Abnormal allocation delivery order processing method and device, equipment and storage medium

Abstract

The application relates to the technical field of data processing, in particular to an abnormal allocation delivery order processing method and device, equipment and a storage medium, the method comprises the following steps: acquiring a real-time cargo image and performing information extraction processing to obtain cargo appearance information; inputting the real-time cargo image and the cargo appearance information into an abnormal type identification model, if it is determined that there is an abnormal problem, inputting the acquired logistics information and the identification result into a judgment model, and inputting the acquired cargo value information, the logistics information and the judgment result output by the judgment model into an amount calculation model to obtain a reward and punishment amount calculation result; generating to-be-audited information based on order information, the judgment result and the reward and punishment amount calculation result; when the to-be-audited information is audited and passed, prepayment information is generated; the method disclosed by the application can realize automatic judgment of abnormal allocation delivery orders, and can correspondingly calculate a judgment amount, and has the advantages of fast processing speed, clear and definite processing flow and high reliability of processing result.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to an abnormal allocation waybill processing method, device, equipment and storage medium. BACKGROUND

[0002] In the modern logistics industry, for abnormal allocation waybills, the common way is to communicate through offline channels by customer service personnel to solve the problem. However, this traditional processing method has several significant defects: first, the communication process is often lengthy and inconvenient, because repeated communication with relevant personnel is needed through telephone, email or other communication tools, which not only consumes a lot of time, but also is extremely inefficient; second, the responsibility of the network is often unclear, leading to frequent responsibility shirking, making it difficult to quickly lock in the root cause of the problem; third, there is a lack of consistency in the standard of fines and rewards for abnormal allocation waybills, different regions or different enterprises may have different regulations, which makes it difficult to meet the needs of all parties in the processing results; finally, the processing flow itself is not clear enough, lacking standardized process guidance, resulting in a lack of transparency and traceability in the entire process of processing abnormal allocation waybills, increasing the complexity and uncertainty of the processing.

[0003] It can be seen that the prior art still needs to be improved and improved. SUMMARY

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present application is to provide an abnormal allocation waybill processing method, which can realize automatic responsibility judgment of abnormal allocation waybills and corresponding calculation of judgment amount, has the advantages of fast processing speed, clear and explicit processing flow and high reliability of processing results.

[0005] The first aspect of the present application provides an abnormal allocation waybill processing method, comprising: acquiring a real-time goods image, sequentially pre-processing and information extraction processing the real-time goods image to obtain goods appearance information; inputting the real-time goods image and the goods appearance information into a pre-trained abnormal type recognition model, and judging whether there is an abnormal problem according to the recognition result output by the abnormal type recognition model; if there is an abnormal problem, acquiring logistics information corresponding to the real-time goods image, and inputting the logistics information and the recognition result into a pre-trained responsibility judgment model to obtain a responsibility judgment result; acquiring goods value information corresponding to the real-time goods image, inputting the goods value information, the logistics information and the responsibility judgment result into a pre-trained amount calculation model to obtain a reward and punishment amount calculation result; acquiring order information corresponding to the real-time goods image, generating to-be-audited information based on the order information, the responsibility judgment result and the reward and punishment amount calculation result; acquiring an audit result corresponding to the to-be-audited information, and if the audit result is passed, generating prepayment information.

[0006] Optionally, in the first implementation manner of the first aspect of the present application, the real-time cargo image is obtained, and the real-time cargo image is sequentially subjected to preprocessing and information extraction processing to obtain cargo appearance information, including: obtaining the real-time cargo image, converting the real-time cargo image into a gray image, and adopting a Gaussian filter to perform denoising processing on the converted gray image to obtain a denoised cargo image; adopting a histogram equalization algorithm to perform contrast enhancement processing on the denoised cargo image to obtain an enhanced cargo image; adopting a Canny edge detection algorithm to identify cargo contour information in the enhanced cargo image to obtain cargo feature information; adopting an OCR technology to identify text content in the enhanced cargo image to obtain cargo label information; and integrating the cargo feature information and the cargo label information to obtain the cargo appearance information.

[0007] Optionally, in the second implementation manner of the first aspect of the present application, the real-time cargo image and the cargo appearance information are input into a pre-trained abnormal type identification model, and whether there is an abnormal problem is determined according to an identification result output by the abnormal type identification model, including: constructing a first training set, the first training set including a plurality of historical cargo images, the plurality of historical cargo images including normal and abnormal historical cargo images, and each historical cargo image being labeled with contour information and label information corresponding to the cargo and whether there is an abnormality; training a convolutional neural network model pre-constructed based on a TensorFlow deep learning framework using the first training set to obtain the abnormal type identification model; using an image labeling tool, labeling the cargo appearance information on the real-time cargo image to obtain a labeled cargo image; based on size information corresponding to the abnormal type identification model, performing size conversion on the labeled cargo image, and inputting the labeled cargo image after size conversion into the abnormal type identification model to obtain the identification result; when the identification result is package damage, label damage or label information error, it is indicated that there is an abnormal problem.

[0008] Optionally, in the third implementation manner of the first aspect of the present application, if there is an abnormal problem, logistics information corresponding to the real-time cargo image is obtained, and the logistics information and the identification result are input into a pre-trained judgment model to obtain a judgment result, including: constructing a second training set, the second training set including involved network points, transportation distances, cargo types and abnormal types, and corresponding responsible parties; dividing the second training set, taking the involved network points, the transportation distances, the cargo types and the abnormal types as feature data, and taking the responsible parties as target data; training a pre-constructed decision tree model based on the second training set after the division processing to obtain the judgment model; if there is an abnormal problem, an API interface is called to obtain logistics information corresponding to the real-time cargo image from a logistics database, the logistics information including involved network point information, transportation route conditions and logistics costs; the logistics information and the identification result are input into the judgment model to obtain the judgment result.

[0009] Optionally, in a fourth implementation form of the first aspect of the present application, the step of obtaining the goods value information corresponding to the real-time goods image, inputting the goods value information, the logistics information and the responsibility result into the pre-trained amount calculation model to obtain the reward and punishment amount calculation result comprises: constructing a third training set, the third training set containing goods value, logistics cost, responsibility result and corresponding reward and punishment amount; performing division processing on the third training set, taking the goods value, the logistics cost and the responsibility result as feature data, and taking the reward and punishment amount as target data; training the pre-constructed linear regression model based on the third training set after the division processing to obtain the amount calculation model; calling an API interface to obtain the goods value information corresponding to the real-time goods image from a logistics database; inputting the goods value information, the logistics cost and the responsibility result into the amount calculation model to obtain the reward and punishment amount calculation result.

[0010] Optionally, in a fifth implementation form of the first aspect of the present application, the step of obtaining the order information corresponding to the real-time goods image, and generating the information to be audited based on the order information, the responsibility result and the reward and punishment amount calculation result comprises: calling an API interface to obtain the order information corresponding to the real-time goods image from a logistics database, the order information comprising an order number; obtaining a pre-constructed audit form, and filling the order information, the responsibility result and the reward and punishment amount calculation result into the audit form by using a direct mapping method to obtain a table to be checked; performing integrity checking on the table to be checked, and performing format checking on the table to be checked based on a pre-set data format checking function, taking the table to be checked passing the integrity checking and the format checking as an audit table; obtaining auditor information, and integrating the audit table and the auditor information to obtain the information to be audited.

[0011] Optionally, in a sixth implementation form of the first aspect of the present application, the step of obtaining the audit result corresponding to the information to be audited, and generating the prepayment information if the audit result is passed comprises: obtaining the audit result corresponding to the information to be audited based on the auditor information, the audit result comprising audit pass and audit fail; when the audit result is passed, obtaining the account information of the responsible party based on the responsibility result, and obtaining the payment contract information corresponding to the account information of the responsible party; integrating the information to be audited, the account information of the responsible party and the payment contract information to obtain the prepayment information; and storing the prepayment information into a pre-constructed prepayment report based on blockchain technology.

[0012] The second aspect of the present application provides an abnormal allocation waybill processing device, comprising: a processing module, configured to obtain a real-time cargo image, sequentially perform preprocessing and information extraction processing on the real-time cargo image, and obtain cargo appearance information; an abnormality identification module, configured to input the real-time cargo image and the cargo appearance information into a pre-trained abnormal type identification model, and determine whether there is an abnormal problem according to an identification result output by the abnormal type identification model; a judgment module, configured to, if there is an abnormal problem, obtain logistics information corresponding to the real-time cargo image, and input the logistics information and the identification result into a pre-trained judgment model to obtain a judgment result; a calculation module, configured to obtain cargo value information corresponding to the real-time cargo image, input the cargo value information, the logistics information, and the judgment result into a pre-trained amount calculation model, and obtain an award and punishment amount calculation result; a first generation module, configured to obtain order information corresponding to the real-time cargo image, and generate to-be-audited information based on the order information, the judgment result, and the award and punishment amount calculation result; and a second generation module, configured to obtain an audit result corresponding to the to-be-audited information, and generate prepayment information if the audit result is passed.

[0013] Optionally, in the first implementation manner of the second aspect of the present application, the processing module comprises: a first processing unit, configured to obtain a real-time cargo image, convert the real-time cargo image into a gray image, and perform denoising processing on the converted gray image by using a Gaussian filter to obtain a denoised cargo image; an enhancement unit, configured to perform contrast enhancement processing on the denoised cargo image by using a histogram equalization algorithm to obtain an enhanced cargo image; an edge detection unit, configured to identify cargo contour information in the enhanced cargo image by using a Canny edge detection algorithm to obtain cargo feature information; an extraction unit, configured to identify text content in the enhanced cargo image by using an OCR technology to obtain cargo label information; and a first integration unit, configured to integrate the cargo feature information and the cargo label information to obtain the cargo appearance information.

[0014] Optionally, in the second implementation manner of the second aspect of the present application, the anomaly identification module comprises: a first construction unit, configured to construct a first training set, wherein the first training set comprises a plurality of historical cargo images, the plurality of historical cargo images comprise normal and abnormal historical cargo images, and each historical cargo image is labeled with corresponding contour information and label information of the cargo and whether there is an anomaly; a first training unit, configured to train a pre-constructed convolutional neural network model based on a TensorFlow deep learning framework by using the first training set, to obtain an anomaly type identification model; a labeling unit, configured to label the cargo appearance information on the real-time cargo image by using an image labeling tool, to obtain a labeled cargo image; a conversion unit, configured to perform size conversion on the labeled cargo image based on size information corresponding to the anomaly type identification model, and input the labeled cargo image after size conversion into the anomaly type identification model, to obtain an identification result; and an anomaly identification unit, configured to indicate that there is an abnormal problem when the identification result is package damage, label damage or label information error.

[0015] Optionally, in the third implementation manner of the second aspect of the present application, the responsibility determination module comprises: a second construction unit, configured to construct a second training set, wherein the second training set comprises involved network points, transportation distances, cargo types, anomaly types, and corresponding responsible parties; a first division unit, configured to perform division processing on the second training set, to take the involved network points, transportation distances, cargo types and anomaly types as feature data, and to take the responsible parties as target data; a second training unit, configured to train a pre-constructed decision tree model based on the second training set after division processing, to obtain a responsibility determination model; a first acquisition unit, configured to call an API interface to obtain logistics information corresponding to the real-time cargo image from a logistics database if there is an abnormal problem, wherein the logistics information comprises involved network point information, transportation route conditions and logistics costs; and a responsibility determination unit, configured to input the logistics information and the identification result into the responsibility determination model, to obtain a responsibility determination result.

[0016] Optionally, in the fourth implementation manner of the second aspect of the present application, the calculation module comprises: a third construction unit, configured to construct a third training set, wherein the third training set comprises cargo values, logistics costs, responsibility determination results, and corresponding reward and punishment amounts; a second division unit, configured to perform division processing on the third training set, to take the cargo values, logistics costs and responsibility determination results as feature data, and to take the reward and punishment amounts as target data; a third training unit, configured to train a pre-constructed linear regression model based on the third training set after division processing, to obtain an amount calculation model; a second acquisition unit, configured to call an API interface to obtain cargo value information corresponding to the real-time cargo image from a logistics database; and a calculation unit, configured to input the cargo value information, the logistics costs and the responsibility determination result into the amount calculation model, to obtain a reward and punishment amount calculation result.

[0017] Optionally, in a fifth implementation form of the second aspect of the present application, the first generating module comprises: a third obtaining unit, configured to call an API interface to obtain order information corresponding to the real-time cargo image from a logistics database, the order information comprising an order number; a filling unit, configured to obtain a pre-constructed audit form, and fill the order information, the responsibility result and the reward and punishment amount calculation result into the audit form by using a direct mapping method to obtain a to-be-verified form; a verifying unit, configured to perform integrity verification on the to-be-verified form, and perform format verification on the to-be-verified form based on a preset data format verification function, and take the to-be-verified form that passes the integrity verification and the format verification as a to-be-audited form; and a second integrating unit, configured to obtain auditor information, integrate the to-be-audited form and the auditor information to obtain to-be-audited information.

[0018] Optionally, in a sixth implementation form of the second aspect of the present application, the second generating module comprises: a fourth obtaining unit, configured to obtain an audit result corresponding to the to-be-audited information based on the auditor information, the audit result comprising audit pass and audit fail; a fifth obtaining unit, configured to obtain responsibility party account information based on the responsibility result when the audit result is audit pass, and obtain payment contract information corresponding to the responsibility party account information; a third integrating unit, configured to integrate the to-be-audited information, the responsibility party account information and the payment contract information to obtain pre-payment information; and a storage unit, configured to store the pre-payment information into a pre-constructed pre-payment report based on a blockchain technology.

[0019] The third aspect of the present application provides an abnormal allocation waybill processing device, comprising a memory and at least one processor, the memory storing instructions; at least one processor calls the instructions in the memory, so that the abnormal allocation waybill processing device executes each step of the abnormal allocation waybill processing method.

[0020] The fourth aspect of the present application provides a computer readable storage medium, the computer readable storage medium storing instructions, the instructions being executed by a processor to implement each step of the abnormal allocation waybill processing method.

[0021] In the technical solution of the present application, the real-time cargo image is acquired and information extraction processing is performed to obtain cargo appearance information; the real-time cargo image and the cargo appearance information are input into an abnormal type recognition model, if it is determined that there is an abnormal problem, the acquired logistics information and the recognition result are input into a judgment model, and the acquired cargo value information, the logistics information and the judgment result output by the judgment model are input into an amount calculation model to obtain a reward and punishment amount calculation result; based on the order information, the judgment result and the reward and punishment amount calculation result, to-be-audited information is generated; when the to-be-audited information is audited and passed, prepayment information is generated; the method disclosed in the present application can realize automatic judgment of abnormal allocation waybills and corresponding calculation of judgment amount, and has the advantages of fast processing speed, clear and explicit processing flow and high reliability of processing result. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 The first flowchart of the abnormal allocation waybill processing method provided for the embodiment of the present application;

[0023] Figure 2 The second flowchart of the abnormal allocation waybill processing method provided for the embodiment of the present application;

[0024] Figure 3 The third flowchart of the abnormal allocation waybill processing method provided for the embodiment of the present application;

[0025] Figure 4 The fourth flowchart of the abnormal allocation waybill processing method provided for the embodiment of the present application;

[0026] Figure 5 The fifth flowchart of the abnormal allocation waybill processing method provided for the embodiment of the present application;

[0027] Figure 6 The sixth flowchart of the abnormal allocation waybill processing method provided for the embodiment of the present application;

[0028] Figure 7 The seventh flowchart of the abnormal allocation waybill processing method provided for the embodiment of the present application;

[0029] Figure 8 The structure schematic diagram of the abnormal allocation waybill processing device provided for the embodiment of the present application;

[0030] Figure 9 The structure schematic diagram of the abnormal allocation waybill processing device provided for the embodiment of the present application; DETAILED DESCRIPTION

[0031] The present application provides an abnormal allocation waybill processing method, device, equipment and storage medium, in the present application, the terms "first", "second", "third", "fourth" and the like (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the term "includes" or "has" and any variation thereof is intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0032] For ease of understanding, the specific flow of the embodiments of the present application is described below. Please refer to Figure 1 One embodiment of the abnormal allocation waybill processing method in the embodiments of the present application includes:

[0033] 101, acquiring a real-time cargo image, sequentially pre-processing and information extraction processing the real-time cargo image to obtain cargo appearance information;

[0034] In this embodiment, by pre-processing and information extraction processing the image, clear and accurate cargo appearance information can be ensured to be obtained, improving the accuracy and reliability of the subsequent output abnormal identification result.

[0035] 102, inputting the real-time cargo image and the cargo appearance information into a pre-trained abnormal type identification model, and judging whether there is an abnormal problem according to the identification result output by the abnormal type identification model;

[0036] In this embodiment, by using the pre-trained abnormal type identification model, the system can automatically analyze the cargo image and quickly identify whether there is an abnormal problem, not only improving the detection accuracy, but also greatly reducing the dependence on manual review, thereby reducing the operating cost.

[0037] 103, if there is an abnormal problem, acquiring logistics information corresponding to the real-time cargo image, and inputting the logistics information and the identification result into a pre-trained judgment model to obtain a judgment result;

[0038] In this embodiment, the judgment model can quickly determine the responsibility attribution according to rich historical data and complex algorithms, providing clear guidance for the calculation of the subsequent responsibility party reward and punishment amount.

[0039] 104. Obtaining cargo value information corresponding to the real-time cargo image, inputting the cargo value information, the logistics information, and the judgment result into a pre-trained amount calculation model to obtain a reward or penalty amount calculation result;

[0040] In this embodiment, the liability judgment model can accurately calculate the amount of rewards and penalties based on factors such as the value of the goods, the degree of damage, and the responsible party, which not only improves the calculation efficiency and accuracy, but also ensures the fairness of rewards and penalties.

[0041] 105. Obtain order information corresponding to the real-time goods image, and generate information to be reviewed based on the order information, the judgment result, and the reward and penalty amount calculation result;

[0042] 106. Obtain the audit result corresponding to the information to be audited. If the audit result is approved, generate prepayment information.

[0043] The present application discloses a method for processing abnormal transfer waybills, which obtains the appearance information of the goods by acquiring real-time cargo images and performing information extraction and processing; the real-time cargo images and cargo appearance information are input into an abnormal type recognition model; if it is determined that an abnormal problem exists, the acquired logistics information and the recognition result are input into a judgment model, and the acquired cargo value information, logistics information and the judgment result output by the judgment model are input into an amount calculation model to obtain a reward and punishment amount calculation result; based on the order information, the judgment result and the reward and punishment amount calculation result, information to be reviewed is generated; when the information to be reviewed is reviewed and approved, prepayment information is generated; the method disclosed in the present application can realize automatic judgment of abnormal transfer waybills and calculate the corresponding judgment amount, and has the advantages of fast processing speed, clear processing flow and high reliability of processing results.

[0044] See also Figure 2 The second embodiment of the method for handling abnormal transfer orders in the embodiment of the present invention includes:

[0045] 201. Acquire a real-time cargo image, convert the real-time cargo image into a grayscale image, and perform denoising on the converted grayscale image using a Gaussian filter to obtain a denoised cargo image;

[0046] In this embodiment, converting the real-time cargo image into a grayscale image not only simplifies the data processing process but also reduces the consumption of computing resources. The application of the Gaussian filter effectively smoothes the image and removes random noise that may interfere with subsequent processing. The denoising process not only improves the overall image quality but also provides a clearer visual basis for subsequent image analysis.

[0047] 202. Using a histogram equalization algorithm to perform contrast enhancement processing on the denoised cargo image to obtain an enhanced cargo image;

[0048] In this embodiment, the use of the histogram equalization algorithm further enhances the contrast of the image, making details that may have been blurred become clearer, thereby providing more accurate input for the edge detection algorithm.

[0049] 203. Using the Canny edge detection algorithm to identify the cargo contour information in the enhanced cargo image, and obtain cargo feature information;

[0050] In this embodiment, the Canny edge detection algorithm can accurately identify the contour information of the goods and accurately extract the shape and features of the goods even in a complex background. By extracting the contour information of the goods, it is convenient to subsequently determine whether there are abnormal problems such as package damage.

[0051] 204. Use OCR technology to recognize and enhance the text content in the cargo image to obtain cargo label information;

[0052] In this embodiment, OCR technology is used to enable the label information of the goods to be quickly captured and processed by the system; by extracting the text content in the image, it is convenient to subsequently determine whether there are abnormal problems such as blurred labels, incorrect label information (such as address errors, time delays), etc.

[0053] 205. Integrate the cargo feature information and the cargo label information to obtain cargo appearance information.

[0054] See also Figure 3 The third embodiment of the method for handling abnormal transfer waybills in the embodiment of the present invention includes:

[0055] 301. Construct a first training set, where the first training set includes a plurality of historical cargo images, including normal and abnormal historical cargo images, and each historical cargo image is annotated with contour information and label information corresponding to the cargo, as well as whether there is an abnormality.

[0056] 302. Use the first training set to train a convolutional neural network model pre-built based on the TensorFlow deep learning framework to obtain an abnormality type recognition model;

[0057] In this embodiment, the anomaly type recognition model is a convolutional neural network model pre-built based on the TensorFlow deep learning framework. As a powerful deep learning platform, TensorFlow can provide efficient computing resources and a rich tool library, making model training faster and more accurate. The trained anomaly recognition model can not only identify common anomaly types, but also adapt to various cargo shapes and sizes, greatly enhancing the model's generalization ability.

[0058] 303、using an image labeling tool, labeling the appearance information of the goods on the real-time goods image to obtain a labeled goods image;

[0059] In this embodiment, the labeled goods image is converted in size to match the input size required by the abnormal type identification model, ensuring that the model can accurately process image data of various sizes.

[0060] 304、based on the size information corresponding to the abnormal type identification model, converting the size of the labeled goods image, and inputting the labeled goods image after size conversion into the abnormal type identification model to obtain an identification result;

[0061] 305、when the identification result is package damage, label damage or label information error, it indicates that there is an abnormal problem.

[0062] Please refer to Figure 4 , the fourth embodiment of the abnormal allocation order processing method in the embodiment of the application comprises:

[0063] 401、constructing a second training set, the second training set containing involved network points, transportation distances, goods types and abnormal types, and corresponding responsible parties;

[0064] In this embodiment, the second training set provides rich learning samples for the decision tree model, enabling the model to learn the responsibility attribution rules in various complex situations, thereby making accurate judgments in actual applications.

[0065] 402、dividing the second training set, taking the involved network points, transportation distances, goods types and abnormal types as feature data, and taking the responsible parties as target data;

[0066] 403、training the pre-constructed decision tree model based on the second training set after division to obtain a judgment model;

[0067] In this embodiment, due to the characteristics of the decision tree model, as the logistics business continues to expand and change, the second training set can be continuously updated and the judgment model can be retrained to adapt to new judgment requirements; this dynamic adjustment capability enables the method to effectively serve the logistics industry for a long time and meet the changing business requirements.

[0068] 404、if there is an abnormal problem, an API interface is called to obtain logistics information corresponding to the real-time goods image from a logistics database, the logistics information including involved network point information, transportation route conditions and logistics costs;

[0069] 405、inputting the logistics information and the identification result into the judgment model to obtain a judgment result.

[0070] Please refer toFigure 5 The fifth embodiment of the abnormal allocation order processing method in the embodiment of the present application comprises:

[0071] 501. A third training set is constructed, which contains the cargo value, the logistics cost, the judgment result and the corresponding reward and punishment amount;

[0072] 502. The third training set is divided and processed, taking the cargo value, the logistics cost and the judgment result as feature data, and taking the reward and punishment amount as target data;

[0073] In this embodiment, the cargo value, the logistics cost and the judgment result are taken as feature data, and the reward and punishment amount is taken as target data for division and processing, which helps the model to deeply learn the complex correlation between these features and the target, so that the model can more accurately capture the key factors affecting the reward and punishment amount, thereby improving the accuracy of the calculation result.

[0074] 503. The pre-constructed linear regression model is trained based on the third training set after division and processing, to obtain an amount calculation model;

[0075] In this embodiment, the linear regression model is selected for training, not only because it is simple and easy to understand, but also because the linear model can often provide clear interpretability when processing such problems. The transparency and interpretability of such a model make it more advantageous in practical applications.

[0076] 504. An API interface is called to obtain cargo value information corresponding to a real-time cargo image from a logistics database;

[0077] 505. The cargo value information, the logistics cost and the judgment result are input into the amount calculation model to obtain a reward and punishment amount calculation result.

[0078] Please refer to Figure 6 The sixth embodiment of the abnormal allocation order processing method in the embodiment of the present application comprises:

[0079] 601. An API interface is called to obtain order information corresponding to a real-time cargo image from a logistics database, the order information including an order number;

[0080] 602. A pre-constructed audit form is obtained, and a direct mapping method is used to fill in the order information, the judgment result and the reward and punishment amount calculation result in the audit form to obtain a to-be-verified form;

[0081] In the embodiment, the order information, the responsibility result and the reward and punishment amount calculation result are filled in the audit form by using the direct mapping method, which not only simplifies the operation process, but also significantly reduces the errors caused by manual input; through accurate matching, it is ensured that each item of data can be accurately reflected in the audit form, thereby improving the accuracy of data processing.

[0082] 603、Perform integrity check on the to-be-verified form, and perform format check on the to-be-verified form based on a preset data format check function, and take the to-be-verified form that passes the integrity check and the format check as a to-be-audited form;

[0083] In the embodiment, the to-be-verified form is checked for format by using the preset data format check function, so as to ensure the standardization and consistency of the form, and make the auditor quickly identify and understand the content of the form during the auditing process, thereby improving the efficiency and accuracy of the auditing.

[0084] 604、Obtain auditor information, integrate the to-be-audited form and the auditor information, and obtain to-be-audited information.

[0085] Please refer to Figure 7 The seventh embodiment of the abnormal allocation order processing method in the embodiment of the application includes:

[0086] 701、Based on the auditor information, obtain an audit result corresponding to the to-be-audited information, the audit result including audit pass and audit fail;

[0087] 702、When the audit result is audit pass, obtain responsibility party account information based on the responsibility result, and obtain payment contract information corresponding to the responsibility party account information;

[0088] In the embodiment, in the process of obtaining the responsibility party account information, the system can quickly locate the related account and accurately extract the payment contract information corresponding thereto, thereby laying a solid foundation for subsequent prepayment operations.

[0089] 703、Integrate the to-be-audited information, the responsibility party account information and the payment contract information, and obtain prepayment information;

[0090] 704、Based on the blockchain technology, store the prepayment information into a pre-prepayment report pre-constructed;

[0091] In this embodiment, the prepayment information is stored in the pre-built prepayment report based on the blockchain technology, which significantly improves the security and tamper resistance of the data; the use of the blockchain technology ensures that the data cannot be changed once recorded, effectively preventing the risk of data leakage or malicious modification; for example, when storing the prepayment information, the blockchain technology can ensure that each record is transparent and irreversible, thereby providing a solid security guarantee for the entire abnormal allocation order processing flow.

[0092] The abnormal allocation order processing method in the embodiments of the application is described above, and the abnormal allocation order processing device in the embodiments of the application is described below. Please refer to Figure 8 An embodiment of the abnormal allocation order processing device in the embodiments of the application includes:

[0093] The processing module 801 is configured to obtain a real-time cargo image, sequentially perform preprocessing and information extraction processing on the real-time cargo image, and obtain cargo appearance information. The abnormality recognition module 802 is configured to input the real-time cargo image and the cargo appearance information into a pre-trained abnormal type recognition model, and determine whether there is an abnormal problem according to a recognition result output by the abnormal type recognition model. The responsibility determination module 803 is configured to, if there is an abnormal problem, obtain logistics information corresponding to the real-time cargo image, and input the logistics information and the recognition result into a pre-trained responsibility determination model to obtain a responsibility determination result. The calculation module 804 is configured to obtain cargo value information corresponding to the real-time cargo image, input the cargo value information, the logistics information, and the responsibility determination result into a pre-trained amount calculation model, and obtain a reward and punishment amount calculation result. The first generation module 805 is configured to obtain order information corresponding to the real-time cargo image, and generate to-be-audited information based on the order information, the responsibility determination result, and the reward and punishment amount calculation result. The second generation module 806 is configured to obtain an audit result corresponding to the to-be-audited information, and generate prepayment information if the audit result is audit passed.

[0094] In this embodiment, the processing module 801 includes: a first processing unit 8011 configured to obtain a real-time cargo image, convert the real-time cargo image into a grayscale image, and perform denoising processing on the converted grayscale image using a Gaussian filter to obtain a denoised cargo image; an enhancement unit 8012 configured to perform contrast enhancement processing on the denoised cargo image using a histogram equalization algorithm to obtain an enhanced cargo image; an edge detection unit 8013 configured to identify cargo contour information in the enhanced cargo image using a Canny edge detection algorithm to obtain cargo feature information; an extraction unit 8014 configured to identify text content in the enhanced cargo image using an OCR technology to obtain cargo label information; and a first integration unit 8015 configured to integrate the cargo feature information and the cargo label information to obtain cargo appearance information.

[0095] In the embodiment, the anomaly identification module 802 includes: a first construction unit 8021 configured to construct a first training set, the first training set including a plurality of historical cargo images, the plurality of historical cargo images including normal and abnormal historical cargo images, and each historical cargo image being labeled with corresponding contour information and label information of the cargo and whether there is an anomaly; a first training unit 8022 configured to train a pre-constructed convolutional neural network model based on a TensorFlow deep learning framework using the first training set to obtain an anomaly type identification model; a labeling unit 8023 configured to use an image labeling tool to label the cargo appearance information on the real-time cargo image to obtain a labeled cargo image; a conversion unit 8024 configured to perform size conversion on the labeled cargo image based on size information corresponding to the anomaly type identification model, and input the labeled cargo image after size conversion into the anomaly type identification model to obtain an identification result; and an anomaly identification unit 8025 configured to indicate that there is an abnormal problem when the identification result is package damage, label damage or label information error.

[0096] In the embodiment, the responsibility determination module 803 includes: a second construction unit 8031 configured to construct a second training set, the second training set including involved network points, transportation distances, cargo types and anomaly types, and corresponding responsible parties; a first division unit 8032 configured to perform division processing on the second training set, taking the involved network points, transportation distances, cargo types and anomaly types as feature data, and taking the responsible parties as target data; a second training unit 8033 configured to train a pre-constructed decision tree model based on the second training set after the division processing to obtain a responsibility determination model; a first acquisition unit 8034 configured to, if there is an abnormal problem, call an API interface to obtain logistics information corresponding to the real-time cargo image from a logistics database, the logistics information including involved network point information, transportation route conditions and logistics costs; and a responsibility determination unit 8035 configured to input the logistics information and the identification result into the responsibility determination model to obtain a responsibility determination result.

[0097] In the embodiment, the computing module 804 comprises: a third construction unit 8041 configured to construct a third training set, the third training set comprising the cargo value, the logistics cost, the judgment result and the corresponding reward and punishment amount; a second division unit 8042 configured to perform division processing on the third training set, taking the cargo value, the logistics cost and the judgment result as feature data and taking the reward and punishment amount as target data; a third training unit 8043 configured to train a pre-constructed linear regression model based on the third training set after the division processing to obtain an amount calculation model; a second acquisition unit 8044 configured to call an API interface to obtain cargo value information corresponding to a real-time cargo image from a logistics database; and a computing unit 8045 configured to input the cargo value information, the logistics cost and the judgment result into the amount calculation model to obtain a reward and punishment amount calculation result.

[0098] In the embodiment, the first generation module 805 comprises: a third acquisition unit 8051 configured to call an API interface to obtain order information corresponding to a real-time cargo image from a logistics database, the order information comprising an order number; a filling unit 8052 configured to obtain a pre-constructed audit form, and fill the order information, the judgment result and the reward and punishment amount calculation result into the audit form by using a direct mapping method to obtain a to-be-verified form; a verification unit 8053 configured to perform integrity verification on the to-be-verified form, and perform format verification on the to-be-verified form based on a preset data format verification function, taking the to-be-verified form that passes the integrity verification and the format verification as a to-be-audited form; and a second integration unit 8054 configured to obtain auditor information, and integrate the to-be-audited form and the auditor information to obtain to-be-audited information.

[0099] In the embodiment, the second generation module 806 comprises: a fourth acquisition unit 8061 configured to obtain an audit result corresponding to the to-be-audited information based on the auditor information, the audit result comprising audit pass and audit fail; a fifth acquisition unit 8062 configured to, when the audit result is audit pass, obtain responsibility party account information based on the judgment result, and obtain payment contract information corresponding to the responsibility party account information; a third integration unit 8063 configured to integrate the to-be-audited information, the responsibility party account information and the payment contract information to obtain pre-payment information; and a storage unit 8064 configured to store the pre-payment information into a pre-constructed pre-payment report based on a blockchain technology.

[0100] Based on the same idea as the method in the above embodiment, the device provided in the present application can implement the method of the above embodiment.

[0101] The above Figure 8The abnormal allocation waybill processing apparatus in the embodiment of the present application is described in detail from the perspective of the modular functional entity, and the abnormal allocation waybill processing device in the embodiment of the present application is described in detail from the perspective of hardware processing.

[0102] Figure 9 Fig. 9 is a structural schematic diagram of an abnormal allocation waybill processing device provided by the embodiment of the present application. The abnormal allocation waybill processing device 900 can be quite different due to different configurations or performances, and can include one or more processors (central processing units, CPU) 910 (for example, one or more processors) and a memory 920, one or more storage media 930 (for example, one or more mass storage devices) for storing application programs 933 or data 932. The memory 920 and the storage media 930 can be temporary storage or persistent storage. The programs stored in the storage media 930 can include one or more modules (not shown in the figure), and each module can include a series of instruction operations in the abnormal allocation waybill processing device 900. Furthermore, the processor 910 can be configured to communicate with the storage media 930 and execute the series of instruction operations in the storage media 930 on the abnormal allocation waybill processing device 900, so as to implement the steps of the abnormal allocation waybill processing method provided by the above-mentioned method embodiments.

[0103] The abnormal allocation waybill processing device 900 can further include one or more power supplies 940, one or more wired or wireless network interfaces 950, one or more input / output interfaces 960, and / or one or more operating systems 931, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that the abnormal allocation waybill processing device 900 can further include other components that are not shown in the figure, such as a bus, a keyboard, a mouse, a display, a hard disk, etc. Figure 9 The abnormal allocation waybill processing device structure shown does not constitute a limitation on the abnormal allocation waybill processing device, and can include more or fewer components than shown, or combine certain components, or different component arrangements.

[0104] The present application also provides a computer readable storage medium, which can be a non-volatile computer readable storage medium or a volatile computer readable storage medium. The computer readable storage medium stores instructions, and when the instructions run on a computer, the computer executes the steps of the abnormal allocation waybill processing method.

[0105] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system or device, unit can refer to the corresponding process in the foregoing method embodiments, which will not be described herein.

[0106] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0107] Finally, it should be noted that: the above only for the preferred examples of the present application, and not for limiting the present application, although the present application is described in detail with reference to the foregoing examples, for those skilled in the art, it still can be modified to the technical solutions recorded in the foregoing embodiments, or equivalent replacement of some technical features. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. An abnormal dispatching waybill processing method, characterized by comprising the steps of: include: Acquire a real-time cargo image, and perform preprocessing and information extraction on the real-time cargo image in sequence to obtain cargo appearance information; specifically, acquire the real-time cargo image, convert the real-time cargo image into a grayscale image, and perform denoising on the converted grayscale image using a Gaussian filter to obtain a denoised cargo image; perform contrast enhancement on the denoised cargo image using a histogram equalization algorithm to obtain an enhanced cargo image; use a Canny edge detection algorithm to identify cargo contour information in the enhanced cargo image to obtain cargo feature information; use optical character recognition technology to identify text content in the enhanced cargo image to obtain cargo label information; and integrate the cargo feature information and the cargo label information to obtain cargo appearance information; The real-time cargo image and cargo appearance information are input into a pre-trained abnormal type recognition model, and whether there is an abnormality problem is determined based on the recognition result output by the abnormal type recognition model; specifically, a first training set is constructed, wherein the first training set includes multiple historical cargo images, the multiple historical cargo images include normal and abnormal historical cargo images, and each historical cargo image is annotated with the corresponding contour information and label information of the cargo and whether there is an abnormality; the first training set is used to train a convolutional neural network model pre-built based on the TensorFlow deep learning framework to obtain an abnormal type recognition model; an image annotation tool is used to annotate the cargo appearance information onto the real-time cargo image to obtain an annotated cargo image; the annotated cargo image is resized based on the size information corresponding to the abnormal type recognition model, and the annotated cargo image after the size conversion is input into the abnormal type recognition model to obtain a recognition result; when the recognition result is that the packaging is damaged, the label is damaged, or the label information is incorrect, it indicates that there is an abnormality problem; If there is an abnormality, the logistics information corresponding to the real-time cargo image is obtained, and the logistics information and recognition results are input into the pre-trained fault judgment model to obtain the fault judgment result; Obtaining cargo value information corresponding to the real-time cargo image, inputting the cargo value information, the logistics information, and the judgment result into a pre-trained amount calculation model to obtain a reward or punishment amount calculation result; Obtaining order information corresponding to the real-time goods image, and generating information to be reviewed based on the order information, the judgment result, and the reward and penalty amount calculation result; Obtain the audit result corresponding to the information to be audited. If the audit result is passed, generate prepayment information.

2. The method of handling an exception waybill according to claim 1, wherein, If an abnormality exists, the logistics information corresponding to the real-time cargo image is obtained, and the logistics information and recognition results are input into the pre-trained judgment model to obtain the judgment results, including: Constructing a second training set, the second training set includes the involved outlets, transportation distances, cargo types, anomaly types, and corresponding responsible parties; The second training set is divided and processed, with the involved outlets, transportation distance, cargo type, and anomaly type as feature data, and the responsible party as target data; The pre-built decision tree model is trained based on the second training set after the partitioning process to obtain a judgment model; If there is an abnormal problem, an API interface is called to obtain logistics information corresponding to the real-time cargo image from a logistics database, the logistics information including involved site information, transportation route conditions and logistics costs; The logistics information and the identification result are input into the judgment model to obtain a judgment result.

3. The method of handling an exception waybill according to claim 2, wherein, The cargo value information corresponding to the real-time cargo image is obtained, the cargo value information, the logistics information and the judgment result are input into a pre-trained amount calculation model to obtain an award and punishment amount calculation result, including: A third training set is constructed, the third training set including cargo value, logistics cost, judgment result and corresponding award and punishment amount; The third training set is divided, the cargo value, the logistics cost and the judgment result are taken as feature data, and the award and punishment amount is taken as target data; A pre-constructed linear regression model is trained based on the third training set after the division to obtain an amount calculation model; An API interface is called to obtain cargo value information corresponding to the real-time cargo image from a logistics database; The cargo value information, the logistics cost and the judgment result are input into the amount calculation model to obtain an award and punishment amount calculation result.

4. The method of handling an exception waybill according to claim 1, wherein, The order information corresponding to the real-time cargo image is obtained, and based on the order information, the judgment result and the award and punishment amount calculation result, to-be-audited information is generated, including: An API interface is called to obtain order information corresponding to the real-time cargo image from a logistics database, the order information including an order number; A pre-constructed audit form is obtained, and a direct mapping method is used to fill the order information, the judgment result and the award and punishment amount calculation result into the audit form to obtain a to-be-checked form; The to-be-checked form is subjected to integrity checking and format checking based on a preset data format checking function, and the to-be-checked form passing the integrity checking and the format checking is taken as a to-be-audited form; Auditor information is obtained, and the to-be-audited form and the auditor information are integrated to obtain to-be-audited information.

5. The method of handling an exception waybill according to claim 4, wherein, The audit result corresponding to the to-be-audited information is obtained, and if the audit result is audit pass, prepayment information is generated, including: The audit result corresponding to the to-be-audited information is obtained based on the auditor information, the audit result including audit pass and audit fail; When the audit result is audit pass, the responsible party account information is obtained based on the judgment result, and the payment contract information corresponding to the responsible party account information is obtained; The to-be-audited information, the responsible party account information and the payment contract information are integrated to obtain prepayment information; The prepayment information is stored in a pre-constructed prepayment report based on blockchain technology.

6. An abnormal dispatching waybill processing apparatus characterized by comprising: including: The processing module is configured to acquire a real-time cargo image, sequentially perform preprocessing and information extraction processing on the real-time cargo image, and obtain cargo appearance information. Specifically, the real-time cargo image is acquired, converted into a grayscale image, and denoised by using a Gaussian filter to obtain a denoised cargo image. The denoised cargo image is subjected to contrast enhancement processing by using a histogram equalization algorithm to obtain an enhanced cargo image. Cargo contour information in the enhanced cargo image is identified by using a Canny edge detection algorithm to obtain cargo feature information. Text content in the enhanced cargo image is identified by using an OCR technique to obtain cargo label information. The cargo feature information and the cargo label information are integrated to obtain the cargo appearance information. The abnormality identification module is configured to input the real-time cargo image and the cargo appearance information into a pre-trained abnormality type identification model, and determine whether there is an abnormal problem according to an identification result output by the abnormality type identification model. Specifically, a first training set is constructed, the first training set includes a plurality of historical cargo images, the plurality of historical cargo images include normal and abnormal historical cargo images, and each historical cargo image is labeled with contour information and label information corresponding to the cargo and whether there is an abnormality. A convolutional neural network model pre-constructed based on a TensorFlow deep learning framework is trained by using the first training set to obtain the abnormality type identification model. An image labeling tool is used to label the cargo appearance information on the real-time cargo image to obtain a labeled cargo image. The labeled cargo image is subjected to size conversion based on size information corresponding to the abnormality type identification model, and the labeled cargo image after size conversion is input into the abnormality type identification model to obtain the identification result. When the identification result is package damage, label damage or label information error, it indicates that there is an abnormal problem. The judgment module is configured to, if there is an abnormal problem, acquire logistics information corresponding to the real-time cargo image, and input the logistics information and the identification result into a pre-trained judgment model to obtain a judgment result. The calculation module is configured to acquire cargo value information corresponding to the real-time cargo image, input the cargo value information, the logistics information and the judgment result into a pre-trained amount calculation model to obtain an award and punishment amount calculation result. The first generation module is configured to acquire order information corresponding to the real-time cargo image, and generate to-be-audited information based on the order information, the judgment result and the award and punishment amount calculation result. The second generation module is configured to acquire an audit result corresponding to the to-be-audited information, and generate prepayment information if the audit result is audit passed.

7. An exception waybill processing apparatus, characterized by comprising: The abnormality allocation waybill processing device includes a memory and at least one processor, and the memory stores instructions. The at least one processor calls the instructions in the memory to enable the abnormality allocation waybill processing device to perform each step of the abnormality allocation waybill processing method of any one of claims 1-5.

8. A computer-readable storage medium having stored thereon instructions, the computer-readable storage medium comprising: The instructions are executed by the processor to implement each step of the abnormality allocation waybill processing method of any one of claims 1-5.

Images