An article storage management system based on the Internet of Things

Abstract

The application relates to the technical field of data processing, in particular to an article storage management system based on the Internet of Things, which comprises a data acquisition module, an authentication storage module, a storage space allocation module and a storage environment sensing module. The article storage management system collects storage information and environment information through the data acquisition module, so that the system can manage the storage cabinet in the future. The article storage management system also authenticates the user identity through the authentication storage module, realizes high-precision recognition according to the dual authentication of the user's iris and palm print, adjusts the dual authentication through the environment information, realizes high-precision recognition, intelligently allocates the storage space of the storage cabinet through the storage space allocation module, and corrects the allocation of the storage cabinet space through the storage environment sensing module, so that the unreasonable use of the storage space is avoided, the user's stored articles can obtain a suitable storage environment, and the reasonable allocation of the storage space is realized.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an Internet of Things (IoT)-based item storage management system. Background Technology

[0002] Traditional baggage storage management systems mostly use barcode identification. While barcode identification is relatively convenient, barcodes are easily damaged or lost. Once a barcode is damaged, it becomes impossible to accurately identify the owner, affecting the normal retrieval of items. Moreover, these traditional methods lack effective mechanisms to prevent unauthorized use of another person's storage tag to retrieve items, resulting in low security.

[0003] For example, Chinese Patent Publication No. CN112184392A discloses a method for reporting the storage space status of a smart locker, including the following steps: After receiving an operation request from a mobile terminal, the smart locker opens the door of the designated storage space according to the server's instructions and reports the status of that storage space to the server, so that the server synchronizes the status of the storage space on the server with the status of the storage space in the smart locker; the mobile terminal periodically queries the server to obtain the status of the storage space on the server; when the status of the storage space on the server does not match the status of the storage space on the smart locker, the mobile terminal reports an anomaly in the status of that storage space to the server. This invention also relates to an apparatus for implementing the above method. The method and apparatus for reporting the storage space status of a smart locker according to this patent have the following beneficial effects: they do not affect the use of the smart locker and increase the security of the items inside the locker. However, this patent cannot be applied to storage locations with high security requirements. When storing important items, it is difficult to avoid the loss of important items due to accidental operation by others or inaccurate identification of the occupant by the storage system. In addition, the patent has a relatively simple classification of storage for important items and does not dynamically adjust the storage space according to factors such as the size and weight of the items, making it difficult to meet the storage needs of important items. Summary of the Invention

[0004] To address this, the present invention provides an Internet of Things-based item storage management system to overcome the problems of low security for valuable items and mismatched storage space for valuable items caused by low identification accuracy and unreasonable allocation of storage space in existing technologies.

[0005] To achieve the above objectives, the present invention provides an Internet of Things (IoT)-based item storage management system, the system comprising: The data acquisition module is used to collect stored information and environmental information; The authentication and storage module is used to obtain the identity similarity of user feature vectors based on the user's iris and palm print in the storage information, and to grant storage permissions to users based on the identity similarity of user feature vectors. It is also used to adjust the process of obtaining the identity similarity of user feature vectors based on the lighting conditions, and to judge the security requirements based on the security requirements of the storage cabinet, and to optimize the adjustment process of obtaining the identity similarity of user feature vectors based on the judgment results. The storage space allocation module is used to intelligently allocate storage space according to a dynamic programming algorithm when granting storage permissions to users, and also to calibrate the judgment process of security requirements based on the importance index of the stored items. The storage environment perception module is used to judge the storage cabinet space allocation based on the space utilization rate in the storage information, and optimize the dynamic programming algorithm based on the judgment result. It is also used to judge the abnormal situation of the storage cabinet based on the number of abnormal current situations, and correct the judgment process of storage cabinet space allocation based on the judgment result.

[0006] Furthermore, the authentication registration module uses the iris feature vector extraction method to obtain the iris feature vector H of the user's iris in the registered information, and uses the palmprint feature vector extraction method to obtain the palmprint feature vector Y of the user's palm in the registered information, and calculates the comprehensive user feature vector K based on the iris feature vector H and the palmprint feature vector Y.

[0007] Furthermore, the authentication storage module uses the Hamming distance method to obtain the identity similarity A between the comprehensive user feature vector and the comprehensive user feature vector in the storage database. It compares the identity similarity A of the comprehensive user feature vector with the preset identity similarity A0, where 0.83≤A0≤0.91. Based on the comparison result, it judges the identity similarity and grants storage permissions to the user based on the judgment result.

[0008] Furthermore, the authentication register module calculates the illumination condition Qt based on the light intensity value Qg and the light angle value Qj in the environmental information, compares the illumination condition Qt with the preset illumination condition Qt0, where 0.73≤Qt0≤0.93, judges the illumination situation based on the comparison result, and adjusts the calculation process of the comprehensive user feature vector based on the judgment result.

[0009] Furthermore, the authentication storage module inputs the pedestrian flow and environmental structure diagram from the environmental information into the expert judgment model, outputs the storage cabinet security requirement value F, compares the storage cabinet security requirement value with the preset requirement value F0, 0.67≤F0≤0.88, and judges the security requirement based on the comparison result, and optimizes the adjustment process of the comprehensive user feature vector calculation process based on the judgment result.

[0010] Furthermore, the register space allocation module uses a dynamic programming algorithm to intelligently allocate register space based on register information.

[0011] Furthermore, the storage space allocation module inputs the image of the stored object from the storage information into the stored object recognition model, outputs the importance index Jz of the stored object, compares the importance index Jz of the stored object with the preset importance index Jz0 of the stored object, 0.42≤Jz0≤0.56, and judges the importance of the stored object based on the comparison result, and calibrates the judgment process of security requirements based on the judgment result.

[0012] Furthermore, the storage space sensing module compares the space utilization rate Ly in the storage information with the preset storage space utilization rate Ly0, where 70%≤Ly0≤88%, and judges the utilization of the storage space based on the comparison result, and records the utilization of the storage cabinet based on the judgment result.

[0013] Furthermore, the storage space sensing module also compares the number of times the storage cabinet has a low utilization rate Lc in the historical utilization data with the preset number of times the low utilization rate Lc0, where 3 times / day ≤ Lc0 ≤ 5 times / day. Based on the comparison results, it judges the storage cabinet space allocation and optimizes the dynamic programming algorithm based on the judgment results.

[0014] Furthermore, the storage space sensing module compares the storage cabinet switching current Ik in the storage information with the preset minimum storage cabinet switching current Ikmin and the preset maximum storage cabinet switching current Ikmax, where Ikmin = 600mA and Ikmax = 800mA. Based on the comparison result, the module judges the storage cabinet current status and records it. The storage space sensing module also compares the number of abnormal storage cabinet current occurrences Ck with the preset number of abnormal storage cabinet occurrences Ck0, where 3 occurrences ≤ Ck0 ≤ 5 occurrences. Based on the comparison result, the module judges the abnormality of the storage cabinet and corrects the judgment process of storage cabinet space allocation based on the judgment result.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: the system is applied to an airport luggage storage management terminal. The system collects luggage and environmental information through a data acquisition module to facilitate subsequent management of the luggage lockers. The system also authenticates user identities through an authentication module, achieving high-precision recognition based on dual authentication of the user's iris and palm print. Furthermore, it adjusts the weighting of iris and palm print recognition based on environmental information such as light intensity, light angle, pedestrian flow, and environmental structure diagrams, adapting to accurate recognition in various environments and situations. Moreover, when dealing with important items, the weighting of iris recognition is increased to improve the accuracy of user identification, thereby ensuring the safe storage of important items. To ensure the security of stored items and enhance the safety of important items, the system intelligently allocates storage space in the lockers through a storage space allocation module, improving the rationality of storage space allocation. Simultaneously, by identifying the importance index of stored items, the system further calibrates the security requirements of stored items, ensuring the safety of valuable items and increasing storage security. The system also uses a storage environment sensing module to judge the storage locker space allocation, avoiding unreasonable use of storage space and achieving reasonable allocation of storage space. This ensures that users' stored items receive a suitable storage environment, preventing damage due to insufficient space and further improving the rationality of storage space allocation. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the IoT-based item storage management system in this embodiment. Detailed Implementation

[0017] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0018] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0019] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0020] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0021] Please see Figure 1 As shown, this is a schematic diagram of the structure of the IoT-based item storage management system in this embodiment. The system includes: The data acquisition module is used to collect stored information and environmental information; The authentication and storage module is used to obtain the identity similarity of user feature vectors based on the user's iris and palm print in the stored information, and to grant storage permissions to the user based on the identity similarity of the user feature vectors. It is also used to adjust the process of obtaining the identity similarity of user feature vectors based on the lighting conditions, and to judge the security requirements based on the security requirements of the storage cabinet, and to optimize the adjustment process of obtaining the identity similarity of user feature vectors based on the judgment results. The authentication and storage module is connected to the data acquisition module. The storage space allocation module is used to intelligently allocate storage space according to a dynamic programming algorithm when granting storage permissions to users. It also calibrates the judgment process of security requirements based on the importance index of stored items. The storage space allocation module is connected to the authentication storage module. The storage environment sensing module is used to judge the storage cabinet space allocation based on the space utilization rate in the storage information, and optimize the dynamic programming algorithm based on the judgment result. It is also used to judge the abnormality of the storage cabinet based on the number of abnormal current occurrences, and correct the judgment process of storage cabinet space allocation based on the judgment result. The storage environment sensing module is connected to the storage space allocation module.

[0022] Specifically, the system is applied to storage environment management terminals with high security requirements, such as jewelry storage management terminals, auction item storage management terminals, and shared transaction storage management terminals. The system collects storage and environmental information through a data acquisition module to facilitate subsequent management of the storage lockers. It also authenticates user identities through an authentication module, achieving high-precision recognition based on dual authentication of the user's iris and palm print. Furthermore, it adjusts the weight of iris and palm print recognition based on environmental information such as light intensity, light angle, pedestrian flow, and environmental structure diagrams to adapt to accurate recognition in various environments and situations. When dealing with important items, the weight of iris recognition is increased to improve user identification accuracy. To ensure the security of stored important items, the system enhances the accuracy of its data allocation. It intelligently allocates storage space through a storage space allocation module, improving the rationality of space allocation. Furthermore, by identifying the importance index of stored items, the system further calibrates the security requirements of the stored items, guaranteeing the safety of valuable items and increasing security. The system also uses a storage environment sensing module to assess the storage space allocation, preventing unreasonable use of storage space and ensuring a suitable environment for users' stored items. This prevents damage due to limited space and further improves the rationality of storage space allocation.

[0023] Specifically, the data acquisition module collects storage information and environmental information through an iris recognition device, a near-infrared light sensor, a light intensity sensor, an angle sensor, an internal camera, a current sensor, and an external camera. The storage information includes the user's iris, palm print, images of stored items, space utilization, locker switch current, the required locker size for user i, the value of items allocated to user i, the total allocated space, the new total allocated space after allocation, the current combination status of different sized lockers in use, and the new locker status after allocation. The data acquisition module collects the user's iris through the iris recognition device, the user's palm print through the near-infrared light sensor, and the locker switch current through the current sensor. The system collects data on various aspects, including: the value of items allocated to user i, the required locker size for user i, the total allocated space, the newly allocated total space, the combination status of currently available lockers of different sizes, and the status of the new lockers after allocation. Environmental information includes light intensity, light angle, pedestrian flow, and environmental structure diagrams. The system also collects light intensity values ​​via a light intensity sensor, light angle via an angle sensor, and pedestrian flow and environmental structure diagrams via an external camera.

[0024] Specifically, the authentication registration module uses iris feature vector extraction to obtain the iris feature vector H of the user's iris in the registered information, and uses palmprint feature vector extraction to obtain the palmprint feature vector Y of the user's palm in the registered information. Based on the iris feature vector H and the palmprint feature vector Y, it calculates the comprehensive user feature vector K, and sets... , , These are the weighting coefficients of the iris feature vector. , The weight coefficients are the palmprint feature vectors.

[0025] Specifically, the iris feature vector extraction method refers to the feature vector transformation method for extracting the user's iris into feature vectors. This embodiment does not limit the specific implementation of the iris feature vector extraction method. Those skilled in the art can set it according to actual needs. For example, Hough transform can be used to find the inner and outer boundaries of the user's iris and determine the center of the user's iris, the radius rs of the inner boundary of the user's iris, and the radius re of the outer boundary of the user's iris. The palmprint feature vector extraction method refers to a method for converting a user's palmprint into a palmprint feature vector. The palmprint feature vector extraction method includes: Step B1: Convert the collected user palm print into grayscale image pixel values ​​(Gray), and set... R is the pixel value corresponding to the red channel of the color image, G is the pixel value corresponding to the green channel of the color image, and B is the pixel value corresponding to the blue channel of the color image. Step B2: Use median filtering to denoise the pixel values ​​of the grayscale image to obtain the denoised palmprint grayscale image values. The median filtering is a non-linear filtering method that sorts the pixel values ​​in the filtering window and takes the median value as the new value of the center pixel of the window. Step B3: Normalize the image values ​​of the palmprint grayscale image after noise reduction to obtain the normalized value Gg, and set... Where Graymin is the minimum gray value of the palmprint image, and Graymax is the minimum gray value of the palmprint image. Step B4: The normalized value Gg is vectorized using principal component analysis to obtain the palmprint feature vector Y. The principal component analysis method is a data analysis method mainly used for feature extraction of data. This embodiment does not limit the specific application of the principal component analysis method. Those skilled in the art can set it according to their needs, as long as the vectorization requirement of the normalized value is met. The user's iris refers to the biometric feature of the user's eyes, and the user's palm print refers to the lines on the surface of the user's palm, which are formed by the combined effects of genetic factors and various factors during embryonic development, and have individual specificity and stability.

[0026] Specifically, by combining the user's iris and palm print, the number of recognition dimensions and features is increased, which can more accurately distinguish different individuals, improve the accuracy of unlocking recognition, reduce the false recognition rate and the rejection rate, and when one feature is stolen or forged, the other feature is difficult to counterfeit at the same time, which can effectively prevent illegal unlocking, increase the security and reliability of the system, and better protect the user's property and privacy.

[0027] Specifically, the authentication storage module uses the Hamming distance method to obtain the identity similarity A between the comprehensive user feature vector and the comprehensive user feature vector in the storage database. It then compares the identity similarity A with a preset identity similarity A0, where 0.83 ≤ A0 ≤ 0.91. Based on the comparison result, it judges the identity similarity and grants storage permissions to the user according to the judgment result. When A < A0, the authentication storage module determines that the identity similarity is not similar and does not grant storage permissions to the user. When A≥A0, the authentication storage module determines that the identity similarity is similar, grants storage permission to the user, matches the user's identity to obtain the user type, and feeds back the user type to the storage space allocation module. The user type includes general users and member users.

[0028] Specifically, the Hamming distance method refers to an information inspection method that measures the degree of difference between two data. This embodiment does not limit the specific usage of the Hamming distance method. Those skilled in the art can set it according to their needs, as long as it meets the requirement of calculating the identity similarity of user feature vectors. The registered database refers to a database that records user information of all registered items. The registered database includes recorded historical comprehensive user feature vectors. The identity similarity refers to the similarity between the comprehensive user feature vector and the recorded historical comprehensive user feature vector in the registered database. The user identity refers to the user information corresponding to the user's iris and palm print. The preset identity similarity refers to a preset value used to judge identity similarity. The identity similarity refers to the similarity between the user feature vector and the feature vector in the database. The identity similarity includes dissimilar identity similarity and similar identity similarity.

[0029] Specifically, the authentication storage module can identify whether a user is a user of the storage locker by judging similar identity situations, and decide whether to open the storage locker based on the recognition result, thereby improving the security of the stored items.

[0030] Specifically, the authentication register module also marks the iris and palm print of the same user collected within the time period t, marking them as users who performed identity recognition within the same time period, and counts the number of users who performed identity recognition within the same time period to obtain the number of users who performed identity recognition simultaneously. This embodiment does not limit the specific value of the time period t. Those skilled in the art can set it according to actual needs, such as setting t=120 seconds.

[0031] Specifically, the authentication register module calculates the illumination condition Qt based on the light intensity value Qg and the light angle value Qj in the environmental information, setting Qt = 0.7 × Qg + 0.3 × Qj. It then compares the illumination condition Qt with the preset illumination condition Qt0, where 0.73 ≤ Qt0 ≤ 0.93. Based on the comparison result, it judges the illumination situation and adjusts the calculation process of the comprehensive user feature vector according to the judgment result. When Qt≥Qt0, the authentication register module determines that the lighting condition is qualified and does not adjust the calculation process of the comprehensive user feature vector; When Qt < Qt0, the authentication register module determines that the lighting condition is unqualified and adjusts the calculation process of the comprehensive user feature vector, including the weighting coefficients of the palmprint feature vector. Weight coefficients of iris feature vectors Adjustments are made to obtain the weight coefficients of the adjusted palmprint feature vector. and the weight coefficients of the adjusted iris feature vector ,set up The weight coefficients of the palmprint feature vector Replace with the weight coefficients of the adjusted palmprint feature vector The weighting coefficients of the iris feature vector Replace with the weight coefficients of the adjusted iris feature vector Then, the comprehensive user feature vector K is recalculated.

[0032] Specifically, the light intensity value refers to the ambient light intensity outside the storage locker. In this embodiment, the light intensity value is assigned according to the range of 0-1. The light angle value refers to the ambient light angle outside the storage locker. In this embodiment, the light angle value is assigned according to the range of 0-1. The light condition refers to the ambient light condition outside the storage locker. The light condition includes light condition being qualified and light condition being unqualified.

[0033] Specifically, the authentication register module adjusts the weight coefficient of the palmprint feature vector by judging the lighting conditions, increasing it and decreasing the weight coefficient of the iris feature vector, so as to avoid collecting too many iris feature vectors when the lighting conditions are not up to standard, which would lead to low user recognition accuracy of the system.

[0034] Specifically, the authentication storage module inputs the pedestrian flow and environmental structure diagram from the environmental information into the expert judgment model, outputs the storage locker security requirement value F, compares the storage locker security requirement value F with the preset requirement value F0, where 0.67≤F0≤0.88, and judges the security requirement based on the comparison result. Based on the judgment result, the adjustment process of the comprehensive user feature vector calculation process is optimized, wherein: When F≤F0, the authentication register module determines that the security requirement is low and does not optimize the adjustment process of the calculation process of the comprehensive user feature vector. When F > F0, the authentication register module determines that the security requirement is high, optimizes the adjustment process of the comprehensive user feature vector calculation, and uses a security coefficient... Weighting coefficients of the adjusted iris feature vector and the weight coefficients of the adjusted palmprint feature vector Optimization is performed to obtain the weight coefficients of the optimized iris feature vector. and the weight coefficients of the optimized palmprint feature vector ,set up The weight coefficients of the adjusted iris feature vector will be... Replace with the weight coefficients of the optimized iris feature vector The weight coefficients of the adjusted palmprint feature vector will be... Replace with the weight coefficients of the optimized palmprint feature vector Then, the comprehensive user feature vector K is recalculated.

[0035] Specifically, the expert judgment model refers to a deep learning model that takes pedestrian flow and environmental structure diagrams as input and outputs the security requirement value of the lockers. This embodiment constructs the expert judgment model using an expert judgment model construction method, which includes: Step C10: Compile historical pedestrian flow and environmental structure diagrams from the expert database, as well as the corresponding security requirement values ​​for the lockers; Step C20: Divide 70% of the data in the expert database into an expert judgment training set and 30% of the data in the expert database into an expert judgment validation set; Step C30: Select a multilayer perceptron as the expert judgment model, initialize the weights and biases of the expert judgment model, input the data from the expert judgment training set into the expert judgment model, and calculate the output of the expert judgment model. Step C40: Calculate the loss function value based on the output and label of the expert judgment model, calculate the gradient through the backpropagation algorithm, and update the weights and biases of the expert judgment model. Repeat the process of forward propagation, loss function calculation and backpropagation. Step C50: Test the accuracy of the expert judgment model using the expert judgment validation set, and output the expert judgment model with an accuracy of 90%. The "personnel flow" refers to the number of people moving around the lockers within a day. The "environmental structure diagram" refers to the environmental structure diagram around the lockers. The "locker security requirement value F" refers to the security requirement value of the lockers. The "preset requirement value" refers to the preset value used to judge the security requirement situation. The "security requirement situation" refers to the security requirement situation of the lockers according to the surrounding environment. The security requirement situation includes a low security requirement situation and a high security requirement situation.

[0036] Specifically, the authentication register module optimizes the weight coefficient of the adjusted iris feature vector by judging the security requirements, increasing the proportion of the iris feature vector in the comprehensive user feature vector. When the security requirements are judged to be high, more user iris data is collected, thereby improving the accuracy of user identification by the system.

[0037] Specifically, the register space allocation module uses a dynamic programming algorithm to intelligently allocate register space based on register information. The dynamic programming algorithm includes: Step S1: Set the allocation state dp[i][j][k], where i is the i-th user, i=1,2,3...m, m is the user order, j is the combination state of different sizes of currently available storage lockers, and k is the total allocated space; Step S2: Construct state transition equations based on the allocation, wherein each state transition equation includes a first state transition equation and a second state transition equation, wherein: When the allocation situation is that user i's item is not allocated to the locker, the storage space allocation module constructs the first state transition equation dp[i][j][k]1 based on the allocation state dp[i][j][k], and sets dp[i][j][k]1=dp[i-1][j][k]; When the allocation situation is to allocate user i's item to the locker, the storage space allocation module constructs a second state transition equation dp[i][j][k]2 based on the allocation state dp[i][j][k], setting dp[i][j][k]2=max(dp[i-1][j][k],dp[i-1][j′][k′]+vi), where vi is the value obtained by allocating the item to user i, si is the size of the locker required by user i, j′ is the new locker state after allocation, and k′ is the new total allocated space after allocation, setting k′=k+si.

[0038] Specifically, the dynamic programming algorithm refers to an algorithmic strategy used to solve optimization problems. Its core principle is to decompose a complex problem into a series of interrelated subproblems. By solving the subproblems and using their solutions to construct the solution to the original problem, a large amount of repetitive calculation of the locker status is avoided. The combination state of currently available lockers of different sizes refers to the combination of all lockers of different sizes. This embodiment does not limit the specific representation of the combination state of currently available lockers of different sizes; those skilled in the art can set it according to actual needs, such as using binary to represent the combination state of currently available lockers of different sizes. The total allocated space refers to the total space obtained by adding the spaces of all allocated lockers. The size of the locker required by user i refers to the size of the locker used by user i. The size of the storage locker required for the items stored by user i is determined by the volume of the items stored by user i. The new storage locker status after allocation refers to the state of the remaining empty storage lockers after user i's items are allocated to the storage lockers. The new total allocated space after allocation refers to the total space size of all allocated storage lockers after user i's items are allocated to the storage lockers. The value obtained by allocating items to user i refers to a quantitative indicator that measures the importance of the item to the space allocation target. This embodiment does not limit the specific criteria for determining the value obtained by allocating items to user i. Those skilled in the art can set it according to actual needs. For example, if the stored items have a clear market price or economic value, then its market price can be directly used as the value obtained by allocating items to user i.

[0039] Specifically, the storage space allocation module can control the storage information of all storage lockers through a dynamic programming algorithm, and allocate the storage lockers reasonably to make full use of the space of each storage locker.

[0040] Specifically, the storage space allocation module compares the number of users simultaneously identifying their identities Sy with the preset number of users simultaneously identifying their identities Sy0, where 10 ≤ Sy0 ≤ 20. Based on the comparison result, it determines the storage locker demand and adjusts the dynamic programming algorithm accordingly. When Sy≤Sy0, the storage space allocation module determines that the storage cabinet demand is low and does not adjust the dynamic programming algorithm. When Sy≤Sy0, the storage space allocation module determines that the storage cabinet demand is high and adjusts the dynamic programming algorithm accordingly. Prioritize the allocation of storage space to users whose user type is member among the number of users whose identity is simultaneously recognized. After the member users are allocated, allocate storage space to the users whose user type is determined to be general among the number of users whose identity is simultaneously recognized. If the allocation of member users is completed, the storage space allocation module stops calculating the storage space for general users when j=0.

[0041] Specifically, the preset number of users who can be identified simultaneously refers to a preset value used to judge the demand for lockers. The demand for lockers refers to the demand for lockers from multiple users at the same time. The demand for lockers includes situations where the demand is low and the demand is high.

[0042] Specifically, the storage space allocation module assesses the demand for storage lockers. When demand is high, it prioritizes meeting the needs of members. When there are insufficient storage lockers, it stops calculating storage space for general users in advance, reducing unnecessary waiting time for users and planning the allocation of storage lockers ahead of time.

[0043] Specifically, the storage space allocation module inputs the image of the stored object from the storage information into the stored object recognition model, outputs the importance index Jz of the stored object, compares the importance index Jz with the preset importance index Jz0 of the stored object (0.42≤Jz0≤0.56), and judges the importance of the stored object based on the comparison result. The judgment process for determining security requirements is then calibrated based on the judgment result. When Jz≤Jz0, the register space allocation module determines that the importance of the registered item is not important and does not calibrate the judgment process of security requirements; When Jz > Jz0, the storage space allocation module determines that the importance of the stored item is important and calibrates the judgment process of security requirements. The preset requirement value F0 is calibrated by the calibration coefficient γ = 1 - (Jz - Jz0 / Jz0) to obtain the calibrated preset requirement value F01. F01 is set to F0 × γ, and the preset requirement value F0 is replaced with the calibrated preset requirement value F01. The security requirements are then re-judged.

[0044] Specifically, the stored object recognition model refers to a convolutional neural network model that takes a stored object image as input and outputs a stored object importance index. The stored space allocation module constructs the stored object recognition model using a stored object recognition model construction method, which includes: 70% of the data recognition learning sample dataset is divided into a data recognition training set, 15% into a data recognition validation set, and 15% into a data recognition test set. A convolutional neural network (CNN) model is trained using the training set, and its weights are updated using a backpropagation algorithm. After each epoch, the CNN model is validated using the validation set to obtain a validation loss value and a validation accuracy. When the validation loss value and the validation accuracy meet preset validation conditions, the CNN model is tested using the test set to obtain a validation test accuracy. When the validation test accuracy reaches a preset accuracy, the CNN model is output as a registered object recognition model. The data recognition learning sample dataset refers to the learning dataset used to train the registered object recognition model. The data recognition learning sample dataset includes historical registered object images and the historical registered object importance index corresponding to those images. An epoch refers to the process of the convolutional neural network model completing one forward and backward propagation on the data recognition training set. The validation loss value refers to the loss value of the loss function in the convolutional neural network model when the data recognition validation set is input into the model for validation. The validation accuracy rate refers to the ratio of the number of data recognition results output by the model that match the data recognition results in the data recognition validation set to the total number of samples in the data recognition validation set. The preset validation condition refers to the condition that the validation loss value does not decrease and the validation accuracy does not increase for five consecutive epochs. The recognition test accuracy rate refers to the ratio of the number of data test results output by the model that match the data recognition results in the data recognition test set to the total number of samples in the data recognition test set. The preset accuracy rate refers to the preset value for the recognition test accuracy rate at which the convolutional neural network model reaches the output standard, such as a preset accuracy rate of 98%. The importance index of the stored item refers to the score of the importance of the stored item. The preset importance index of the stored item refers to the preset value used to judge the importance of the stored item. The importance of the stored item refers to the importance of the stored item. The importance of the stored item includes the stored item being unimportant and the stored item being important.

[0045] Specifically, the storage space allocation module adjusts the preset requirement value by judging the importance of the stored items, thereby reducing the preset requirement value. This makes it easier to identify stored items as important items, improves their storage security, makes identity verification more difficult and accurate, and thus increases the security of storing valuable items.

[0046] Specifically, the storage space sensing module compares the space utilization rate Ly in the storage information with the preset storage space utilization rate Ly0, where 70%≤Ly0≤88%, and judges the utilization of the storage space based on the comparison result, and records the storage cabinet utilization based on the judgment result, wherein: When Ly > Ly0, the storage space sensing module determines that the utilization rate of the storage space is high and does not record the storage cabinet utilization. When Ly≤Ly0, the storage space sensing module determines that the storage space utilization is low, records the storage cabinet utilization as one instance of low utilization, and also compares the number of times the storage cabinet has been used (Lc) with the preset number of times it has been used (Lc0) in the historical utilization data, where 3 times / day ≤ Lc0 ≤ 5 times / day. Based on the comparison result, the module judges the storage cabinet space allocation and optimizes the dynamic programming algorithm accordingly. When Lc≤Lc0, the storage space sensing module determines that the storage cabinet space allocation is normal and does not optimize the dynamic programming algorithm. When Lc > Lc0, the storage space sensing module determines that the storage cabinet space allocation is abnormal and optimizes the dynamic programming algorithm. The allocation state dp[i][j][k] is optimized by the algorithm optimization coefficient Ω = (Ly - Ly0) / (Lc - Lc0) to obtain the optimized allocation state dp[i][j][k]'. The allocation state dp[i][j][k]' is set to dp[i][j][k] - ((Ly - Ly0) / (Lc - Lc0)), and the allocation state dp[i][j][k] is replaced with the optimized allocation state dp[i][j][k]'. The first state transition equation and the second state transition equation are then redefined.

[0047] Specifically, the space utilization rate refers to the proportion of space occupied by stored items in the locker; the preset storage space utilization rate is a preset value used to judge the utilization of storage space; the storage space utilization refers to the utilization of storage space by stored items in the locker; the storage space utilization includes high utilization and low utilization; the historical utilization data refers to the total data recorded by the storage space sensing module regarding the utilization of the locker; the number of times the locker has a low utilization rate refers to the number of times the locker is recorded as having a low utilization rate within a day; the preset number of times the locker has a low utilization rate is a preset value used to judge the space allocation of the locker; the locker space allocation refers to the allocation of the locker space within a day in which it is reasonably utilized; the locker space allocation includes normal space allocation and abnormal space allocation.

[0048] Specifically, the storage space sensing module judges the utilization of the storage space of the lockers by monitoring the space utilization rate and records the utilization of the lockers. When the number of times the storage locker has a low utilization rate in the historical utilization rate data is greater than the preset number of times it has a low utilization rate, the storage space sensing module determines that the storage locker space allocation is abnormal, that is, the allocation of the storage lockers is unreasonable, and optimizes the dynamic programming algorithm to make the allocation of the storage lockers more reasonable and more in line with the actual volume of the stored items.

[0049] Specifically, the storage space sensing module compares the storage cabinet switching current Ik in the storage information with the preset minimum storage cabinet switching current Ikmin and the preset maximum storage cabinet switching current Ikmax, where Ikmin = 600mA and Ikmax = 800mA. Based on the comparison result, it judges the storage cabinet current situation and records the judgment result. When Ikmin<Ik≤Ikmax, the storage space sensing module determines that the storage cabinet current is normal and does not record the storage cabinet current. When Ikmin=Ik, the storage space sensing module determines that the storage cabinet current is normal and does not record the storage cabinet current. When Ikmin>Ik, the storage space sensing module determines that the storage cabinet current is abnormal, records the storage cabinet current, and records it as a current abnormality. When Ikmax < Ik, the storage space sensing module determines that the storage cabinet current is abnormal, records the storage cabinet current, and records it as a current abnormality.

[0050] Specifically, the locker switch current refers to the current magnitude when controlling the locker to open and close; the preset minimum locker switch current refers to a preset minimum value used to judge the locker current status; the preset maximum locker switch current refers to a preset maximum value used to judge the locker current status; the locker current status refers to the normal and abnormal conditions of the locker switch current, including normal and abnormal conditions.

[0051] Specifically, the storage space sensing module monitors the current status of the storage cabinets, records abnormal current conditions, and adjusts subsequent plans based on the number of abnormal current conditions to enhance the rationality of space allocation.

[0052] Specifically, the storage space sensing module compares the number of abnormal storage cabinet current events Ck with the preset number of abnormal storage cabinet events Ck0, where 3 events ≤ Ck0 ≤ 5 events. Based on the comparison result, it judges the abnormality of the storage cabinet and corrects the judgment process of storage cabinet space allocation based on the judgment result. When Ck≤Ck0, the storage space sensing module determines that the abnormal situation of the storage cabinet is infrequent and does not correct the judgment process of storage cabinet space allocation. When Ck > Ck0, the storage space sensing module determines that the abnormal situation of the storage cabinet is abnormally frequent, and corrects the judgment process of storage cabinet space allocation by using a correction coefficient ES = 1.4 - 0.4 × e -（Ck-Ck0） The preset number of low utilizations Lc0 is corrected to obtain the corrected number of low utilizations Lc0k. Lc0k is set to Lc0×ES. The preset number of low utilizations Lc0 is replaced with the corrected number of low utilizations Lc0k, and the storage cabinet space allocation is re-evaluated.

[0053] Specifically, the number of abnormal current conditions in the storage cabinet refers to the number of times the current condition of the storage cabinet is determined to be abnormal. The preset number of abnormal storage cabinet conditions refers to the abnormal value used to judge the abnormal condition of the storage cabinet. The abnormal condition of the storage cabinet refers to whether the abnormal condition of the storage cabinet is frequent. The abnormal condition of the storage cabinet includes the abnormal condition of the storage cabinet being infrequent and the abnormal condition of the storage cabinet being frequent.

[0054] Specifically, the storage space sensing module monitors abnormal conditions of the storage cabinets. When abnormal conditions occur frequently, it corrects the judgment process of storage cabinet space allocation and increases the preset number of low utilization rates. This reduces the difficulty of judging the storage cabinet space allocation and avoids the possibility that low storage cabinet utilization is caused by too many abnormal current conditions, thereby improving the rationality of storage cabinet space allocation.

[0055] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. An Internet of Things (IoT) based item storage management system, characterized in that, The system includes: The data acquisition module is used to collect stored information and environmental information; The authentication and storage module is used to obtain the identity similarity of user feature vectors based on the user's iris and palm print in the storage information, and to grant storage permissions to users based on the identity similarity of user feature vectors. It is also used to adjust the process of obtaining the identity similarity of user feature vectors based on the lighting conditions, and to judge the security requirements based on the security requirements of the storage cabinet, and to optimize the adjustment process of obtaining the identity similarity of user feature vectors based on the judgment results. The storage space allocation module is used to intelligently allocate storage space according to a dynamic programming algorithm when granting storage permissions to users, and also to calibrate the judgment process of security requirements based on the importance index of the stored items. The storage environment perception module is used to judge the storage cabinet space allocation based on the space utilization rate in the storage information, and optimize the dynamic programming algorithm based on the judgment result. It is also used to judge the abnormal situation of the storage cabinet based on the number of abnormal current situations, and correct the judgment process of storage cabinet space allocation based on the judgment result. The storage space sensing module determines the abnormality of the storage cabinet based on the number of abnormal current occurrences, and corrects the judgment process of storage cabinet space allocation based on the judgment result. This includes comparing the number of abnormal current occurrences Ck of the storage cabinet with the preset number of abnormal occurrences Ck0, where 3 times ≤ Ck0 ≤ 5 times. Based on the comparison result, the module determines the abnormality of the storage cabinet, and corrects the preset number of low utilization rates in the judgment process of storage cabinet space allocation based on the judgment result and correction coefficient.

2. The IoT-based item storage management system according to claim 1, characterized in that, The authentication registration module uses the iris feature vector extraction method to obtain the iris feature vector H of the user's iris in the registered information, and uses the palmprint feature vector extraction method to obtain the palmprint feature vector Y of the user's palm in the registered information. Based on the iris feature vector H and the palmprint feature vector Y, the module calculates the comprehensive user feature vector K.

3. The IoT-based item storage management system according to claim 2, characterized in that, The authentication storage module uses the Hamming distance method to obtain the identity similarity A between the comprehensive user feature vector and the comprehensive user feature vector in the storage database. The identity similarity A of the comprehensive user feature vector is compared with the preset identity similarity A0. 0.83≤A0≤0.

91. The identity similarity is judged based on the comparison result, and storage permissions are granted to the user based on the judgment result.

4. The IoT-based item storage management system according to claim 3, characterized in that, The authentication register module calculates the illumination condition Qt based on the light intensity value Qg and the light angle value Qj in the environmental information, compares the illumination condition Qt with the preset illumination condition Qt0, and determines the illumination condition based on the comparison result (0.73≤Qt0≤0.93). Based on the result, the module also adjusts the calculation process of the comprehensive user feature vector.

5. The IoT-based item storage management system according to claim 4, characterized in that, The authentication storage module inputs the pedestrian flow and environmental structure diagram from the environmental information into the expert judgment model, outputs the storage cabinet security requirement value F, compares the storage cabinet security requirement value with the preset requirement value F0, 0.67≤F0≤0.88, and judges the security requirement based on the comparison result. Based on the judgment result, the adjustment process of the calculation process of the comprehensive user feature vector is optimized.

6. The IoT-based item storage management system according to claim 5, characterized in that, The storage space allocation module uses a dynamic programming algorithm to intelligently allocate storage space based on storage information.

7. The IoT-based item storage management system according to claim 6, characterized in that, The storage space allocation module inputs the image of the stored object from the storage information into the stored object recognition model, outputs the importance index Jz of the stored object, compares the importance index Jz of the stored object with the preset importance index Jz0 of the stored object, 0.42≤Jz0≤0.56, and judges the importance of the stored object based on the comparison result, and calibrates the judgment process of security requirements based on the judgment result.

8. The IoT-based item storage management system according to claim 7, characterized in that, The storage space sensing module compares the space utilization rate Ly in the storage information with the preset storage space utilization rate Ly0, where 70%≤Ly0≤88%, and judges the utilization of the storage space based on the comparison result, and records the utilization of the storage cabinet based on the judgment result.

9. The Internet of Things-based item storage management system according to claim 8, characterized in that, The storage space sensing module also compares the number of times the storage cabinet has a low utilization rate (Lc) in the historical utilization data with the preset number of times the low utilization rate (Lc0), where 3 times / day ≤ Lc0 ≤ 5 times / day. Based on the comparison results, it judges the storage cabinet space allocation and optimizes the dynamic programming algorithm based on the judgment results.

10. The Internet of Things-based item storage management system according to claim 9, characterized in that, The storage space sensing module compares the storage cabinet switching current Ik in the storage information with the preset minimum storage cabinet switching current Ikmin and the preset maximum storage cabinet switching current Ikmax, where Ikmin = 600mA and Ikmax = 800mA. Based on the comparison results, the module judges the storage cabinet current status and records the current status based on the judgment results.

Images