Internet of Things warehouse receipt management method and system with optimized efficiency

Abstract

The invention relates to an efficiency-optimized Internet of Things warehouse receipt management method and system, and relates to the technical field of Internet of Things warehouses, and the method comprises the steps: obtaining a current cargo scheduling number of an Internet of Things warehouse, and obtaining a cargo scheduling recognition image sequence in a past preset time length; randomly configuring a cargo scheduling identification speed, and carrying out identification success rate prediction to obtain a first identification success rate; printing quality identification and label pasting quality identification of the cargo label are carried out, a printing quality parameter and a label pasting quality parameter are obtained, identification influence prediction is carried out, an identification influence coefficient is obtained, and correction calculation is carried out to obtain a second identification success rate; and calculating to obtain an efficiency improvement score, optimizing a cargo scheduling identification speed, obtaining an optimal cargo scheduling identification speed, and performing warehouse receipt management. According to the invention, the technical effect of improving the efficiency of warehouse cargo scheduling and warehouse receipt information identification management of the Internet of Things is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things warehouses, and in particular to an efficiency-optimized Internet of Things warehouse receipt management method and system. Background Art

[0002] The warehouse receipt management efficiency of IoT warehouses is an important indicator that needs to be improved in the current logistics warehouse field. Warehouse receipt management is performed by recording cargo information when leaving and entering the warehouse. In order to improve the efficiency of warehouse management, automatic recognition technology based on barcodes and QR code labels is widely used to quickly complete warehouse receipt information collection, update and management by scanning cargo labels.

[0003] In order to further improve the efficiency of warehouse receipt management, the efficiency is also improved by increasing the transmission speed of the cargo conveyor belt. However, a higher transmission speed may lead to a decrease in the quality of label image acquisition, affecting the success rate of warehouse receipt information recognition, which in turn leads to a decrease in efficiency. Summary of the invention

[0004] The present invention aims to solve the technical problem of low efficiency in warehouse receipt information collection and management in the prior art of the Internet of Things warehouse, and provides an efficiency-optimized Internet of Things warehouse receipt management method and system.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: In a first aspect, the present invention provides an efficiency-optimized IoT warehouse receipt management method, comprising: obtaining the current cargo dispatch quantity of the IoT warehouse, and obtaining a cargo dispatch identification image sequence within a preset time period in the past; randomly configuring the cargo dispatch identification speed, predicting the identification success rate, and obtaining a first identification success rate; according to the cargo dispatch identification image sequence, performing printing quality identification and label pasting quality identification of the cargo label, obtaining printing quality parameters and label pasting quality parameters, and performing identification impact prediction to obtain an identification impact coefficient, performing correction calculation on the first identification success rate, and obtaining a second identification success rate; according to the second identification success rate, the cargo dispatch quantity, and the cargo dispatch identification speed, calculating an efficiency improvement score, optimizing the cargo dispatch identification speed, obtaining the optimal cargo dispatch identification speed, and performing warehouse receipt management.

[0006] In the second aspect, the present invention provides an efficiency-optimized Internet of Things warehouse receipt management system, comprising: a cargo dispatch data acquisition module, used to obtain the current cargo dispatch quantity of the Internet of Things warehouse, and obtain a cargo dispatch identification image sequence within a preset time period in the past; a recognition success rate prediction module, used to randomly configure the cargo dispatch identification speed, perform recognition success rate prediction, and obtain a first recognition success rate; a recognition success rate correction module, used to perform printing quality recognition and label pasting quality recognition of the cargo label according to the cargo dispatch identification image sequence, obtain printing quality parameters and label pasting quality parameters, and perform recognition impact prediction to obtain a recognition impact coefficient, perform correction calculation on the first recognition success rate, and obtain a second recognition success rate; a warehouse receipt management optimization module, used to calculate the efficiency improvement score according to the second recognition success rate, the cargo dispatch quantity and the cargo dispatch identification speed, and optimize the cargo dispatch identification speed to obtain the optimal cargo dispatch identification speed for warehouse receipt management.

[0007] The beneficial effects of the present invention are as follows: by analyzing historical cargo dispatch data and real-time images, the recognition success rate of the configured cargo dispatch recognition speed is predicted, and dynamic correction is performed based on the fluctuation influence coefficient of printing quality and pasting quality. Specifically, by combining the cargo dispatch recognition image sequence and the printing quality and pasting quality parameter analysis of the cargo label, the recognition success rate and dispatch efficiency in the warehouse receipt management process are significantly improved, and a more accurate second recognition success rate is obtained. Secondly, by randomly configuring and optimizing the cargo dispatch recognition speed, a comprehensive calculation and evaluation is performed in combination with the cargo dispatch quantity and the recognition success rate, ensuring that the dispatch speed of the outbound or inbound process is optimized while improving the recognition efficiency, thereby maximizing the overall efficiency. Compared with the prior art, the present invention can ensure the label recognition success rate on the basis of maximizing the dispatch speed of cargo in and out of the warehouse, and taking into account the impact of the recognition success rate caused by the fluctuation of label quality or the change of label pasting, the efficiency, intelligence level and operation stability of warehouse receipt information recognition management and warehousing management are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A schematic diagram of a flow chart of an efficiency-optimized IoT warehouse receipt management method provided by the present invention; Figure 2 A structural schematic diagram of an efficiency-optimized Internet of Things warehouse receipt management system provided by the present invention.

[0009] In the accompanying drawings, the components represented by the reference numerals are described as follows: Cargo dispatching data acquisition module 11, recognition success rate prediction module 12, recognition success rate correction module 13, warehouse receipt management optimization module 14. DETAILED DESCRIPTION

[0010] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0011] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0012] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0013] Embodiment 1, as Figure 1 As shown, the embodiment of the present invention provides an efficiency-optimized IoT warehouse receipt management method, which specifically includes the following steps: S1: Obtain the current cargo dispatch quantity of the IoT warehouse and obtain the cargo dispatch recognition image sequence within the past preset time period; In the embodiment of the present application, the current cargo dispatch quantity of the IoT warehouse is first obtained. The larger the current cargo dispatch quantity, the greater the warehouse receipt information management efficiency is required to improve the overall warehouse management efficiency.

[0014] And obtain the cargo dispatch recognition image sequence within the preset time period in the past, which includes the image of the cargo dispatched by the most recent recognition information, and includes the label image of the cargo, which includes the warehouse receipt information of the cargo, and manages the entry and exit of the warehouse through the recognition label. The printing quality and pasting position of the label image will affect the acquisition quality of the label image. For example, the offset of the pasting position will cause the label image to be distorted. Therefore, the cargo dispatch recognition image sequence within the preset time period in the past is collected to analyze the printing quality and pasting quality of the label image, and the success rate of cargo label recognition is analyzed in combination with the cargo dispatch recognition speed to analyze the quality of warehouse receipt management.

[0015] Step S1 of the method provided in the embodiment of the present application includes: Obtain the number of goods currently to be dispatched in the IoT warehouse; Retrieving cargo dispatch identification images within a preset time period in the past to obtain a historical cargo dispatch identification image set; In the historical cargo dispatch recognition image set, the recognition images when the cargo labels are recognized are extracted to obtain a cargo dispatch recognition image sequence.

[0016] In the embodiment of the present application, the total number of goods that need to be dispatched is detected in real time by an IoT sensor (such as an infrared counter). The number of goods dispatched that need to be dispatched can also be obtained based on the warehouse cargo in and out task list. The cargo dispatch is, for example, cargo out or in and the corresponding warehouse receipt management. For example, the number of goods that need to be managed for outbound or inbound warehouse receipts in the next day is obtained, such as 1,000 pieces.

[0017] Furthermore, the cargo dispatch identification image data within the past preset time period in the warehouse database is called. The past preset time period can be set to one day or one week. For example, the cargo dispatch identification images within the past day, that is, the cargo images for identifying cargo information are collected as cargo dispatch identification images, which include images of labels attached to the cargo to identify cargo information in the labels, such as batch, place of origin, order number, etc. In this way, a historical cargo dispatch identification image set is obtained. For example, the historical cargo dispatch identification image set within the past 24 hours may contain 5,000 frames of images.

[0018] Furthermore, in the historical cargo dispatch identification image set, image frames directly related to cargo label identification are screened out, such as cargo dispatch identification images of image frames containing cargo label images. Cargo label identification images refer to images that contain cargo labels (such as barcodes, QR codes, or RFID tags) and can be identified by acquisition devices. Among them, cargo dispatch identification images that include label images can be manually screened out, and cargo dispatch identification images that do not include label images, such as conveyor belt images or images of cargo parts that do not include labels, can be deleted. It is also possible to obtain historical cargo dispatch identification images of the time frame in which cargo label information is identified based on the records collected and identified by an acquisition device (such as a QR code acquisition and identification device).

[0019] Then, the screened cargo dispatch identification images are sorted in chronological order to obtain a cargo dispatch identification image sequence.

[0020] Through the implementation of the above steps, the current cargo dispatch quantity and cargo dispatch identification image sequence can be efficiently obtained, providing solid data support for the subsequent optimization of cargo dispatch identification speed. For example, by combining the cargo dispatch identification image sequence and the current cargo quantity, the dispatch speed can be evaluated and the label quality can be analyzed, thereby significantly improving the operating efficiency and quality of warehouse receipt identification management.

[0021] S2: Randomly configure the cargo scheduling recognition speed, predict the recognition success rate, and obtain the first recognition success rate; In the embodiment of the present application, the transmission speed of warehouse cargo scheduling is optimized to improve the warehouse receipt recognition efficiency and management efficiency.

[0022] First, the cargo dispatch identification speed is randomly configured, which is specifically the speed of the cargo dispatch transmission equipment, for example, the transmission speed of the cargo on the conveyor belt when the label warehouse receipt information is identified and managed, for example, 1.0 m / s.

[0023] Although increasing the speed of cargo delivery can improve efficiency, it will reduce the image acquisition time, exposure time, and image acquisition quality, thus reducing the success rate of label image recognition, increasing the probability of misidentification or failure to recognize, and reducing warehouse receipt management efficiency. Therefore, it is necessary to predict the label recognition success rate based on the cargo dispatch recognition speed to obtain the first recognition success rate. In this way, the probability of successful label information recognition and warehouse receipt information management under the cargo dispatch recognition speed is reflected.

[0024] Step S2 in the method provided in the embodiment of the present application includes: Obtain the transmission speed range of the transmission equipment for cargo scheduling identification, and randomly generate a transmission speed as the cargo scheduling identification speed; According to the cargo dispatch identification data of different batches of cargo, a plurality of different cargo identification prediction paths are constructed to obtain a cargo identification predictor; The cargo scheduling recognition speed is input into a plurality of cargo recognition prediction paths in the cargo recognition predictor, a plurality of recognition success rates are predicted, and a first recognition success rate is calculated.

[0025] In the embodiment of the present application, the transmission speed range of the transmission device is first obtained and the scheduling speed is randomly generated. For example, the transmission speed range of the transmission device (such as a conveyor belt for transmitting goods) can be obtained through the parameter interface of the transmission device. For example, the speed range of the conveyor belt is 0.5 m / s to 1.5 m / s, and the operating speed range of the AGV (automatic guided vehicle) is 1.0 m / s to 2.0 m / s.

[0026] Then, a transmission speed is randomly generated within the transmission speed interval as the cargo dispatch identification speed. For example, a transmission speed of 1 m / s is randomly generated within 0.5 m / s to 1.5 m / s as the cargo dispatch identification speed, which is used as the initial solution for optimizing the cargo transmission identification tag speed.

[0027] Furthermore, based on the cargo dispatching identification data of different batches of cargo, different multiple cargo identification prediction paths are constructed to obtain a cargo identification predictor for predicting the success rate of cargo label identification at different cargo dispatching identification speeds. Among them, the accuracy and robustness of the predicted identification success rate can be improved by constructing based on the cargo dispatching identification data of different batches of cargo.

[0028] The step of "constructing multiple different cargo identification prediction paths according to cargo scheduling identification data of different batches of cargo to obtain a cargo identification predictor" in the method provided in the embodiment of the present application includes: Randomly select multiple batches of goods in the historical time and collect multiple sample goods scheduling recognition speed sets as input features; The success rates of cargo label recognition under different sample cargo scheduling recognition speeds are collected and marked as multiple sample success rate sets as output features; The plurality of sample cargo scheduling recognition speed sets and the plurality of sample success rate sets are respectively used to construct a plurality of cargo recognition prediction paths to obtain a cargo recognition predictor.

[0029] In the embodiment of the present application, in the cargo dispatch identification data of the IoT warehouse in the historical time, multiple batches of goods are randomly selected, for example, multiple batches of goods are randomly selected within multiple time periods, for example, multiple batches of goods within multiple 24-hour periods. When the multiple batches of goods are transmitted for tag identification, there may be an adjustment of the cargo dispatch identification speed. Based on this, the cargo dispatch identification speeds configured for the multiple batches of goods are collected respectively, and multiple sample cargo dispatch identification speed sets are obtained as input features for the subsequent construction of the cargo identification predictor. Exemplarily, the sample cargo dispatch identification speed can be 0.8 m / s, 1.2 m / s, 1.5 m / s, etc.

[0030] Furthermore, the success rate of cargo label recognition at different sample cargo dispatching and identification speeds is collected, for example, the proportion of cargo that successfully and correctly transmits cargo identification labels to all cargo at different sample cargo dispatching and identification speeds. For example, if the sample cargo dispatching and identification speed is 0.8 m / s, 100 pieces of cargo are transmitted at this speed, and the number of cargo that successfully and accurately identifies cargo labels is 98 pieces, then the success rate of cargo label recognition at the sample cargo dispatching and identification speed is 98%, and the success rate of cargo label recognition at the sample cargo dispatching and identification speed of 1.2 m / s is 95%. For example, multiple sample success rate sets corresponding to multiple sample cargo dispatching and identification speed sets are obtained by labeling, and used as the output features for constructing a cargo identification predictor in the subsequent construction.

[0031] In the embodiment of the present application, the multiple sample cargo scheduling identification speed sets and the multiple sample success rate sets are respectively used as construction data to construct multiple cargo identification prediction paths.

[0032] Exemplarily, a plurality of sample cargo scheduling recognition speed sets are used as input features, and a plurality of sample success rate sets are used as output features, and a mapping relationship between each sample cargo scheduling recognition speed and the corresponding sample success rate is constructed to form a mapping table of sample cargo scheduling recognition speed and sample success rate as a plurality of cargo identification prediction paths, and then the plurality of cargo identification prediction paths are combined to obtain a cargo identification predictor.

[0033] Optionally, a network architecture of multiple cargo identification prediction paths can be constructed based on a feedforward neural network in machine learning, including an input layer, a hidden layer, a fully connected layer, and an output layer. Then, multiple sample cargo scheduling recognition speed sets are used as input features, and multiple sample success rate sets are used as output features. Multiple cargo identification prediction paths are supervised and trained to reduce the loss until the accuracy meets the requirements, for example, reaches 95%, and the training of multiple cargo identification prediction paths is completed. Multiple cargo identification prediction paths that have been trained are combined to obtain a cargo identification predictor.

[0034] Based on the constructed cargo identification predictor, the currently configured cargo scheduling identification speed is input into the multiple cargo identification prediction paths in the cargo identification predictor, and the identification success rate corresponding to the cargo scheduling identification speed mapping is obtained in the mapping table in the multiple cargo identification prediction paths, so that multiple identification success rates are obtained by mapping.

[0035] Among them, if there is a sample cargo dispatching recognition speed that is the same as the cargo dispatching recognition speed in the mapping table, the corresponding sample recognition success rate is mapped as the predicted recognition success rate. If there is no sample cargo dispatching recognition speed that is the same as the cargo dispatching recognition speed, the sample recognition success rate corresponding to the sample cargo dispatching recognition speed mapping that is closest to the currently configured cargo dispatching recognition speed is obtained as the predicted recognition success rate.

[0036] Alternatively, the current cargo scheduling recognition speed is input into a plurality of cargo recognition prediction paths based on feedforward neural network training, and the prediction output obtains a plurality of recognition success rates.

[0037] Furthermore, the average of multiple recognition success rates is calculated to obtain the success rate of identifying cargo labels and managing warehouse receipt information when transporting cargo at the current cargo scheduling recognition speed, which is used as the prediction result of the recognition success rate.

[0038] The embodiment of the present application establishes an accurate recognition success rate prediction path by collecting success rate and error rate data, and can obtain the probability of successful cargo label recognition under different cargo scheduling recognition speeds, so as to quantify the impact of improving warehouse receipt information recognition efficiency on the recognition success rate, and then use it as reference data for optimizing efficiency, and then conduct comprehensive optimization of efficiency in the future.

[0039] S3: According to the cargo dispatch identification image sequence, the printing quality identification and label pasting quality identification of the cargo label are performed to obtain the printing quality parameters and the label pasting quality parameters, and the identification influence prediction is performed to obtain the identification influence coefficient, and the first identification success rate is corrected and calculated to obtain the second identification success rate; In the embodiment of the present application, the printing quality and pasting quality of the cargo label will affect the quality of label image acquisition. The printing equipment of the cargo label may have errors, and there may also be position errors when the label is pasted, causing the label image to change and distort. Combined with the speed of cargo transmission, it affects the quality of label image acquisition and thus affects the success rate of label recognition.

[0040] Therefore, according to the cargo dispatch identification image sequence within the preset time in the past, the printing quality identification and label pasting quality identification of the cargo label are performed to obtain the cargo label printing quality parameters and label pasting quality parameters in the recent time. Then, according to the cargo label printing quality parameters and label pasting quality parameters, the influence on the label recognition success rate is analyzed to obtain the recognition influence coefficient, and the first recognition success rate based on the cargo dispatch identification speed prediction in the above content is corrected and calculated to obtain the second recognition success rate. In this way, the speed influence and the label printing and pasting influence are combined to obtain the final recognition label success rate, that is, the success rate of warehouse receipt information recognition.

[0041] Step S3 of the method provided in the embodiment of the present application includes: According to the cargo dispatch label recognition records in the historical time, a set of sample cargo dispatch recognition images is collected, and the deviation between the label image in each sample cargo dispatch recognition image and the standard printed label image is marked to obtain a set of sample printing quality parameters; Mark the movement range between the label position in the label image and the standard label position in each sample cargo dispatch recognition image to obtain a set of sample label pasting quality parameters; The sample cargo dispatch identification image set is used as input features, the sample printing quality parameter set and the sample label pasting quality parameter set are used as output features, and a convolutional neural network is used to train a cargo label feature identifier; Inputting the cargo dispatch identification images in the cargo dispatch identification image sequence into the cargo label feature identifier respectively, and identifying and obtaining a plurality of basic printing quality parameters and a plurality of basic label pasting quality parameters; The quality fluctuations of the multiple basic printing quality parameters and the multiple basic label pasting quality parameters are calculated, and the fluctuation compensation is performed on the mean to obtain the printing quality parameters and the label pasting quality parameters.

[0042] In the embodiment of the present application, based on the cargo dispatch label recognition records in the historical time, the cargo dispatch image data generated during the recognition process is extracted to form a sample cargo dispatch recognition image set. The set contains images of cargo when it is being transported on the transport equipment, including label images, which is the basis for analyzing label quality.

[0043] In the sample cargo dispatch identification image set, the label image in each sample cargo dispatch identification image is compared and analyzed with the standard printed label image generated in the pre-printing device, and the deviation range is marked to generate a sample printing quality parameter set.

[0044] Among them, the deviation amplitude within each sample printing quality parameter includes the graphic size check of the label image in the sample cargo dispatch identification image and the standard printed label image, such as the sum of the deviation ratios of the length and width of multiple barcode lines in the barcode. For example, the width of a barcode line in the standard printed label image is 1mm, and the width of the barcode line in the label image printed in the actual sample cargo dispatch identification image is 1.1mm, then the deviation ratio is 0.1. In this way, the sample printing quality parameter set is obtained by annotation. Among them, the width and length of the barcode lines of the barcode are generally fixed, while the errors caused by printing are relatively random and there are many kinds, so the length and width errors of the barcode lines can be obtained by identifying the cargo dispatch identification image to obtain the printing quality parameters.

[0045] Furthermore, the label pasting position on the goods in the sample cargo dispatch recognition image is further analyzed to further analyze the label position offset. By comparing with the standard label position, the label movement amplitude is marked to generate a sample label pasting quality parameter set.

[0046] For example, to obtain the pasting position of the label on the goods in the sample cargo scheduling recognition image, the label position coordinates of the label in the sample cargo scheduling recognition image can be identified according to the camera coordinate system of the camera that collects the cargo scheduling recognition image, and then the standard label position coordinates of the standard label pasted on the goods are obtained, and then the distance between the actual label position coordinates and the standard label position coordinates is calculated as the movement amplitude, that is, the position deviation amplitude of the pasted label, which is marked as the sample label pasting quality parameter set, which corresponds to the sample cargo scheduling recognition image set.

[0047] For example, if the standard label position coordinates are (5,5), and the label position coordinates of the center of the label in the actual sample cargo scheduling recognition image are (5,4), then the label movement amplitude is 1, which will be marked as the sample label pasting quality parameter set.

[0048] Furthermore, a set of sample cargo scheduling recognition images is used as input features, and a set of sample printing quality parameters and a set of sample label pasting quality parameters are used as output features respectively, and a convolutional neural network is used to train a cargo label feature identifier that recognizes label printing quality parameters and label pasting quality parameters in cargo scheduling recognition images.

[0049] Convolutional neural network (CNN) can extract edge features, texture features and shape features of sample cargo scheduling recognition images through multi-layer convolution operations, and automatically learn the nonlinear logical relationship between sample printing quality parameters and sample label pasting quality parameters and sample cargo scheduling recognition images. Specifically, a convolutional neural network is used to construct a cargo label feature identifier, which includes an input layer, a convolution layer, a pooling layer, a fully connected layer and an output layer. The input layer sets the size of the input image, for example, the size of the sample cargo scheduling recognition image is 224×224. The convolution layer designs multi-layer convolution operations to extract local features of the image, such as edges, textures, etc. For example, the convolution layer can use 32 3×3 convolution kernels to extract features. The pooling layer reduces the feature dimension of the sample cargo scheduling recognition image, extracts features of different scales by convolution layer by layer, and the fully connected layer classifies the extracted features to obtain printing quality parameters and label pasting quality parameters. The loss function of the cargo label feature identifier can adopt the cross entropy loss function. During the training process, the sample cargo scheduling recognition image is input to obtain the output printing quality parameters and label pasting quality parameters, and then the loss with the real sample printing quality parameters and sample label pasting quality parameters is calculated. Then, the network parameters such as weights are adjusted, and multiple rounds of iterative training with multiple sets of training data are performed to improve the accuracy of the cargo label feature identifier until the requirements are met, such as the prediction error of the printing quality parameters is less than ±5%, and the prediction error of the label pasting quality parameters is less than ±10%. After the training is completed, verification and testing can also be performed. If the accuracy still meets the requirements, the training is completed. If not, iterative training continues.

[0050] After the cargo label feature identifier is trained, multiple cargo scheduling recognition images in the past cargo scheduling recognition image sequence are respectively input into the cargo label feature identifier for recognition, and multiple basic printing quality parameters and multiple basic label pasting quality parameters are respectively recognized and output.

[0051] Furthermore, the quality fluctuations of multiple basic printing quality parameters and multiple basic label pasting quality parameters are calculated, and the quality fluctuations reflect the possible fluctuation ranges of multiple basic printing quality parameters and multiple basic label pasting quality parameters. Exemplarily, two groups of several basic printing quality parameters are randomly selected as several first basic printing quality parameters and several second basic printing quality parameters, and then the difference between the mean of several first basic printing quality parameters and the mean of several second basic printing quality parameters is calculated, and the ratio of the absolute value of the difference to the mean of several first basic printing quality parameters is used as the printing quality fluctuation range. And using the same method, the pasting quality fluctuation range of multiple basic label pasting quality parameters is calculated, and the average of the printing quality fluctuation range and the pasting quality fluctuation range is calculated as the quality fluctuation.

[0052] Furthermore, the greater the quality fluctuation, the greater the fluctuation range of the actual printing quality parameters and label pasting quality parameters may be. Therefore, by compensating and adjusting the predicted multiple basic printing quality parameters and multiple basic label pasting quality parameters, more conservative printing quality parameters and label pasting quality parameters with poorer printing quality and label pasting quality can be obtained, thereby improving the recognition success rate in subsequent efficiency optimization.

[0053] Exemplarily, the means of multiple basic printing quality parameters and multiple basic label pasting quality parameters are calculated, and the two means are compensated for fluctuations using the sum of 1 and the quality fluctuation degree. Specifically, the two means are directly multiplied to obtain printing quality parameters and label pasting quality parameters. In this way, the means of multiple basic printing quality parameters and multiple basic label pasting quality parameters are magnified to obtain printing quality parameters and label pasting quality parameters with poor printing quality and poor pasting quality.

[0054] Through the above steps, the quality of label printing and pasting in the recent period can be identified, providing a data basis for the impact of label printing and printing quality on the recognition success rate, thereby obtaining a more accurate label recognition success rate, and conducting quality fluctuation analysis and compensation, magnifying the possible fluctuations that may cause a decline in label quality and recognition success rate, and ensuring the quality of warehouse receipt information recognition management as much as possible while optimizing efficiency.

[0055] In the embodiment of the present application, after the print quality parameters and label pasting quality parameters of the goods labels within the past preset time period are identified, the print quality parameters and label pasting quality parameters will affect the label image acquisition quality, and thus affect the label information recognition success rate. Therefore, the recognition influence degree is predicted based on the print quality parameters and label pasting quality parameters, and the recognition influence coefficient affecting the recognition success rate is obtained. The first recognition success rate initially predicted in the above content is corrected and calculated to obtain the second recognition success rate.

[0056] Step S3 of the method provided in the embodiment of the present application further includes: According to the cargo dispatch identification data in the historical time, a sample printing quality parameter set and a sample label pasting quality parameter set are collected; Collect the success rate change ratio of cargo label recognition under different sample printing quality parameters or sample label pasting quality parameters, and obtain a set of sample printing recognition influence coefficients and a set of sample pasting recognition influence coefficients; Constructing a mapping table between the sample printing quality parameter set and the sample printing identification influence coefficient set, and constructing a mapping table between the sample label pasting quality parameter set and the sample pasting identification influence coefficient set; Input the printing quality parameter and the label pasting quality parameter into the classification, obtain the printing recognition influence coefficient and the pasting recognition influence coefficient, and calculate the recognition influence coefficient; The first recognition success rate is corrected and calculated using the recognition influence coefficient to obtain a second recognition success rate.

[0057] In the embodiment of the present application, based on the cargo dispatch identification data in the historical time, specifically based on the previous label identification data, the printing quality parameters and label pasting quality parameters of other labels in the previous time are collected to obtain a sample printing quality parameter set and a sample label pasting quality parameter set. For example, the sample printing quality parameter can be 5%, 9%, 15%, etc., and the sample label pasting quality parameter can be 1, 2, 4, etc.

[0058] Furthermore, the change ratio of the cargo label recognition success rate under different printing quality parameters or pasting quality parameters is collected. For example, the test is firstly conducted to collect the cargo label recognition success rate under the conditions of the highest printing quality and the highest pasting quality, that is, the cargo label recognition success rate under the conditions of meeting the standard requirements, the printing quality parameters and the pasting quality parameters being both 0, for example, 100%.

[0059] Then, the success rate of cargo label recognition under different sample printing quality parameters or sample label pasting quality parameters is collected. For example, when the sample printing quality parameter is 5%, the success rate of cargo label recognition is 99%, and the change ratio of cargo label recognition success rate is 1%. The change ratio is marked as the sample printing recognition influence coefficient, and thus, the sample printing recognition influence coefficient set is obtained. And, for example, when the sample label pasting quality parameter is 2, the success rate of cargo label recognition is 99%, and the change ratio of cargo label recognition success rate is 1%. The change ratio is marked as the sample pasting recognition influence coefficient, and the sample pasting recognition influence coefficient set is obtained.

[0060] Furthermore, a mapping relationship between the sample printing quality parameter set and the sample printing identification influence coefficient set is constructed to obtain a mapping table. And a mapping relationship between the sample label pasting quality parameter set and the sample pasting identification influence coefficient set is constructed to obtain a mapping table.

[0061] Then, the current printing quality parameters and label pasting quality parameters are respectively input into two mapping tables for mapping classification to obtain the sample printing recognition influence coefficient and sample pasting recognition influence coefficient corresponding to the same or closest sample printing quality parameters and sample label pasting quality parameters, which are used as the printing recognition influence coefficient and pasting recognition influence coefficient of the current label printing and pasting quality affecting the label recognition success rate.

[0062] The sum of the printing recognition influence coefficient and the pasting recognition influence coefficient is calculated as the recognition influence coefficient, so as to obtain the sum of the influence of the two dimensions of label printing and pasting quality on the label recognition success rate.

[0063] The recognition influence coefficient is used to modify the first recognition success rate based on speed prediction in the above content, for example, the difference of 1-recognition influence coefficient is multiplied by the first recognition success rate to obtain the second recognition success rate. The greater the influence of label printing quality and pasting quality on recognition, the smaller the second recognition success rate.

[0064] In this way, the impact of label printing quality and pasting quality on label recognition success rate is combined, and the recognition success rate is corrected to improve the accuracy of recognition success rate calculation, thereby ensuring the quality and stability of warehouse receipt information recognition management in subsequent efficiency optimization.

[0065] S4: Calculate the efficiency improvement score according to the second recognition success rate, the cargo dispatch quantity and the cargo dispatch recognition speed, optimize the cargo dispatch recognition speed, obtain the optimal cargo dispatch recognition speed, and perform warehouse receipt management.

[0066] In the embodiment of the present application, in order to optimize the efficiency of warehouse warehouse receipt information entry management, the randomly configured cargo dispatch recognition speed is evaluated, and the efficiency improvement score of the cargo dispatch recognition speed is calculated based on the current second recognition success rate, cargo dispatch quantity and cargo dispatch recognition speed. The cargo dispatch recognition speed is evaluated from the two dimensions of efficiency and information recognition quality, and then as the evaluation basis, the cargo dispatch recognition speed is optimized to obtain the optimal cargo dispatch recognition speed, that is, the cargo dispatch recognition speed with the largest efficiency improvement score. Carrying out cargo dispatch and warehouse receipt information recognition management according to the optimal cargo dispatch recognition speed can effectively improve efficiency and ensure the success rate of information recognition.

[0067] Step S4 in the method provided in the embodiment of the present application includes: Calculating an efficiency improvement score according to the second recognition success rate, the cargo dispatch quantity, and the cargo dispatch recognition speed; Continue to randomly configure the cargo dispatching recognition speed, perform iterative optimization of the cargo dispatching recognition speed, retain the optimal cargo dispatching recognition speed with the highest efficiency improvement score, and perform warehouse receipt management.

[0068] In the embodiment of the present application, the efficiency improvement score is calculated according to the second recognition success rate, the cargo dispatch quantity and the cargo dispatch recognition speed, as shown in the following formula: ; in, It is an efficiency improvement score, which reflects the pros and cons of the current efficiency improvement and cargo dispatch identification speed. The larger the efficiency improvement score, the better the efficiency improvement effect. is the second recognition success rate, is the current cargo dispatch quantity, The preset cargo dispatch quantity, such as the average cargo dispatch quantity per day in the warehouse in the past month. Identify speed for cargo dispatch, The preset cargo dispatch recognition speed is, for example, the average cargo dispatch recognition speed configured by the warehouse in the past month. Be a small positive number to avoid denominator being 0.

[0069] Second, the size of the recognition success rate, the size of the cargo dispatch quantity, and the cargo dispatch recognition speed are positively correlated with the size of the efficiency improvement score. The larger the three are, the greater the efficiency improvement is, the more stable the recognition success rate is, and the better the randomly configured cargo dispatch recognition speed is.

[0070] In this way, the cargo scheduling recognition speed is optimized, the cargo scheduling recognition speed continues to be randomly configured, and the efficiency improvement scores of other randomly configured cargo scheduling recognition speeds are calculated based on the steps in the above content, and optimization is performed until convergence. For example, when the randomly configured cargo scheduling recognition speeds reach a preset number, such as 100, the optimization is stopped, and the optimal cargo scheduling recognition speed with the largest efficiency improvement score is output to perform cargo in and out warehousing scheduling and warehouse receipt information identification management.

[0071] Optionally, an optimization algorithm such as a genetic algorithm may be used to optimize the cargo scheduling identification speed.

[0072] An efficiency-optimized IoT warehouse receipt management method provided by an embodiment of the present invention has at least the following technical effects: The embodiment of the present invention predicts the recognition success rate of the configured cargo scheduling recognition speed by analyzing historical cargo scheduling data and real-time images, and performs dynamic correction based on the fluctuation influence coefficient of printing quality and pasting quality. Specifically, by combining the cargo scheduling recognition image sequence and the printing quality and pasting quality parameter analysis of the cargo label, the recognition success rate and scheduling efficiency in the warehouse receipt management process are significantly improved, and a more accurate second recognition success rate is obtained. Secondly, by randomly configuring and optimizing the cargo scheduling recognition speed, a comprehensive calculation and evaluation is performed in combination with the cargo scheduling quantity and the recognition success rate, ensuring that the scheduling speed of the outbound or inbound process is optimized while improving the recognition efficiency, thereby maximizing the overall efficiency. Compared with the prior art, the present invention can ensure the label recognition success rate on the basis of maximizing the cargo outbound and inbound scheduling speed, and considers the impact of the recognition success rate caused by the fluctuation of the label quality or the change of the label pasting, thereby improving the efficiency, intelligence level and operation stability of the warehouse receipt information recognition management and warehousing management.

[0073] Embodiment 2, as Figure 2 As shown, based on the same inventive concept of an efficiency-optimized IoT warehouse receipt management method provided in Embodiment 1, an embodiment of the present invention further provides an efficiency-optimized IoT warehouse receipt management system, including: The cargo dispatch data acquisition module 11 is used to obtain the current cargo dispatch quantity of the IoT warehouse and obtain the cargo dispatch identification image sequence within a preset time period in the past; The recognition success rate prediction module 12 is used to randomly configure the cargo scheduling recognition speed, perform recognition success rate prediction, and obtain a first recognition success rate; The recognition success rate correction module 13 is used to perform printing quality recognition and label pasting quality recognition of the cargo label according to the cargo dispatch recognition image sequence, obtain printing quality parameters and label pasting quality parameters, perform recognition influence prediction, obtain recognition influence coefficient, perform correction calculation on the first recognition success rate, and obtain a second recognition success rate; The warehouse receipt management optimization module 14 is used to calculate the efficiency improvement score according to the second recognition success rate, the cargo dispatch quantity and the cargo dispatch recognition speed, and optimize the cargo dispatch recognition speed to obtain the optimal cargo dispatch recognition speed and perform warehouse receipt management.

[0074] Furthermore, the cargo dispatch data acquisition module 11 is also used to: obtain the cargo dispatch quantity currently to be dispatched by the IoT warehouse; Retrieving cargo dispatch identification images within a preset time period in the past to obtain a historical cargo dispatch identification image set; In the historical cargo dispatch recognition image set, the recognition images when the cargo labels are recognized are extracted to obtain a cargo dispatch recognition image sequence.

[0075] Furthermore, the recognition success rate prediction module 12 is also used to: obtain the transmission speed interval of the transmission equipment for cargo scheduling recognition, and randomly generate a transmission speed as the cargo scheduling recognition speed; According to the cargo dispatch identification data of different batches of cargo, a plurality of different cargo identification prediction paths are constructed to obtain a cargo identification predictor; The cargo scheduling recognition speed is input into a plurality of cargo recognition prediction paths in the cargo recognition predictor, a plurality of recognition success rates are predicted, and a first recognition success rate is calculated.

[0076] Among them, according to the cargo scheduling identification data of different batches of cargo, different multiple cargo identification prediction paths are constructed to obtain a cargo identification predictor, including: Randomly select multiple batches of goods in the historical time and collect multiple sample goods scheduling recognition speed sets as input features; The success rates of cargo label recognition under different sample cargo scheduling recognition speeds are collected and marked as multiple sample success rate sets as output features; The plurality of sample cargo scheduling recognition speed sets and the plurality of sample success rate sets are respectively used to construct a plurality of cargo recognition prediction paths to obtain a cargo recognition predictor.

[0077] Furthermore, the recognition success rate correction module 13 is also used to: collect a set of sample cargo dispatch recognition images according to the cargo dispatch label recognition records in the historical time, mark the deviation between the label image in each sample cargo dispatch recognition image and the standard printed label image, and obtain a set of sample printing quality parameters; Mark the movement range between the label position in the label image and the standard label position in each sample cargo dispatch recognition image to obtain a set of sample label pasting quality parameters; The sample cargo dispatch identification image set is used as input features, the sample printing quality parameter set and the sample label pasting quality parameter set are used as output features, and a convolutional neural network is used to train a cargo label feature identifier; Inputting the cargo dispatch identification images in the cargo dispatch identification image sequence into the cargo label feature identifier respectively, and identifying and obtaining a plurality of basic printing quality parameters and a plurality of basic label pasting quality parameters; The quality fluctuations of the multiple basic printing quality parameters and the multiple basic label pasting quality parameters are calculated, and the fluctuation compensation is performed on the mean to obtain the printing quality parameters and the label pasting quality parameters.

[0078] Furthermore, the recognition success rate correction module 13 is also used for: According to the cargo dispatch identification data in the historical time, a sample printing quality parameter set and a sample label pasting quality parameter set are collected; Collect the success rate change ratio of cargo label recognition under different sample printing quality parameters or sample label pasting quality parameters, and obtain a set of sample printing recognition influence coefficients and a set of sample pasting recognition influence coefficients; Constructing a mapping table between the sample printing quality parameter set and the sample printing identification influence coefficient set, and constructing a mapping table between the sample label pasting quality parameter set and the sample pasting identification influence coefficient set; Input the printing quality parameter and the label pasting quality parameter into the classification, obtain the printing recognition influence coefficient and the pasting recognition influence coefficient, and calculate the recognition influence coefficient; The first recognition success rate is corrected and calculated using the recognition influence coefficient to obtain a second recognition success rate.

[0079] Furthermore, the warehouse receipt management optimization module 14 is further used to calculate the efficiency improvement score according to the second recognition success rate, the cargo dispatch quantity and the cargo dispatch recognition speed, as shown in the following formula: ; in, Score efficiency improvements. is the second recognition success rate, is the cargo dispatch quantity, To preset the quantity of goods dispatched, Identify speed for cargo dispatch, To preset the cargo dispatch identification speed, is a small positive number; Continue to randomly configure the cargo dispatching recognition speed, perform iterative optimization of the cargo dispatching recognition speed, retain the optimal cargo dispatching recognition speed with the highest efficiency improvement score, and perform warehouse receipt management.

[0080] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0081] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. An efficiency-optimized IoT warehouse receipt management method, characterized in that: The method comprises: Obtain the current cargo dispatch quantity of the IoT warehouse and obtain the cargo dispatch identification image sequence within the past preset time period; Randomly configure the cargo scheduling recognition speed, predict the recognition success rate, and obtain the first recognition success rate; According to the cargo dispatch identification image sequence, the printing quality identification and label pasting quality identification of the cargo label are performed to obtain the printing quality parameters and the label pasting quality parameters, and the identification influence prediction is performed to obtain the identification influence coefficient, and the first identification success rate is corrected and calculated to obtain the second identification success rate; According to the second recognition success rate, the cargo dispatch quantity and the cargo dispatch recognition speed, the efficiency improvement score is calculated, and the cargo dispatch recognition speed is optimized to obtain the optimal cargo dispatch recognition speed for warehouse receipt management.

2. The efficiency-optimized IoT warehouse receipt management method according to claim 1, characterized in that: Get the current cargo dispatch quantity of the IoT warehouse and obtain cargo dispatch identification images within the past preset time, including: Obtain the number of goods currently to be dispatched in the IoT warehouse; Retrieving cargo dispatch identification images within a preset time period in the past to obtain a historical cargo dispatch identification image set; In the historical cargo dispatch recognition image set, the recognition images when the cargo labels are recognized are extracted to obtain a cargo dispatch recognition image sequence.

3. The efficiency-optimized IoT warehouse receipt management method according to claim 1 is characterized in that: Randomly configure the cargo scheduling recognition speed, predict the recognition success rate, and obtain the first recognition success rate, including: Obtain the transmission speed range of the transmission equipment for cargo scheduling identification, and randomly generate a transmission speed as the cargo scheduling identification speed; According to the cargo dispatch identification data of different batches of cargo, a plurality of different cargo identification prediction paths are constructed to obtain a cargo identification predictor; The cargo scheduling recognition speed is input into a plurality of cargo recognition prediction paths in the cargo recognition predictor, a plurality of recognition success rates are predicted, and a first recognition success rate is calculated.

4. The efficiency-optimized IoT warehouse receipt management method according to claim 3 is characterized in that: According to the cargo scheduling identification data of different batches of cargo, different multiple cargo identification prediction paths are constructed to obtain a cargo identification predictor, including: Randomly select multiple batches of goods in the historical time and collect multiple sample goods scheduling recognition speed sets as input features; The success rates of cargo label recognition under different sample cargo scheduling recognition speeds are collected and marked as multiple sample success rate sets as output features; The plurality of sample cargo scheduling recognition speed sets and the plurality of sample success rate sets are respectively used to construct a plurality of cargo recognition prediction paths to obtain a cargo recognition predictor.

5. The efficiency-optimized IoT warehouse receipt management method according to claim 1, characterized in that: According to the cargo dispatch identification image sequence, the printing quality identification and label pasting quality identification of the cargo label are performed to obtain the printing quality parameters and label pasting quality parameters, including: According to the cargo dispatch label recognition records in the historical time, a set of sample cargo dispatch recognition images is collected, and the deviation between the label image in each sample cargo dispatch recognition image and the standard printed label image is marked to obtain a set of sample printing quality parameters; Mark the movement range between the label position in the label image and the standard label position in each sample cargo dispatch recognition image to obtain a set of sample label pasting quality parameters; The sample cargo dispatch identification image set is used as input features, the sample printing quality parameter set and the sample label pasting quality parameter set are used as output features, and a convolutional neural network is used to train a cargo label feature identifier; Inputting the cargo dispatch identification images in the cargo dispatch identification image sequence into the cargo label feature identifier respectively, identifying and obtaining a plurality of basic printing quality parameters and a plurality of basic label pasting quality parameters; The quality fluctuations of the multiple basic printing quality parameters and the multiple basic label pasting quality parameters are calculated, and the fluctuation compensation is performed on the mean to obtain the printing quality parameters and the label pasting quality parameters.

6. The efficiency-optimized IoT warehouse receipt management method according to claim 5 is characterized in that: Performing recognition influence prediction to obtain a recognition influence coefficient, and performing correction calculation on the first recognition success rate to obtain a second recognition success rate, including: According to the cargo dispatch identification data in the historical time, a sample printing quality parameter set and a sample label pasting quality parameter set are collected; Collect the success rate change ratio of cargo label recognition under different sample printing quality parameters or sample label pasting quality parameters, and obtain a set of sample printing recognition influence coefficients and a set of sample pasting recognition influence coefficients; Constructing a mapping table between the sample printing quality parameter set and the sample printing identification influence coefficient set, and constructing a mapping table between the sample label pasting quality parameter set and the sample pasting identification influence coefficient set; Input the printing quality parameter and the label pasting quality parameter into the classification, obtain the printing recognition influence coefficient and the pasting recognition influence coefficient, and calculate the recognition influence coefficient; The first recognition success rate is corrected and calculated using the recognition influence coefficient to obtain a second recognition success rate.

7. The efficiency-optimized IoT warehouse receipt management method according to claim 1, characterized in that: According to the second recognition success rate, the cargo dispatch quantity and the cargo dispatch recognition speed, the efficiency improvement score is calculated, and the cargo dispatch recognition speed is optimized to obtain the optimal cargo dispatch recognition speed, including: According to the second recognition success rate, the number of cargo dispatches, and the cargo dispatch recognition speed, the efficiency improvement score is calculated as follows: ； in, Score efficiency improvements. is the second recognition success rate, is the cargo dispatch quantity, To preset the quantity of goods dispatched, Identify speed for cargo dispatch, To preset the cargo dispatch identification speed, is a small positive number; Continue to randomly configure the cargo dispatching recognition speed, perform iterative optimization of the cargo dispatching recognition speed, retain the optimal cargo dispatching recognition speed with the highest efficiency improvement score, and perform warehouse receipt management.

8. An efficiency-optimized IoT warehouse receipt management system, characterized in that: The steps for implementing the efficiency-optimized IoT warehouse receipt management method according to any one of claims 1 to 7 include: The cargo dispatch data acquisition module is used to obtain the current cargo dispatch quantity of the IoT warehouse and obtain the cargo dispatch identification image sequence within a preset time period in the past; The recognition success rate prediction module is used to randomly configure the cargo scheduling recognition speed, predict the recognition success rate, and obtain the first recognition success rate; a recognition success rate correction module, configured to perform printing quality recognition and label pasting quality recognition of the cargo label according to the cargo dispatch recognition image sequence, obtain printing quality parameters and label pasting quality parameters, perform recognition influence prediction, obtain recognition influence coefficient, perform correction calculation on the first recognition success rate, and obtain a second recognition success rate; The warehouse receipt management optimization module is used to calculate the efficiency improvement score according to the second recognition success rate, the cargo dispatch quantity and the cargo dispatch recognition speed, and optimize the cargo dispatch recognition speed to obtain the optimal cargo dispatch recognition speed and perform warehouse receipt management.

Images