An ultra-high frequency graphene electronic tag multi-scene adaptive identification system and method

Abstract

The application discloses a kind of ultra-high frequency graphene electronic tag multi-scene adaptive identification system and method, specifically related to electronic tag identification field, including in current scene to the graphene electronic tag set that passes through reading area is executed continuous reading, form current position sequence, and from current position sequence intercept target label corresponding current sequence, target label is formed in previous scene and current sequence is compared bit by bit with previous sequence, and the bit that exists in previous sequence and is missing in current sequence, and the bit that exists in previous sequence and current sequence but bit order is different is written in target deviation result;The application executes scene stripping, bit order rearrangement and order rewriting to the sequence change of target label in adjacent scene, to solve the problem that the recognition instability caused by the difficulty in distinguishing label state change and external scene disturbance in the prior art, and the difficulty in adaptively determining subsequent identification action accordingly.

Description

Technical Field

[0001] This invention relates to the field of electronic tag recognition technology, and more specifically, to an ultra-high frequency graphene electronic tag multi-scenario adaptive recognition system and method. Background Technology

[0002] In the process of UHF electronic tag identification, the existing processing methods are usually focused on improving the integrity of batch reading and enhancing the consistency of identification under different usage environments. The common processing path is to set the corresponding identification parameters according to the deployment conditions of the reading and writing area, and adjust the transmission power, polling order or supplementary reading time based on the reading results. Some solutions also introduce reinforcement learning to correct subsequent reading and writing actions based on real-time feedback. In the circulation of cold chain medical consumables, graphene electronic tags are often placed on flexible packaging, sealing films, or flexible surfaces. The tags need to complete continuous identification under conditions of stacked boxes, low temperature and moisture, local pressure, proximity to metal frames, and rapid door-to-door reading. They are also limited by the limited dwell time, the inability to check each tag individually, and the inability to frequently interrupt circulation. Under these conditions, it can be directly observed that the same tag can respond continuously at the previous reading position, but after entering the next reading position, there are significant increases in echo fluctuations, changes in the order of response, and continuous loss of local rounds. Existing processing can usually only make uniform adjustments based on the current reading result, making it difficult to determine whether the recognition instability is caused by changes in the working state of the tag itself due to bending, pressure, or moisture, or by scene disturbances caused by external media, obstruction, or changes in spatial position. This results in a lack of specificity in subsequent power adjustment, polling rearrangement, and supplementary reading actions, and the recognition result continues to fluctuate during scene switching. Therefore, the technical problem to be solved by this application is: how to distinguish between the tag's own state changes and the recognition instability caused by external scene disturbances during the multi-scene recognition process of ultra-high frequency graphene electronic tags, and adaptively determine subsequent recognition actions accordingly. Summary of the Invention

[0003] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an ultra-high frequency graphene electronic tag multi-scene adaptive recognition system and method, which solves the problems mentioned in the background art by performing scene stripping, position rearrangement and order rewriting on the sequence changes of the target tag in adjacent scenes.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a multi-scene adaptive identification method for ultra-high frequency graphene electronic tags, comprising: S1. In the current scenario, continuously read the set of graphene electronic tags that have passed through the reading area to form the current position sequence, and extract the current sequence corresponding to the target tag from the current position sequence; S2. Compare the target label in the previous sequence with the current sequence position by position, and write the positions that exist in the previous sequence but are missing in the current sequence, as well as the positions that exist in both the previous sequence and the current sequence but are different in order, into the target deviation result. S3. Compare the remaining preceding sequences formed by the remaining labels in the previous scene with the remaining current sequences corresponding to the current position sequence, and write the missing positions and position changes that occur repeatedly in the comparison process between the remaining preceding sequences and the remaining current sequences into the scene deviation result. Delete each position from the scene deviation result from the target deviation result to generate the label deviation result. S4. Based on the label deviation result, rearrange the reading order of the target label in the current position sequence so that each bit corresponding to the label deviation result is read again before the other bits. If the target label is not obtained in the second reading, the target label is read separately to generate the recognition result. S5. Based on the recognition results, rewrite the initial reading position of the target label in the next scene so that when the target label re-enters the next scene, it forms the corresponding current sequence according to the rewritten initial reading position.

[0005] In a preferred embodiment, S1 includes: S1-1. In the current scenario, perform multiple rounds of reading on the reading area. The graphene electronic tag identifiers obtained in each round of reading are arranged into a corresponding round sequence according to their appearance order. The corresponding round sequences are then spliced ​​together according to the reading order to form the current position sequence. S1-2. Take the position of each target label in the current position sequence as the target position, and extract the adjacent label before and after each target position to form a corresponding segment with the target label to form a target segment sequence. S1-3. Compare adjacent segments in the target segment sequence one by one. If the target label of two adjacent segments is the same and the adjacent label labels are the same, retain the next corresponding segment. Then, splice the retained corresponding segments in the order of their appearance to form the current sequence corresponding to the target label.

[0006] In a preferred embodiment, S2 includes: S2-1. Write the digits of the previous sequence into the previous position according to their order of appearance, and write the digits of the current sequence into the current position according to their order of appearance, to form the previous position sequence and the current position sequence. S2-2. Starting from the first position of the preceding position sequence, read the corresponding tag identifier of each position one by one, and search for the same unmatched position in the current position sequence in ascending order of the current position to form the matching position corresponding to each preceding position. S2-3. Determine the preceding positions that have not formed a matching position as missing positions, and determine the preceding positions that have formed a matching position and whose preceding position is different from the current position corresponding to the matching position as shift positions, forming a deviation position sequence. Arrange each missing position and each shift position in the deviation position sequence in ascending order of the preceding position to generate the target deviation result.

[0007] In a preferred embodiment, S3 includes: S3-1. Compare each of the remaining preceding sequences with each of the remaining current sequences, determine the missing positions and position changes for each remaining label, and write the missing positions and position changes corresponding to each remaining label into the remaining deviation results according to the deviation type, the preceding position and the current position. S3-2. Take each bit in the remaining deviation results as the reference bit, and search for the corresponding bit in the remaining deviation results that has the same deviation type, the same previous position, and the same current position as the reference bit. When at least one corresponding bit is found, write the reference bit and each corresponding bit into the same position scenario result; otherwise, write the reference bit into the pending result. S3-3. Determine the displacement difference for each bit in the result to be judged. Record the displacement difference of the bit whose current position is empty as empty, and record the displacement difference of the bit whose current position is not empty as the current position minus the previous position. Search for the corresponding bit in the result to be judged one by one with the same deviation type, the same displacement difference and adjacent previous position. When the corresponding bit is found, connect the bit with each corresponding bit according to the previous position and write it into the connected bit scenario result. Otherwise, write the bit into the reserved result.

[0008] In a preferred embodiment, S3 further includes: S3-4. Compare each bit in the target deviation result with the same-position scene result and the consecutive-position scene result one by one. If there is a corresponding bit in the same-position scene result with the same deviation type, the same preceding position, and the same current position, write the bit into the scene deviation result. Otherwise, if there is a corresponding bit in the consecutive-position scene result with the same deviation type, the same displacement difference, and the same preceding position, and the previous or next position of the bit is also written into the consecutive-position scene result in the target deviation result, write the bit into the scene deviation result. Otherwise, write the bit into the label retention result. S3-5. Sort the label retention results in ascending order of their previous positions to generate the label deviation results.

[0009] In a preferred embodiment, S4 includes: S4-1. Map each missing position in the label deviation result to the same position in the current position sequence according to the previous position. Map each position change position in the label deviation result to the corresponding position in the current position sequence according to the current position. After removing duplicates from the mapped positions, arrange them in ascending order of position to generate a rearranged position sequence. S4-2. Write the first digit of the rearranged position sequence as the starting position, read the label corresponding to the starting position from the current position sequence, and if the reading result is not the target label, execute the reading in the subsequent positions of the rearranged position sequence in turn until the reading result is the target label or the rearranged position sequence is executed, and generate the reading result again.

[0010] In a preferred embodiment, S4 further includes: S4-3. When the result of the second reading is not the target label, delete the label identifiers in the current position sequence that are outside the rearranged position sequence in the original position sequence in the current position sequence. After deleting each label identifier, perform a second reading on the rearranged position sequence until the reading result is the target label or the deletion of the label identifiers outside the rearranged position sequence ends, and a separate reading result is generated. S4-4. When the result of the second reading is the target label, write the hit position corresponding to the target label as the recognition result. When the result of the single reading is the target label, write the hit position corresponding to the target label and the current position sequence before the deletion ends as the recognition result. When neither the result of the second reading nor the result of the single reading is the target label, write the miss as the recognition result.

[0011] In a preferred embodiment, S5 includes: S5-1. When the recognition result is a confirmed result, the reread position of the target label in the current position sequence is written as the rewritten position. When the recognition result is a verification result, the previous position of the target label in the previous scene is written as the rewritten position. S5-2. Compare the rewritten position with the original initial read position of the target tag in the next scene. If the rewritten position is less than the original initial read position, shift each position in the next scene between the rewritten position and the position before the original initial read position one position to the right and write it into the rewritten position. If the rewritten position is greater than the original initial read position, shift each position in the next scene between the position after the original initial read position and the rewritten position one position to the left and write it into the rewritten position to generate the rewritten initial read position.

[0012] In a preferred embodiment, S5 further includes: S5-3. Write the rewritten initial read position into the read order of the target tag in the next scene, and rearrange the read positions of the remaining tags in the next scene in ascending order of read position to generate the initial read position sequence in the next scene. S5-4. When the target label re-enters the next scene, the reading area is continuously read according to the initial reading position sequence, and when the target label is read, the corresponding current sequence is generated using the sequence fragment corresponding to the position of the target label.

[0013] In a preferred embodiment, a multi-scenario adaptive identification method for ultra-high frequency graphene electronic tags includes: The sequence extraction module is used to continuously read the graphene electronic tags in the current scene, form the current position sequence, and extract the current sequence corresponding to the target tag from the current position sequence; The deviation generation module is used to compare the previous sequence of the target label formed in the previous scene with the current sequence position by position, and write the positions that exist in the previous sequence but are missing in the current sequence, as well as the positions that exist in both the previous sequence and the current sequence but are different in order, into the target deviation result. The scene stripping module is used to compare the remaining preorder sequences formed by the remaining labels in the previous scene with the corresponding remaining current sequences in the current position sequence, write the missing positions and position change positions that repeatedly appear in the comparison process between the remaining preorder sequences and the remaining current sequences into the scene deviation result, and delete each position in the scene deviation result from the target deviation result to generate the label deviation result. The position rearrangement module rearranges the position of the target label in the current position sequence according to the label deviation result, so that the position corresponding to the label deviation result is read again before the other positions. If the target label is not obtained in the second reading, the target label is read separately to generate the recognition result. The sequence rewriting module rewrites the initial reading position of the target label in the next scene based on the recognition result, so that when the target label re-enters the next scene, it forms the corresponding current sequence according to the rewritten initial reading position.

[0014] The technical effects and advantages of this invention are as follows: By comparing the preceding sequence of the target label with the current sequence position by position, and combining the same position changes and consecutive position changes of other labels to perform scene stripping, it is possible to distinguish between the label's own state changes and the recognition instability caused by scene disturbances, thereby making subsequent recognition actions more targeted and relatively reducing recognition fluctuations during scene switching. By continuously performing multiple rounds of reading on the reading area and extracting the current sequence around the target label, the local arrangement relationship of the target label at the current position can be preserved, thereby reducing the interference of repeated segments on subsequent position judgments during continuous reading; By establishing a unique matching relationship between the preceding position sequence and the current position sequence, and generating missing and shifted positions accordingly, the position changes of the target label across scenes can be transformed into a structurally clear target deviation result, which relatively improves the executability of deviation recognition. By organizing the deviations of multiple other labels into co-position scene results and contiguous scene results, and then removing common scene changes from the target deviation results, the deviation of the target label itself can be preserved, thereby suppressing the interference of occasional scene changes on the target label judgment. By rearranging the reading positions based on the label deviation results and gradually deleting non-target interference items before performing individual readings when a match is missed, the target label can transition from priority reading to convergent reading, thereby relatively improving the target label's re-hitting capability. By rewriting the initial reading position of the target label in the next scene based on the recognition result, and forming the corresponding current sequence according to the rewritten reading order when entering the next scene again, the recognition conclusion of the current scene can be carried over to the next scene, thus improving the continuity of cross-scene recognition. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method steps of the present invention.

[0016] Figure 2 This is a schematic diagram of the system modules of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Refer to the instruction manual appendix Figure 1-2 The present invention provides a multi-scenario adaptive identification method for ultra-high frequency graphene electronic tags, comprising: S1. In the current scenario, continuously read the set of graphene electronic tags that have passed through the reading area to form the current position sequence, and extract the current sequence corresponding to the target tag from the current position sequence; In this embodiment, the current position sequence and the current sequence are used to provide a unified input for the generation of subsequent target deviation results, scene deviation results and tag deviation results. Therefore, it is necessary to first organize the continuous reading results of the graphene electronic tag set in the reading area under the current scene into a current position sequence with a clear sequential relationship, then extract the target segment sequence that can characterize its current position adjacency relationship from the current position sequence around the target tag, and finally delete duplicate segments through continuous comparison between adjacent target segments to form the current sequence corresponding to the target tag. The purpose of this processing is to ensure that the current sequence retains the local permutation relationship of the target label at the current position, while avoiding writing the same local permutation that repeatedly appears during continuous reading as multiple independent results into the subsequent processing chain. This implementation process includes the following steps: First, multiple rounds of reading are performed continuously on the reading area in the current scenario. The graphene electronic tag identifiers obtained in each round of reading are written into the corresponding round sequence according to their appearance order. Then, the corresponding round sequences are concatenated in the order of their formation to form the current position sequence. When the same graphene electronic tag identifier appears in different rounds of reading, it is written into different positions according to its actual appearance position. If no graphene electronic tag identifier is obtained in a round of reading, the corresponding round sequence is marked as empty and the concatenation is skipped. Through this process, the current position sequence completely preserves the appearance order of the target tag and its adjacent tags in multiple rounds of reading, providing a unified sequence basis for subsequent target position truncation. Subsequently, each position in the current position sequence that matches the target label is identified as a target position, and corresponding segments are extracted around each target position. When the target position is in the middle of the current position sequence, the label identifier of the position before the target position, the target label identifier of the target position, and the label identifier of the position after the target position are extracted to form a corresponding segment. When the target position is at the beginning of the current position sequence, the target label identifier of the target position and the label identifier of the position after the target position are extracted to form a corresponding segment. When the target position is at the end of the current position sequence, the label identifier of the position before the target position and the target label identifier of the target position are extracted to form a corresponding segment. When only the target label identifier exists in the current position sequence, only the target label identifier is used to form a corresponding segment. Each corresponding segment is written into the target segment sequence in the order in which the target position appears in the current position sequence, so that each corresponding segment in the target segment sequence corresponds one-to-one with a specific target position in the current position sequence. Furthermore, the corresponding segments in the target segment sequence are compared segment by segment. When the target label identifiers of two corresponding segments are the same and the common adjacent label identifiers of the two corresponding segments are the same, the preceding corresponding segment is deleted and the following corresponding segment is retained. Then, the retained following corresponding segment is compared with the next corresponding segment. When the target label identifiers of two corresponding segments are the same but the common adjacent label identifiers are different, both the preceding and following corresponding segments are retained. Here, the common adjacent label identifiers refer to comparing the preceding adjacent label identifiers when both preceding and following corresponding segments have the preceding adjacent label identifiers, or comparing the following adjacent label identifiers when both preceding and following segments have the following adjacent label identifiers. After all comparisons are completed, the retained corresponding segments are spliced ​​together in the order of their appearance to form the current sequence corresponding to the target label. Through the above processing, the discrete results formed by continuous reading in the current scenario can be organized into a position sequence that can directly participate in the bit-by-bit comparison, and stable local arrangement information can be extracted around the target label, thereby reducing the interference of repeated segments on subsequent deviation judgment during continuous reading. In practical applications: When the target tag passes through the reading area of ​​the current scene with the turnover vehicle, if three consecutive readings yield "Label A, Target Tag, Label B", "Label A, Target Tag, Label B", and "Label C, Target Tag, Label B", the results of the three readings are first concatenated into the current position sequence according to their formation order, and then corresponding segments are formed around each target position. Since the target tag identifiers in the first two corresponding segments are the same and the adjacent tag identifiers that coexist are also the same, the first corresponding segment is deleted and the second corresponding segment is retained. The third corresponding segment has a different adjacent tag identifier than the first retained segment, so it is retained as well. Finally, the current sequence is formed by concatenating the retained corresponding segments, and it serves as the input for generating the target deviation result.

[0019] S2. Compare the target label in the previous sequence with the current sequence position by position, and write the positions that exist in the previous sequence but are missing in the current sequence, as well as the positions that exist in both the previous sequence and the current sequence but are different in order, into the target deviation result. In this embodiment, the target deviation result is used to characterize the positional change of the target label between the previous scene and the current scene. Therefore, it is necessary to first convert the previous sequence and the current sequence into a positional sequence that can be directly compared, and then find the unique corresponding position in the current sequence according to the position in the previous sequence. Finally, based on whether the corresponding position is found and whether the position of the corresponding position changes, missing positions and shifted positions are distinguished, and the target deviation result is generated accordingly. The focus of this process is not on simply comparing whether the label appears, but on matching the specific positional relationship of the same label in the two sequences position by position, so as to avoid mismatch when the same label appears repeatedly, and also to avoid misjudging positions that only change in position as missing positions. To ensure consistency in the generation basis of subsequent scene deviation results and label deviation results, the preceding position sequence, current position sequence, matching position, missing position, shift position, and target deviation result are all organized using the same positional caliber; this implementation process includes the following steps: First, the elements in the preceding sequence are numbered sequentially according to their order of appearance, with the first element designated as the preceding position 1, and each subsequent element incremented by 1 until the end of the preceding sequence, forming the preceding position sequence. Simultaneously, the elements in the current sequence are numbered sequentially according to their order of appearance, with the first element designated as the current position 1, and each subsequent element incremented by 1 until the end of the current sequence, forming the current position sequence. Each element in the preceding position sequence includes its preceding position and a corresponding tag, and each element in the current position sequence includes its current position and a corresponding tag. When a sequence contains only one tag, that unique position is directly designated as position 1. When the current sequence is empty, no current position is written, and the preceding position sequence continues to be numbered consecutively according to its actual length. Through this process, the preceding and current sequences are uniformly converted into position sequences that can be used for positional lookups, providing a unified input for subsequent matching position determination. Subsequently, starting from the first position of the preceding position sequence, the corresponding tag identifiers are read position by position in ascending order of the preceding position. Then, in the current position sequence, positions with the same tag identifier that are not matched are searched in ascending order of the current position. When multiple positions with the same tag identifier that are not matched are found, the position with the smallest current position is taken as the matching position corresponding to that preceding position, and this matching position is marked as matched, so that it no longer participates in the search of subsequent preceding positions. When no positions with the same tag identifier that are not matched are found, that preceding position does not form a matching position. Following the above rules, the process continues from the first position of the preceding position sequence to the last position to obtain the matching results corresponding to each preceding position. Using the ascending order of the current position and matching each position only once ensures that when the same tag identifier appears repeatedly in the current position sequence, it always first establishes a correspondence with the preceding position with the smaller preceding position, thus avoiding multiple preceding positions repeatedly occupying a current position and preventing later-appearing current positions from being used prematurely for preceding position matching, which would lead to overall correspondence disorder. Furthermore, the preceding positions that do not form a matching position are identified as missing positions, and the preceding positions that have formed a matching position and whose preceding position is different from the current position corresponding to the matching position are identified as shift positions. Missing positions retain at least the corresponding label identifier and preceding position, while shift positions retain at least the corresponding label identifier, preceding position, and current position. Preceding positions that have formed a matching position and whose preceding position is the same as the current position are not written into the deviation position sequence. After determining all preceding positions, the missing positions and shift positions are arranged in ascending order of preceding position to form the target deviation result. If multiple consecutive preceding positions in the preceding position sequence do not form a matching position, these multiple consecutive preceding positions are identified as multiple missing positions and written consecutively in order of preceding position. If the same label identifier appears repeatedly in both the preceding position sequence and the current position sequence, each preceding position is independently identified as a missing position, shift position, or not written, based solely on its own matching result, thus ensuring that each bit in the target deviation result has a unique source and a unique destination. Through the above processing, the target label sequence relationship in the previous scene and the current scene can be transformed into target deviation results that can directly participate in the subsequent scene stripping process. This gives the formation process of missing and shifted positions a clear numbering basis, a unique matching rule, and a unique result destination, thereby reducing the mismatch impact when the same label identifier appears repeatedly, and establishing the generation of subsequent scene deviation results and label deviation results on a unified and stable positional relationship. In practical applications: If the preceding sequence is "Label A, Target Label, Label B, Target Label", then the preceding positions of the four bits in the preceding sequence are 1, 2, 3, and 4 respectively; if the current sequence is "Label A, Label B, Target Label", then the current positions of the three bits in the current sequence are 1, 2, and 3 respectively; during processing, label A corresponding to preceding position 1 is found in the current sequence at position 1, forming a matching position; the target label corresponding to preceding position 2 is found in the current sequence at position 3, forming a matching position. The label B corresponding to the previous position 3 is found in the current position 2 in the current position sequence, forming a matching position; the target label corresponding to the previous position 4 does not form a matching position because the current position 3 has already been matched, and there are no more unmatched positions with the same label identifier; therefore, the previous positions 2 and 3 are determined to be shift positions because the previous positions are different from the current positions, and the previous position 4 is determined to be a missing position because no matching position is formed. Then, the positions are arranged in ascending order of the previous positions to obtain the target deviation result, which can be directly called for subsequent scene deviation result stripping processing.

[0020] S3. Compare the remaining preceding sequences formed by the remaining labels in the previous scene with the remaining current sequences corresponding to the current position sequence, and write the missing positions and position changes that occur repeatedly in the comparison process between the remaining preceding sequences and the remaining current sequences into the scene deviation result. Delete each position from the scene deviation result from the target deviation result to generate the label deviation result. In this embodiment, the scene deviation result and the label deviation result are used to further break down the deviation position of the target label in the current scene into the common scene change part and the target label's own retained part. Therefore, it is necessary to first uniformly organize the sequence changes of each other label between the previous scene and the current scene, and then extract the repeated same position changes and consecutive position changes from the sequence changes of each other label. Subsequently, these two types of changes are matched with the target deviation result position by position to determine which deviation positions belong to the common changes that occur in the current scene and which deviation positions are only retained in the target label. The focus of this process is not to observe a single other label, but to cross-compare and progressively filter the deviation positions of multiple other labels, so that the scene deviation result has both the basis for repetition and the basis for the succession of previous and subsequent positions, thereby avoiding the direct writing of occasional deviations into the scene deviation result and avoiding the retention of the common scene changes that should be separated into the label deviation result. The implementation process includes the following steps: First, perform a position-by-position comparison for each remaining label to generate the corresponding remaining deviation results. Specifically, for each remaining label, first generate a remaining preceding sequence and a remaining current sequence. Then, according to the generation criteria of the target deviation results, search for positions in the remaining current sequence that have the same label identifier but are not matched, position by position. When a matching position is found, compare the preceding position with the current position. Positions that differ from the current position are identified as position change positions; positions that are the same as the current position are not written into the remaining deviation results. When no matching position is found, the preceding position is identified as a missing position, and the current position is recorded as empty. For each remaining label, write all missing positions and all position change positions corresponding to that label into a remaining deviation result in ascending order of the preceding position. Each position must include at least the deviation type, the preceding position, and the current position. After processing all remaining labels in this manner, multiple remaining deviation results are generated, providing a unified input for subsequent extraction of recurring deviation positions. Subsequently, each bit in each of the remaining deviation results is used as a reference bit. Deviation bits that are completely identical in position among the multiple remaining deviation results are extracted to form the same-position scene result. Specifically, the deviation type, previous position, and current position of each reference bit are read sequentially. Then, the corresponding bits that are the same as the reference bit in terms of deviation type, previous position, and current position are searched one by one in the remaining deviation results. When at least one corresponding bit is found, the reference bit and all corresponding bits are written into the same-position scene result, and the group of bits is marked as written, so that the group of bits will not be repeatedly written into the same-position scene result as a new reference bit in the subsequent process. When no corresponding bit is found, the reference bit is written into the pending result. Here, "at least one corresponding bit" means that in addition to the remaining deviation result where the reference bit is located, there is at least one other remaining deviation result with a bit that is completely identical to the reference bit. This ensures that each bit written into the same-position scene result is not a change unique to a single label, but a change that appears together in multiple other labels. After all reference bits are processed, the same-position scene result retains the deviation bits that are completely identical in position, and the pending result retains the remaining deviation bits that are not completely identical in position. Furthermore, the displacement difference is calculated for each bit in the result to be judged, and the consecutive position changes with continuous preceding positions and consistent displacement directions are extracted to form the consecutive position scene result. In specific processing, for each bit in the result to be judged, if the current position is empty, the displacement difference of that bit is recorded as empty; if the current position is not empty, the displacement difference of that bit is determined as the current position minus the preceding position. Subsequently, using the deviation type and displacement difference of that bit as the comparison basis, the corresponding bit in the result to be judged is searched one by one for the bit with the same deviation type, the same displacement difference and adjacent preceding positions. Here, adjacent preceding positions refer to the absolute value of the difference between the preceding positions of the two bits. =1; if the corresponding bit is found, the bit is connected to each corresponding bit in ascending order of the preceding position and written into the linked position scenario result; if the corresponding bit is not found, the bit is written into the reserved result; for missing bits, when the current position of both bits is empty, the displacement difference between the two bits is considered the same; for the case where the current position of one bit is empty and the current position of the other bit is not empty, the displacement difference between the two bits is considered different; through the above processing, the linked position scenario result retains the deviation bits that, although they do not form a complete same-position repetition, show the same type of deviation in multiple other tags and are continuously expanded along the preceding position, while the reserved result retains the deviation bits that do not form a same-position repetition and do not form a linked position continuation; Based on this, each bit in the target deviation result is compared sequentially with the corresponding scene result and the consecutive scene result to determine the scene deviation result and the label retention result. Specifically, first, one bit in the target deviation result is read, and the corresponding bit with the same deviation type, the same preceding position, and the same current position is searched in the corresponding scene result. If found, the bit is written into the scene deviation result. If not found, the corresponding bit with the same deviation type, the same displacement difference, and the same preceding position is searched in the consecutive scene result. When the corresponding bit is found, it is further checked whether the bit adjacent to the preceding position of the bit in the target deviation result also exists and is written into the consecutive scene result. Specifically, the process proceeds by first checking the previous position, and then checking the next position if the previous position does not meet the requirements. If either the previous or next position contains a bit that should be written into the connected scene result, that bit is written into the scene deviation result. If neither the previous nor the next position contains a bit that should be written into the connected scene result, that bit is written into the label retention result. If no corresponding bit is found in the connected scene result, that bit is directly written into the label retention result. After processing all bits in the target deviation result according to this rule, the scene deviation result retains the scene change bits that appear together with the other tags, while the label retention result retains the target tag's own change bits that are not absorbed by the common scene changes. Finally, the label retention results are sorted in ascending order of their preceding positions to generate label deviation results. During sorting, the deviation type, preceding position, and current position of each bit are not changed; only their order within the result sequence is adjusted so that the label deviation results can be directly used as input for subsequent position rearrangement. If the label retention result is empty, the label deviation result is recorded as an empty sequence. If the label retention result contains only one bit, that unique bit directly constitutes the label deviation result. If the label retention result contains multiple bits, the bit with the smaller preceding position is placed first, followed by the bit with the larger preceding position. Through this processing, the label deviation results maintain consistency with the aforementioned target deviation results and scene deviation results in terms of field scope, retaining only the deviation bits that were not absorbed by the common changes in the scene, thus providing a direct basis for subsequent position rearrangement based on the label deviation results. Through the above processing, without changing the generation criteria of the target deviation result, the deviation positions of multiple other labels can be organized into same-position scene results and consecutive-position scene results. Based on this, the common changes of the scene can be extracted from the target deviation result, while the changes of the target label itself can be retained. This makes the formation process of the scene deviation result and the label deviation result have clear repetition, continuation and diversion criteria. This can reduce the interference of occasional deviations of individual other labels on scene judgment, and also avoid incorrectly retaining deviation positions that should belong to the common changes of the scene in the label deviation result. This ensures that subsequent reading of position reordering is based on the label's own deviation. In practical applications: if the remaining deviation results corresponding to the three remaining labels are "missing position of the previous position 2, position change from the previous position 4 to the current position 5", "missing position of the previous position 2, position change from the previous position 4 to the current position 5", and "position change from the previous position 3 to the current position 4, position change from the previous position 4 to the current position 5", then the missing position of the previous position 2 and the position change from the previous position 4 to the current position 5 can both be written into the same-position scenario result because they appear repeatedly in multiple remaining deviation results; if one of the positions to be judged is the position change from the previous position 5 to the current position 6, and another position to be judged is the position change from the previous position 6 to the current position 7, then... If both have the same deviation type, a displacement difference of 1, and adjacent preceding positions, they can be connected and written into the linked position scenario result. If the target deviation result contains a missing position of preceding position 2, a position change position from preceding position 4 to current position 5, and a position change position from preceding position 7 to current position 9, then the first two positions can be found in the same position scenario result and written into the scenario deviation result. The position change positions from preceding position 7 to current position 9 cannot be found in the same position scenario result and do not meet the continuation condition in the linked position scenario result, so they are written into the label retention result. Finally, the label retention result is sorted by preceding position in ascending order to form the label deviation result, which is used as the input for subsequent position rearrangement.

[0021] S4. Based on the label deviation result, rearrange the reading order of the target label in the current position sequence so that each bit corresponding to the label deviation result is read again before the other bits. If the target label is not obtained in the second reading, the target label is read separately to generate the recognition result. In this implementation, the label deviation result is used to guide the order of rereading the target label in the current position sequence. Therefore, it is necessary to first convert each missing bit and each position change bit in the label deviation result into a rearranged position that can be directly applied to the current position sequence, and then perform rereading according to the order of the rearranged positions. If the target label is still not obtained after rereading, non-target interference items are gradually eliminated in the current position sequence, and the rearranged positions are reread after each elimination, so that the target label transitions from a group reading state to an individual reading state, and finally forms a recognition result. The key point of this process is to first use the label deviation result to determine the priority processing position, and then compress the interference range in the current position sequence by gradually deleting the label identifiers outside the rearranged position sequence, so that the hit position of the target label in the current position sequence can be stably determined, and provide a direct basis for rewriting the initial reading position in the next scenario. The implementation process includes the following steps: First, each missing position and each position change position in the label deviation result is mapped to its corresponding position in the current position sequence to generate a rearranged position sequence. Specifically, for each missing position, its preceding position is read, and it is checked whether a position with the same preceding position value exists in the current position sequence. If it exists, the missing position is mapped to that position; otherwise, it is mapped to an appended position after the end of the current position sequence, where the appended position is determined by adding one to the position of the last position in the current position sequence. For each position change position, its current position is read, and the position change is... The delimited bits are mapped to the corresponding positions in the current position sequence that have the same frequency value as the current position. If multiple missing bits or multiple position change bits are mapped to the same position, only one of the positions is retained. After all mappings are completed, the deduplication of each position obtained by mapping is removed, and the positions are arranged in ascending order to form a rearranged position sequence. Through this process, the label deviation result is converted into a position order that can be directly used for rereading, where missing bits are preferentially placed in the position corresponding to their previous position, and position change bits are preferentially placed in the position corresponding to their current position, so that subsequent rereadings will revolve around the position where the target label has deviated. Subsequently, the rearranged position sequence guides the reading process again to determine if the target tag can be read again within the priority position range. Specifically, the first position in the rearranged position sequence is determined as the starting position, and the tag identifier corresponding to the starting position is read from the current position sequence. If the reading result is the target tag, the starting position is determined as the hit position, and the reading ends. If the reading result is not the target tag, the reading continues sequentially according to the subsequent positions in the rearranged position sequence. Each time a reading is performed, the tag identifier corresponding to the current reading position is compared with the target tag. If the comparison results are the same, immediately stop reading subsequent positions and determine the current position as the hit position; if the comparison results are different, continue processing the next position; if the target label is not read after processing all positions in the rearranged position sequence, the result of the second reading is determined as a miss; the second reading here means reading each corresponding position in the current position sequence in the order given by the rearranged position sequence, rather than regenerating a new current position sequence, so as to ensure that the position correspondence between the rearranged position sequence and the current position sequence remains unchanged. Furthermore, when the read result is a miss, the tag identifiers outside the rearranged position sequence in the current position sequence are deleted sequentially in their original position order. After each tag identifier is deleted, the rearranged position sequence is read again to generate a separate read result. Here, deletion refers to removing the data record corresponding to the tag identifier from the current position sequence, so that the data record will no longer participate in subsequent reads, without involving the physical removal of the graphene electronic tag. In specific processing, first, all tag identifiers outside the rearranged position sequence are found in the current position sequence and arranged in ascending order of their original position to form a sequence to be deleted. Then, starting from the first position of the sequence to be deleted, the data records corresponding to each tag identifier are deleted sequentially. Each time a tag identifier is deleted... After identification, the original count value in the rearranged sequence remains unchanged, and the deleted data records are skipped. The rearranged sequence is read again. When the target label is read after the re-execution, the current hit position is determined as the hit position in the individual read result, and subsequent deletion stops. If the target label is still not read after the re-execution, the next label identifier in the sequence to be deleted is deleted. If all label identifiers in the sequence to be deleted are deleted and the target label is still not read, the individual read result is determined as a miss. Through this process, non-target interference items in the current position sequence are gradually eliminated. If the target label is not displayed in the re-read due to interference from adjacent labels, it can be reconfirmed after the current position sequence gradually converges. Finally, the recognition result is generated based on the rereading result and the individual reading result. In specific processing, when the rereading result is the target label, the hit position obtained during the rereading process is written as the recognition result; when the rereading result is not the target label but the individual reading result is the target label, the hit position obtained during the individual reading process and the current position sequence before the deletion ends are written as the recognition result, where the current position sequence before the deletion ends refers to the data record sequence retained before the last deletion is completed and the rereading is performed; when neither the rereading result nor the individual reading result is the target label, the miss is written as the recognition result. After adopting the above writing method, the recognition result can reflect whether the target label is directly hit within the rearranged position range or is hit again after gradually deleting interference items, and can also provide a unified result caliber for subsequent review and rewriting of the initial reading position in the next scenario when there is a miss. Through the above processing, the label deviation result can be directly transformed into the basis for rearranging the position for rereading. When the reordering read fails, the current position sequence is compressed by deleting non-target interference items, so that the target label gradually transitions from priority position reading to individual reading, thereby improving the certainty of the recognition result. At the same time, the hit position and the current position sequence before the deletion ends are explicitly written into the recognition result, so that the subsequent initial reading position rewriting has a directly callable input basis. In practical applications: If the label identifiers corresponding to each position in the current position sequence are "Label A, Label B, Target Label, Label C, Label D" in sequence, and the label deviation result contains the missing position corresponding to the previous position 2 and the position change position corresponding to the current position 3, then first map the missing position to position 2 in the current position sequence, map the position change position to position 3 in the current position sequence, and after deduplication, form the rearranged position sequence "2, 3"; then, perform the reading again according to position 2 and position 3 in sequence. If the reading result corresponding to position 2 is label B and the reading result corresponding to position 3 is the target label... If a label is found, then position 3 is written as the recognition result. If the target label is still not found after reading again by position 2 and position 3, then the label identifiers corresponding to positions 1, 4, and 5 are deleted from the current position sequence in the original position order, and after each label identifier is deleted, the reading is performed again by position 2 and position 3. For example, if the target label is not found after deleting label A, and the target label is found again by reading again by position 3 after deleting label C, then position 3 and the current position sequence after deleting label C and before deleting label D are written together as the recognition result, so that the initial reading position rewriting can be directly called in the next scenario.

[0022] S5. Rewrite the initial reading position of the target label in the next scene according to the recognition result, so that when the target label enters the next scene again, it forms the corresponding current sequence according to the rewritten initial reading position. In this embodiment, the recognition result is used to transform the recognition conclusion in the current scene into the basis for adjusting the reading order in the next scene, so that when the target tag re-enters the next scene, it can participate in continuous reading with an updated initial reading position and form a current sequence corresponding to that position. This process does not only record the recognition situation of the previous scene, but also rewrites the hit information or miss information obtained in the current scene into position adjustment actions in the next scene, so that the entry reading position of the target tag in the next scene changes accordingly, while maintaining the continuity and clear order of the position relationship of the other tags in the next scene. To ensure that the reading order in the subsequent scene can be directly used for subsequent recognition, the rewritten position is first determined based on the recognition result. Then, the rewritten position is compared with the original initial reading position in the subsequent scene and the position is shifted. Subsequently, the initial reading position sequence in the subsequent scene is generated. When the target label re-enters the subsequent scene, the corresponding current sequence is formed according to this initial reading position sequence. This implementation process includes the following steps: First, the rewrite position is determined based on the recognition result. Specifically, when the recognition result is a confirmation result, the target label's reread position in the current position sequence is read and written as the rewrite position. Here, the reread position refers to the position in the current position sequence where the target label is actually hit after being read again or read separately. When the recognition result is a verification result, the target label's preceding position in the previous scene is read and written as the rewrite position. The position already formed by the target label in the previous scene is used as the basis for backtracking in the next scene. Through the above processing, both the confirmation and verification results are converted into numerical positions that can directly participate in the rewrite of positions in the next scene, thus ensuring that subsequent position comparisons are based on the same value. Subsequently, the rewritten position is compared with the original initial read position of the target tag in the next scene, and the rewritten initial read position is generated based on the comparison result. Specifically, the original initial read position in the next scene is first determined. If the target tag has already established a read order in the next scene, the position corresponding to the target tag in that read order is determined as the original initial read position. If the target tag has not yet established a read order in the next scene, the position after the last position of the current read order in the next scene is determined as the original initial read position. Then, a position comparison is performed: if the rewritten position is less than the original initial read position, the positions in the next scene between the rewritten position and the position preceding the original initial read position are compared. The positions are shifted one position to the right in ascending order, and then the rewritten position is written to the target label. When the rewritten position is greater than the original initial read position, all positions in the next scene between the original initial read position and the rewritten position are shifted one position to the left in ascending order, and then the rewritten position is written to the target label. When the rewritten position is equal to the original initial read position, the position of the target label in the next scene remains unchanged, and this position is directly determined as the rewritten initial read position. Through this process, the position adjustment of the target label in the next scene is implemented as a clear position shift process, and the position changes of other labels have a unique destination, so that the same position will not be repeatedly occupied or the vacancy will not be filled. Then, the rewritten initial read positions are written into the read order of the target tag in the next scene, and the read positions of the remaining tags in the next scene are sorted to generate the initial read position sequence in the next scene. In specific processing, the target tag is first written into the rewritten initial read positions, and then the remaining tags are rearranged in ascending order to the remaining positions except for the rewritten initial read positions. For the relative order between the remaining tags, the original order before the position rewriting remains unchanged, and only the vacant positions after the target tag position change are filled in sequentially. This yields a complete initial reading position sequence containing the target label and the remaining labels. If the target label's position moves forward, the remaining labels originally located between the rewritten position and the position before the original initial reading position move backward. If the target label's position moves backward, the remaining labels originally located between the position after the original initial reading position and the rewritten position move forward. After all positions are sorted, each label in the next scene corresponds to a unique reading position, thus forming the initial reading position sequence in the next scene, which serves as the basis for the reading order when the target label re-enters the next scene. Finally, when the target tag re-enters the next scene, the reading area is continuously read according to the initial reading position sequence, and the current sequence is generated based on the sequence fragment corresponding to the position of the target tag when the target tag is read. Specifically, the tags in the next scene are read continuously according to their positions in the initial reading position sequence. When the target tag is read, its position is first determined, and then the corresponding sequence fragment is extracted around that position. When the target tag's position is in the middle of the sequence, the tag identifier of the position before that position, the target tag identifier of that position, and the tag identifier of the position after that position are extracted to form the corresponding sequence fragment. When the target tag's position is at the beginning, the target tag identifier of that position and the tag identifier of the position after that position are extracted to form the corresponding sequence fragment. The label identifiers in the position form the corresponding sequence fragments; when the target label is at the last position, the label identifier in the previous position and the target label identifier in this position are extracted to form the corresponding sequence fragments; when there are no other labels before and after the target label, only the target label identifier is used to form the corresponding sequence fragments; if the target label appears multiple times during continuous reading, the sequence fragments that appear later and have the same adjacent label identifiers are retained according to the segment retention rules corresponding to the current position sequence, and the corresponding current sequence is generated from the retained sequence fragments; thus, the reading order adjustment result in the latter scenario is further implemented in the current sequence generation process of the target label, and is consistent with the organization of the current position sequence, current sequence, and previous sequence mentioned above. Through the above processing, the recognition result in the current scene can be transformed into the position adjustment result in the next scene, and further implemented into the initial reading position sequence and the corresponding current sequence in the next scene, so that the reading order of the target label in different scenes forms a continuous relationship. In practical applications: If the target label's recognition result in the current scene is a confirmed result, and its read position in the current position sequence is 2, while the target label's original initial read position in the next scene is 5, then position 2 is first written as the rewritten position. Then, the remaining labels originally located at positions 2, 3, and 4 in the next scene are sequentially moved to positions 3, 4, and 5, and the target label is written into position 2, forming the rewritten initial read position. Subsequently, the original order of the remaining labels is kept unchanged, generating the initial read position sequence in the next scene. If the target label re-enters the next scene, continuous reading is performed according to this initial read position sequence, and the target label is read at position 2, then the label identifiers corresponding to the position before and after the target label at position 2 are extracted to form a sequence fragment. The corresponding current sequence is then generated from this sequence fragment for subsequent position-by-position comparison between the previous sequence and the current sequence.

[0023] Furthermore, a multi-scenario adaptive recognition method for ultra-high frequency graphene electronic tags includes: The sequence extraction module is used to continuously read the graphene electronic tags in the current scene, form the current position sequence, and extract the current sequence corresponding to the target tag from the current position sequence; The deviation generation module is used to compare the previous sequence of the target label formed in the previous scene with the current sequence position by position, and write the positions that exist in the previous sequence but are missing in the current sequence, as well as the positions that exist in both the previous sequence and the current sequence but are different in order, into the target deviation result. The scene stripping module is used to compare the remaining preorder sequences formed by the remaining labels in the previous scene with the corresponding remaining current sequences in the current position sequence, write the missing positions and position change positions that repeatedly appear in the comparison process between the remaining preorder sequences and the remaining current sequences into the scene deviation result, and delete each position in the scene deviation result from the target deviation result to generate the label deviation result. The position rearrangement module rearranges the position of the target label in the current position sequence according to the label deviation result, so that the position corresponding to the label deviation result is read again before the other positions. If the target label is not obtained in the second reading, the target label is read separately to generate the recognition result. The sequence rewriting module rewrites the initial reading position of the target label in the next scene based on the recognition result, so that when the target label re-enters the next scene, it forms the corresponding current sequence according to the rewritten initial reading position.

[0024] Working Principle: This solution first analyzes the reading performance of the target tag in different scenarios, then separates the changes truly belonging to the target tag itself from the common changes in the scenarios, and finally adjusts the recognition order accordingly. Specifically, it first continuously reads the set of graphene electronic tags passing through the reading area in the current scenario to form the current position sequence, and extracts the current sequence corresponding to the target tag from it; then, it compares the current sequence with the preceding sequence of the target tag in the previous scenario position by position to find missing positions and shift positions, forming the target deviation result; next, it calculates the changes of other tags in adjacent scenarios simultaneously, and filters out the isotopic changes that occur together by multiple tags. The transformation and linking changes are treated as scene deviation results, which are then separated from the target deviation results, leaving the label deviation results. Subsequently, the reading positions are rearranged according to the label deviation results, prioritizing the reading of positions more likely to re-hit the target label. If it still fails to hit, non-target interference items are gradually removed, and individual readings are performed to obtain the recognition results. Finally, the initial reading positions of the target label in the next scene are rewritten based on the recognition results, so that the target label is read earlier and more accurately in the next scene. The whole process is connected and is essentially an adaptive recognition process that continuously corrects the reading positions, which can also provide a continuous result feedback basis for reinforcement learning. For example, in a cold chain medical consumables turnover scenario, the same graphene electronic tag may first pass through the packing station, then through the turnover channel, and finally enter the sorting station. Due to box stacking, tag bending, nearby metal frames, or crowding of surrounding tags, the order in which the target tag appears in different stations may change, or even disappear for a short time. This solution first records the previous sequence of the target tag at the packing station. After the target tag enters the next station, it extracts the current sequence from the current position sequence and compares it position by position with the previous sequence. If the position change of the target tag is found to also appear in multiple other tags, it is judged that this is more like a perturbation of the overall scene. If it only appears in the target tag, it is retained as a tag deviation result. Then the system prioritizes reading the target tag again near these deviation positions. If it is still not read, it gradually eliminates the interference of other tags in the current position sequence and then confirms the target tag individually. Once confirmed, the position of this hit is rewritten into the initial reading order of the next scene. In this way, when the target tag enters a similar scene again, the system will place it in the priority reading position earlier, thereby continuously providing usable state changes and action feedback for reinforcement learning.

[0025] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-scenario adaptive recognition method for ultra-high frequency graphene electronic tags, characterized in that, include: S1. In the current scenario, continuously read the set of graphene electronic tags that have passed through the reading area to form the current position sequence, and extract the current sequence corresponding to the target tag from the current position sequence; S2. Compare the target label in the previous sequence with the current sequence position by position, and write the positions that exist in the previous sequence but are missing in the current sequence, as well as the positions that exist in both the previous sequence and the current sequence but are different in order, into the target deviation result. S3. Compare the remaining preceding sequences formed by the remaining labels in the previous scene with the remaining current sequences corresponding to the current position sequence, and write the missing positions and position changes that occur repeatedly in the comparison process between the remaining preceding sequences and the remaining current sequences into the scene deviation result. Delete each position from the scene deviation result from the target deviation result to generate the label deviation result. S4. Based on the label deviation result, rearrange the reading order of the target label in the current position sequence so that the bits corresponding to the label deviation result are read again before the other bits. If the target label is not obtained in the second reading, the target label is read separately to generate the recognition result. S5. Based on the recognition results, rewrite the initial reading position of the target label in the next scene so that when the target label re-enters the next scene, it forms the corresponding current sequence according to the rewritten initial reading position.

2. The multi-scene adaptive recognition method for ultra-high frequency graphene electronic tags according to claim 1, characterized in that: S1 includes: S1-1. In the current scenario, perform multiple rounds of reading on the reading area. The graphene electronic tag identifiers obtained in each round of reading are arranged into a corresponding round sequence according to their appearance order. The corresponding round sequences are then spliced ​​together according to the reading order to form the current position sequence. S1-2. Take the position of each target label in the current position sequence as the target position, and extract the adjacent label before and after each target position to form a corresponding segment with the target label to form a target segment sequence. S1-3. Compare adjacent segments in the target segment sequence one by one. If the target label of two adjacent segments is the same and the adjacent label is the same, retain the next corresponding segment. Then, splice the retained corresponding segments in the order of their appearance to form the current sequence corresponding to the target label.

3. The multi-scene adaptive recognition method for ultra-high frequency graphene electronic tags according to claim 2, characterized in that: S2 includes: S2-1. Write the digits of the previous sequence into the previous position according to their order of appearance, and write the digits of the current sequence into the current position according to their order of appearance, to form the previous position sequence and the current position sequence. S2-2. Starting from the first position of the preceding position sequence, read the corresponding tag identifier of each position one by one, and search for the same unmatched position in the current position sequence in ascending order of the current position to form the matching position corresponding to each preceding position. S2-3. Determine the preceding positions that have not formed a matching position as missing positions, and determine the preceding positions that have formed a matching position and whose preceding position is different from the current position corresponding to the matching position as shift positions, forming a deviation position sequence. Arrange each missing position and each shift position in the deviation position sequence in ascending order of the preceding position to generate the target deviation result.

4. The multi-scene adaptive recognition method for ultra-high frequency graphene electronic tags according to claim 3, characterized in that: S3 includes: S3-1. Compare each of the remaining preceding sequences with each of the remaining current sequences, determine the missing positions and position changes for each remaining label, and write the missing positions and position changes corresponding to each remaining label into the remaining deviation results according to the deviation type, the preceding position and the current position. S3-2. Take each bit in the remaining deviation results as the reference bit, and search for the corresponding bit in the remaining deviation results that has the same deviation type, the same previous position, and the same current position as the reference bit. When at least one corresponding bit is found, write the reference bit and each corresponding bit into the same position scenario result; otherwise, write the reference bit into the pending result. S3-3. Determine the displacement difference for each bit in the result to be judged. Record the displacement difference of the bit whose current position is empty as empty, and record the displacement difference of the bit whose current position is not empty as the current position minus the previous position. Search for the corresponding bit in the result to be judged one by one with the same deviation type, the same displacement difference and adjacent previous position. When the corresponding bit is found, connect the bit with each corresponding bit according to the previous position and write it into the connected bit scenario result. Otherwise, write the bit into the reserved result.

5. The multi-scene adaptive recognition method for ultra-high frequency graphene electronic tags according to claim 4, characterized in that: S3 further includes: S3-4. Compare each bit in the target deviation result with the same-position scene result and the consecutive-position scene result one by one. If there is a corresponding bit in the same-position scene result with the same deviation type, the same preceding position, and the same current position, write the bit into the scene deviation result. Otherwise, if there is a corresponding bit in the consecutive-position scene result with the same deviation type, the same displacement difference, and the same preceding position, and the previous or next position of the bit is also written into the consecutive-position scene result in the target deviation result, write the bit into the scene deviation result. Otherwise, write the bit into the label retention result. S3-5. Sort the label retention results in ascending order of their previous positions to generate the label deviation results.

6. The multi-scene adaptive recognition method for ultra-high frequency graphene electronic tags according to claim 5, characterized in that: S4 includes: S4-1. Map each missing position in the label deviation result to the same position in the current position sequence according to the previous position. Map each position change position in the label deviation result to the corresponding position in the current position sequence according to the current position. After removing duplicates from the mapped positions, arrange them in ascending order of position to generate a rearranged position sequence. S4-2. Write the first digit of the rearranged position sequence as the starting position, read the label corresponding to the starting position from the current position sequence, and if the reading result is not the target label, execute the reading in the subsequent positions of the rearranged position sequence in turn until the reading result is the target label or the rearranged position sequence is executed, and generate the reading result again.

7. The multi-scene adaptive recognition method for ultra-high frequency graphene electronic tags according to claim 6, characterized in that: S4 further includes: S4-3. When the result of the second reading is not the target label, delete the label identifiers in the current position sequence that are outside the rearranged position sequence in the original position sequence in the current position sequence. After deleting each label identifier, perform a second reading on the rearranged position sequence until the reading result is the target label or the deletion of the label identifiers outside the rearranged position sequence ends, and a separate reading result is generated. S4-4. When the result of the second reading is the target label, write the hit position corresponding to the target label as the recognition result. When the result of the single reading is the target label, write the hit position corresponding to the target label and the current position sequence before the deletion ends as the recognition result. When neither the result of the second reading nor the result of the single reading is the target label, write the miss as the recognition result.

8. The multi-scene adaptive recognition method for ultra-high frequency graphene electronic tags according to claim 7, characterized in that: S5 includes: S5-1. When the recognition result is a confirmed result, the reread position of the target label in the current position sequence is written as the rewritten position. When the recognition result is a verification result, the previous position of the target label in the previous scene is written as the rewritten position. S5-2. Compare the rewritten position with the original initial read position of the target tag in the next scene. If the rewritten position is less than the original initial read position, shift each position in the next scene between the rewritten position and the position before the original initial read position one position to the right and write it into the rewritten position. If the rewritten position is greater than the original initial read position, shift each position in the next scene between the position after the original initial read position and the rewritten position one position to the left and write it into the rewritten position to generate the rewritten initial read position.

9. The multi-scene adaptive recognition method for ultra-high frequency graphene electronic tags according to claim 8, characterized in that: The S5 also includes: S5-3. Write the rewritten initial read position into the read order of the target tag in the next scene, and rearrange the read positions of the remaining tags in the next scene in ascending order of read position to generate the initial read position sequence in the next scene. S5-4. When the target label re-enters the next scene, the reading area is continuously read according to the initial reading position sequence, and when the target label is read, the corresponding current sequence is generated using the sequence fragment corresponding to the position of the target label.

10. A multi-scene adaptive identification system for ultra-high frequency graphene electronic tags, characterized in that, include: The sequence extraction module is used to continuously read the graphene electronic tags in the current scene, form the current position sequence, and extract the current sequence corresponding to the target tag from the current position sequence; The deviation generation module is used to compare the previous sequence of the target label formed in the previous scene with the current sequence position by position, and write the positions that exist in the previous sequence but are missing in the current sequence, as well as the positions that exist in both the previous sequence and the current sequence but are different in order, into the target deviation result. The scene stripping module is used to compare the remaining preorder sequences formed by the remaining labels in the previous scene with the corresponding remaining current sequences in the current position sequence, write the missing positions and position change positions that repeatedly appear in the comparison process between the remaining preorder sequences and the remaining current sequences into the scene deviation result, and delete each position in the scene deviation result from the target deviation result to generate the label deviation result. The position rearrangement module rearranges the position of the target label in the current position sequence according to the label deviation result, so that the position corresponding to the label deviation result is read again before the other positions. If the target label is not obtained in the second reading, the target label is read separately to generate the recognition result. The sequence rewriting module rewrites the initial reading position of the target label in the next scene based on the recognition result, so that when the target label re-enters the next scene, it forms the corresponding current sequence according to the rewritten initial reading position.

Images