification tasks, resulting in inefficiency and limited accuracy. Especially during peak hours, manual labor is difficult to keep up with the speed of robot handling and is easily affected by human factors.

Method used

Using a robot distribution system, the object images are collected through the camera and combined with image processing and semantic analysis algorithms based on artificial intelligence and deep learning, the delivery items are automatically detected and checked, and the item information semantic encoding and attribute information analysis is used to determine the item accuracy to realize the delivery items. automatic check.

Benefits of technology

It improves the efficiency and reliability of automatic distribution of warehousing and logistics, reduces manual intervention, improves the overall degree of automation and intelligence, and avoids the problems of low efficiency and low accuracy in traditional solutions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118823686B_ABST
    Figure CN118823686B_ABST
Patent Text Reader

Abstract

The present application relates to an automatic distribution system and method for warehousing logistics. The method includes: in the task of using a robot to automatically detect and check the delivered items, the images of the delivered items are collected by the robot's camera, and an image processing and semantic analysis algorithm based on artificial intelligence and deep learning is introduced at the back end to analyze and understand the images of the delivered items and the order information of the items, so as to learn and capture the correlation relationship and matching features between the semantics of the item information in the order information and the semantics of the item attribute information in the item image, so as to detect and check the delivered items to confirm whether the delivered items are accurate. In this way, the delivered items can be automatically checked based on the feature matching and comparison between the order information of the items and the images of the delivered items, avoiding the low efficiency and low accuracy of the tasks of manual detection and verification in the traditional solution, and improving the efficiency and reliability of automatic distribution of warehousing logistics.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of warehousing and logistics, and in particular, to an automatic distribution system and method for warehousing and logistics. Background Art

[0002] With the development of robot technology, existing robots can complete more and more complex tasks and can efficiently complete a large number of tasks that originally required manual work. At present, in the logistics industry, robots can be used for specific tasks, for example, robots can replace warehouse staff to achieve automatic delivery of warehouse goods.

[0003] In the existing warehousing logistics distribution mode, although robots can perform the task of moving warehouse goods, in most cases, the inspection and verification of delivery items still need to be done manually to ensure the accuracy of delivery and avoid errors. However, this warehouse logistics distribution method that requires manual inspection and verification has a low degree of intelligence and may become an efficiency bottleneck in the entire automation process, especially during peak hours, when the speed of manual operation may not keep up with the speed of robot handling. At the same time, manual inspection and verification tasks for warehouse logistics may be affected by human fatigue, negligence or subjective factors, resulting in incorrect inspection and verification of delivery items, affecting the accuracy of delivery.

[0004] Therefore, an optimized warehouse logistics automatic distribution system is desired. Summary of the invention

[0005] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0006] In a first aspect, the present application provides a warehouse logistics automatic distribution system, the system comprising: a distribution controller, a robot and a logistics distribution tray connected in communication;

[0007] The delivery controller is used to send control instructions and order information to the robot;

[0008] The robot is used to confirm whether the delivery items are accurate based on the order information and the delivery item images captured by the camera, wherein the delivery items are placed on the logistics delivery tray;

[0009] The robot is further configured to transport the logistics delivery tray to a predetermined location in response to the delivery item being accurate.

[0010] Optionally, the robot includes: an item information extraction module, used to extract item information from order information; an item information semantic encoding module, used to semantically encode the item information to obtain an item information semantic encoding feature vector; an item attribute information description generation module, used to input the delivery item image into an item attribute information descriptor based on the AIGC model to obtain item attribute information; an item attribute information word granularity semantic understanding module, used to semantically encode the item attribute information based on word granularity to obtain a sequence of item attribute information word granularity semantic encoding feature vectors; an item attribute information word granularity semantic feature enhancement module, used to input the item attribute information ... The sequence of information word granularity semantic coding feature vectors is input into a semantic association enhancement module based on feature energy level measurement association radiation to obtain a sequence of enhanced item attribute information word granularity semantic coding feature vectors; a query response adaptive matching module is used to use the item information semantic coding feature vector as the query feature vector, and perform feature query response adaptive encoding based on fine-grained decoupling on the query feature vector and the sequence of enhanced item attribute information word granularity semantic coding feature vectors to obtain a query response adaptive matching result representation matrix; a delivery item accuracy detection module is used to determine whether the delivery item is accurate based on the query response adaptive matching result representation matrix.

[0011] Optionally, the item attribute information word granularity semantic understanding module is used to: perform word segmentation processing on the item attribute information and then pass it through an item attribute information semantic encoder including a word embedding layer to obtain a sequence of item attribute information word granularity semantic encoding feature vectors.

[0012] Optionally, the item attribute information word granularity semantic feature enhancement module includes: a feature energy level coefficient calculation unit, used to calculate the feature energy level coefficient of each item attribute information word granularity semantic coding feature vector in the sequence of the item attribute information word granularity semantic coding feature vector to obtain a sequence of item attribute information word granularity semantic feature energy level coefficients; a vector arrangement unit, used to arrange the sequence of item attribute information word granularity semantic feature energy level coefficients into an item attribute information word granularity semantic feature energy level coefficient input vector; a one-dimensional convolution encoding unit, used to perform one-dimensional convolution encoding on the item attribute information word granularity semantic feature energy level coefficient input vector to obtain an item attribute information word granularity semantic feature energy level radiation association vector; and a point convolution encoding unit, used to The method comprises the following steps: performing point convolution encoding on the item attribute information word granularity semantic feature energy level radiation association vector to obtain the item attribute information word granularity semantic feature energy level radiation convolution association vector; a probabilistic mapping unit, which is used to input the item attribute information word granularity semantic feature energy level radiation convolution association vector into a sigmoid function for probabilistic mapping to obtain the item attribute information word granularity semantic feature energy level radiation association weight vector; and a vector calculation unit, which is used to use the feature value of each position in the item attribute information word granularity semantic feature energy level radiation association weight vector as the weight value, perform vector-by-vector point multiplication on the sequence of item attribute information word granularity semantic encoding feature vectors by position, and add the sequence of item attribute information word granularity semantic encoding feature vectors to obtain the sequence of enhanced item attribute information word granularity semantic encoding feature vectors.

[0013] Optionally, the feature energy level coefficient calculation unit is used to: extract the feature maximum value of each position in the granular semantic coding feature vector of the item attribute information word to obtain the maximum value of the granular semantic feature of the item attribute information word; respectively calculate the mean and variance of the granular semantic coding feature vector of the item attribute information word to obtain the mean of the granular semantic feature of the item attribute information word and the variance of the granular semantic feature of the item attribute information word; multiply the value obtained by adding the variance of the granular semantic feature of the item attribute information word and a preset hyperparameter by four to obtain the first energy level coefficient of the granular semantic feature of the item attribute information word; calculate the maximum value of the granular semantic feature of the item attribute information word The square of the difference between the value and the mean value of the granular semantic feature of the item attribute information word is obtained to obtain the granular semantic feature difference value of the item attribute information word; the modulated item attribute information word granular semantic feature variance obtained by multiplying the item attribute information word granular semantic feature variance by two is added to the value obtained by multiplying the preset hyperparameter by two and the granular semantic feature difference value of the item attribute information word to obtain the second energy level coefficient of the granular semantic feature of the item attribute information word; the first energy level coefficient of the granular semantic feature of the item attribute information word is divided by the second energy level coefficient of the granular semantic feature of the item attribute information word to obtain the energy level coefficient of the granular semantic feature of the item attribute information word.

[0014] Optionally, the query response adaptive matching module includes: a fine-grained decoupling unit, which is used to perform fine-grained decoupling on the sequence of the enhanced item attribute information word granularity semantic encoding feature vectors to obtain a set of subsequences of the enhanced item attribute information word granularity semantic library feature vectors; a semantic enhancement encoding unit, which is used to input each subsequence of the enhanced item attribute information word granularity semantic library feature vector in the set of subsequences of the enhanced item attribute information word granularity semantic library feature vectors into a subsequence semantic enhancement encoder based on the Bi-LSTM model to obtain a set of enhanced item attribute information word granularity semantic library feature fine-grained decoupling local correlation feature vectors; a covariance matrix calculation unit, which is used to calculate the query feature vector and each enhanced item attribute information word granularity semantic library feature fine-grained decoupling local correlation feature vector in the set of enhanced item attribute information word granularity semantic library feature vectors. The library feature fine-grained decoupling of the covariance matrix between the local correlation feature vectors to obtain a set of item information semantic query response covariance matrices; a matrix trace calculation unit, used to calculate the matrix trace of each item information semantic query response covariance matrix in the set of item information semantic query response covariance matrices to obtain a set of item information semantic query response trace measurement factors; an activation unit, used to input the set of item information semantic query response trace measurement factors into a Sigmoid activation function to obtain a set of item information semantic query response trace attention weights; a weighted sum calculation unit, used to calculate the weighted sum of the set of item information semantic query response covariance matrices using each item information semantic query response trace attention weight in the set of item information semantic query response trace attention weights as a weight to obtain the query response adaptive matching result representation matrix.

[0015] Optionally, the delivery item accuracy detection module is used to: input the query response adaptive matching result representation matrix into a classifier-based delivery identifier to obtain a recognition result, wherein the recognition result is used to indicate whether the delivery item is accurate.

[0016] In a second aspect, the present application provides a method for automatic warehousing logistics distribution, the method comprising:

[0017] Use the delivery controller to send control instructions and order information to the robot;

[0018] Using a robot to confirm whether the delivered items are accurate based on the order information and an image of the delivered items captured by a camera, wherein the delivered items are placed on a logistics delivery tray;

[0019] The robot is used to accurately transport the logistics delivery tray to a predetermined location in response to the delivery items.

[0020] Optionally, the robot includes: extracting item information from order information; semantically encoding the item information to obtain an item information semantic encoding feature vector; inputting the delivered item image into an item attribute information descriptor based on an AIGC model to obtain item attribute information; semantically encoding the item attribute information based on word granularity to obtain a sequence of item attribute information word granularity semantic encoding feature vectors; inputting the sequence of item attribute information word granularity semantic encoding feature vectors into a semantic association enhancement module based on feature energy level metric association radiation to obtain a sequence of enhanced item attribute information word granularity semantic encoding feature vectors; using the item information semantic encoding feature vector as a query feature vector, performing feature query response adaptive encoding based on fine-grained decoupling on the query feature vector and the sequence of enhanced item attribute information word granularity semantic encoding feature vectors to obtain a query response adaptive matching result representation matrix; and determining whether the delivered item is accurate based on the query response adaptive matching result representation matrix.

[0021] Optionally, the item attribute information is semantically encoded based on word granularity to obtain a sequence of item attribute information word granularity semantic encoding feature vectors, including: performing word segmentation processing on the item attribute information and passing it through an item attribute information semantic encoder including a word embedding layer to obtain a sequence of item attribute information word granularity semantic encoding feature vectors.

[0022] By adopting the above technical solution, in the task of using a robot to automatically detect and check the delivered items, the robot's camera collects the delivered item image, and introduces an image processing and semantic analysis algorithm based on artificial intelligence and deep learning at the back end to analyze and understand the delivered item image and the item order information, so as to learn and capture the correlation relationship and matching features between the item information semantics in the order information and the item attribute information semantics in the item image, so as to detect and check the delivered items to confirm whether the delivered items are accurate. In this way, the delivered items can be automatically checked based on the feature matching and comparison between the item order information and the delivered item image, avoiding the low efficiency and low accuracy of the traditional solution that requires manual detection and verification tasks, and improving the efficiency and reliability of automatic delivery of warehousing logistics.

[0023] Other features and advantages of the present application will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings:

[0025] Figure 1 It is a schematic diagram of a warehousing logistics automatic distribution system according to an exemplary embodiment.

[0026] Figure 2 It is a block diagram of the robot of a warehousing logistics automatic distribution system according to an exemplary embodiment.

[0027] Figure 3 The present invention is a flowchart of a method for automatic distribution of warehousing logistics according to an exemplary embodiment.

[0028] Figure 4 It is a block diagram of an electronic device according to an exemplary embodiment.

[0029] Figure 5 It is an application scenario diagram of a warehousing logistics automatic distribution system according to an exemplary embodiment. DETAILED DESCRIPTION

[0030] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be construed as being limited to the embodiments described herein. Instead, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.

[0031] It should be understood that the various steps described in the method implementation of the present application can be performed in different orders and / or performed in parallel. In addition, the method implementation may include additional steps and / or omit the steps shown. The scope of the present application is not limited in this respect.

[0032] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0033] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0034] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0035] The names of the messages or information exchanged between multiple devices in the embodiments of the present application are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0036] The specific implementation of the present application is described in detail below with reference to the accompanying drawings.

[0037] Figure 1 FIG. 1 is a schematic diagram of an automatic warehousing logistics distribution system according to an exemplary embodiment. Figure 1 As shown, the system 100 includes:

[0038] The delivery controller 1, robot 2 and logistics delivery tray 3 are communicatively connected;

[0039] The delivery controller 1 is used to send control instructions and order information to the robot 2;

[0040] The robot 2 is used to confirm whether the delivery items are accurate based on the order information and the delivery item images captured by the camera, wherein the delivery items are placed on the logistics delivery tray 3;

[0041] The robot 2 is further used to transport the logistics delivery tray 3 to a predetermined location in response to the accuracy of the delivery items.

[0042] In the above-mentioned automatic warehousing and logistics delivery system, in the task of using a robot to automatically detect and check the delivered items, the robot's camera collects the images of the delivered items, and introduces an image processing and semantic analysis algorithm based on artificial intelligence and deep learning at the back end to analyze and understand the images of the delivered items and the order information of the items, so as to learn and capture the correlation relationship and matching features between the semantics of the item information in the order information and the semantics of the item attribute information in the item image, so as to detect and check the delivered items to confirm whether the delivered items are accurate. In this way, the delivery items can be automatically checked based on the feature matching and comparison between the order information of the items and the images of the delivered items, avoiding the low efficiency and low accuracy of the tasks of manual detection and verification in the traditional solution, and improving the efficiency and reliability of automatic warehousing and logistics delivery. In this way, the whole process automation from warehousing and logistics order processing to item delivery is realized, reducing manual intervention and improving the overall automation level and intelligence level.

[0043] Figure 2FIG. 1 is a block diagram of a robot in a warehousing logistics automatic distribution system according to an exemplary embodiment. Figure 2 As shown, the robot 2 comprises:

[0044] An item information extraction module 101 is used to extract item information from order information;

[0045] An item information semantic encoding module 102, configured to perform semantic encoding on the item information to obtain an item information semantic encoding feature vector;

[0046] An item attribute information description generating module 103, configured to input the delivery item image into an item attribute information descriptor based on an AIGC model to obtain item attribute information;

[0047] The item attribute information word granularity semantic understanding module 104 is used to perform word granularity-based semantic encoding on the item attribute information to obtain a sequence of item attribute information word granularity semantic encoding feature vectors;

[0048] The item attribute information word granularity semantic feature strengthening module 105 is used to input the sequence of the item attribute information word granularity semantic encoding feature vectors into the semantic association strengthening module based on feature energy level measurement association radiation to obtain a sequence of strengthened item attribute information word granularity semantic encoding feature vectors;

[0049] A query response adaptive matching module 106 is used to use the item information semantic encoding feature vector as a query feature vector, and perform feature query response adaptive encoding based on fine-grained decoupling on the sequence of the query feature vector and the enhanced item attribute information word granularity semantic encoding feature vector to obtain a query response adaptive matching result representation matrix;

[0050] The delivery item accuracy detection module 107 is used to determine whether the delivery item is accurate based on the query response adaptive matching result representation matrix.

[0051] Specifically, in the technical solution of the present application, in the process of confirming whether the delivered items are accurate, first, the item information is extracted from the order information, and the item information is semantically encoded to extract the implicit semantic features in the item order information, thereby obtaining a semantic encoding feature vector of the item information.

[0052] Next, considering that the delivery item image contains various feature information about the warehousing logistics delivery items, these feature information include features and semantic information such as color, shape and texture of the delivery items, which plays an important role in the subsequent delivery item correctness detection. Based on this, in the technical solution of the present application, the delivery item image is input into the item attribute information descriptor based on the AIGC model to obtain the item attribute information. Through the processing of the item attribute information descriptor based on the AIGC model, the large model can be used to automatically generate the attribute information of the item according to various types of features and semantics of the delivery item in the image, so as to facilitate obtaining a more comprehensive item description for the subsequent feature matching process.

[0053] Furthermore, after word segmentation, the item attribute information is encoded in an item attribute information semantic encoder including a word embedding layer to extract word-granular attribute semantic features about the delivery items in the item attribute information, thereby obtaining a sequence of item attribute information word-granular semantic encoding feature vectors.

[0054] In one embodiment of the present application, the item attribute information word granularity semantic understanding module is used to: perform word segmentation processing on the item attribute information and then pass it through an item attribute information semantic encoder including a word embedding layer to obtain a sequence of item attribute information word granularity semantic encoding feature vectors.

[0055] Then, since each item attribute information word granularity semantic coding feature vector in the sequence of item attribute information word granularity semantic coding feature vectors represents the word granularity semantic coding features of the item attribute information, these word semantics will have different contribution and importance in the process of overall understanding of the item attribute information. Based on this, in the technical solution of the present application, the sequence of item attribute information word granularity semantic coding feature vectors is further input into a semantic association enhancement module based on feature energy level measurement association radiation to obtain a sequence of enhanced item attribute information word granularity semantic coding feature vectors. Through the processing of the semantic association enhancement module based on feature energy level measurement association radiation, the expressiveness of the item attribute information word granularity semantic features and the semantic understanding ability of the model can be significantly improved through detailed feature energy level analysis and word granularity semantic association enhancement. This means that the attribute information of the delivered items can be captured and represented more accurately for subsequent matching and correctness detection tasks.

[0056] Specifically, the semantic association enhancement module based on the characteristic energy level measurement association radiation first calculates the characteristic energy level coefficients of the sequence of item attribute information word granularity semantic coding feature vectors, and quantifies the energy distribution of each item attribute information word granularity semantic coding feature vector. In the technical solution of the present application, the characteristic energy level coefficient is related to the statistical characteristics of the scale and characteristic distribution of each item attribute information word granularity semantic coding feature vector, which is used to represent the semantic information content density of each item attribute information word granularity semantic coding feature vector itself and its influence on the semantic information of other feature vectors in the surrounding neighborhood. Furthermore, one-dimensional convolutional coding is applied to capture the local dependencies in the characteristic energy level coefficient sequence of item attribute information word granularity semantics, and point convolution coding is used to further enhance the expression ability of spatial features. On this basis, a function-based probabilistic mapping is used to generate the item attribute information word granularity semantic feature energy level radiation association weight vector to achieve adaptive weighting of feature energy levels. In particular, the item attribute information word granularity semantic feature energy level radiation association weight vector is used as the weight, and the feature vector sequence is weighted and residual processed to obtain the sequence of the enhanced item attribute information word granularity semantic encoding feature vectors. This process not only strengthens the semantic association in the sequence of item attribute information word granularity semantic encoding feature vectors, but also improves the significance of the features, and enhances the model's ability to capture key item attribute information word granularity semantic information. Therefore, by dynamically adjusting feature weights and strengthening feature representation, the model can more flexibly cope with the diversity and complexity of attribute information of different delivery items, thereby achieving better performance in various delivery item detection tasks.

[0057] In one embodiment of the present application, the item attribute information word granularity semantic feature enhancement module includes: a feature energy level coefficient calculation unit, which is used to calculate the feature energy level coefficient of each item attribute information word granularity semantic coding feature vector in the sequence of the item attribute information word granularity semantic coding feature vector to obtain a sequence of item attribute information word granularity semantic feature energy level coefficients; a vector arrangement unit, which is used to arrange the sequence of item attribute information word granularity semantic feature energy level coefficients into an item attribute information word granularity semantic feature energy level coefficient input vector; a one-dimensional convolution encoding unit, which is used to perform one-dimensional convolution encoding on the item attribute information word granularity semantic feature energy level coefficient input vector to obtain an item attribute information word granularity semantic feature energy level radiation association vector; and a point convolution encoding unit. A unit is used to perform point convolution encoding on the item attribute information word granularity semantic feature energy level radiation association vector to obtain the item attribute information word granularity semantic feature energy level radiation convolution association vector; a probabilistic mapping unit is used to input the item attribute information word granularity semantic feature energy level radiation convolution association vector into a sigmoid function for probabilistic mapping to obtain the item attribute information word granularity semantic feature energy level radiation association weight vector; a vector calculation unit is used to use the feature value of each position in the item attribute information word granularity semantic feature energy level radiation association weight vector as the weight value, perform vector-by-vector point multiplication on the sequence of the item attribute information word granularity semantic encoding feature vectors by position, and add the sequence of the item attribute information word granularity semantic encoding feature vectors to obtain the sequence of the enhanced item attribute information word granularity semantic encoding feature vectors.

[0058] Furthermore, in one embodiment of the present application, the feature energy level coefficient calculation unit is used to: extract the feature maximum value of each position in the granular semantic coding feature vector of the item attribute information word to obtain the maximum value of the granular semantic feature of the item attribute information word; respectively calculate the mean and variance of the granular semantic coding feature vector of the item attribute information word to obtain the mean of the granular semantic feature of the item attribute information word and the variance of the granular semantic feature of the item attribute information word; multiply the value obtained by adding the granular semantic feature variance of the item attribute information word and the preset hyperparameter by four to obtain the first energy level coefficient of the granular semantic feature of the item attribute information word; calculate the granular semantic feature of the item attribute information word The square of the difference between the maximum value of the semantic feature and the mean value of the semantic feature of the item attribute information word granularity is obtained to obtain the difference value of the semantic feature of the item attribute information word granularity; the modulated item attribute information word granularity semantic feature variance obtained by multiplying the item attribute information word granularity semantic feature variance by two is added to the value obtained by multiplying the preset hyperparameter by two and the difference value of the semantic feature of the item attribute information word granularity to obtain the second energy level coefficient of the semantic feature of the item attribute information word granularity; the first energy level coefficient of the semantic feature of the item attribute information word granularity is divided by the second energy level coefficient of the semantic feature of the item attribute information word granularity to obtain the energy level coefficient of the semantic feature of the item attribute information word granularity.

[0059] Specifically, the sequence of the granular semantic encoding feature vectors of the item attribute information words is input into the semantic association enhancement module based on the feature energy level measurement association radiation for processing using the following semantic association enhancement formula to obtain the sequence of the enhanced item attribute information word granular semantic encoding feature vectors;

[0060] The semantic association strengthening formula is:

[0061]

[0062]

[0063]

[0064]

[0065] in, is a sequence of granular semantic encoding feature vectors of the item attribute information words, and are the first and The granular semantic encoding feature vector of item attribute information words, The number of feature vectors in the sequence representing the granular semantic encoding feature vectors of the item attribute information word, and are the first The mean and variance of the granular semantic encoding feature vector of item attribute information words, For the The maximum value of the features at each position in the granular semantic encoding feature vector of item attribute information words, is the preset hyperparameter, It is The characteristic energy level coefficient of the granular semantic encoding feature vector of item attribute information words, and They are and The characteristic energy level coefficient of the granular semantic encoding feature vector of item attribute information words, is the granular semantic feature energy level coefficient input vector of the item attribute information word, is a one-dimensional convolutional code, is the scale of the one-dimensional convolution kernel, is the granular semantic feature energy level radiation association vector of the item attribute information word, is point convolutional coding, is the function, is the granular semantic feature energy level radiation association weight vector of the item attribute information word, It means point multiplication by position. is a sequence of granular semantic encoding feature vectors of the reinforced item attribute information words, It is the weight vector associated with the energy level radiation of the granular semantic feature of item attribute information words.

[0066] It should be understood that, since the semantic encoding feature vector of the item information contains the semantic understanding features of the delivery item information extracted from the order information, each enhanced item attribute information word granularity semantic encoding feature vector in the sequence of the enhanced item attribute information word granularity semantic encoding feature vector contains the item attribute information word granularity semantic encoding features extracted from the actual delivery item image after feature enhanced significant expression. Therefore, in order to achieve the correct recognition and detection of the delivery item, it is necessary to associate and match the semantic features of the two. Specifically, in the technical solution of the present application, the item information semantic encoding feature vector is further used as the query feature vector, and the query feature vector and the sequence of the enhanced item attribute information word granularity semantic encoding feature vector are subjected to feature query response adaptive encoding based on fine-grained decoupling to obtain a query response adaptive matching result representation matrix. Through the processing method of feature query response adaptive encoding based on fine-grained decoupling, the matching accuracy and robustness between the query semantic features and the item attribute information word granularity semantic encoding features can be improved through fine-grained decoupling and adaptive encoding.

[0067] Specifically, first, the sequence of the granular semantic encoding feature vectors of the enhanced item attribute information words is fine-grained decoupled to obtain a subsequence set of the granular semantic library feature vectors of the item attribute information words. Fine-grained decoupling can capture the fine-grained local information and details of the item attribute information words by decomposing the granular semantic encoding feature vectors of the enhanced item attribute information words into smaller subsequences, thereby improving the expressiveness of the semantic features of the item attribute information, so that the subsequent matching process can more accurately capture the subtle differences between the granular semantic features of the item attribute information words and the query semantic features.

[0068] Next, the subsequences of the feature vectors of the word granularity semantic library of the enhanced item attribute information are input into the subsequence semantic enhancement encoder based on the Bi-LSTM model to obtain a set of fine-grained decoupled local association feature vectors of the word granularity semantic library of the enhanced item attribute information. Bi-LSTM (bidirectional long short-term memory network) can capture the contextual information in the sequence data and enhance its semantic expression by encoding the subsequence. In this way, the granular semantic information of the item attribute word of the subsequence is enhanced, making the feature vector more expressive and relevant at the local level.

[0069] Then, the covariance matrix between the query feature vector and each fine-grained decoupled local association feature vector of the granular semantic library of enhanced item attribute information words is calculated to obtain a set of query response covariance matrices. The covariance matrix can capture the linear relationship between the two semantic feature vectors. By calculating the covariance matrix, the correlation between the query feature vector and the fine-grained decoupled local association feature vector of the granular semantic library of enhanced item attribute information words can be quantified. This step provides a quantitative measure that reflects the degree of match between the query feature vector and the fine-grained decoupled local association feature vector of the granular semantic library of enhanced item attribute information words.

[0070] Next, the matrix trace of each item information semantic query response covariance matrix is ​​calculated to obtain a set of query response trace measurement factors. The matrix trace is the sum of the diagonal elements of the matrix, reflecting the overall characteristics of the matrix. By calculating the matrix trace, the overall energy of each item information semantic query response covariance matrix can be reflected, which is used to evaluate the overall correlation between the query feature vector and the library feature vector. Of course, by calculating the matrix trace, the representation of the covariance matrix can also be simplified, and its main features can be extracted to facilitate subsequent processing.

[0071] Then, the set of item information semantic query response trace metric factors is input into the Sigmoid activation function to obtain the set of query response trace attention weights. The Sigmoid function maps the input value to between 0 and 1 as the attention weight, reflecting the relative importance of each item information semantic query response trace metric factor. Through the attention mechanism, different query response trace metric factors are given different weights to highlight important word granularity semantic features and suppress minor features.

[0072] Finally, the weighted sum of the set of item information semantic query response covariance matrices is calculated using the attention weights of each item information semantic query response trace in the set of item information semantic query response trace attention weights as the weight, and the query response adaptive matching result representation matrix is ​​obtained. The weighted sum is obtained by weighting and summing each covariance matrix according to the weight, and synthesizing the contribution of the granular semantic matching features of each item attribute information word to obtain the final matching result representation. In this way, by synthesizing the contribution of each feature, an adaptive matching result representation matrix is ​​obtained, which improves the matching accuracy and robustness.

[0073] In one embodiment of the present application, the query response adaptive matching module includes: a fine-grained decoupling unit, which is used to perform fine-grained decoupling on the sequence of the granular semantic encoding feature vectors of the enhanced item attribute information words to obtain a set of subsequences of the granular semantic library feature vectors of the enhanced item attribute information words; a semantic reinforcement encoding unit, which is used to input each subsequence of the granular semantic library feature vectors of the enhanced item attribute information words in the set of subsequences of the granular semantic library feature vectors of the enhanced item attribute information words into a subsequence semantic reinforcement encoder based on a Bi-LSTM model to obtain a set of fine-grained decoupled local correlation feature vectors of the granular semantic library features of the enhanced item attribute information words; and a covariance matrix calculation unit, which is used to calculate the correlation between the query feature vector and each of the fine-grained decoupled local correlation feature vectors of the granular semantic library features of the enhanced item attribute information words. The granular semantic library features fine-grained decoupling of the covariance matrix between the local correlation feature vectors to obtain a set of item information semantic query response covariance matrices; a matrix trace calculation unit, used to calculate the matrix traces of each item information semantic query response covariance matrix in the set of item information semantic query response covariance matrices to obtain a set of item information semantic query response trace measurement factors; an activation unit, used to input the set of item information semantic query response trace measurement factors into a Sigmoid activation function to obtain a set of item information semantic query response trace attention weights; a weighted sum calculation unit, used to calculate the weighted sum of the set of item information semantic query response covariance matrices using each item information semantic query response trace attention weight in the set of item information semantic query response trace attention weights as a weight to obtain the query response adaptive matching result representation matrix.

[0074] Specifically, the item information semantic encoding feature vector is used as the query feature vector, and the query feature vector and the sequence of the enhanced item attribute information word granularity semantic encoding feature vector are subjected to feature query response adaptive encoding based on fine-grained decoupling using the following query response adaptive matching formula to obtain the query response adaptive matching result representation matrix;

[0075] The query response adaptive matching formula is:

[0076]

[0077]

[0078]

[0079]

[0080]

[0081] in, is a sequence of granular semantic encoding feature vectors of the reinforced item attribute information words, Represents fine-grained decoupled operations, To strengthen the collection of subsequences of feature vectors of the granular semantic library of item attribute information, is the first in the set of subsequences of the feature vectors of the granular semantic library of the enhanced item attribute information word A subsequence of the feature vector of the granular semantic library of the word that enhances the item attribute information, represents the Bi-LSTM encoder, For the The subsequence of the granular semantic library feature vectors of the enhanced item attribute information word corresponds to the fine-grained decoupling local association feature vector of the enhanced item attribute information word granular semantic library feature, is the query feature vector, represents the calculation of the covariance matrix, For the The query response covariance matrix corresponding to the subsequence of the feature vector of the granular semantic library of enhanced item attribute information words, represents the calculation matrix trace, For the The query response trace metric corresponding to the subsequence of the feature vector of the granular semantic library of the enhanced item attribute information word, represents the Sigmoid function, The said The query response trace attention weight corresponding to the subsequence of the feature vector of the granular semantic library of the enhanced item attribute information word, An adaptive matching result representation matrix is ​​generated for the query response.

[0082] Then, the query response adaptive matching result representation matrix is ​​input into the distribution identifier based on the classifier to obtain the recognition result, and the recognition result is used to indicate whether the distribution item is accurate. In other words, the query response adaptive matching result representation information between the semantic features of the item information and the granular semantic features of the item attribute information word is used for classification processing, so as to detect and check the distribution items to confirm whether the distribution items are accurate. In this way, the distribution items can be automatically checked based on the feature matching and comparison between the order information of the items and the images of the distribution items, avoiding the low efficiency and low accuracy of the tasks of manual detection and verification in the traditional solution, and improving the efficiency and reliability of automatic distribution of warehousing logistics. In this way, the whole process automation from warehousing logistics order processing to item distribution is realized, reducing manual intervention and improving the overall automation level and intelligence.

[0083] In one embodiment of the present application, the delivery item accuracy detection module is used to: input the query response adaptive matching result representation matrix into a classifier-based delivery identifier to obtain a recognition result, and the recognition result is used to indicate whether the delivery item is accurate.

[0084] Preferably, inputting the query response adaptive matching result representation matrix into a classifier-based delivery identifier to obtain an identification result comprises:

[0085] Expanding the query response adaptive matching result representation matrix into a query response adaptive matching result representation vector;

[0086] Cascading the sequence of the enhanced item attribute information word granularity semantic coding feature vectors into an enhanced item attribute information word granularity semantic coding cascade feature vector, and converting the item information semantic coding feature vector into an item information semantic coding interpolation feature vector having the same length as the enhanced item attribute information word granularity semantic coding cascade feature vector through linear interpolation;

[0087] Calculating the first norm and second norm of the mean vector of the interpolation feature vector of the semantic coding of the item information and the cascade feature vector of the granular semantic coding of the enhanced item attribute information word;

[0088] Calculate the weighted sum of the inverse of the square root of the second norm and the first norm, and perform dot multiplication with the dot-added sum vector of the interpolation feature vector of the semantic encoding of the item information and the cascade feature vector of the semantic encoding of the enhanced item attribute information word granularity to obtain a first query response adaptive matching syndrome vector;

[0089] The second query response adaptive matching syndrome vector is obtained by performing a dot multiplication of the dot product vector of the item information semantic encoding interpolation feature vector and the cascade feature vector of the enhanced item attribute information word granularity semantic encoding and the square root of the length of the query response adaptive matching result representation vector;

[0090] Calculating a weighted sum of the first query response adaptive matching syndrome vector and the second query response adaptive matching syndrome vector to obtain a query response adaptive matching correction vector;

[0091] After linearly transforming the query response adaptive matching correction vector to have the same length as the query response adaptive matching vector representation matrix, performing point multiplication with the query response adaptive matching result representation vector to obtain a corrected query response adaptive matching result representation vector; and

[0092] The corrected query response adaptive matching result representation matrix is ​​input into the classifier-based delivery identifier to obtain the identification result.

[0093] The query response adaptive matching correction vector is expressed as: and are respectively the interpolation feature vector of the semantic coding of the item information and the cascade feature vector of the semantic coding of the word granularity of the enhanced item attribute information, is its mean vector, is the length of the query response adaptive matching result representation vector, and and is the weighted sum weight as a hyperparameter, is the corrected query response adaptive matching result representation vector, It is added by location point. It is the point multiplication by position. and They represent the first norm and second norm of the mean vector respectively.

[0094] Here, considering that the sequence of the item information semantic encoding feature vector, the query feature vector and the enhanced item attribute information word granularity semantic encoding feature vector respectively represent the encoded semantic features of item information and item attribute information, when performing feature query response adaptive encoding based on fine-grained decoupling, it is expected to further compensate for the imbalance in matching domain shift caused by the imbalance in correspondence ratio under the fine-grained distribution of the sequence, so as to improve the expression effect of the query response adaptive matching result representation matrix, thereby improving the accuracy of the classification result.

[0095] Therefore, based on the mean vector norm between the sequence of the item information semantic encoding feature vector and the query feature vector and the enhanced item attribute information word granularity semantic encoding feature vector for the constrained representation of the structured foreground and background distinction of the supermanifold of the sequence of the item information semantic encoding feature vector and the query feature vector and the enhanced item attribute information word granularity semantic encoding feature vector, the feature-level key correspondence between the feature vectors of the item information semantic encoding feature vector and the query feature vector and the enhanced item attribute information word granularity semantic encoding feature vector is modeled, and the global association relationship between the corresponding features is adjusted, so as to make positive fine-grained correspondence suggestions by controlling the unbalanced ratio between the corresponding feature values ​​of the sequence of the item information semantic encoding feature vector and the query feature vector and the enhanced item attribute information word granularity semantic encoding feature vector, so as to avoid the imbalance of matching domain offset between the feature vectors of the item information semantic encoding feature vector and the query feature vector and the enhanced item attribute information word granularity semantic encoding feature vector by focusing on the focus.

[0096] In this way, by correcting the query response adaptive matching result representation vector obtained by expanding the query response adaptive matching result representation matrix through the query response adaptive matching correction vector, the expression effect of the query response adaptive matching result representation vector can be improved, thereby improving the accuracy of the recognition result obtained by inputting it into the distribution identifier based on the classifier. In this way, the whole process automation from warehousing and logistics order processing to goods distribution is realized, manual intervention is reduced, and the overall automation level and intelligence level are improved.

[0097] In summary, the above scheme is adopted, in the task of using robots to automatically detect and check the delivered items, the images of the delivered items are collected through the robot's camera, and the image processing and semantic analysis algorithms based on artificial intelligence and deep learning are introduced at the back end to analyze and understand the images of the delivered items and the order information of the items, so as to learn and capture the correlation relationship and matching features between the semantics of the item information in the order information and the semantics of the item attribute information in the item image, so as to detect and check the delivered items to confirm whether the delivered items are accurate. In this way, the delivered items can be automatically checked based on the feature matching and comparison between the order information of the items and the images of the delivered items, avoiding the low efficiency and low accuracy of the tasks of manual detection and verification in the traditional scheme, and improving the efficiency and reliability of automatic delivery of warehousing logistics.

[0098] Figure 3 is a flow chart of a method for automatic warehousing logistics distribution according to an exemplary embodiment. Figure 3 As shown, the method includes:

[0099] Step S201: Use the delivery controller to send control instructions and order information to the robot;

[0100] Step S202: using a robot to confirm whether the delivered items are accurate based on the order information and the images of the delivered items captured by the camera, wherein the delivered items are placed on a logistics delivery tray;

[0101] Step S203: using the robot to deliver the logistics delivery tray to a predetermined location in response to the delivery of the items.

[0102] In one embodiment of the present application, the robot includes: extracting item information from order information; semantically encoding the item information to obtain an item information semantic encoding feature vector; inputting the delivery item image into an item attribute information descriptor based on an AIGC model to obtain item attribute information; semantically encoding the item attribute information based on word granularity to obtain a sequence of item attribute information word granularity semantic encoding feature vectors; inputting the sequence of item attribute information word granularity semantic encoding feature vectors into a semantic association enhancement module based on feature energy level metric association radiation to obtain a sequence of enhanced item attribute information word granularity semantic encoding feature vectors; using the item information semantic encoding feature vector as a query feature vector, performing feature query response adaptive encoding based on fine-grained decoupling on the query feature vector and the sequence of enhanced item attribute information word granularity semantic encoding feature vectors to obtain a query response adaptive matching result representation matrix; and determining whether the delivery item is accurate based on the query response adaptive matching result representation matrix.

[0103] In one embodiment of the present application, the item attribute information is semantically encoded based on word granularity to obtain a sequence of item attribute information word granularity semantic encoding feature vectors, including: performing word segmentation processing on the item attribute information and passing it through an item attribute information semantic encoder including a word embedding layer to obtain a sequence of item attribute information word granularity semantic encoding feature vectors.

[0104] Reference below Figure 4 , which shows a schematic diagram of the structure of an electronic device 600 suitable for implementing the embodiment of the present application. The terminal device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0105] like Figure 4 As shown, the electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0106] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 4 The electronic device 600 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0107] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present application are executed.

[0108] It should be noted that the computer-readable medium mentioned above in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0109] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0110] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0111] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0112] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0113] The modules involved in the embodiments described in this application can be implemented by software or hardware. The name of the module does not limit the module itself in some cases. For example, the test parameter acquisition module can also be described as a "module for acquiring device test parameters corresponding to the target device".

[0114] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0115] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0116] Figure 5 FIG. 1 is an application scenario diagram of a warehousing logistics automatic distribution system according to an exemplary embodiment. Figure 5 As shown, in this application scenario, first, the product information is extracted from the order information (for example, Figure 5 Then, the obtained item information is input into a server that is deployed with a warehouse logistics automatic distribution algorithm (for example, Figure 5 In S) shown in , the server can process the item information based on the warehouse logistics automatic distribution algorithm to determine whether the distributed items are accurate.

[0117] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

[0118] In addition, although each operation is described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or to be performed in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the application. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0119] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely exemplary forms of implementation. With respect to the apparatus in the above-mentioned embodiments, the specific manner in which each module performs the operation has been described in detail in the embodiments related to the method, and will not be described in detail here.

Claims

1. A warehousing logistics automatic distribution system, characterized in that: include: Communicatively connected delivery controller, robot and logistics delivery tray; The delivery controller is used to send control instructions and order information to the robot; The robot is used to confirm whether the delivery items are accurate based on the order information and the delivery item images captured by the camera, wherein the delivery items are placed on the logistics delivery tray; The robot is further used to transport the logistics delivery tray to a predetermined location in response to the delivery item being accurate; Wherein, the robot comprises: An item information extraction module, used to extract item information from order information; An item information semantic encoding module, used to perform semantic encoding on the item information to obtain an item information semantic encoding feature vector; An item attribute information description generation module, used for inputting the delivery item image into an item attribute information descriptor based on an AIGC model to obtain item attribute information; An item attribute information word granularity semantic understanding module, used to perform word granularity-based semantic encoding on the item attribute information to obtain a sequence of item attribute information word granularity semantic encoding feature vectors; An item attribute information word granularity semantic feature enhancement module, used for inputting the sequence of item attribute information word granularity semantic encoding feature vectors into a semantic association enhancement module based on feature energy level measurement association radiation to obtain a sequence of enhanced item attribute information word granularity semantic encoding feature vectors; A query response adaptive matching module is used to use the item information semantic encoding feature vector as a query feature vector, and perform feature query response adaptive encoding based on fine-grained decoupling on the sequence of the query feature vector and the enhanced item attribute information word granularity semantic encoding feature vector to obtain a query response adaptive matching result representation matrix; A delivery item accuracy detection module, used to determine whether the delivery item is accurate based on the query response adaptive matching result representation matrix; Wherein, the query response adaptive matching module includes: A fine-grained decoupling unit, used to perform fine-grained decoupling on the sequence of the granular semantic encoding feature vectors of the enhanced item attribute information words to obtain a set of subsequences of the granular semantic library feature vectors of the enhanced item attribute information words; A semantic enhancement encoding unit, used for inputting each subsequence of the feature vector of the enhanced item attribute information word granularity semantic library in the set of the feature vector subsequences of the enhanced item attribute information word granularity semantic library into a subsequence semantic enhancement encoder based on the Bi-LSTM model to obtain a set of fine-grained decoupled local correlation feature vectors of the enhanced item attribute information word granularity semantic library; A covariance matrix calculation unit, used to calculate the covariance matrix between the query feature vector and each of the enhanced item attribute information word granularity semantic library feature fine-grained decoupled local association feature vectors in the set of enhanced item attribute information word granularity semantic library feature fine-grained decoupled local association feature vectors to obtain a set of item information semantic query response covariance matrices; a matrix trace calculation unit, configured to calculate the matrix trace of each item information semantic query response covariance matrix in the set of item information semantic query response covariance matrices to obtain a set of item information semantic query response trace measurement factors; An activation unit, used for inputting the set of item information semantic query response trace measurement factors into a Sigmoid activation function to obtain a set of item information semantic query response trace attention weights; A weighted sum calculation unit is used to calculate the weighted sum of the set of item information semantic query response covariance matrices using each item information semantic query response trace attention weight in the set of item information semantic query response trace attention weights as a weight to obtain the query response adaptive matching result representation matrix.

2. The automatic warehousing logistics distribution system according to claim 1 is characterized in that: The item attribute information word granularity semantic understanding module is used to: perform word segmentation processing on the item attribute information and then pass it through an item attribute information semantic encoder including a word embedding layer to obtain a sequence of item attribute information word granularity semantic encoding feature vectors.

3. The automatic warehousing logistics distribution system according to claim 2 is characterized in that: The item attribute information word granularity semantic feature enhancement module includes: A characteristic energy level coefficient calculation unit is used to calculate the characteristic energy level coefficient of each item attribute information word granularity semantic coding feature vector in the sequence of item attribute information word granularity semantic coding feature vectors to obtain a sequence of item attribute information word granularity semantic characteristic energy level coefficients; A vector arrangement unit, used for arranging the sequence of the item attribute information word granularity semantic feature energy level coefficients into an item attribute information word granularity semantic feature energy level coefficient input vector; A one-dimensional convolutional coding unit, used for performing one-dimensional convolutional coding on the item attribute information word granularity semantic feature energy level coefficient input vector to obtain an item attribute information word granularity semantic feature energy level radiation association vector; A point convolution encoding unit, used for performing point convolution encoding on the item attribute information word granularity semantic feature energy level radiation association vector to obtain the item attribute information word granularity semantic feature energy level radiation convolution association vector; A probabilistic mapping unit, used for inputting the item attribute information word granularity semantic feature energy level radiation convolution association vector into a sigmoid function for probabilistic mapping to obtain an item attribute information word granularity semantic feature energy level radiation association weight vector; A vector calculation unit is used to use the feature value of each position in the weight vector associated with the granular semantic feature energy level radiation of the item attribute information word as the weight value, perform vector-by-vector point multiplication on the sequence of the granular semantic coding feature vectors of the item attribute information word, and add the sequence of the granular semantic coding feature vectors of the item attribute information word to obtain the sequence of the enhanced granular semantic coding feature vectors of the item attribute information word.

4. The automatic warehousing logistics distribution system according to claim 3 is characterized in that: The characteristic energy level coefficient calculation unit is used to: Extracting the maximum value of the feature at each position in the granular semantic encoding feature vector of the item attribute information word to obtain the maximum value of the granular semantic feature of the item attribute information word; Calculating the mean and variance of the granular semantic encoding feature vectors of the item attribute information words respectively to obtain the mean of the granular semantic features of the item attribute information words and the variance of the granular semantic features of the item attribute information words; The value obtained by adding the variance of the granular semantic feature of the item attribute information word and the preset hyperparameter is multiplied by four to obtain the first energy level coefficient of the granular semantic feature of the item attribute information word; Calculating the square of the difference between the maximum value of the granular semantic feature of the item attribute information word and the mean value of the granular semantic feature of the item attribute information word to obtain the granular semantic feature difference value of the item attribute information word; The second energy level coefficient of the granular semantic feature of the item attribute information word is obtained by adding the modulated granular semantic feature variance of the item attribute information word obtained by multiplying the granular semantic feature variance of the item attribute information word by two, the value obtained by multiplying the preset hyperparameter by two, and the granular semantic feature difference value of the item attribute information word; The first energy level coefficient of the granular semantic feature of the item attribute information word is divided by the second energy level coefficient of the granular semantic feature of the item attribute information word to obtain the energy level coefficient of the granular semantic feature of the item attribute information word.

5. The automatic warehousing logistics distribution system according to claim 4 is characterized in that: The delivery item accuracy detection module is used to: input the query response adaptive matching result representation matrix into a classifier-based delivery identifier to obtain a recognition result, and the recognition result is used to indicate whether the delivery item is accurate.

6. A method for automatic distribution of warehousing logistics, characterized in that: include: Use the delivery controller to send control instructions and order information to the robot; Using a robot to confirm whether the delivered items are accurate based on the order information and an image of the delivered items captured by a camera, wherein the delivered items are placed on a logistics delivery tray; Using the robot to accurately deliver the logistics delivery tray to a predetermined location in response to the delivery item; Wherein, the robot comprises: Extract item information from order information; Performing semantic encoding on the item information to obtain a semantic encoding feature vector of the item information; Inputting the delivery item image into an item attribute information descriptor based on an AIGC model to obtain item attribute information; Performing word-granularity-based semantic coding on the item attribute information to obtain a sequence of word-granularity semantic coding feature vectors of the item attribute information; Inputting the sequence of the granular semantic encoding feature vectors of the item attribute information words into a semantic association strengthening module based on feature energy level measurement association radiation to obtain a sequence of strengthened granular semantic encoding feature vectors of the item attribute information words; Taking the item information semantic encoding feature vector as a query feature vector, performing feature query response adaptive encoding based on fine-grained decoupling on a sequence of the query feature vector and the enhanced item attribute information word granularity semantic encoding feature vector to obtain a query response adaptive matching result representation matrix; Determining whether the delivered item is accurate based on the query response adaptive matching result representation matrix; The method uses the semantic encoding feature vector of the item information as a query feature vector, performs feature query response adaptive encoding based on fine-grained decoupling on the sequence of the query feature vector and the granular semantic encoding feature vector of the enhanced item attribute information word to obtain a query response adaptive matching result representation matrix, including: Fine-grained decoupling is performed on the sequence of the granular semantic encoding feature vectors of the enhanced item attribute information words to obtain a set of subsequences of the granular semantic library feature vectors of the enhanced item attribute information words; Inputting each subsequence of the feature vector of the granular semantic library of the enhanced item attribute information word in the set of the subsequences of the feature vector of the granular semantic library of the enhanced item attribute information word into a subsequence semantic enhancement encoder based on the Bi-LSTM model to obtain a set of fine-grained decoupled local association feature vectors of the granular semantic library of the enhanced item attribute information word; Calculating the covariance matrix between the query feature vector and each of the enhanced item attribute information word granularity semantic library feature fine-grained decoupled local association feature vectors in the set of enhanced item attribute information word granularity semantic library feature fine-grained decoupled local association feature vectors to obtain a set of item information semantic query response covariance matrices; Calculating the matrix trace of each item information semantic query response covariance matrix in the set of item information semantic query response covariance matrices to obtain a set of item information semantic query response trace measurement factors; Inputting the set of item information semantic query response trace measurement factors into a Sigmoid activation function to obtain a set of item information semantic query response trace attention weights; Taking each item information semantic query response trace attention weight in the set of item information semantic query response trace attention weights as a weight, the weighted sum of the set of item information semantic query response covariance matrices is calculated to obtain the query response adaptive matching result representation matrix.

7. The automatic warehousing logistics distribution method according to claim 6, characterized in that: The item attribute information is semantically encoded based on word granularity to obtain a sequence of item attribute information word granularity semantic encoding feature vectors, including: after word segmentation processing of the item attribute information, passing it through an item attribute information semantic encoder including a word embedding layer to obtain a sequence of item attribute information word granularity semantic encoding feature vectors.