A circulating token method based on a node hash chain

Abstract

The application belongs to the technical field of data processing, and particularly relates to a circulation evidence method based on a node hash chain. The method acquires a business field, a packaging field and an initial voucher digest, and generates a circulation object identity anchor; generates a node challenge code based on a previous node hash, the circulation object identity anchor and a subject state vector, and binds the node challenge code with a multi-format voucher to generate a challenge code verification result; converts the multi-format voucher into a typed evidence leaf node and constructs a cross-format evidence root; determines a cross-format evidence closed residual error according to a consistency result and the challenge code verification result; generates a handover responsibility edge and a responsibility edge tension value in combination with a handover state, and generates a node hash or an abnormal bifurcation node for replay verification according to the same. The method can provide replayable positioning basis for chain breaking, mismatching, replacement and responsibility attribution.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a circulation and evidence preservation method based on node hash chains. Background Technology

[0002] Circulation record keeping is used to solidify the handover records of goods, materials, or vouchers during cross-entity circulation, enabling subsequent verification to trace back to specific circulation nodes. Existing technologies typically upload and archive information such as order numbers, waybill numbers, sender and recipient entities, images, videos, PDFs, manifests, or waybills separately, or generate a summary for a single file and write it into the record keeping record. Some solutions also use blockchain or hash values ​​to fix the file content, proving that the document has not been tampered with after submission.

[0003] However, in the continuous circulation process of manufacturer shipment, logistics receipt, logistics transshipment, and distributor receipt, multi-format vouchers are not isolated files, but rather collectively describe the same circulating object, the same packaging state, and the same handover behavior. Existing methods often use a single voucher, a single business number, or a single upload record as the evidence storage object, lacking an object foundation that binds business fields, packaging fields, and the initial voucher summary together. This leads to vouchers subsequently uploaded by different entities being valid only at the file level, failing to prove they all point to the same circulating object. Existing methods also lack a unified closed-loop judgment on the consistency of fields, regions, and challenge code verification results among image vouchers, video vouchers, PDF vouchers, manifest vouchers, and waybill vouchers, making it difficult to promptly detect issues such as old voucher reuse, field mismatches, or inconsistencies in seals, waybills, and box areas. Furthermore, existing evidence storage records typically only retain the upload time and file summary, failing to establish a replayable handover responsibility relationship between adjacent circulation nodes regarding quantity confirmation results, seal status, outer packaging status, and objection status. This results in unstable criteria for locating broken chains, mismatches, replacements, or attribution of responsibility. Summary of the Invention

[0004] This invention provides a circulation evidence preservation method based on node hash chains, which is used to at least solve the problems of how to uniformly preserve the continuity of evidence, consistency of documents, and handover responsibility of the same circulating object under the conditions of multi-subject, multi-node, and multi-format document circulation, and to locate chain breaks, mismatches, replacements, or responsibility attribution.

[0005] This invention provides a circulation evidence storage method based on a node hash chain. The method includes: obtaining business fields, encapsulated fields, and an initial credential digest to generate a circulation object identity anchor; generating a node challenge code based on the previous node hash, the circulation object identity anchor, and the subject state vector; binding the node challenge code to a multi-format credential and generating a challenge code verification result; converting the multi-format credential into a typed evidence leaf node and constructing a cross-format evidence root; generating a consistency result based on the typed evidence leaf node, and determining the cross-format evidence closure residual based on the consistency result and the challenge code verification result; generating a handover responsibility edge and responsibility edge tension value based on the cross-format evidence closure residual and the handover state; determining the sealing state based on the cross-format evidence closure residual and the responsibility edge tension value; generating a node hash to access the node hash chain when the sealing state meets the sealing conditions; generating an abnormal fork node when the sealing state does not meet the sealing conditions; and performing replay verification based on the node hash chain and the handover responsibility edge. In the case of generating an abnormal fork node, determining the chain break, mismatch, replacement, or responsibility attribution result based on the abnormal fork node.

[0006] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows: By generating identity anchors for circulation objects through business fields, encapsulation fields, and initial voucher summaries, unified binding of circulation objects in different circulation nodes is achieved, enabling subsequent multi-format vouchers to be stored and verified around the same object base.

[0007] By generating a node challenge code using the hash of the previous node, the identity anchor of the circulating object, and the state vector of the subject, and binding the node challenge code to a multi-format certificate, the submission of the certificate of the current circulating node can be associated with the operation sequence of the previous circulating node, the circulating object, and the circulating subject, thereby reducing the verification deviation caused by the reuse of old certificates or cross-node replacement.

[0008] By converting multi-format vouchers into typed evidence leaf nodes and constructing cross-format evidence roots, image vouchers, video vouchers, PDF vouchers, manifest vouchers, and waybill vouchers can be incorporated into a unified data structure for subsequent processing.

[0009] By determining the cross-format evidence closure residuals through consistency results and challenge code verification results, inconsistencies in fields, image regions, and node challenge codes can be transformed into the same judgment criteria.

[0010] By handing over the responsibility edge and the responsibility edge tension value, the internal consistency deviation of the voucher and the handover status deviation of adjacent circulation nodes are used together to determine the sealing status, so that the node hash chain and abnormal fork nodes can provide a stable basis for replay verification. Attached Figure Description

[0011] Figure 1 This is a schematic flowchart of the method of the present invention; Figure 2 This is a record diagram of the generation of identity anchors and node challenge codes for circulating objects in this embodiment of the invention; Figure 3 This is a diagram showing the construction of typified evidence leaf nodes and cross-format evidence root nodes in an embodiment of the present invention. Figure 4 This is a diagram showing the field consistency matrix and region consistency results in an embodiment of the present invention; Figure 5 This is a diagram showing the handover responsibility edge and sealing status determination in an embodiment of the present invention. Detailed Implementation

[0012] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of one or more embodiments. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of this invention.

[0013] In this embodiment, the node hash chain uses the node hashes formed by the circulating object at each circulating node as the chained processing object, with each node hash corresponding to the evidence storage result of a circulating node. The data sources of the circulating nodes include business fields, encapsulation fields, initial certificate digests, multi-format certificates, subject state vectors, handover states, and the hash of the previous node. The circulation evidence storage method generates the circulating object identity anchor, node challenge code, typed evidence leaf node, cross-format evidence root, cross-format evidence closure residual, handover responsibility edge, and responsibility edge tension value within each circulating node, and determines the node hash or abnormal fork node that the current circulating node connects to the node hash chain based on the closure state. The node hash chain serves as the basis for subsequent replay verification, used to verify the hash inheritance relationship between nodes in the order of circulating nodes, and combined with the handover responsibility edge and abnormal fork node to determine chain breakage, mismatch, replacement, or responsibility attribution.

[0014] like Figure 1 As shown, a circulation evidence storage method based on node hash chains includes the following processing steps.

[0015] During the stage of generating the identity anchor of the circulation object, the circulation evidence storage method receives business fields, encapsulation fields, and initial credential digest.

[0016] The business field is used to describe the transaction and product attributes of the circulating object, the encapsulation field is used to describe the external encapsulation and logistics identification of the circulating object, and the initial document summary is used to fix the content status of the initial document.

[0017] After business fields and encapsulated fields undergo standardized processing, field names, field order, character format, and null value markers are unified under the same processing rules, preventing different identity anchors from being generated when different entities upload fields due to differences in capitalization, spaces, separators, or field order.

[0018] After normalization, the normalized fields are combined with the initial credential digest in a fixed order, and the digest is used to generate the circulation object identity anchor. The circulation object identity anchor serves as the common binding object for subsequent node challenge codes, cross-format evidence roots, and node hash chains, enabling multi-format credentials submitted by different circulation nodes to continuously point to the same circulation object.

[0019] In one embodiment, the order number, batch number, product model and quantity are extracted from the business field, and the box number, seal number and waybill number are extracted from the packaging field.

[0020] The order number is used to distinguish the transaction to which the circulating object belongs; the batch number is used to distinguish the production or outbound batches under the same transaction; the product model is used to define the product attributes of the circulating object; and the quantity is used to define the quantity boundary of the circulating object. The box number is used to define the packaging unit where the circulating object is located; the seal number is used to define the sealing status of the packaging unit; and the waybill number is used to define the external identification of the circulating object during the logistics process.

[0021] After field extraction, standardization is performed according to the principles of fixed field names, fixed field values, and fixed field order. Fixed field names mean that order number, batch number, product model, quantity, carton number, seal number, and waybill number are all stored using pre-defined field names to avoid name differences when the same field is uploaded by different distribution entities. Fixed field values ​​mean that leading and trailing spaces, full-width characters, half-width characters, uppercase and lowercase characters, date separators, and consecutive whitespace in field values ​​are uniformly processed, and quantities are stored using the same unit of measurement and the same numerical format. Fixed field order means that the fields are arranged in the order of order number, batch number, product model, quantity, carton number, seal number, and waybill number to form a field sequence.

[0022] In cases of missing fields, the normalization process does not delete the missing fields but instead writes a missing field marker at the same location. The missing field marker is used to preserve the structural stability of the field sequence, enabling subsequent circulation nodes to identify the missing location and incorporating the missing information into the judgment basis for subsequent cross-format evidence closure residuals. If any field among order number, batch number, box number, seal number, and waybill number has multiple values, the circulation evidence preservation method determines the valid value based on the matching relationship between submission time, submitting entity, and initial document summary; if a valid value cannot be determined, multiple values ​​are retained and a conflict marker is written, which is then used in subsequent field consistency judgments.

[0023] The initial document summary is obtained by summarizing the initial document. The initial document can originate from the outbound list, packing image, seal image, waybill document, or electronic confirmation document. Before or simultaneously with generating the initial document summary, the circulation evidence preservation method extracts the initial image region summary from the initial document and associates the initial image region summary with the initial document summary, writing it into the initial document record. During subsequent region consistency checks, the circulation evidence preservation method reads the initial image region summary from the initial document record, without extracting the image region content from the initial document summary. The initial document summary does not replace business fields and encapsulation fields; instead, it serves as a basis for binding content beyond field combinations, preventing the repeated generation of the same circulation object identity anchor based solely on fields.

[0024] The identity anchor of the circulating object can be generated in the following way:

[0025] in, As an identity anchor for the circulating object, The function is used to calculate the summary. This is the normalized sequence of business fields. This is the normalized encapsulated field sequence. For the initial voucher summary, This is a sequential join operation.

[0026] This computational relationship is used to anchor business fields, encapsulation fields, and initial document summaries to the same identity anchor. When any field value, field order, or initial document summary changes, the circulating object's identity anchor changes accordingly.

[0027] After the identity anchor of the circulating object is generated, it is written into the evidence storage record of the current circulating object and passed to the node challenge code generation stage. When the circulating node submits multi-format credentials subsequently, the node challenge code, the typed evidence leaf node, the cross-format evidence root, and the node hash chain are all bound using the same circulating object identity anchor, thereby ensuring that multi-format credentials have a unified object basis across different circulating nodes.

[0028] During the node challenge code generation and binding phase, the circulation evidence storage method receives the hash of the previous node, the identity anchor of the circulation object, and the subject state vector.

[0029] The previous node hash is used to define the chain relationship between the current circulation node and the previous circulation node, the circulation object identity anchor is used to define the circulation object processed by the current circulation node, and the subject state vector is used to define the operation order of the circulation subject on the current circulation object.

[0030] During the node challenge code generation phase, the hash of the previous node, the identity anchor of the circulating object, and the subject state vector are input into the digest calculation function to form the node challenge code used by the current circulating node. After the node challenge code is generated, it is bound to multi-format credentials, enabling electronic confirmation credentials, image credentials, and video keyframes to carry the node identification content of the current circulating node.

[0031] After the node challenge code is bound, the multi-format voucher enters the stage of generating typified evidence leaf nodes. Subsequent stages involve identifying the node challenge codes in the multi-format voucher and generating challenge code verification results. These results are used to determine the closure residuals of cross-format evidence.

[0032] Node challenge codes can be generated in the following way:

[0033] in, For the first The node challenge code of each circulation node. The function is used to calculate the summary. The hash of the previous node. As an identity anchor for the circulating object, For the first The main state vector corresponding to each circulation node This is a sequential join operation.

[0034] This computational relationship allows the node challenge code to simultaneously inherit the previous node's hash, the circulating object's identity anchor, and the subject's state vector. When the previous node's hash changes, the circulating object's identity anchor changes, or the subject's state vector changes, the node challenge code changes accordingly, thus enabling the node challenge code to serve as the binding basis for subsequent verification of multi-format credentials.

[0035] In one embodiment, the main state vector is used to record manufacturer state counts, logistics state counts, distributor state counts, and platform state counts.

[0036] The manufacturer status count corresponds to the number of operations performed by the manufacturer on the current circulation object; the logistics status count corresponds to the number of operations performed by the logistics entity on the current circulation object; the distributor status count corresponds to the number of operations performed by the distributor on the current circulation object; and the platform status count corresponds to the number of verification, confirmation, or anomaly registration operations performed by the platform on the current circulation object. These four types of status counts are written into the entity status vector in a fixed order. This fixed order is established after the circulation object's identity anchor is generated and remains unchanged across all circulation nodes of the same circulation object.

[0037] The entity status vector is updated when the distribution entity submits vouchers, confirms handover, raises objections, or provides supplementary confirmation. When a manufacturer submits an outbound list, packing photos, seal photos, or electronic confirmation vouchers, the manufacturer status count increases. When a logistics entity submits pickup vouchers, transit vouchers, transportation vouchers, or handover confirmations, the logistics status count increases. When a distributor submits receiving vouchers, quantity confirmations, outer packaging status confirmations, or provides supplementary confirmations, the distributor status count increases. When the platform completes voucher integrity verification, objection registration, or supplementary confirmation registration, the platform status count increases.

[0038] The update of the subject state vector is performed on a unit basis for the current circulation object, and the subject state vectors of different circulation objects are independent of each other. After the operation of the circulation subject is received, the circulation evidence storage method reads the operation type, submitting subject identifier, and circulation object identity anchor, and assigns the operation to one of the following count positions: manufacturer status count, logistics status count, distributor status count, or platform status count. If the same operation includes both submitting vouchers and confirming handover, the circulation evidence storage method updates one count position according to the main operation type in the operation record, and includes the other operation content as additional content in the subsequent multi-format voucher processing, avoiding duplicate increments of the subject state vector due to the same operation.

[0039] The updated result of the main state vector is written to the pending sealing record of the current circulation node and used as input for generating the node challenge code. After the node challenge code is generated, a one-to-one binding relationship is formed between the node challenge code and the main state vector. During subsequent replay verification at the circulation node, the circulation evidence preservation method reconstructs the main state vector according to the same update rule and recalculates the node challenge code. If the recalculated node challenge code is inconsistent with the node challenge code identified in the multi-format voucher, the challenge code verification result is recorded as failed; if the recalculated node challenge code is consistent with the node challenge code identified in the multi-format voucher, the challenge code verification result is recorded as passed.

[0040] The subject state vector is also used to constrain the operational sequence of the distribution entities. If the logistics entity has submitted transit documents, but the manufacturer's status count still does not reflect an update record of document submission or handover confirmation, the distribution evidence preservation method marks the current operation as an order anomaly and writes the order anomaly into the objection status as the basis for platform registration of objections. The objection status enters the subsequent handover responsibility edge generation stage, serving as input content when determining the responsibility edge tension value. Through this processing, the subject state vector not only participates in the node challenge code generation but also transmits the operational sequence of the distribution entities to the determination process of the handover responsibility edge and the sealing status.

[0041] In one embodiment, when a node challenge code is bound to a multi-format certificate, the circulation and evidence preservation method selects different binding methods according to the certificate type.

[0042] For electronic confirmation documents, the node challenge code is written to a fixed field within the electronic confirmation document. This fixed field can be set in the confirmation content area, receipt content area, or handover content area of ​​the electronic confirmation document. Before the electronic confirmation document generates a summary, the fixed field is stored together with the order number, batch number, box number, seal number, and waybill number. When retrieving the node challenge code, if the fixed field is readable and matches the recalculated node challenge code, the challenge code verification result is recorded as passed. If the fixed field is not readable but contains verifiable character fragments, the challenge code verification result is recorded as pending review. If the fixed field is missing or inconsistent with the recalculated node challenge code, the challenge code verification result is recorded as failed, and the node challenge code verification result is written to the corresponding typed evidence leaf node.

[0043] For image-based vouchers, the node challenge code is not only stored as a backend field but also needs to appear in the image voucher along with the seal, waybill, or box. When collecting image vouchers, the collection interface displays the node challenge code, and the circulation entity places the node challenge code, seal, waybill, or box in the same field of view during the shooting. After receiving the image voucher, the circulation evidence storage method identifies the node challenge code, seal number, waybill number, or box area in the image voucher and records the identification status of the node challenge code in the image voucher. If the node challenge code can be completely identified and is consistent with the node challenge code recalculated by the current circulation node, the challenge code verification result is recorded as passed; if the node challenge code has partial occlusion, is blurry, or has a recognition confidence level lower than the stable recognition requirement but still retains comparable character fragments, the challenge code verification result is recorded as pending review; if the node challenge code does not appear in the image voucher, cannot be identified, or the identified node challenge code is inconsistent with the recalculated node challenge code, the challenge code verification result is recorded as failed.

[0044] For video credentials, the circulation and evidence preservation method selects keyframes from the video credential. Keyframes can be selected based on time intervals or the degree of change in the image. Each keyframe must simultaneously contain a node challenge code and at least one of the following objects: a seal, a waybill, or a container. The circulation and evidence preservation method performs the same recognition processing on the video keyframes as on image credentials and aggregates the recognition status of multiple keyframes to generate the challenge code verification result corresponding to the video credential. If at least one keyframe can reliably recognize the node challenge code, and the recognized challenge code matches the recalculated node challenge code of the current circulation node, the challenge code verification result for the video credential is recorded as passed. If multiple keyframes can only recognize partial character fragments of the node challenge code, or the recognition results of different keyframes are unstable but do not show a clear inconsistency, the challenge code verification result for the video credential is recorded as pending review. If none of the keyframes can recognize the node challenge code, or the recognized node challenge code is inconsistent with the recalculated node challenge code, the challenge code verification result for the video credential is recorded as failed.

[0045] After the challenge code verification result is generated, it enters the typified evidence leaf node generation stage along with the voucher type, original text summary, field summary, and image region summary of the multi-format voucher. The challenge code verification result does not independently determine the validity of the multi-format voucher, but rather serves as one of the input contents for the cross-format evidence closure residual. Field consistency, region consistency, and the challenge code verification result together enter the cross-format evidence closure residual determination stage, enabling the formation of anomalous evidence in the subsequent sealing status determination stage when old images, old videos, or old electronic confirmation vouchers are reused.

[0046] During the multi-format certificate conversion stage, the circulation and storage method receives multi-format certificates that have been bound to node challenge codes.

[0047] Multi-format vouchers can originate from image vouchers, video vouchers, PDF vouchers, manifest vouchers, and waybill vouchers. The circulation and evidence preservation method identifies the voucher type for each multi-format voucher, generates an original text summary, and extracts field summaries, image region summaries, and submitter identifiers from the multi-format voucher that can participate in subsequent verification.

[0048] After extraction, the multi-format vouchers are converted into typed evidence leaf nodes. Typed evidence leaf nodes are used to unify vouchers of different formats into data units that are sortable, can be digested, and can be entered into a hash tree.

[0049] Multiple typified evidence leaf nodes are sorted by document type, submission time, and document number before entering the hash tree construction process. The root hash of the hash tree is determined as the cross-format evidence root. The cross-format evidence root is passed to the cross-format evidence closure residual determination stage, which serves as the basis for constraining the calculation of subsequent field consistency, region consistency, and challenge code verification results.

[0050] In one embodiment, the typified evidence leaf nodes are set as image evidence leaf nodes, video evidence leaf nodes, PDF evidence leaf nodes, manifest evidence leaf nodes, and waybill evidence leaf nodes. Different types of typified evidence leaf nodes are all stored using a unified data structure, so that multi-format vouchers of different formats have the same data entry point when constructing the hash tree subsequently.

[0051] The image evidence leaf node is derived from the image certificate. After receiving the image certificate, the circulation evidence preservation method generates a summary of the original image certificate, identifies the seal, waybill, or box area in the image certificate, and writes the identified image area summary into the image evidence leaf node. If the image certificate also identifies the order number, batch number, product model, quantity, box number, seal number, or waybill number, the circulation evidence preservation method organizes the identified field content into a field summary. The identity of the circulation entity submitting the image certificate is entered into the submitting entity identifier. The image evidence leaf node is formed by the certificate type, original summary, field summary, image area summary, and submitting entity identifier.

[0052] The video evidence leaf node is derived from the video credential. After receiving the video credential, the circulation and evidence preservation method selects keyframes based on the submission time or the degree of image change, and performs the same region and field recognition on these keyframes as on image credentials. The original text summary of the video credential is generated from the complete file content of the video credential. Seals, waybills, or box areas identified in the video keyframes form image region summaries, and business fields or encapsulation fields identified in the video keyframes form field summaries. The video evidence leaf node carries the credential type, original text summary, field summary, image region summary, and submitting entity identifier, representing the evidence content of the video credential in the current circulation node.

[0053] PDF evidence leaf nodes are derived from PDF credentials. Upon receiving a PDF credential, the circulation and evidence preservation method generates a summary of the original PDF credential and extracts identifiable fields from the order number, batch number, product model, quantity, box number, seal number, and waybill number. The extracted fields are then normalized to form a field summary. If the PDF credential contains a signature area, receipt area, or confirmation area, the circulation and evidence preservation method can organize the area location, area summary, and area identification results into an image area summary. Each PDF evidence leaf node carries at least the credential type, original summary, and field summary, as well as the submitting entity identifier.

[0054] The manifest evidence leaf nodes are derived from manifest documents. Manifest documents typically describe the relationships between product models, quantities, batch numbers, and box numbers. After generating the original text summary of the manifest document using the circulation evidence preservation method, the line item fields, quantity fields, and box number fields in the manifest document are organized into a field summary. If the manifest document is in image or scanned form, the circulation evidence preservation method can also generate an image region summary. The manifest evidence leaf nodes are then incorporated into the subsequent field consistency matrix to verify the consistency of quantities and box numbers between different types of evidence leaf nodes.

[0055] The leaf node for waybill evidence is derived from the waybill document. After receiving the waybill document, the circulation evidence preservation method generates a summary of the original text of the waybill document, extracting the waybill number, sender / receiver entity, box number, seal number, and logistics identifier, and forming a field summary. If the waybill document originates from an image document, the image area summary of the waybill area is simultaneously written into the waybill evidence leaf node. The waybill evidence leaf node is used in the subsequent handover responsibility edge generation stage to associate the sender / receiver entity identifier and the receiver / receiver identifier.

[0056] Each typified evidence leaf node carries the document type and original text summary. Field summaries, image region summaries, and submitter identifiers are written according to the document type and the actual identifiable content. When a field summary or image region summary cannot be identified, the circulation evidence preservation method writes a missing marker at the corresponding location. Missing markers do not participate in the majority agreement of field values, but they are included in the necessary evidence integrity assessment and participate in the determination of field conflict residuals based on the field missing ratio, ensuring that missing information can form constraints in the subsequent cross-format evidence closure residuals.

[0057] In one embodiment, cross-format evidence roots are constructed from multiple typified evidence leaf nodes. After the generation of typified evidence leaf nodes is completed at the current circulation node, the circulation evidence preservation method performs sorting processing on all typified evidence leaf nodes. The sorting criteria are set as document type, submission time, and document number. Document type ensures that image evidence leaf nodes, video evidence leaf nodes, PDF evidence leaf nodes, manifest evidence leaf nodes, and waybill evidence leaf nodes have a fixed arrangement position within the same circulation node. Submission time distinguishes the order of submission under the same document type. Document number provides a stable sorting criterion when document type and submission time are the same.

[0058] After sorting, the circulation evidence preservation method generates a leaf node summary for each type of evidence leaf node. The content of the leaf node summary comes from the document type, original text summary, field summary, image region summary, and submitting entity identifier. If a certain type of evidence leaf node has a missing marker, the missing marker and the identified content participate in the leaf node summary calculation, thus fixing the missing information in the cross-format evidence root.

[0059] Leaf node summaries can be generated as follows:

[0060] in, For the first Leaf node summary of each typed evidence leaf node The function is used to calculate the summary. For the first The credential type of a categorized evidence leaf node. For the first Original text summary of each typified evidence leaf node, For the first A field summary of each typed evidence leaf node. For the first Image region summary of each typified evidence leaf node. For the first The submitting entity identifier of each typed evidence leaf node, This is a sequential join operation.

[0061] This computational relationship is used to simultaneously fix the voucher type, original text summary, field summary, image region summary, and submitting entity identifier in the leaf nodes of typified evidence. When any of these elements changes, the leaf node summary changes accordingly.

[0062] The circulation evidence preservation method uses the sorted leaf node digests as the bottom-level nodes of the hash tree. Adjacent leaf node digests are concatenated according to their sorting order, and the upper-level digest is calculated. The upper-level digests continue to be calculated layer by layer according to the same rules until a single root hash is formed. When unpaired nodes exist at the same level, the circulation evidence preservation method either promotes the unpaired node directly to the next level or copies the unpaired node to participate in the upper-level digest calculation, according to the pre-defined processing method for the same circulation object. The processing method for unpaired nodes remains consistent across all circulation nodes of the same circulation object, avoiding inconsistencies in cross-format evidence roots due to different processing methods for different circulation nodes.

[0063] Cross-format evidence roots can be determined in the following way:

[0064] in, For the first Cross-format evidence roots of each circulation node, For the hash tree root hash generation function, For the first Sorted leaf node summaries from each circulation node. For the first The number of typified evidence leaf nodes in each circulation node.

[0065] After the cross-format evidence root is generated, it is written into the pending closure record of the current circulation node. In the subsequent cross-format evidence closure residual determination stage, instead of directly rereading all the original multi-format vouchers, the typified evidence leaf node set corresponding to the cross-format evidence root is used to calculate field consistency, region consistency, and challenge code verification results. The cross-format evidence root also enters the handover responsibility edge and node hash generation stage, used to fix the multi-format vouchers of the current circulation node as a whole into the node hash chain.

[0066] The cross-format evidence closure residual determination stage receives the leaf nodes of categorized evidence, the challenge code verification results, and the node set corresponding to the root of cross-format evidence.

[0067] The circulation evidence preservation method extracts the fields to be verified and the image region summary from the leaf nodes of the typified evidence. The fields to be verified are used to compare whether the field descriptions of the same circulation object are consistent across different format documents, and the image region summary is used to compare whether the seal, waybill, or box area of ​​the current circulation node matches the initial image region summary in the initial document record.

[0068] The field consistency matrix is ​​formed by the field values ​​of the field to be verified in different types of evidence leaf nodes, and the region consistency result is formed by the matching relationship between image region summaries. The field consistency matrix and the region consistency result together constitute the consistency result.

[0069] The circulation evidence preservation method combines the consistency results and challenge code verification results to generate cross-format evidence closure residuals. The cross-format evidence closure residuals then enter the stage of determining the handover responsibility edge and the responsibility edge tension value, which is used to constrain whether the current circulation node can form a stable sealing state.

[0070] In one embodiment, the circulation evidence preservation method extracts the fields to be verified and image region summaries from multiple typified evidence leaf nodes, and generates a field consistency matrix and a region consistency result.

[0071] The fields to be verified can be selected as order number, batch number, product model, quantity, box number, seal number, and waybill number. The selection of these fields is based on the components of the identity anchor of the circulating object and the fields that can be reliably identified in multi-format documents. These fields are used to establish horizontal comparison relationships for the same field among image evidence leaf nodes, video evidence leaf nodes, PDF evidence leaf nodes, manifest evidence leaf nodes, and waybill evidence leaf nodes.

[0072] The field consistency matrix is ​​generated with the fields to be verified as rows and the typed evidence leaf nodes as columns. Matrix cells record field values, field sources, and identification confidence levels. Field values ​​originate from field summaries in the typed evidence leaf nodes. Field sources indicate whether the field value comes from image evidence leaf nodes, video evidence leaf nodes, PDF evidence leaf nodes, manifest evidence leaf nodes, or waybill evidence leaf nodes. Identification confidence reflects the stability of the field value's identification from multi-format documents. Identification confidence can be determined by combining character recognition results, field format matching results, field position matching results, and duplicate identification results, and is used to constrain the calculation weight of field conflict residuals.

[0073] When a certain type of evidence leaf node lacks a field to be verified, a missing field marker is written into the corresponding cell in the field consistency matrix. The missing field marker is not used as a field value in the majority consensus judgment, but it is retained in the field consistency matrix and included in the integrity judgment of cross-format evidence closure residuals. If the same field to be verified has different values ​​in multiple type of evidence leaf nodes, the field consistency matrix retains all field values ​​and identification confidence levels, without directly deleting the minority values. Subsequent field conflict residuals determine the degree of conflict based on the distribution of field values ​​and identification confidence levels.

[0074] The regional consistency result is generated by matching relationships between image region summaries. Image region summaries can originate from image evidence leaf nodes, video evidence leaf nodes, or image regions within PDF evidence leaf nodes, manifest evidence leaf nodes, or waybill evidence leaf nodes. The image region summaries target the seal area, waybill area, or box area. The image region summary of the current circulation node is compared with the initial image region summary read from the initial voucher record. The comparison includes region category, region location, region texture, and region identifier. The region category distinguishes between seals, waybills, and boxes; the region location describes the relative position of the image region within the voucher image; the region texture describes the seal texture, box texture, or waybill edge; and the region identifier describes the seal number, waybill number, or box number.

[0075] The regional consistency result can be set to pass, pending review, or fail. Pass indicates that the image region summary of the current circulation node matches the initial image region summary read from the initial voucher record; pending review indicates that the image region summary of the current circulation node has occlusion, blurring, or angular deviation, but still retains some matching content; fail indicates that the region category, region identifier, or main region texture cannot match. The regional consistency result, together with the field consistency matrix, constitutes the consistency result and is passed to the cross-format evidence closure residual generation process.

[0076] In one embodiment, the circulation evidence preservation method determines field conflict residuals based on the field values ​​and identification confidence levels of the same field to be verified in different typified evidence leaf nodes.

[0077] For any field to be verified, the circulation evidence preservation method counts the number of valid field values ​​in the field consistency matrix and determines the field value that appears most frequently. A valid field value is one that has not been marked as missing and whose identification confidence meets the field identification requirements. Field identification requirements can be set according to the document type and the stability of the field to be verified. Order numbers, batch numbers, box numbers, seal numbers, and waybill numbers typically have higher identification requirements, while product models and quantities can have their identification requirements determined by combining field format and frequency of occurrence. Field identification requirements are used to exclude obviously misidentified fields and prevent invalid field values ​​from amplifying field conflict residuals.

[0078] Field conflict residuals can be determined as follows:

[0079] in, For field conflict residuals, The number of fields to be verified. This is the sequence number of the field to be verified. For the first The field weight of each field to be verified. For the first The number of the most frequently occurring field values ​​among the fields to be verified. For the first The number of valid field values ​​for each field to be verified. For the first Mean confidence score of each field to be verified For the first The number of missing markers in each field to be verified. For the first The number of typed evidence leaf nodes corresponding to each field to be verified. , and These are the weights for the field conflict ratio, identification confidence, and field missing ratio, respectively, and all are non-negative coefficients.

[0080] Field weights are set based on the importance of the field to be verified in identifying the distribution object. Order number, batch number, box number, seal number, and waybill number are used to identify the distribution object and packaging unit; these fields have higher weights than product model and quantity. Field weights are kept consistent across all distribution nodes for the same distribution object to ensure the comparability of field conflict residuals across different distribution nodes. At that time, the method of evidence preservation in circulation will be the first The field conflict ratio item for the field to be verified is set to 1, and the first field is set to 1. The mean confidence level of the identification of each field to be verified is set to 0.

[0081] When field values ​​in leaf nodes of different types of evidence are highly consistent, the number of most frequently occurring field values ​​approaches the number of valid field values, and the field conflict residuals tend to decrease. When multiple field values ​​are scattered, or when field values ​​are consistent but the identification confidence is low, the field conflict residuals increase. Field conflict residuals are incorporated into the calculation basis of cross-format evidence closure residuals and participate in the subsequent determination of responsibility edge tension values.

[0082] In one embodiment, the circulation evidence preservation method determines the image matching residual based on the image region summary of the current circulation node and the initial image region summary read from the initial credential record.

[0083] The current circulating node's image region summary originates from the seal area, waybill area, or box area within the image evidence leaf node and video evidence leaf node. The initial image region summary originates from the box image, seal image, waybill, or other document content that can fix the initial packaging state, used to generate the initial document summary, and is stored in the initial document record in association with the initial document summary. Both types of image region summaries enter the region consistency result generation process, and the region consistency result is used to record the region matching status. The image matching residual is used to convert the region matching status into a numerical value that can enter the cross-format evidence closure residual.

[0084] Image matching residuals can be determined as follows:

[0085] in, To match residuals to the image, For regional identifier differences, This represents the regional location difference value. This represents the regional texture difference value. The weights for the regional identifier differences. The weights for the regional location differences. The weights for the regional texture differences.

[0086] The regional identification difference value is determined based on whether the seal number, waybill number, or box number matches. The regional position difference value is determined based on the relative position difference of the seal area, waybill area, or box area in the document image. The regional texture difference value is determined based on the difference between the seal texture, box texture, or waybill edge features. The three weights are set according to the degree of influence of the regional object on the identification of the circulating object. The seal area is used to determine whether the packaging has been replaced, and the regional identification difference value and regional texture difference value can be assigned higher weights; the waybill area is used to determine whether the external logistics markings are consistent, and the regional identification difference value can be assigned higher weights; the box area is used to determine whether the outer packaging status is continuous, and the regional position difference value and regional texture difference value can be used as auxiliary criteria.

[0087] When the challenge code verification result enters the cross-format evidence closure residual, it can be converted into challenge code residuals according to pass, pending review, and fail. Pass corresponds to a challenge code residual of 0, pending review corresponds to a challenge code residual of 0.5, and fail corresponds to a challenge code residual of 1. When adjustments are needed based on the circulation node type, the challenge code residual for pending review can be set within the range of 0.3 to 0.7 and fixed within the same circulation node type for the same circulation object. Pending review can be generated when the node challenge code is partially obscured, video keyframe recognition is unstable, or the readability of electronic confirmation certificate fields is insufficient.

[0088] Cross-format evidence closure residuals can be generated as follows:

[0089] in, For cross-format evidence closure residuals, For field conflict residuals, To match residuals to the image, To challenge the residual code, The weights of the field conflict residuals, Weights for image matching residuals, The weights are for the challenge code residuals.

[0090] The three weights are configured based on the circulation node type and certificate type. The weights of the field conflict residual, the image matching residual, and the challenge code residual are non-negative coefficients. They can be calibrated based on the contribution of historical normal evidence samples and historical abnormal evidence samples to the differentiation of field conflict, image matching, and challenge code verification. After calibration, they are fixed in the same circulation node type of the same circulation object.

[0091] During the manufacturer's shipment phase, the focus is typically on the complete binding of initial fields and image region summaries. During the logistics transit phase, the focus is typically on the verification results of seals, waybills, and challenge codes. During the distributor's receiving phase, the focus is typically on quantity, outer packaging condition, and challenge code verification results. Weighting is kept consistent across the same circulation node type for the same circulation object to ensure that cross-format evidence closure residuals have a stable basis for comparison among similar nodes.

[0092] After the cross-format evidence closure residual is generated, it is written into the pending sealing record of the current circulation node and passed to the handover responsibility edge and responsibility edge tension value determination stage. The higher the cross-format evidence closure residual, the greater the consistency deviation between different format documents. The subsequent sealing status is determined jointly based on the cross-format evidence closure residual and the responsibility edge tension value, enabling field conflicts, image mismatches, and challenge code verification anomalies to enter the generation logic of node hash or abnormal fork nodes.

[0093] During the handover responsibility generation phase, the cross-format evidence closure residual and handover status are received.

[0094] The handover status originates from the handover and receipt operations between adjacent circulation nodes, and at least records the quantity confirmation result, seal status, outer packaging status, and objection status. The circulation evidence preservation method combines the hash of the previous node, cross-format evidence root, handover entity identifier, receiving entity identifier, and key content in the handover status to generate a handover responsibility edge.

[0095] After the handover responsibility edge is generated, the circulation evidence preservation method reads the state differences of adjacent circulation nodes in terms of quantity, seal status, and outer packaging status, and determines the responsibility edge tension value by combining the cross-format evidence closure residual and objection status.

[0096] The responsibility edge tension value is used to characterize the degree of deviation in the handover state between adjacent flow nodes and is passed to the sealing state determination stage. The sealing state is determined based on the cross-format evidence closure residual and the responsibility edge tension value, thereby constraining the generation of node hashes or anomalous fork nodes.

[0097] In one embodiment, the handover responsibility edge is formed by the previous node hash, cross-format evidence root, handover subject identifier, receiving subject identifier, quantity confirmation result, seal status, outer packaging status, and objection status.

[0098] The previous node hash is used to limit the previous circulation node to which the handover responsibility edge takes over. The cross-format evidence root is used to limit the overall content of multi-format vouchers in the current circulation node. The handover entity identifier is used to limit the circulation entity that completed the handover operation. The receiving entity identifier is used to limit the circulation entity that completed the receiving operation. The quantity confirmation result is used to record the relationship between the quantity confirmed by the receiving entity and the quantity recorded by the handover entity. The seal status is used to record whether the seal is intact, damaged, missing, or unidentifiable. The outer packaging status is used to record whether the outer packaging is intact, damaged, contaminated, deformed, or unidentifiable. The objection status is used to record whether the receiving entity has raised objections to the quantity, seal status, or outer packaging status, or whether the platform has registered an objection based on the abnormal operation sequence of the circulation entity.

[0099] The handover of responsibility edges can be generated in the following way:

[0100] in, For the first The handover responsibility edge of each circulation node. The function is used to calculate the summary. The hash of the previous node. For the first Cross-format evidence roots of each circulation node, To hand over the main logo, For the receiving entity identifier, To confirm the quantity results, It is currently sealed. In its outer packaging condition. The status is in dispute. This is a sequential join operation.

[0101] The calculation relationship is used to simultaneously fix the previous circulation node, the multi-format voucher of the current circulation node, the handover entity, the receiving entity, and the handover status in the handover responsibility edge. When any input content changes, the handover responsibility edge changes, and the handover responsibility edge summary also changes accordingly. The handover responsibility edge summary enters the node hash generation stage, which is used to fix the handover responsibility relationship into the node hash chain.

[0102] The quantity confirmation result is obtained by comparing the quantity submitted by the submitter and the quantity confirmed by the receiver. If they match, the quantity confirmation result is recorded as consistent; if they do not match, the quantity confirmation result is recorded as inconsistent, and the quantity submitted by the submitter, the quantity confirmed by the receiver, and the submission time are retained. The seal status and outer packaging status are extracted from the image, video, or electronic confirmation documents submitted by the receiver, or can be obtained by comparing the submission documents submitted by the submitter with the receipt documents submitted by the receiver. The objection status is determined by the receiver raising an objection or the platform registering an objection.

[0103] The responsibility edge tension value is determined based on the cross-format evidence closure residual, the state differences between adjacent circulation nodes in terms of quantity, seal status, and outer packaging status, and the objection status. The state differences between adjacent circulation nodes describe whether the output status of the previous circulation node is consistent with the input status of the current circulation node. The quantity difference is obtained by comparing the quantity of the previous circulation node with the quantity of the current circulation node. The seal status difference is obtained by comparing the seal status of the previous circulation node with the seal status of the current circulation node. The outer packaging status difference is obtained by comparing the outer packaging status of the previous circulation node with the outer packaging status of the current circulation node.

[0104] Differences in seal status and outer packaging status can be determined by combining status category, status recognition confidence, and image region summary differences. When the status category is consistent and recognition is stable, the difference value is 0. When the status category is consistent but there is a slight image deviation or a decrease in recognition confidence, the difference value is greater than 0 but does not exceed the upper limit of slight differences. When the status categories are inconsistent, the difference value is higher than the upper limit of slight differences.

[0105] The tension value on the responsible side can be determined as follows:

[0106] in, For the first The responsibility edge tension value corresponding to each circulation node. For the first Cross-format evidence closure residuals at each circulation node The difference in the number of adjacent circulation nodes. The difference in seal status between adjacent circulation nodes. The difference in the outer packaging condition between adjacent distribution nodes The status is in dispute. For the weight of the cross-format evidence closure residual, Weights for quantity differences As the weight for differences in seal status, Weighting for differences in outer packaging condition. The weight of the objection status.

[0107] Each weight is set according to the type of circulation node and the attributes of the circulation object. The weights of cross-format evidence closure residuals, quantity differences, seal status differences, outer packaging status differences, and objection status are non-negative coefficients. They can be calibrated based on the contribution of each input item in the historical handover samples to distinguish between normal and abnormal handovers. After calibration, they are fixed for use in the same type of circulation nodes of the same circulation object.

[0108] Manufacturer shipping nodes and logistics transit nodes can increase the weight of differences in seal status, while distributor receiving nodes can increase the weight of differences in quantity and outer packaging status. The weight of objection status is used to constrain handover records with clear objections to enter the sealing status determination stage. The weights are kept consistent across similar circulation nodes for the same type of object, providing a stable basis for comparison of the responsibility edge tension value.

[0109] After the responsibility edge tension value is generated, it is written into the pending sealing record of the current circulation node and enters the sealing status determination stage together with the cross-format evidence closure residual. A low responsibility edge tension value indicates that the deviation in the handover status between adjacent circulation nodes is small. A high responsibility edge tension value indicates that there are large deviations in quantity, seal status, outer packaging status, or objection status. The sealing status is determined jointly based on the cross-format evidence closure residual and the responsibility edge tension value, so that deviations in internal consistency of documents and deviations in handover between circulation nodes can be used in the same sealing judgment.

[0110] During the stage of determining the sealing status, the closure residual and the responsibility edge tension value of cross-format evidence are received.

[0111] Cross-format evidence closure residuals are used to reflect consistency deviations among multi-format documents within the current circulation node, while the responsibility edge tension value is used to reflect handover deviations between adjacent circulation nodes. The circulation evidence preservation method compares these two types of deviation results with sealing conditions to form a sealing status.

[0112] When the sealing condition is met, the current circulating node is allowed to generate a node hash and join the node hash chain. When the sealing condition is not met, the current circulating node does not join the node hash chain, but instead generates an abnormal fork node.

[0113] The node hash chain, the handover responsibility edge, and the abnormal fork node enter the replay verification process. The replay verification verifies the node hash according to the connection order of the node hash chain, and combines the handover responsibility edge and the abnormal fork node to determine the chain break, mismatch, replacement, or responsibility attribution.

[0114] In one embodiment, the sealing condition is formed jointly by cross-format evidence closure residual constraints and responsibility edge tension value constraints. The cross-format evidence closure residual constraints are determined based on the document type of the current circulation node, the circulation node type, and the residual distribution in historical normal evidence samples, and are used to constrain consistency deviations of multi-format documents within the current circulation node. The responsibility edge tension value constraints are determined based on the state difference distribution of adjacent circulation nodes in terms of quantity, sealing status, and outer packaging status, and are used to constrain handover state deviations between adjacent circulation nodes. Both types of constraints maintain consistency across similar circulation nodes of the same circulation object, providing a stable basis for judging the sealing status.

[0115] The sealing status can be determined in the following ways:

[0116] in, For the first The sealing status of each circulation node For the first Cross-format evidence closure residuals at each circulation node For the first The responsibility edge tension value of each circulation node To limit the residual closure for cross-format evidence, The tension value is limited for the responsible side. This indicates that the sealing condition has been met. This indicates that the sealing condition is not met.

[0117] When the sealing status meets the sealing conditions, the circulation evidence preservation method reads the digests of the previous node hash, the circulation object identity anchor, the cross-format evidence root, and the handover responsibility edge, and combines them in a fixed order to calculate the node hash:

[0118] in, For the first The node hash of each circulating node. The function is used to calculate the summary. The hash of the previous node. As an identity anchor for the circulating object, For the first Cross-format evidence roots of each circulation node, For the first A summary of the handover responsibility edge for each circulation node. This is a sequential join operation. After the node hash is generated, it is added to the node hash chain and becomes the input for subsequent circulating nodes to generate node challenge codes and node hashes.

[0119] When the sealing status does not meet the sealing conditions, the circulation evidence preservation method queries the node hash chain for the node hash that most recently met the sealing conditions, and determines the query result as the last valid node hash. The last valid node hash is used to fix the valid chain position to which the abnormal fork node belongs, preventing abnormal circulation nodes from directly overwriting valid nodes in the node hash chain.

[0120] Abnormal fork nodes can be generated in the following way:

[0121] in, For the first Abnormal fork node of each circulation node The hash of the last valid node. For the first Cross-format evidence roots of each circulation node, For the first Cross-format evidence closure residuals at each circulation node For the first The responsibility edge tension value of each circulating node. After an abnormal fork node is generated, it does not connect to the main connection relationship of the node hash chain, but instead establishes a fork association with the last valid node hash and enters the replay verification process.

[0122] The replay verification reads the node hash chain, handover responsibility edges, and abnormal fork nodes. The node hash chain is used to verify the hash inheritance relationship between adjacent circulation nodes. The handover responsibility edge is used to verify the handover entity identifier, receiving entity identifier, quantity confirmation result, sealing status, outer packaging status, and objection status. Abnormal fork nodes are used to locate circulation nodes whose sealing status does not meet the sealing conditions. A broken chain is determined when the hash inheritance relationship is interrupted; a mismatch is determined when there is inconsistency between the typified evidence leaf nodes corresponding to cross-format evidence roots; a replacement is determined when the node hash, cross-format evidence root, or handover responsibility edge digest is inconsistent with the recalculation result; abnormal fork nodes and adjacent handover responsibility edges are used together to determine responsibility attribution.

[0123] This embodiment takes the process of a batch of motor controllers being shipped from the manufacturer, received by logistics, and then verified by the distributor before receipt as the processing object. At the outbound node, the manufacturer submits business fields, packaging fields, and an initial document summary. The business fields include order number ORD-240716-081, batch number B2407-16, product model MTR-A9, and quantity 120; the packaging fields include box number BOX-07-18, seal number SEAL-73A2, and waybill number WB-889102; the initial document summary includes outbound manifest summary H-LST-8F31, packing image summary H-IMG-42B0, and seal image summary H-SEAL-19D2.

[0124] After receiving the aforementioned fields, the circulation evidence preservation method performs normalization processing on the business fields and encapsulated fields. Normalization unifies field names, character formats, and field order, and preserves missing fields. After normalization, the normalized business fields, normalized encapsulated fields, and the initial credential digest together generate the circulation object identity anchor A-7D3C9E21. This circulation object identity anchor enters the node challenge code generation process and serves as the object binding basis for subsequent cross-format evidence roots, handover responsibility edges, and node hash chains.

[0125] like Figure 2 As shown, Figure 2 This is a record diagram of the generation of identity anchors and node challenge codes for circulating objects in an embodiment of the present invention. Figure 2 The left side displays the business fields, encapsulation fields, and initial voucher summary; the middle section displays the normalization process and the main state vector; and the right side displays the previous node hash, circulating object identity anchor, node challenge code, and multi-format voucher. Figure 2 The previous node hash is C0-GENESIS-6E2A, and the main state vector includes manufacturer state count 1, logistics state count 0, distributor state count 0, and platform state count 1. The circulation evidence storage method generates a node challenge code R-1-4F9C based on the previous node hash C0-GENESIS-6E2A, the circulation object identity anchor A-7D3C9E21, and the main state vector.

[0126] After generation, the node challenge code R-1-4F9C is bound to the multi-format certificate. When the manufacturer submits an image certificate, the node challenge code appears together with the seal number SEAL-73A2 in the image certificate; when the manufacturer submits a video certificate, the node challenge code appears together with the waybill number WB-889102 in the video keyframe; when the manufacturer submits an electronic confirmation certificate, the node challenge code is written into the confirmation field of the electronic confirmation certificate. The circulation evidence preservation method identifies the node challenge code in the multi-format certificate. If the identification result is consistent with the recalculated node challenge code, the challenge code verification result is recorded as passed.

[0127] The current circulation node accepts five types of multi-format certificates: image certificates, video certificates, PDF certificates, manifest certificates, and waybill certificates. Image certificates correspond to certificate number IMG-001, with original digest P-73A, field digest F-21C, and image area digest G-SEAL. Video certificates correspond to certificate number VID-001, with original digest P-65B, field digest F-20A, and image area digest G-WB. PDF certificates correspond to certificate number PDF-001, with original digest P-4E9, field digest F-88D, and the submitting entity identifier is manufacturer F01. Manifest certificates correspond to certificate number LST-001, with original digest P-1C7, field digest F-QTY120, and the submitting entity identifier is manufacturer F01. Waybill certificates correspond to certificate number WB-001, with original digest P-2B6, field digest F-WB889102, and image area digest G-WB.

[0128] The circulation and evidence preservation method converts the aforementioned multi-format documents into categorized evidence leaf nodes. Image documents are converted into image evidence leaf nodes, video documents into video evidence leaf nodes, PDF documents into PDF evidence leaf nodes, manifest documents into manifest evidence leaf nodes, and waybill documents into waybill evidence leaf nodes. Each categorized evidence leaf node carries the document type and original text summary, and, based on the document content, also carries a field summary, image area summary, or submitting entity identifier. The circulation and evidence preservation method sorts multiple categorized evidence leaf nodes according to document type, submission time, and document number, forming leaf node summaries L1, L2, L3, L4, and L5, where L1 corresponds to image evidence leaf nodes, L2 to video evidence leaf nodes, L3 to PDF evidence leaf nodes, L4 to manifest evidence leaf nodes, and L5 to waybill evidence leaf nodes.

[0129] like Figure 3 As shown, Figure 3 This is a diagram showing the construction of typified evidence leaf nodes and cross-format evidence roots in an embodiment of the present invention. Figure 3 The left side displays five types of typified evidence leaf nodes, the middle displays the sorting key and leaf node summaries, and the right side displays the hash tree structure from the leaf node summaries, generating cross-format evidence roots. The sorting key includes the credential type, submission time, and credential number. After the leaf node summaries enter the hash tree, L1 and L2 form H12, L3 and L4 form H34, unpaired L5 forms H5 according to fixed rules, H12 and H5 form H125, and H34 forms an upgraded H34 according to fixed rules. H125 and the upgraded H34 together form the cross-format evidence root E1-8C21A. Unpaired nodes are processed using fixed rules; L5 can be directly upgraded or copied to participate in the upper-level summary calculation, maintaining consistency within the same circulating object. The root hash of the hash tree is determined as the cross-format evidence root E1-8C21A.

[0130] After cross-format evidence roots are generated, the circulation evidence preservation method extracts the fields to be verified and image region summaries from multiple typified evidence leaf nodes. The fields to be verified include order number, batch number, product model, quantity, box number, seal number, and waybill number. The field consistency matrix uses the fields to be verified as rows and image evidence leaf nodes, video evidence leaf nodes, PDF evidence leaf nodes, manifest evidence leaf nodes, and waybill evidence leaf nodes as columns. Matrix cells record consistency, missing, or conflicting data. The region consistency results record the matching status of the seal region, waybill region, and box region.

[0131] like Figure 4 As shown, Figure 4 This is a diagram showing the field consistency matrix and region consistency results in an embodiment of the present invention. Figure 4 In the evidence presented, the order number is consistent across the image evidence, video evidence, PDF evidence, and manifest evidence leaf nodes, but missing in the waybill evidence leaf node. The batch number distribution is the same as the order number. The product model is missing in the image evidence, video evidence, and waybill evidence leaf nodes, but consistent in the PDF evidence and manifest evidence leaf nodes. The quantity is consistent across the image evidence, PDF evidence, and manifest evidence leaf nodes, but missing in the video evidence and waybill evidence leaf nodes. The box number is consistent across all five types of typified evidence leaf nodes. The seal number is consistent across the image evidence, video evidence, and waybill evidence leaf nodes, but missing in the PDF evidence and manifest evidence leaf nodes. The waybill number is consistent across the image evidence, video evidence, PDF evidence, and waybill evidence leaf nodes, but missing in the manifest evidence leaf node.

[0132] Figure 4 The regional consistency results and residual results are displayed below. The regional consistency results are: seal area passed, waybill area pending review, and box area passed. The field conflict residual is 0.083, the image matching residual is 0.120, and the challenge code residual is 0.000. With weights of 0.78, 0.22, and 0.10 for the field conflict residual, image matching residual, and challenge code residual, respectively, the cross-format evidence closure residual is approximately 0.091. The field conflict residual is determined by the field value and recognition confidence of the same field to be verified in different typified evidence leaf nodes. The image matching residual is determined by comparing the image region summary of the current circulation node with the initial image region summary read from the initial voucher record. The challenge code residual is determined by the challenge code verification result. The cross-format evidence closure residual of 0.091 proceeds to the process of determining the handover responsibility edge and responsibility edge tension value.

[0133] Manufacturer F01 is the delivery party, and logistics company L02 is the receiving party. Logistics company L02 confirms the quantity as 120, and manufacturer F01 submits the quantity as 120, resulting in a quantity confirmation result of 120 / 120. The image voucher and video keyframes submitted by logistics company L02 show that the seal number SEAL-73A2 is identifiable, and the seal status is complete. The outer packaging image area summary matches the initial box area summary in the initial voucher record, indicating that the outer packaging status is complete. Logistics company L02 did not raise any objections, and the objection status is "none". The circulation evidence preservation method combines the previous node hash C0-GENESIS-6E2A, the cross-format evidence root E1-8C21A, the delivery party identifier manufacturer F01, the receiving party identifier logistics company L02, the quantity confirmation result 120 / 120, the seal status complete, the outer packaging status complete, and the objection status "none" to generate the handover responsibility edge B1.

[0134] The responsibility edge tension value is determined based on the cross-format evidence closure residual, the differences in quantity, seal status, and outer packaging status between adjacent circulation nodes, and the objection status. In this embodiment, the quantity difference is 0, the seal status difference is 0.02, the outer packaging status difference is 0.03, and the objection status is 0. With the weights of the cross-format evidence closure residual, quantity difference, seal status difference, outer packaging status difference, and objection status being 1.00, 0.50, 0.40, 0.17, and 0.60, respectively, the responsibility edge tension value is approximately 0.104. The sealing condition is set to a cross-format evidence closure residual not exceeding 0.20 and a responsibility edge tension value not exceeding 0.25. This sealing condition is determined based on the distribution of cross-format evidence closure residual and responsibility edge tension value in historical normal evidence samples of similar circulation nodes, and is used to limit the consistency deviation of multi-format documents within the current circulation node and the handover status deviation between adjacent circulation nodes. Since the cross-format evidence closure residual of 0.091 is not greater than 0.20 and the responsibility edge tension value of 0.104 is not greater than 0.25, the sealing status meets the sealing conditions.

[0135] like Figure 5 As shown, Figure 5 This is a diagram showing the handover responsibility edge and sealing status determination in an embodiment of the present invention. Figure 5 The upper part displays the normal sealing path. Input results include the cross-format evidence closure residual of 0.091, the responsibility edge tension value of 0.104, and the sealing conditions D≤0.20 and T≤0.25. After the sealing state meets the sealing conditions, the circulation evidence preservation method generates a node hash C1 based on the previous node hash C0, the circulation object identity anchor A, the cross-format evidence root E1, and the handover responsibility edge B1, and connects the node hash C1 to the node hash chain, forming a chain relationship from C0 to C1. Figure 5The middle section displays the handover responsibility edge B1 and the handover status. Handover responsibility edge B1 includes the previous node hash C0-GENESIS-6E2A, the cross-format evidence root E1-8C21A, the handover entity identifier manufacturer F01, and the receiving entity identifier logistics L02. The handover status includes quantity confirmation result 120 / 120, seal status intact, outer packaging status intact, and no objection status. The replay verification result shows the node hash chain is from C0 to C1, the handover responsibility edge is B1, there are no abnormal fork nodes, and the responsibility attribution is a normal transfer.

[0136] Figure 5 The lower section displays anomaly sample N2. Anomaly sample N2 has a cross-format evidence closure residual of 0.340, a responsibility edge tension value of 0.410, a damaged seal status, and an objection status of "raised." Since the cross-format evidence closure residual of 0.340 is greater than 0.20, and the responsibility edge tension value of 0.410 is greater than 0.25, the sealing status does not meet the sealing conditions. The circulation evidence preservation method determines the last valid node hash C1 in the node hash chain that most recently met the sealing conditions as the last valid node hash, and generates an anomalous fork node X2 based on the last valid node hash C1, the cross-format evidence root E2, the cross-format evidence closure residual D2, and the responsibility edge tension value T2. The anomalous fork node X2 forms a fork association with C1, and the responsibility attribution is marked as pending verification.

[0137] During the replay verification phase, the circulation evidence preservation method reads the node hash chain, the handover responsibility edge, and the abnormal fork node. For normal sealing paths, the circulation evidence preservation method recalculates the cross-format evidence root E1, the handover responsibility edge B1, and the node hash C1, and compares them with the records in the node hash chain. If the recalculation results are consistent, the current circulation node has not experienced chain breakage, mismatch, or replacement, and the responsibility attribution result is a normal transfer. For abnormal sample N2, the circulation evidence preservation method reads the cross-format evidence closure residual and responsibility edge tension value in the abnormal fork node X2, and combines this with the damaged seal status, the raised objection status, and the last valid node hash C1 to determine that the anomaly occurred in the handover interval after C1, and the responsibility attribution enters the pending verification state.

[0138] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0139] The various embodiments in this invention are described in a progressive manner. For the same or similar parts between the various embodiments, please refer to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A circulation and evidence preservation method based on node hash chains, characterized in that, The method includes: Retrieve business fields, encapsulated fields, and initial credential summary to generate identity anchors for circulation objects; A node challenge code is generated based on the hash of the previous node, the identity anchor of the circulating object, and the subject state vector. The node challenge code is bound to a multi-format voucher, and a challenge code verification result is generated based on the recognition status of the node challenge code in the multi-format voucher. The multi-format credentials are converted into typed evidence leaf nodes and cross-format evidence roots are constructed. A consistency result is generated based on the typified evidence leaf nodes, and the cross-format evidence closure residual is determined based on the consistency result and the challenge code verification result. Based on the cross-format evidence closure residual and handover status, generate the handover responsibility edge and responsibility edge tension value; The closure status is determined based on the cross-format evidence closure residual and the responsibility edge tension value; if the closure status meets the closure conditions, a node hash of the access node hash chain is generated; if the closure status does not meet the closure conditions, an abnormal fork node is generated; and replay verification is performed based on the node hash chain and the handover responsibility edge, and if the abnormal fork node is generated, the chain break, mismatch, replacement, or responsibility attribution result is determined in conjunction with the abnormal fork node.

2. The method according to claim 1, characterized in that, The generation of the circulation object identity anchor includes: The order number, batch number, product model and quantity are extracted from the business field, and the box number, seal number and waybill number are extracted from the packaging field. The extracted fields are normalized and then combined with the initial voucher summary to generate the circulation object identity anchor.

3. The method according to claim 1, characterized in that, The main state vector includes manufacturer state count, logistics state count, distributor state count, and platform state count. The main state vector is updated in response to the operation of the circulation entity submitting vouchers, confirming handover, raising objections, or supplementing confirmation, and serves as the input for generating the node challenge code.

4. The method according to claim 1, characterized in that, The step of binding the node challenge code to a multi-format credential includes: The node challenge code is written into the electronic confirmation certificate, or the node challenge code appears together with the seal, waybill, or box in the image certificate or video keyframe, and the challenge code verification result is generated based on the recognition status of the node challenge code in the multi-format certificate.

5. The method according to claim 1, characterized in that, The typified evidence leaf nodes include image evidence leaf nodes, video evidence leaf nodes, PDF evidence leaf nodes, list evidence leaf nodes, and waybill evidence leaf nodes. Each of the typified evidence leaf nodes carries the voucher type and original text summary, and carries at least one of the following: field summary, image region summary, and submitting entity identifier, in order to construct the cross-format evidence root.

6. The method according to claim 5, characterized in that, The construction of cross-format evidence roots includes: The leaf nodes of the categorized evidence are sorted according to the document type, submission time, and document number. A hash tree is constructed based on the sorted leaf nodes of the categorized evidence, and the root hash of the hash tree is determined as the cross-format evidence root.

7. The method according to claim 1, characterized in that, The step of generating a consistency result based on the typed evidence leaf nodes, and determining the cross-format evidence closure residual based on the consistency result and the challenge code verification result, includes: Extract the fields to be verified and image region summaries from multiple types of evidence leaf nodes, generate a field consistency matrix and a region consistency result, determine the field consistency matrix and the region consistency result as the consistency result, and generate the cross-format evidence closure residual based on the consistency result and the challenge code verification result.

8. The method according to claim 7, characterized in that, The generation of the cross-format evidence closure residual includes: The field conflict residual is determined based on the field values ​​and identification confidence of the same field to be verified in different typified evidence leaf nodes. The image matching residual is determined based on the image region summary of the current circulation node and the initial image region summary extracted from the initial document that generates the initial document summary and bound to the initial document summary. The challenge code residual is determined based on the challenge code verification result. The cross-format evidence closure residual is generated based on the field conflict residual, the image matching residual and the challenge code residual.

9. The method according to claim 1, characterized in that, The handover responsibility edge includes the previous node hash, the cross-format evidence root, the handover subject identifier, the receiving subject identifier, the quantity confirmation result, the seal status, the outer packaging status, and the objection status. The responsibility edge tension value is determined based on the cross-format evidence closure residual, the state differences of adjacent circulation nodes in quantity, seal status, and outer packaging status, and the objection status, and is used to determine the sealing status.

10. The method according to claim 1, characterized in that, The generation of the node hash of the access node hash chain, or the generation of abnormal fork nodes, includes: When the sealing state meets the sealing conditions, a node hash is generated based on the digest of the previous node hash, the identity anchor of the circulating object, the cross-format evidence root, and the handover responsibility edge, and the node hash is connected to the node hash chain; If the sealing state does not meet the sealing conditions, the node hash that most recently met the sealing conditions in the node hash chain is determined as the last valid node hash, and an abnormal fork node is generated based on the last valid node hash, the cross-format evidence root, the cross-format evidence closure residual, and the responsibility edge tension value.

Images