Warehouse material identification method based on data analysis

Abstract

The invention relates to the technical field of intelligent warehousing, and particularly discloses a warehouse material identification method based on data analysis, and the method comprises the steps: collecting operation event data and material coordinate information in real time through a UHF-RFID reader-writer array and a visual positioning system, and generating an operation event time sequence; calculating a material mixing risk value, an access frequency factor and an access confusion degree index, analyzing coordinate data to detect a displacement sudden change and a boundary crossing event, generating a position anomaly degree index, fusing multiple indexes to obtain a goods allocation inspection priority score, generating an anomaly inspection instruction when the score exceeds a threshold value, and marking anomaly types such as mixing, deviation and boundary crossing. According to the invention, real-time acquisition and multi-dimensional analysis are carried out on the operation event data and the material coordinate position information of the warehouse goods allocation, and the operation event time sequence characteristics and the position dynamic change are combined to construct a quantitative index system to identify the abnormity of the goods allocation, so that accurate identification of warehouse materials and timely early warning of the abnormal state are realized, and the working efficiency is improved. Therefore, the refinement level and the operation efficiency of warehouse management are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent warehousing, in particular to a warehouse material identification method based on data analysis. BACKGROUND

[0002] In the modern warehousing management system, efficient and accurate identification and management of warehouse materials are the key links to ensure smooth logistics and reduce operating costs. However, traditional warehouse material management and identification methods have many drawbacks and cannot meet the growing demand for warehousing business and the requirements of fine management.

[0003] In information collection, barcodes or manual records are relied on. Barcodes are easily damaged and need to be scanned at close range, which is low in efficiency. Manual records are prone to errors and cannot track material locations in real time, resulting in inaccurate inventory data and affecting subsequent links. In terms of material control, it is difficult to real-time early warning of mixed storage risk, and manual inventory cycle is long and prone to missing mixed storage problems. There is a lack of position monitoring, and traditional methods cannot intervene in real time when material displacement mutates or exceeds the boundary, which has high safety risks. In terms of data utilization and inspection planning, multi-source data is isolated and difficult to analyze collaboratively. Manual inspection has no quantitative standard, and resources are not allocated reasonably, which may miss high-risk storage locations.

[0004] Therefore, there is an urgent need for a warehouse material identification method based on data analysis to solve the above problems. SUMMARY

[0005] The present application aims to provide a warehouse material identification method based on data analysis, comprising the following steps:

[0006] Real-time collection of operation event data and material coordinate position information of warehouse storage locations by UHF-RFID reader array and visual positioning system, the operation event data including material SKU code and operation type, and sorting the continuously collected operation event data by timestamp to generate operation event time series;

[0007] According to the difference between adjacent event SKU codes and the operation type in the operation time series, the material mixing risk value is calculated;

[0008] Obtain the number of operation events per unit time to generate the access frequency factor, and fuse the material mixing risk value and the access frequency factor to generate the access confusion degree index;

[0009] Analyze the continuous coordinate position data, detect the displacement mutation event, compare the material coordinate position with the preset electronic fence boundary, and obtain the boundary crossing event mark level;

[0010] Fuse the displacement mutation event intensity and the boundary crossing event mark to generate the position abnormality degree index, and obtain the storage location inspection priority score based on the access confusion degree index, the position abnormality degree index and the access frequency factor.

[0011] When the priority score exceeds a preset threshold, a storage location anomaly checking instruction is generated and an anomaly type label is marked, the anomaly type including at least one of material mixed storage anomaly, position offset anomaly, and out-of-bound stacking anomaly.

[0012] Further, the step of calculating the material mixing risk value according to the difference in SKU code of adjacent events in the operation time sequence and the operation type includes:

[0013] Extract all operation events of the target storage location within a set time window, and calculate the average access frequency per unit time;

[0014] Analyze the material SKU code of adjacent operation events pair by pair, and calculate the proportion of the same SKU code as the material similarity;

[0015] According to the proportion of the total volume of the material in the adjacent operation events to the maximum capacity of the storage location, calculate the space load coefficient;

[0016] The material similarity and the space load coefficient are inversely related to generate a mixed risk weighting value;

[0017] Based on the product operation of the average access frequency and the mixed risk weighting value, a material mixing risk value is generated.

[0018] Further, the step of obtaining the number of operation events per unit time, generating an access frequency factor, and fusing the material mixing risk value and the access frequency factor to generate an access confusion degree index includes:

[0019] Based on the operation event time sequence, the total number of operation events of the target storage location within a preset period is counted, and the operation frequency per unit time is calculated;

[0020] Obtain the standard volume and the standard weight of a single piece of material;

[0021] Calculate a first proportion value of the material standard volume to the maximum volume of the storage location, and a second proportion value of the material standard weight to the safe bearing threshold of the storage location;

[0022] Obtain a material attribute influence coefficient according to the first proportion value and the second proportion value;

[0023] Obtain an access frequency factor according to the material attribute influence coefficient and the operation frequency per unit time;

[0024] Convert the material mixing risk value to a relative risk proportion value, and convert the access frequency factor to a relative frequency proportion value;

[0025] Obtain an access confusion degree index according to the relative risk proportion value and the relative frequency proportion value.

[0026] Further, the step of analyzing the continuous coordinate position data, detecting displacement mutation events, comparing the material coordinate position with the preset electronic fence boundary, and obtaining the boundary crossing event marking level comprises: dividing the continuous three-dimensional coordinate data stream collected by the visual positioning system into trajectory segments at fixed time intervals, and calculating the linear distance change between adjacent time points;

[0027] Calculating the average moving speed of each trajectory segment, if the average moving speed exceeds the preset speed threshold, marking it as a displacement mutation event, if the average moving speed does not exceed the preset speed threshold, not marking it;

[0028] Calculating the speed change rate of adjacent trajectory segments as the instantaneous acceleration;

[0029] Obtaining the boundary vertex coordinate set of the preset electronic fence, and calculating the vertical distance from the material coordinate to the nearest fence boundary in real time;

[0030] Generating a boundary crossing mark when the vertical distance is less than zero, generating a warning mark when the vertical distance is within a safe warning range, and synchronously triggering the UHF-RFID reader when a displacement mutation event is detected;

[0031] Verifying whether the current material SKU code matches the storage location registration information;

[0032] If the SKU code does not match, the warning level of the boundary crossing event mark is raised; if the SKU code matches, the credibility weight of the displacement mutation event is reduced.

[0033] Further, the step of fusing the displacement mutation event intensity and the boundary crossing event mark to generate a position abnormality index, and obtaining a storage location inspection priority score based on the access confusion index, the position abnormality index, and the access frequency factor comprises:

[0034] Based on the displacement mutation event, extracting the instantaneous acceleration absolute value and the total duration of the event;

[0035] Combining the instantaneous acceleration absolute value and the duration according to a preset proportion to generate a displacement event intensity value;

[0036] According to the boundary crossing event mark level, a corresponding weight coefficient is obtained;

[0037] According to the displacement event intensity value multiplied by the weight coefficient, a position abnormality index is obtained;

[0038] According to the access confusion index, the position abnormality index, and the access frequency factor, a storage location inspection priority score is obtained.

[0039] Further, when the inspection priority score exceeds the preset threshold, a storage location abnormality inspection instruction is generated and an abnormality type label is marked, the abnormality type including at least one of material mixed storage abnormality, position offset abnormality, and out-of-bound stacking abnormality.

[0040] When the storage location inspection priority score exceeds the preset threshold, multi-source consistency verification is performed according to the displacement mutation event coordinates, the RFID verification result, the position abnormality degree index, and the access confusion degree index.

[0041] Based on the continuous three-dimensional coordinate data stream and the SKU code, the following spatial relationship analysis is performed:

[0042] The X, Y and Z axis offset amounts of the actual position of the material from the preset coordinates are calculated.

[0043] It is detected whether the distance between the surfaces of adjacent materials is less than a safe operation threshold.

[0044] It is detected whether the SKU codes of materials in the same region are the same; an abnormality type label is marked according to the multi-source verification result, and a risk level is divided according to the position abnormality degree index.

[0045]

[0046] The application also discloses a warehouse material identification system based on data analysis, which comprises:

[0047] A generation module is configured to collect operation event data and material coordinate position information of a warehouse storage location in real time through a UHF-RFID reader-writer array and a visual positioning system, the operation event data including a material SKU code and an operation type, and sort the continuously collected operation event data according to timestamps to generate an operation event time sequence.

[0048] A calculation module is configured to calculate a material mixing risk value according to the SKU code difference degree and the operation type of adjacent events in the operation time sequence.

[0049] A fusion module is configured to obtain the number of operation events in a unit time to generate an access frequency factor, fuse the material mixing risk value and the access frequency factor, and generate an access confusion degree index.

[0050] An analysis module is configured to analyze continuous coordinate position data, detect displacement mutation events, compare the material coordinate position with a preset electronic fence boundary, and obtain an out-of-bound event label level.

[0051] An acquisition module is configured to fuse the displacement mutation event intensity and the out-of-bound event label to generate a position abnormality degree index, and obtain a storage location inspection priority score based on the access confusion degree index, the position abnormality degree index and the access frequency factor.

[0052] ​The output module is configured to generate a storage location abnormality checking instruction and mark an abnormality type label when the checking priority score exceeds a preset threshold, the abnormality type including at least one of material mixed storage abnormality, position offset abnormality, and out-of-bound stacking abnormality.

[0053] Further, the calculation module comprises:

[0054] The extraction unit is configured to extract all operation events of the target storage location within a set time window, and count an average access frequency per unit time;

[0055] The analysis unit is configured to analyze, pair by pair, material SKU codes of adjacent operation events, and calculate a proportion of the same SKU codes as a material similarity;

[0056] The calculation unit is configured to calculate a space load coefficient according to a proportion of a total volume of the materials in the adjacent operation events to a maximum capacity of the storage location;

[0057] The first generation unit is configured to inversely relate the material similarity and the space load coefficient to generate a mixed risk weighting value;

[0058] The second generation unit is configured to generate a material mixing risk value based on a product operation of the average access frequency and the mixed risk weighting value.

[0059] The present application also provides a computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the above warehouse material identification method based on data analysis when executing the computer program.

[0060] The present application also provides a computer readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the steps of the above warehouse material identification method based on data analysis.

[0061] The present application has the following beneficial effects:

[0062] The present application realizes accurate identification of warehouse materials and timely early warning of abnormal states by real-time collection and multi-dimensional analysis of operation event data and material coordinate position information of warehouse storage locations through the cooperative operation of a UHF-RFID reader array and a visual positioning system, combined with operation event timing characteristics and position dynamic changes, to construct a quantitative index system to identify storage location abnormalities, thereby improving the fine management level and operation efficiency of warehouse management. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 The method flowchart is provided for an embodiment of the present application.

[0064] Figure 2 The system structure diagram is provided for an embodiment of the present application.

[0065] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0066] It should be understood that the specific embodiments described herein are merely exemplary and are not intended to limit the present application.

[0067] As shown in the drawings, the present application provides a warehouse material identification method based on data analysis, comprising the following steps: Figure 1

[0068] S1, through the UHF-RFID reader array and the visual positioning system, real-time collection of warehouse location operation event data and material coordinate position information, the operation event data including material SKU code and operation type, and sorting the continuously collected operation event data according to the time stamp to generate operation event time sequence;

[0069] S2, according to the difference of adjacent event SKU code and operation type in operation time sequence, calculate the material mixed risk value;

[0070] S3, get the number of operation events per unit time, generate access frequency factor, fuse material mixed risk value and access frequency factor, generate access confusion index;

[0071] S4, analyze the continuous coordinate position data, detect displacement mutation event, compare the material coordinate position with the preset electronic fence boundary, and get the boundary crossing event mark level;

[0072] S5, fuse displacement mutation event intensity and boundary crossing event mark, generate position abnormality index, and get the location check priority score based on access confusion index, position abnormality index and access frequency factor;

[0073] S6, when the check priority score exceeds the preset threshold, generate the location abnormality check instruction and mark the abnormal type mark, the abnormal type including at least one of material mixed abnormality, position offset abnormality and boundary crossing accumulation abnormality.

[0074] As described above in steps S1-S6, the present application realizes the real-time collection and multi-dimensional analysis of warehouse location operation event data and material coordinate position information through the cooperative operation of UHF-RFID reader array and visual positioning system, combines the operation event time sequence characteristics and position dynamic change, constructs a quantitative index system to identify the location abnormality, and finally realizes the accurate identification of warehouse material and timely warning of abnormal state, so as to improve the fine level and operation efficiency of warehouse management.

[0075] ​In the warehouse management scenario, the material access operation has high frequency and dynamics, the storage location state will change continuously during the operation process, and the difference of SKU code may cause material mixing, the deviation or out-of-bound of coordinate position may cause space utilization confusion and even safety risk. Real-time capture of these dynamic changes and accurate assessment of their impact on the storage location state are the core needs to ensure the orderly operation of the warehouse and reduce management costs.

[0076] In the traditional warehouse material management method, data collection relies on manual recording or single barcode scanning, which is not only inefficient, but also cannot associate operation events and location information in real time, lacks time sequence analysis for material mixing risk assessment, and cannot predict mixing trend through SKU difference of adjacent operations, and the detection of position abnormality only depends on static boundary judgment, ignoring the dynamic characteristics of displacement mutation, resulting in late abnormality discovery. The present application systematically solves these problems through multi-source data fusion and quantitative index construction.

[0077] In one embodiment, the step of calculating the material mixing risk value according to the SKU code difference degree of adjacent events in the operation time sequence and the operation type comprises:

[0078] S21, extracting all operation events of the target storage location within a set time window, and counting the average access frequency per unit time;

[0079] S22, analyzing the material SKU codes of adjacent operation events in pairs, and calculating the proportion of the same SKU code as the material similarity;

[0080] S23, calculating the space load coefficient according to the proportion of the total volume of the materials in the adjacent operation events to the maximum capacity of the storage location;

[0081] S24, inversely associating the material similarity and the space load coefficient to generate a mixing risk weighted value;

[0082] S25, generating a material mixing risk value based on the product operation of the average access frequency and the mixing risk weighted value.

[0083] As described in steps S21-S25 above, the present application calculates the material mixing risk value by performing time sequence analysis on the operation events of the target storage location within a specific time range, combining the material type difference, the storage space occupation condition and the operation frequency, so as to quantitatively evaluate the possibility of material mixing in the storage location, and provide accurate risk basis for subsequent identification of material mixing abnormality.

[0084] In actual warehouse operation, the risk of mixed storage of materials is not isolated, and is closely related to the operation frequency of the storage location, the change of material types, and the utilization status of the storage location space. When the material storage and retrieval operation of a storage location is more frequent, each operation may introduce different types of materials, increasing the probability of mixed storage. The greater the difference in material types in adjacent operations, the higher the risk of management confusion and sorting errors after mixed storage. At the same time, if the occupancy rate of the storage location space is high, the materials are closely stacked, and once mixed storage occurs, it is difficult to quickly distinguish, and it may also exacerbate the disorder of space utilization, further increasing the negative impact of mixed storage. Therefore, it is necessary to quantitatively analyze these factors to build an index that can accurately reflect the risk of mixed storage of materials.

[0085] Traditional material mixed storage risk assessment mainly relies on manual inspection and experience judgment. This method is not only inefficient and cannot follow the dynamic changes of storage location operation in real time, but also cannot comprehensively consider factors such as operation frequency, material difference, and space occupancy, resulting in a lag and subjectivity in the judgment of mixed storage risk, which is often discovered when the mixed storage problem is already serious, increasing the cost of subsequent arrangement and error correction. The present application realizes dynamic and objective evaluation of the risk of mixed storage of materials by systematic analysis of the time series of operation events, extraction of key influencing factors and quantitative calculation, effectively making up for the shortcomings of traditional methods.

[0086] Specifically, all operation events of the target storage location within a set time window are extracted, and the average access frequency per unit time is counted. The set time window can be determined according to the operation rhythm of the warehouse, for example, 1 hour, 8 hours (one shift) or 24 hours are selected as the analysis period, all storage, retrieval, shifting and other operation events occurring at the target storage location within the time window are collected by UHF-RFID reader array, and then the total number of operation events is divided by the duration of the time window to obtain the average access frequency. This frequency value directly reflects the busy degree of the storage location, for example, a storage location has 10 operations in 1 hour, the average access frequency is 10 times / hour, high-frequency operation means that materials are frequently moved, and the potential opportunity for mixed storage is more, which is an important basic factor constituting the risk of mixed storage.

[0087] The SKU codes of the materials of adjacent operation events are analyzed pair by pair, and the proportion of the same SKU codes is calculated as the material similarity. In the time sequence of operation events, whether the SKU codes of the materials involved in the two adjacent operations (such as the previous operation is warehousing A material and the next operation is delivery B material) are the same directly reflects the continuity of the material types. By traversing each pair of adjacent events in the time sequence, the number of pairs of adjacent events with the same SKU codes is counted, and the number is divided by the total number of pairs of adjacent events to obtain the material similarity. For example, in 10 pairs of adjacent operation events, 3 pairs have the same SKU codes, and the material similarity is 30%. The lower the material similarity, the higher the proportion of different types of materials introduced by adjacent operations, and the greater the possibility of mixed storage. This index provides a judgment basis for mixed risk from the perspective of material type difference.

[0088] According to the proportion of the total volume of materials in adjacent operation events to the maximum capacity of the storage location, the space load coefficient is calculated. For each pair of adjacent operation events, the volume data of the materials involved in the operation (which can be retrieved from the warehouse material information database through the SKU code of the material, and each SKU code corresponds to a unique standard volume of the material) is obtained, and the total volume of the materials in the storage location after the two operations is calculated. Then, the total volume is divided by the maximum volume (a preset parameter) of the storage location to obtain the space load coefficient. For example, the total volume of the materials after the adjacent operations is 5 cubic meters, and the maximum capacity of the storage location is 8 cubic meters, so the space load coefficient is 62.5%. The higher the space load coefficient, the more occupied the space of the storage location, and at this time, if there are different types of materials, the difficulty of arranging and the risk of chaos after mixed storage will be higher. This coefficient quantifies the mixed risk from the perspective of space utilization.

[0089] The material similarity and the space load coefficient are inversely related to generate a mixed risk weighted value. Since the material similarity is negatively correlated with the mixed risk (low similarity means high risk), and the space load coefficient is positively correlated with the mixed risk (high coefficient means high risk), an inverse correlation method is adopted, that is, by multiplying (1-material similarity) and the space load coefficient, the mixed risk weighted value is obtained.

[0090] The calculation formula of the mixed risk weighted value is:

[0091] ;

[0092] Wherein, the Q represents the mixed risk weighted value, represents the material similarity, The space load coefficient represents the space load coefficient, for example, when the material similarity is 30% and the space load coefficient is 62.5%, the mixed risk weighted value is (1-30%) x 62.5%=43.75%. This association makes the material type difference and the space occupation tense situation get a higher weighted value, which accurately reflects the potential mixed risk level under the joint action of the two factors.

[0093] Based on the product operation of the average access frequency and the mixed risk weighted value, the material mixed risk value is generated. The average access frequency reflects the frequency of operation, and the mixed risk weighted value reflects the potential risk under the combination of single operation. The product of the two comprehensively reflects the influence of the two factors on the final mixed risk. For example, the average access frequency is 10 times / hour, and the mixed risk weighted value is 43.75%, then the material mixed risk value is 10 x 43.75%=4.375. The higher the value, the greater the risk of mixed storage of the storage location under the current operation mode. Through this quantitative index, the material mixed storage risk of the storage location can be objectively and timely evaluated, which provides a key basis for subsequent generation of access confusion degree index and determination of storage location inspection priority, and effectively improves the early identification ability and management efficiency of the warehouse to the material mixed storage abnormality.

[0094] In one embodiment, the step of obtaining the number of operation events per unit time, generating an access frequency factor, and fusing the material mixed risk value and the access frequency factor to generate an access confusion degree index includes:

[0095] S31, based on the operation event time sequence, the total number of operation events of the target storage location in a preset period is counted, and the operation frequency per unit time is calculated;

[0096] S32, obtaining the standard volume and the standard weight of a single piece of material;

[0097] S33, calculating a first proportion value of the material standard volume to the maximum volume of the storage location, and calculating a second proportion value of the material standard weight to the safety bearing threshold of the storage location;

[0098] S34, obtaining a material attribute influence coefficient according to the first proportion value and the second proportion value;

[0099] S35, obtaining an access frequency factor according to the material attribute influence coefficient and the operation frequency per unit time;

[0100] S36, converting the material mixed risk value into a relative risk proportion value, and converting the access frequency factor into a relative frequency proportion value;

[0101] S37, obtaining an access confusion degree index according to the relative risk proportion value and the relative frequency proportion value.

[0102] As described in steps S31-S37 above, the application generates an access disorder index by combining the operation frequency of the target storage location, the impact of the material's own properties on the occupancy of the storage location, and the material mixing risk, to quantitatively reflect the overall disorder degree of the storage location in the material access process, and to provide a comprehensive evaluation basis for subsequent storage location anomaly recognition.

[0103] In warehouse management, the access disorder degree of a storage location is not only determined by the operation frequency, but also closely related to the volume, weight and other properties of the material. When the storage location is operated frequently, the material is repeatedly moved, and if the material has a large volume and high weight, its placement in the storage location is easily affected by the operation and becomes disordered. At the same time, if there is already a certain material mixing risk, frequent operation will further exacerbate the disorder. Therefore, the operation frequency, material property impact and mixing risk need to be comprehensively considered to accurately judge the access disorder state of the storage location.

[0104] Traditional methods for evaluating the disorder degree of a storage location mostly only focus on the number of operations, ignoring the impact of material properties on the occupancy of the storage location, resulting in a deviation in the judgment of the disorder degree. For example, two storage locations with the same operation frequency, one storing a material with small volume and light weight, and the other storing a material with large volume and weight close to the storage location's load threshold, the latter is more likely to cause placement disorder due to space limitations during operation, but the traditional method cannot distinguish this difference, thereby affecting the accuracy of the judgment of the storage location anomaly. The application introduces a material property impact coefficient, combines the operation frequency with the actual occupancy pressure of the material on the storage location, and further integrates the mixing risk, to realize the accurate quantification of the access disorder degree, and make up for the shortcomings of the traditional method.

[0105] Specifically, based on the operation event time series, the total number of operation events of the target storage location in a preset period is counted, and the operation frequency per unit time is calculated. The preset period can be set according to the warehouse management requirements, such as 1 hour, 4 hours, etc. All warehouse-in, warehouse-out, and shift operations of the target storage location in the preset period are collected by the UHF-RFID reader array, and the total number of operation events is divided by the length of the preset period to obtain the operation frequency per unit time. For example, if the preset period is 2 hours and the total number of operation events is 16, the operation frequency per unit time is 8 times / hour, which directly reflects the operation intensity of the storage location and is a basic parameter for measuring the disorder degree.

[0106] The standard volume and standard weight of a single piece of material are obtained. These data are stored in the material information database of the warehouse and can be retrieved through the SKU code of the material. Each SKU code corresponds to unique material specification parameters, such as a material with a SKU code corresponding to a standard volume of 0.5 m³ and a standard weight of 50 kg. This provides basic data for subsequent analysis of the impact of material on the occupancy of the storage location.

[0107] The first ratio of the material's standard volume to the maximum volume of the storage location, and the second ratio of the material's standard weight to the storage location's safe load threshold, are calculated. The maximum volume and safe load threshold are preset parameters. For example, if a storage location has a maximum volume of 10m³ and a safe load threshold of 1000kg, the material's standard volume of 0.5m³ accounts for 5% of the maximum volume (the first ratio), and its standard weight of 50kg accounts for 5% of the safe load threshold (the second ratio). These two ratios reflect the basic degree of space occupancy and load-bearing pressure of a single material in the storage location, respectively.

[0108] The material attribute impact coefficient is obtained based on the first and second ratio values. This coefficient is generated by comprehensively analyzing the impact of material volume and weight on the storage location. It is usually calculated as the maximum or weighted average of the two ratio values. For example, if the first and second ratio values ​​are both 5%, the material attribute impact coefficient is 5% (0.05). If the first ratio value of a material is 30% and the second ratio value is 20%, the coefficient can be 30% (0.3) to highlight the greater impact of volume occupancy on storage location operations. This coefficient directly reflects the degree to which material attributes restrict storage and access operations. Therefore, a higher coefficient indicates that the material is more likely to cause storage location disorganization during operation due to volume or weight.

[0109] The access frequency factor is derived from the material attribute influence coefficient and the operation frequency per unit time. By multiplying the two, we obtain a comprehensive frequency index that takes into account both operation intensity and the influence of material attributes. For example, when the operation frequency per unit time is 8 times / hour and the material attribute influence coefficient is 0.05, the access frequency factor is 0.4. When the operation frequency remains at 8 times / hour and the coefficient is 0.3, the factor is 2.4. This result shows that, at the same operation frequency, materials with a higher volume or weight ratio will produce a higher access frequency factor, which more accurately reflects the actual impact of operations on the level of storage chaos.

[0110] Convert the material mixing risk value into a relative risk ratio value, and convert the access frequency factor into a relative frequency ratio value. The relative risk ratio value is the ratio of the risk value to the preset maximum risk value. For example, if the mixing risk value of a certain cargo location is 6 and the preset maximum value is 10, then the relative risk ratio value is 60%; the relative frequency ratio value of the access frequency factor is calculated in a similar way. For example, if the factor is 2.4 and the preset maximum value is 5, then the relative frequency ratio value is 48%. This conversion can eliminate the differences in the dimensions of different indicators and make the two comparable.

[0111] According to the relative risk ratio value and the relative frequency ratio value, an access confusion degree index is obtained, and by performing a weighted operation on the two ratio values (the weights can be set according to the importance of the warehouse to the mixed risk and the operation frequency, for example, 50% each), an index that comprehensively reflects the confusion degree of the storage location access can be obtained. For example, the relative risk of 60% and the relative frequency of 48% are weighted and calculated to obtain an access confusion degree index of 54%. The higher the access confusion degree index, the more chaotic the access state of the storage location under the combined action of material mixing and operation frequency.

[0112] Through the above steps, the storage location access confusion degree is quantitatively evaluated, the frequency of operation is considered, the actual influence of the material attribute on the storage location is taken into account, and the potential risk of material mixing is integrated. Compared with the traditional evaluation method that only relies on the number of operations, the real chaotic state of the storage location can be more accurately reflected, a scientific basis is provided for the subsequent judgment of the priority of the storage location inspection, and the identification efficiency of the warehouse for access abnormalities is effectively improved.

[0113] In one embodiment, the step of analyzing continuous coordinate position data, detecting displacement mutation events, comparing the material coordinate position with the preset electronic fence boundary, and obtaining the grade of the out-of-bound event marker comprises: S41, continuously three-dimensional coordinate data stream collected by the visual positioning system is divided into trajectory segments at a fixed time interval, and the linear distance change between adjacent time points is calculated;

[0114] S42, the average moving speed of each trajectory segment is calculated, and if the average moving speed exceeds the preset speed threshold, the displacement mutation event is marked, and if the average moving speed does not exceed the preset speed threshold, no marking is performed;

[0115] S43, the speed change rate of adjacent trajectory segments is calculated as the instantaneous acceleration;

[0116] S44, the boundary vertex coordinate set of the preset electronic fence is obtained, and the vertical distance from the material coordinate to the nearest fence boundary is calculated in real time;

[0117] S45, when the vertical distance is less than zero, an out-of-bound marker is generated, when the vertical distance is within a safe warning range, a warning marker is generated, and when a displacement mutation event is detected, a UHF-RFID reader is triggered synchronously;

[0118] S46, whether the current material SKU code matches the storage location registration information is verified;

[0119] S47, if the SKU code does not match, the warning level of the out-of-bound event marker is raised; and S48, if the SKU code matches, the credibility weight of the displacement mutation event is reduced.

[0120] As described in steps S41-S48 above, the application dynamically analyzes the continuous coordinate position data of the material through the visual positioning system, detects the displacement mutation event, obtains the boundary crossing event mark by comparing the preset electronic fence boundary, and adjusts the warning level by verifying the material identity through the UHF-RFID reader, thereby realizing accurate identification and grading of material position abnormalities, and providing a reliable basis for the generation of subsequent position abnormality indicators.

[0121] In the warehouse scenario, the position state of the material directly affects the space utilization efficiency and operation safety. The material may suddenly displace (displacement mutation) due to forklift collision, unstable stacking, etc., or exceed the preset storage location boundary (boundary crossing) due to improper placement. If these situations are not discovered in time, it will lead to chaotic storage location space, insufficient distance between adjacent materials, and even safety accidents. At the same time, whether the material belongs to the current storage location (i.e., whether the SKU code and storage location registration match) will directly affect the severity of the position abnormality, and the risk level of temporary displacement of the material in the current storage location is completely different from that of the foreign material crossing the boundary. Therefore, the dynamic displacement characteristics, boundary compliance, and identity information of the material need to be comprehensively considered to accurately determine the nature and level of the position abnormality.

[0122] Traditional position monitoring methods mostly rely on static boundary checking (such as only judging whether the boundary is crossed), which cannot capture the mutation characteristics in the displacement process (such as a violent displacement of 1 meter in 0.5 seconds), nor does it combine material identity verification, resulting in one-sidedness in abnormality judgment. For example, a material in the current storage location may temporarily cross the boundary due to normal handling, which is significantly different from the risk of a foreign material crossing the boundary for a long time, but the traditional method may mark them equally, resulting in insufficient warning accuracy. Meanwhile, the neglect of displacement mutation will lead to delayed discovery of potential hazards such as material tilting caused by collision. The application realizes all-round and accurate identification of position abnormalities through the cooperation of dynamic trajectory analysis, boundary comparison, and identity verification, thereby overcoming the shortcomings of traditional methods.

[0123] Specifically, the continuous three-dimensional coordinate data stream collected by the visual positioning system is divided into trajectory segments at fixed time intervals, and the linear distance change between adjacent time points is calculated. The visual positioning system collects the three-dimensional coordinates of the material in real time (such as (x, y, z)), and the fixed time interval can be set according to the monitoring accuracy requirements (such as 0.1 seconds, 0.5 seconds), and a segment of coordinate data is intercepted as a trajectory segment every time the interval passes. For example, with an interval of 0.1 seconds, the coordinates at time t are set as (x1, y1, z1), , , ), the coordinates at time t+1 are set as (x2, y2, z2), , , ), then the calculation formula of the linear distance change between adjacent time points is:

[0124] ;

[0125] For example, when the moment coordinates are (1.2, 0.8, 0.5), the moment is (1.3, 0.9, 0.5), the linear distance change of adjacent time points is [(1.3-1.2)²+(0.9-0.8)²+(0.5-0.5)²]≈0.14 meters, which reflects the moving distance of the material in a short time and is the basic data for detecting displacement mutation.

[0126] The average moving speed of each trajectory segment is calculated, and if the average moving speed exceeds the preset speed threshold, it is marked as a displacement mutation event. The average moving speed is the linear distance change of the trajectory segment divided by the fixed time interval, and the preset speed threshold is set according to the material characteristics and operation specifications (such as 0.5 meters / second, used to distinguish between normal slow movement and sudden and severe displacement). For example, the linear distance change of a certain trajectory segment is 0.3 meters, the time interval is 0.1 seconds, the average moving speed is 3 meters / second, which exceeds the threshold of 0.5 meters / second, and is marked as a displacement mutation event. This marking can accurately capture the rapid movement of the material caused by accidents and provide key signals for subsequent risk assessment.

[0127] The speed change rate of adjacent trajectory segments is calculated as the instantaneous acceleration, which is obtained by dividing the difference between the average moving speeds of the adjacent two trajectory segments by the time interval. For example, the average speed of the previous trajectory segment is 0.2 meters / second, the current trajectory segment is 3 meters / second, and the time interval is 0.1 seconds. Then the instantaneous acceleration is (3-0.2) / 0.1=28 meters / second². The instantaneous acceleration reflects the degree of change in displacement speed, and the larger the value, the more significant the external force acting on the material (such as a violent collision), which is a core parameter for measuring the intensity of displacement mutation.

[0128] The boundary vertex coordinate set of the preset electronic fence is obtained, and the vertical distance from the material coordinates to the nearest fence boundary is calculated in real time. The electronic fence is a three-dimensional boundary set according to the size of the storage location, and its vertex coordinate set is a preset parameter (such as the x-axis boundary of a certain storage location electronic fence is 0-3 meters, the y-axis is 0-2 meters, and the z-axis is 0-1.8 meters, and the vertex coordinates include (0, 0, 0), (3, 0, 0), (3, 2, 0), etc.). The vertical distance from the material coordinates to the nearest fence boundary is the shortest distance from the current coordinates of the material to each boundary. For example, the material coordinates are (3.2, 1.5, 1.2), the vertical distance to the x-axis 3-meter boundary is -0.2 meters (negative sign indicates exceeding the boundary), and the distance to the y-axis 0-meter boundary is 1.5 meters. Therefore, the nearest vertical distance is -0.2 meters, which directly reflects whether the material has crossed the boundary and the degree of crossing.

[0129] The out-of-bound marker or early warning marker is generated according to the vertical distance, and the UHF-RFID reader is triggered synchronously when the displacement mutation event is detected. When the vertical distance is less than zero (for example, -0.2 m), it indicates that the material has exceeded the electronic fence, and the out-of-bound marker is generated; when the vertical distance is within the safe early warning range (for example, 0-0.3 m, that is, close to the boundary but not out of bounds), the early warning marker is generated; and when the displacement mutation event is detected, the system immediately triggers the UHF-RFID reader array to read the SKU code of the material, preparing for subsequent identity verification, and ensuring that the material identity information can be quickly associated when the position anomaly occurs.

[0130] The current material SKU code is verified to match the storage location registration information. The storage location registration information is stored in the warehouse management system and records the material SKU codes allowed to be stored in the storage location (for example, the A storage location is registered as SKU001 and SKU002). By comparing the material SKU code read by the RFID with the registration information, it can be determined whether the material belongs to the current storage location.

[0131] The early warning level is adjusted according to the matching result: if the SKU code does not match (for example, SKU003 appears in the A storage location), it indicates that it is an out-of-bound material, and the early warning level of the out-of-bound event marker needs to be raised (for example, from general out-of-bound to serious out-of-bound); if the SKU code matches (for example, SKU001 in the A storage location), it indicates that it is a material in the current storage location, which may be temporarily displaced due to normal operation, so the credibility weight of the displacement mutation event is reduced (for example, the influence weight on risk assessment is reduced from 0.8 to 0.3), avoiding excessive early warning for normal operation.

[0132] Through the above steps, the present application realizes dynamic and multi-dimensional analysis of material position anomalies, capturing the mutation characteristics of displacement (speed, acceleration) and judging the boundary compliance (vertical distance), and further distinguishing the nature of anomalies by combining material identity verification. Compared with the traditional method of relying only on static boundary checking, the accuracy and pertinence of position anomaly identification are significantly improved, providing a comprehensive and reliable basis for subsequent risk classification, effectively reducing the operational risks and safety hazards caused by position anomalies.

[0133] In one embodiment, the step of fusing the displacement mutation event intensity and the out-of-bound event marker to generate a position anomaly degree index, and obtaining a storage location inspection priority score based on the access confusion degree index, the position anomaly degree index, and the access frequency factor, comprises:

[0134] S51, based on the displacement mutation event, extracting the instantaneous acceleration absolute value and the total duration of the event;

[0135] S52, combining the instantaneous acceleration absolute value and the duration according to a preset ratio to generate a displacement event intensity value;

[0136] S53, obtaining a corresponding weight coefficient according to the out-of-bound event marking level;

[0137] S54, obtaining a position abnormality degree index according to the displacement event intensity value and the weight coefficient;

[0138] S55, obtaining a storage location inspection priority score according to the access confusion degree index, the position abnormality degree index and the access frequency factor.

[0139] As described in steps S51-S55, the application extracts key dynamic parameters of displacement mutation events, combines the marking level of out-of-bound events, generates a quantitative position abnormality degree index, and then fuses the access confusion degree index and the access frequency factor to calculate a storage location inspection priority score, thereby realizing accurate quantification and sorting of the position abnormality risk of the storage location and providing a scientific basis for the priority allocation of warehouse resources.

[0140] In warehouse management, the position abnormality risk of materials is not only related to whether it is out of bounds, but also closely related to the severity of the displacement process. The greater the acceleration of displacement mutation and the longer the duration, the more significant the impact on the stability of the material and the surrounding environment. At the same time, the severity level of the out-of-bound event (such as a slight warning or serious out-of-bound) directly reflects the degree of spatial violation. Relying on a single dimension (such as only looking at whether it is out of bounds) cannot comprehensively evaluate the actual risk of position abnormality. The inspection priority of the storage location also needs to be combined with its access confusion state and operation frequency. The position abnormality of a storage location with high confusion and frequent operation may cause more serious chain problems and needs to be handled first. Therefore, the fusion of multi-dimensional indexes is needed to realize accurate quantification of the position abnormality risk and the inspection priority.

[0141] Traditional methods for evaluating position abnormality mostly rely on qualitative judgment (such as "out of bounds" or "not out of bounds"), which neither quantifies the intensity of displacement mutation nor combines the out-of-bound level for comprehensive analysis, resulting in a rough judgment of the risk. At the same time, the inspection priority of the storage location relies on human experience and lacks objective quantitative standards, often resulting in missed detection of high-risk storage locations and excessive inspection of low-risk storage locations. The application solves these problems by constructing a position abnormality degree index and fusing multiple factors to calculate the priority, thereby realizing fine evaluation of the position abnormality risk and rational allocation of inspection resources.

[0142] Based on the displacement mutation event, the absolute value of instantaneous acceleration and the total duration of the event are extracted. The displacement mutation event is a marked result whose average moving speed exceeds a preset threshold. The instantaneous acceleration is the rate of change of speed of adjacent trajectory segments in the event (for example, in a certain displacement mutation event, the average speed of the previous trajectory segment is 0.3 m / s, the current trajectory segment is 3.5 m / s, the time interval is 0.1 second, and the instantaneous acceleration is (3.5-0.3) / 0.1=32 m / s², and the absolute value is 32). The total duration of the event is the time span from the beginning to the end of the displacement mutation (for example, the event starts at t=10.2s and ends at t=10.5s, and the total duration is 0.3 seconds). These two parameters respectively reflect the severity and influence range of the displacement mutation, so the greater the acceleration, the more violent the external force on the material, the longer the duration, and the greater the possibility of the material deviating from the normal position. Both of them together constitute the core dimension of evaluating the impact of displacement mutation.

[0143] The absolute value of instantaneous acceleration and the duration are combined according to a preset proportion to generate a displacement event intensity value. The preset proportion is set according to the importance of acceleration (reflecting impact strength) and duration (reflecting cumulative impact) in the warehouse. In this embodiment, the acceleration proportion is 70%, and the duration proportion is 30%.

[0144] The calculation formula of the displacement event intensity value is:

[0145] ;

[0146] Among them, represents the displacement event intensity value, represents the absolute value of instantaneous acceleration, represents the acceleration proportion, represents the duration, represents the duration proportion.

[0147] Taking the above data as an example, the absolute value of instantaneous acceleration is 32, and the duration is 0.3 seconds. Then the displacement event intensity value = 32x0.7+0.3x0.3=22.4+0.09=22.49. The greater the value, the more significant the disturbance of the displacement mutation to the material position, which provides a quantitative dynamic displacement parameter for subsequent risk assessment.

[0148] The corresponding weight coefficient is obtained according to the out-of-bound event marking level. The out-of-bound event marking level comes from the above classification result (for example, the "warning mark" of the vertical distance in the safety warning range is level 1, and the "out-of-bound mark" of the vertical distance less than zero is divided into level 2 and level 3 according to the exceeding degree), different levels correspond to different weight coefficients (for example, level 1 weight 0.4, level 2 weight 0.7, level 3 weight 1.0), the higher the level, the greater the weight, so as to highlight the contribution of serious out-of-bound to position anomaly. For example, a certain material out-of-bound mark is level 3, and the corresponding weight coefficient is 1.0, which directly reflects the high risk of space violation.

[0149] The position anomaly degree index is obtained by multiplying the displacement event intensity value by the weight coefficient. The position anomaly degree index integrates dynamic displacement intensity and static out-of-bound level, and comprehensively reflects the actual risk of position anomaly. For example, the displacement event intensity value is 22.49, the out-of-bound mark is level 3 (weight 1.0), and the position anomaly degree index = 22.49 x 1.0 = 22.49; if another material displacement event intensity value is 15, the out-of-bound mark is level 1 (weight 0.4), and the index = 15 x 0.4 = 6, obviously the former position anomaly risk is higher, and the index provides a comparable quantitative standard for the position anomaly risk of different storage locations.

[0150] The storage location inspection priority score is obtained according to the access confusion degree index, the position anomaly degree index and the access frequency factor. The three are weighted and summed by a pre-set weight (for example, position anomaly degree accounts for 40%, access confusion degree accounts for 30%, and access frequency factor accounts for 30%) to obtain the priority score. For example, the position anomaly degree of a certain storage location is 22.49, the access confusion degree is 8.5, and the access frequency factor is 5.2, then the priority score = 22.49 x 0.4 + 8.5 x 0.3 + 5.2 x 0.3 = 8.996 + 2.55 + 1.56 = 13.106, the higher the score, the higher the urgency of the storage location that needs to be checked due to position anomaly, access confusion and frequent operation.

[0151] Through the above steps, the present application realizes multi-dimensional quantification of position anomaly risk, which not only captures the intensity of dynamic displacement through displacement event intensity, but also reflects the severity of space violation through out-of-bound level weight, and finally integrates the overall confusion state and operation frequency of the storage location to generate an objective inspection priority score. Compared with traditional qualitative evaluation and experience judgment, this method can accurately distinguish the risk level of different storage locations, ensure that high-risk storage locations are inspected first, and significantly improve the efficiency and pertinence of warehouse anomaly processing.

[0152] In one embodiment, when the inspection priority score exceeds a pre-set threshold, a storage location anomaly inspection instruction is generated and an anomaly type mark is labeled, the anomaly type including at least one of material mixed storage anomaly, position offset anomaly and out-of-bound stacking anomaly, the step comprising:

[0153] When the storage location inspection priority score exceeds a preset threshold value, multi-source consistency verification is performed according to the displacement mutation event coordinates and RFID verification results, the position abnormality degree index, and the access confusion degree index.

[0154] S61, based on the continuous three-dimensional coordinate data stream and the SKU code, the following spatial relationship analysis is performed:

[0155] S62, the X, Y, and Z axis offset amounts of the actual position of the material from the preset coordinates are calculated, and if the axis offset amounts are greater than a design threshold value, it is determined that there is a position offset anomaly.

[0156] S63, it is detected whether the distance between the surfaces of adjacent materials is less than a safe operation threshold value, the safe operation threshold value being 0.5 m, and if the distance is less than the safe operation threshold value, it is determined that there is an out-of-bound stacking anomaly.

[0157] S64, it is detected whether the SKU codes of materials in the same region are the same, and if the SKU codes of materials in the same region are different, it is determined that there is a material mixing anomaly; S65, an abnormal type label is labeled according to the multi-source verification results, and a risk level is divided according to the position abnormality degree index

[0158]

[0159] As described in steps S61-S65, the present application performs multi-source data consistency verification and three-dimensional spatial relationship analysis on storage locations with priority score exceeding the threshold value, accurately identifies abnormal types such as material mixing, position offset, and out-of-bound stacking, divides risk levels based on the position abnormality degree index, generates targeted abnormality inspection instructions, and realizes accurate positioning and hierarchical disposal of warehouse abnormalities, improving the efficiency and accuracy of abnormality processing.

[0160] In warehouse operation, when the storage location inspection priority score exceeds a preset threshold value, it indicates that the storage location has a high abnormal risk, but the specific abnormal type (mixing, offset, or out-of-bound) and the risk level still need to be determined in order to take effective intervention measures. Different abnormal types have different causes and effects: material mixing may cause sorting errors, position offset may affect space utilization, and out-of-bound stacking may cause safety accidents. If accurate distinction cannot be made, the treatment measures will lack pertinence, and a single data source may have errors (such as occasional deviation in visual positioning), so multi-source data cross verification is needed to ensure the reliability of abnormality judgment. Therefore, multi-dimensional verification and analysis are needed to accurately identify abnormal types and risk levels.

[0161] ​The traditional method relies on manual on-site inspection after discovering high-risk storage locations, which lacks multi-source data collaborative verification (such as relying only on visual judgment of boundary crossing while ignoring RFID identity information) and cannot systematically analyze spatial relationships (such as ignoring axial offset or adjacent spacing), resulting in ambiguous abnormal type judgment, arbitrary risk level classification, and low processing efficiency. The present application systematically solves these problems through multi-source consistency verification and structured spatial analysis, achieving automation and precision of abnormal identification.

[0162] When the storage location inspection priority score exceeds the preset threshold, multi-source consistency verification is performed according to the displacement mutation event coordinates, RFID verification results, position abnormality index, and access confusion index. The storage location inspection priority score comes from the above calculation results, and the preset threshold is set according to the warehouse risk tolerance (such as 10 points, exceeding which triggers inspection); the displacement mutation event coordinates are the three-dimensional coordinates when the marked mutation event occurs (such as (3.5, 1.2, 0.8)); the RFID verification result is the matching of the material SKU code and the storage registration information (such as not matching); the position abnormality index is 22.49, and the access confusion index is 8.5. Multi-source consistency verification is achieved by comparing whether these data point to the same abnormal trend: if the displacement mutation coordinates show boundary crossing, RFID verification does not match, the position abnormality index is high, and the access confusion index is high, then the verification is consistent, confirming that the abnormality really exists; if there is a contradiction in one of the data (such as high position abnormality index but RFID verification matches and no displacement mutation), recheck is required to avoid misjudgment.

[0163] Based on continuous three-dimensional coordinate data stream, spatial relationship analysis is performed. The continuous three-dimensional coordinate data stream is collected in real time by the visual positioning system and includes the (x, y, z) coordinates of the material at each time. The specific analysis includes three aspects: first, calculate the X, Y, Z axial offset of the actual position of the material and the preset coordinates, the preset coordinates are the standard storage location of the storage planning (such as (2.0, 1.0, 0.5)), the difference between the actual position and the preset coordinates is the offset (such as X axis offset 0.8m, Y axis offset -0.3m, Z axis offset 0.2m), which reflects the degree of deviation of the material from the standard position; second, detect whether the distance between the surfaces of adjacent materials is less than the safe operation threshold (≥0.5m), by calculating the three-dimensional spatial distance of the current material and the nearest surrounding material (such as 0.3m), to judge whether it affects the operation safety; third, detect whether the SKU codes of the materials in the same area are the same, the storage safety limit is the preset Z axis maximum height (such as 1.5m), if the Z coordinate of the top of the material is 1.8m, it is judged as over-limit.

[0164] According to the multi-source verification result, an abnormal type label is marked, and a risk level is divided according to a position abnormality index. The abnormal type label is determined based on the checking and analysis results: if the access confusion degree is high and the RFID verification does not match the SKU, the label is marked as “material mixed storage abnormality”; if the axial offset exceeds the preset allowed range (such as ±0.5m), the label is marked as “position offset abnormality”; if the boundary is crossed and the stacking height is over the limit, the label is marked as “boundary stacking abnormality” (multiple types can be marked at the same time). The risk level is divided according to the position abnormality index, and the specific division method is as follows:

[0165] Position abnormality index ∈ [0, 0.3) → I-level risk (yellow warning);

[0166] Position abnormality index ∈ [0.3, 0.7) → II-level risk (orange alert);

[0167] Position abnormality index ≥ 0.7 → III-level risk (red alert).

[0168] The higher the level, the stronger the urgent processing demand of the abnormality.

[0169] Through the above steps, the present application realizes a closed loop from high-risk storage location identification to specific abnormal type confirmation, multi-source consistency checking ensures the reliability of abnormal judgment, avoids misjudgment caused by single data error, three-dimensional space relationship analysis comprehensively captures details such as position offset, insufficient spacing and stacking over-limit, provides an objective basis for abnormal type labeling, and risk level division clearly defines the processing priority. Compared with the traditional manual inspection, this method significantly improves the accuracy and efficiency of abnormal identification, so that the warehouse management personnel can quickly locate the problem and take targeted measures, effectively reducing the operational risk caused by abnormality.

[0170] The present application also discloses a warehouse material identification system based on data analysis, comprising:

[0171] A generation module 1 is used to collect operation event data and material coordinate position information of the warehouse storage location in real time through a UHF-RFID reader-writer array and a visual positioning system, the operation event data includes material SKU code and operation type, and the continuously collected operation event data is sorted according to time stamp to generate an operation event time sequence;

[0172] A calculation module 2 is used to calculate a material mixing risk value according to the SKU code difference and operation type of adjacent events in the operation time sequence;

[0173] A fusion module 3 is used to obtain the number of operation events per unit time to generate an access frequency factor, fuse the material mixing risk value and the access frequency factor, and generate an access confusion degree index;

[0174] Analysis module 4 is used to analyze the continuous coordinate position data, detect displacement mutation events, compare the material coordinate position with the preset electronic fence boundary, and obtain the cross-border event mark level;

[0175] Acquisition module 5 is used to fuse the displacement mutation event intensity and the cross-border event mark to generate a location abnormality index, and obtain the cargo location inspection priority score based on the access chaos index, the location abnormality index and the access frequency factor;

[0176] The output module 6 is used to generate a storage location abnormality inspection instruction and mark the abnormality type mark when the inspection priority score exceeds a preset threshold. The abnormality type includes at least one of material mixing abnormality, position offset abnormality, and out-of-bounds stacking abnormality.

[0177] Furthermore, the calculation module includes:

[0178] The extraction unit is used to extract all operation events of the target storage location within the set time window and calculate the average access frequency per unit time;

[0179] The analysis unit is used to analyze the material SKU codes of adjacent operation events one by one and calculate the proportion of the same SKU codes as the material similarity;

[0180] A calculation unit, for calculating a space load factor according to a ratio of the total volume of materials in adjacent operation events to the maximum capacity of the storage location;

[0181] The first generating unit is used to inversely correlate the material similarity with the spatial load coefficient to generate a confounding risk weighted value;

[0182] The second generating unit is used to generate a material mixing risk value based on a product operation of the average access frequency and the mixing risk weighted value.

[0183] The present application also provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned warehouse material identification method based on data analysis when executing the computer program.

[0184] The present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned warehouse material identification method based on data analysis.

[0185] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiment methods. Any reference to memory, storage, value library or other medium provided by the present application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0186] It should be noted that in this document, the terms "comprising", "including", or any other variant thereof are intended to cover a non-exclusive inclusion, so that a process, device, article or method that includes a list of elements not only includes those elements, but also includes other elements not expressly listed or inherent to such process, device, article or method. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, device, article or method that includes the element.

[0187] The above description is only the preferred embodiment of the present application, and does not limit the scope of the present application. Any equivalent result or equivalent process transformation obtained by using the content of the present application specification and drawings, or directly or indirectly applied to other related technical fields, is also included in the protection scope of the present application.

Claims

1. A warehouse material identification method based on data analysis, characterized in that: The following steps are involved: Through the UHF-RFID reader array and visual positioning system, the operation event data and material coordinate location information of the warehouse location are collected in real time. The operation event data includes the material SKU code and operation type. The continuously collected operation event data are sorted by timestamp to generate an operation event time series. Calculate the material mix-up risk value based on the SKU code differences of adjacent events in the operation time series and the operation type; Obtain the number of operation events per unit time, generate the access frequency factor, and integrate the material mixing risk value and the access frequency factor to generate the access chaos index; Analyze continuous coordinate position data, detect sudden displacement events, compare the material coordinate position with the preset electronic fence boundary, and obtain the cross-border event mark level; The intensity of displacement mutation events and out-of-bounds event markers are integrated to generate a location anomaly index. The location inspection priority score is then obtained based on the access chaos index, location anomaly index, and access frequency factor. When the inspection priority score exceeds the preset threshold, a cargo location abnormality inspection instruction is generated and an abnormality type mark is marked. The abnormality type includes at least one of material mixing abnormality, position offset abnormality, and out-of-bounds stacking abnormality.

2. The warehouse material identification method based on data analysis according to claim 1 is characterized in that: The step of calculating the material mixing risk value based on the SKU code differences of adjacent events in the operation time series and the operation type includes: Extract all operation events of the target location within the set time window and calculate the average access frequency per unit time; Analyze the material SKU codes of adjacent operation events one by one, and calculate the proportion of identical SKU codes as the material similarity; Calculate the space load factor based on the ratio of the total volume of materials in adjacent operation events to the maximum capacity of the storage location; The material similarity is inversely correlated with the spatial load factor to generate a confounding risk weighted value; The material mix-up risk value is generated based on the product of the average access frequency and the mix-up risk weighted value.

3. The warehouse material identification method based on data analysis according to claim 1 is characterized in that: The steps of obtaining the number of operation events per unit time, generating an access frequency factor, fusing the material mixing risk value and the access frequency factor, and generating an access chaos index include: Based on the time series of operation events, the total number of operation events at the target location within the preset time period is counted to calculate the operation frequency per unit time; Get the standard volume and weight of a single piece of material; Calculate a first ratio of the standard volume of the material to the maximum volume of the cargo space, and calculate a second ratio of the standard weight of the material to the safe load-bearing threshold of the cargo space; Obtaining a material property influence coefficient according to the first ratio value and the second ratio value; Obtaining an access frequency factor according to the material attribute influence coefficient and the operation frequency per unit time; Convert the material mixing risk value into a relative risk ratio value, and convert the access frequency factor into a relative frequency ratio value; An access disorder index is obtained according to the relative risk ratio value and the relative frequency ratio value.

4. The warehouse material identification method based on data analysis according to claim 1 is characterized in that: The steps of analyzing the continuous coordinate position data, detecting displacement mutation events, comparing the material coordinate position with the preset electronic fence boundary, and obtaining the cross-border event marking level include: dividing the continuous three-dimensional coordinate data stream collected by the visual positioning system into trajectory segments at fixed time intervals, and calculating the linear distance change between adjacent time points; Calculate the average moving speed of each trajectory segment. If the average moving speed exceeds the preset speed threshold, it is marked as a displacement mutation event. If the average moving speed does not exceed the preset speed threshold, it is not marked. Calculate the velocity change rate of adjacent trajectory segments as the instantaneous acceleration; Obtain the coordinate set of the boundary vertices of the preset electronic fence, and calculate the vertical distance from the material coordinates to the nearest fence boundary in real time; When the vertical distance is less than zero, an out-of-bounds mark is generated; when the vertical distance is within the safety warning range, an early warning mark is generated; when a sudden displacement event is detected, the UHF-RFID reader is triggered synchronously; Verify whether the current material SKU code matches the cargo location registration information; If the SKU code does not match, the warning level of the out-of-bounds event mark is increased; if the SKU code matches, the credibility weight of the displacement mutation event is reduced.

5. The warehouse material identification method based on data analysis according to claim 1 is characterized in that: The steps of fusing the displacement mutation event intensity and the cross-border event mark to generate a location abnormality index, and obtaining a cargo location inspection priority score based on the access chaos index, the location abnormality index, and the access frequency factor include: Based on the displacement mutation event, extract the instantaneous acceleration absolute value and the total duration of the event; The absolute value of the instantaneous acceleration and the duration are combined in a preset ratio to generate a displacement event intensity value; Obtaining a corresponding weight coefficient according to the cross-border event mark level; Obtaining a position anomaly index by multiplying the displacement event intensity value by the weight coefficient; The cargo location inspection priority score is obtained according to the access disorder index, the location abnormality index and the access frequency factor.

6. The warehouse material identification method based on data analysis according to claim 4 is characterized in that: The step of generating a cargo location abnormality inspection instruction and marking an abnormality type mark when the inspection priority score exceeds a preset threshold, wherein the abnormality type includes at least one of material mixing abnormality, position offset abnormality, and out-of-bounds accumulation abnormality, includes: When the cargo location inspection priority score exceeds a preset threshold, a multi-source consistency check is performed based on the displacement mutation event coordinates and RFID verification results, location anomaly index, and access chaos index; Based on the continuous 3D coordinate data stream and SKU code, the following spatial relationship analysis is performed: Calculate the X, Y, and Z axis offsets between the actual material position and the preset coordinates; Detect whether the distance between adjacent material surfaces is less than the safe operation threshold; Check whether the SKU codes of materials in the same area are the same; mark the abnormal type mark according to the multi-source verification results, and classify them according to the abnormality index of the location Risk level.

7. Warehouse material identification system based on data analysis, characterized by: include: A generation module is used to collect operation event data and material coordinate location information of warehouse locations in real time through a UHF-RFID reader array and a visual positioning system. The operation event data includes the material SKU code and operation type, and the continuously collected operation event data is sorted by timestamp to generate an operation event time series; The calculation module is used to calculate the material mixing risk value based on the SKU code differences of adjacent events in the operation time series and the operation type; The fusion module is used to obtain the number of operation events per unit time, generate the access frequency factor, and fuse the material mixing risk value and the access frequency factor to generate the access chaos index; The analysis module is used to analyze continuous coordinate position data, detect displacement mutation events, compare the material coordinate position with the preset electronic fence boundary, and obtain the cross-border event mark level; The acquisition module is used to integrate the displacement mutation event intensity and the out-of-bounds event marker to generate a location anomaly index, and obtain the cargo location inspection priority score based on the access chaos index, location anomaly index and access frequency factor; The output module is used to generate a storage location abnormality inspection instruction and mark the abnormality type mark when the inspection priority score exceeds a preset threshold. The abnormality type includes at least one of material mixing abnormality, position offset abnormality, and out-of-bounds stacking abnormality.

8. The warehouse material identification system based on data analysis according to claim 7 is characterized in that: The calculation module includes: The extraction unit is used to extract all operation events of the target storage location within the set time window and calculate the average access frequency per unit time; The analysis unit is used to analyze the material SKU codes of adjacent operation events one by one and calculate the proportion of the same SKU codes as the material similarity; A calculation unit, for calculating a space load factor according to a ratio of the total volume of materials in adjacent operation events to the maximum capacity of the storage location; The first generating unit is used to inversely correlate the material similarity with the spatial load coefficient to generate a confounding risk weighted value; The second generating unit is used to generate a material mixing risk value based on a product operation of the average access frequency and the mixing risk weighted value.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images