Surveying and mapping geographic information data collection method, device and medium based on cloud computing

Abstract

This invention relates to the field of surveying and mapping geographic information data processing technology, and provides a cloud-based surveying and mapping geographic information data acquisition method, equipment, and medium, including: generating device identifiers and session identifiers in the cloud, distributing parameters, forming an observation stream according to a unified time base, and generating event identifiers to establish a unified identifier index; calculating metric summaries and generating signatures for the observation streams at the acquisition end, forming chain evidence nodes and uploading them with the data; performing chain consistency checks on the chain evidence nodes, writing data that passes the check to the event ledger, and sending data that fails the check to the isolation zone and triggering a retransmission task; performing real-time positioning and map building processing based on LiDAR data and IMU data in a compact combination to obtain residual indices and pose covariance matrices, and calculating quality scores; generating version differential patches and performing incremental merging in the cloud based on the tile set and patch index of the acquisition area, generating conflict packets and outputting rollback pointers when version conflicts are detected.

Description

Technical Field

[0001] This invention relates to the field of surveying and mapping geographic information data processing technology, specifically to a cloud computing-based surveying and mapping geographic information data acquisition method, equipment, and medium. Background Technology

[0002] Surveying and mapping geographic information refers to a complete set of technologies, data, results, and related industries that acquire, process, manage, analyze, visualize, and apply geospatial data through surveying, remote sensing, computer technology, and other means. It includes geographic features, coordinates, attributes, and information on temporal changes.

[0003] In mapping and data acquisition scenarios where the Global Navigation Satellite System (GNSS) is unavailable, existing integrated mobile mapping and data acquisition for underground spaces and indoor and outdoor spaces typically uses Simultaneous Localization and Mapping (SLAM) and post-collection methods to form point clouds and trajectories from LiDAR data and IMU data. However, due to the lack of unified identification indexes for data acquisition segments, quantification of time consistency errors, expression of trajectory uncertainty, and verifiable data acquisition evidence chains in the cloud, it is difficult to stably generate version differences for multiple acquisitions of the same area, automatic location and rollback of conflicting segments, and failure to trigger supplementary acquisition in closed loops due to unqualified acquisition quality. Summary of the Invention

[0004] To at least partially address the problems existing in the prior art, this invention proposes a cloud computing-based method, equipment, and medium for collecting surveying and mapping geographic information data.

[0005] In a first aspect, the cloud computing-based surveying and mapping geographic information data acquisition method provided by the present invention includes: Establish an event log and collection identifiers, acquire the observation stream, and generate event identifiers based on the collection identifiers and observation streams; A metric digest is calculated on the observation stream. A signature is generated using the metric digest and the event identifier as payloads. The metric digest, signature, and event identifier are encapsulated into a chain of evidence nodes. A consistency check is performed on the chain of evidence nodes. If the check passes, the data is written into the event ledger. The observation stream is processed to construct a location, resulting in a residual index and a pose covariance matrix. The residual index and pose covariance matrix are then weighted and fused to obtain a quality score. When the quality score is not less than the preset gating threshold, the observation stream is encapsulated into a qualified acquisition segment. Using the event log as an index, the qualified acquisition segments are differentially merged to generate a version patch.

[0006] In a preferred embodiment, the event identifier generation includes: The acquisition end divides the observation stream into windows according to a preset time window length, aggregating multi-source observations within the same time window into a single observation stream segment. For each observation stream segment, the acquisition end serializes the observation records into a byte sequence according to a fixed field order. The acquisition end calculates an observation summary for the byte sequence and uses the observation summary, along with the time window start and end points, tile number, device identifier, and session identifier, to constitute the input elements of an atomic event. The input elements are concatenated in sequence to form a byte string to be identified, and a hash operation is performed to obtain the event identifier. The event identifier calculation formula is as follows: , Where k represents the time window index. This represents the event identifier for the k-th time window, where H represents the hash function. This indicates the device identifier issued by the cloud. This indicates a session identifier generated and distributed from the cloud. Indicates the time window identifier. This indicates the tile identifier within the set of tiles in the region. This represents the observation stream summary calculated from the byte sequence.

[0007] As a preferred implementation, the generated event identifier also includes establishing a unified identifier index: The acquisition end writes the event identifier into the header of the observation stream segment and simultaneously establishes a unified identifier index table locally. The index table includes the event identifier, session identifier, time window start and end, tile number, observation summary, and local file offset address. When the acquisition end performs uploading or resuming of interrupted transmission of the observation stream segment, the acquisition end carries the event identifier and segment sequence number in the metadata of each uploaded segment and records the confirmation result of the corresponding segment in the cloud receipt, enabling the acquisition end to locate the specific observation stream segment and its segment boundary based on the unified identifier index table.

[0008] In a preferred embodiment, the metric summary calculated for the observed flow includes: The observation stream segment is serialized, and the serialized byte sequence is used as the object to be digested. A preset digest algorithm is used to calculate the metric digest of the time window, and the metric digest is concatenated with the metadata of the time window to form a message to be signed. The metric digest refers to the hash value calculated after re-serializing the header and data body of the complete observation stream segment.

[0009] In a preferred embodiment, performing consistency verification on the chain of evidence nodes includes: After writing the chained evidence nodes into the local evidence cache table and establishing a correspondence with the observation stream fragments, the upload is performed. During the upload, the chained evidence nodes, the block data of the observation stream fragments, and the block sequence number are used as metadata for the same upload transaction. After receiving the data, the cloud first reads the public key and algorithm identifier of the corresponding device according to the device registration table and verifies the signature value. After the verification is successful, the chained consistency check is performed on the chained evidence nodes. The previous evidence node is queried using the chain pointer value as an index, and the existence of the previous evidence node and the consistency of the session identifier are verified. The cloud verifies that the time window sequence number is incremented and that the start and end of the time window meet the non-overlapping rule. The cloud re-serializes the received observation stream fragments according to the same fixed field order as the acquisition end and calculates the metric digest, verifying that the metric digest is completely consistent with the metric digest in the chained evidence nodes.

[0010] As a preferred implementation, the writing of the event log includes: When all the verification and consistency checks pass, the cloud writes the chain evidence node into the event ledger. The ledger is saved using an append-only method. When any check fails, the chain evidence node and its corresponding observation data are written into the isolation area in blocks. At the same time, a retransmission task is generated and returned to the acquisition end. Based on the retransmission task, the corresponding block is located from the local evidence cache table and the block cache and retransmission is performed.

[0011] In a preferred embodiment, the calculation process of the quality score includes: The acquisition unit inputs the residual index and pose covariance matrix into the quality scoring function to calculate the quality score, and compares it with the gating threshold. The quality score calculation formula is as follows: , in This represents the quality score for the k-th time window, with values ​​ranging from 0 to 1. Represents an exponential function. Represents the residual index. Let represent the pose covariance matrix of the k-th time window. The trace of the covariance matrix is ​​represented. and These are the weight coefficients for the corresponding items.

[0012] In a preferred embodiment, the quality score and comparison with the gating threshold includes: When the quality score is greater than or equal to the gate threshold, the time window is encapsulated as a qualified acquisition segment and enters the incremental patching process. When the quality score is less than the gate threshold, a re-acquisition instruction is generated and sent back to the acquisition end for execution, so that the acquisition end can re-acquire the same spatial unit and form a replaceable time window segment.

[0013] In a second aspect, this application provides an electronic device, including: a processor; and a memory storing program instructions that, when executed by the processor, cause the electronic device to implement one or more embodiments of the first aspect described above.

[0014] In a third aspect, this application provides a storage medium having computer-readable instructions stored thereon, which, when executed by one or more processors, implement one or more embodiments of the first aspect described above.

[0015] Compared with the prior art, the present invention has the following advantages: 1. Event identifiers are generated by combining device identifiers, session identifiers, time window identifiers, tile identifiers, and observation summaries, and a unified identifier index table is established. This ensures that any observation stream segment and its blocks can be uniquely located and reused during uploading, retransmission, and retransmission, avoiding the problems of duplicate uploading and mismatched merging caused by the lack of unique anchor points in existing technologies.

[0016] 2. The acquisition end uses a tight combination of LiDAR and IMU data for real-time positioning and mapping, outputs residual index and pose covariance matrix, and calculates quality score based on these. By using a gating threshold, time window segments are judged as qualified or re-acquisition commands are triggered, so that quality control is changed from manual screening after the fact to real-time gating during acquisition, avoiding the problem of low-quality segments being mixed in and causing rework of the overall results in existing technologies. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation on the scope of this application.

[0018] Figure 1 This is a schematic diagram of the cloud computing-based surveying and mapping geographic information data acquisition method provided in an embodiment of the present invention; Figure 2 This is a comparison chart of the effects of the cloud computing-based surveying and mapping geographic information data acquisition method provided in this embodiment of the invention with existing technologies, where gray bars represent existing technologies and black bars represent the present invention. Detailed Implementation

[0019] To make the technical means, creative features, objectives, and effects of this invention easier to understand, the invention is further described below with reference to specific embodiments. However, the following embodiments are merely preferred embodiments of this invention and not all of them. Other embodiments obtained by those skilled in the art based on the embodiments described herein without creative effort are all within the protection scope of this invention. Example 1

[0020] To achieve the above objectives, please refer to Figures 1 to 2 This invention provides a cloud computing-based method for collecting surveying and mapping geographic information data, the method comprising the following steps: Step S1: Create an acquisition session and distribute parameters to form an observation stream with a unified time base; Step S2: Calculate the metric digest to generate the evidence chain, and perform consistency verification in the cloud; Step S3: Generate acquisition segments based on trajectory covariance matrix and quality gating; Step S4: Generate version differential patch and roll back conflicts, and package and deliver session results.

[0021] This method is implemented in the order of S1–S4, and its overall process is as follows: Establish an event log and collection identifiers, acquire the observation stream, and generate event identifiers based on the collection identifiers and observation streams; A metric digest is calculated on the observation stream. A signature is generated using the metric digest and the event identifier as payloads. The metric digest, signature, and event identifier are encapsulated into a chain of evidence nodes. A consistency check is performed on the chain of evidence nodes. If the check passes, the data is written into the event ledger. The observation stream is processed to construct a location, resulting in a residual index and a pose covariance matrix. The residual index and pose covariance matrix are then weighted and fused to obtain a quality score. When the quality score is not less than the preset gating threshold, the observation stream is encapsulated into a qualified acquisition segment. Using the event log as an index, the qualified acquisition segments are differentially merged to generate a version patch.

[0022] S1. The specific steps for creating an acquisition session and distributing parameters to form an observation stream with a unified time base are as follows: When creating a data acquisition task, the cloud receives device registration information reported by the acquisition terminal. The cloud refers to the surveying and mapping geographic information data acquisition and management system deployed on a cloud computing resource pool, and the acquisition terminal refers to the surveying and mapping data acquisition terminal equipment and its edge computing unit deployed at the acquisition site. The registration information includes the device serial number, sensor list, and fingerprint values ​​of its calibration files. The cloud generates a device identifier using the device serial number and registration time as input, and establishes a one-to-one correspondence between the device identifier and the registration information, writing it into the device registration table. When the acquisition terminal initiates the data acquisition start request, the cloud generates a session identifier for this task and creates a session record. The session record includes the device identifier, task number, start time, and parameter package version number. During session initialization, the cloud creates an event ledger using the device identifier and session identifier as keys to record the index information of verified evidence nodes and qualified acquisition segments in this acquisition session. Furthermore, the cloud generates a set of tiles for the collection area based on the coverage of the collection task. Specifically, the task coverage area is first projected onto a unified coordinate system, and the coverage area is divided into regular grids according to the preset tile side length, for example, a preferred tile side length of 20m. A tile number is generated for each grid, and the spatial boundary of the tile is recorded. The cloud combines all tile numbers and their spatial boundaries into a tile set, and generates a parameter package that encapsulates the sensor sampling configuration and the tile set together, which is then sent to the collection end. The sensor sampling configuration includes the lidar frame rate, inertial measurement sampling rate, image frame rate, observation segment time window length, and upload block size. After receiving the parameter package, the collection end performs a checksum verification on the parameter package. If the verification passes, the device identifier, session identifier, tile set, and sampling configuration are written into the local session header file, and the unified time reference establishment process is initiated accordingly. Specifically, the parameter packet verification process involves first reading the verification value and hash algorithm identifier written by the cloud from the packet header; then, the acquisition end concatenates all byte sequences in the parameter packet except for the verification value field in a fixed order to form a byte string to be verified, and uses the hash algorithm to calculate a local verification digest; the acquisition end compares the local verification digest with the packet header verification value byte by byte. If they are completely consistent, the parameter packet is determined to be unaltered and transmitted completely; otherwise, the verification fails and the parameter packet is discarded. At the same time, a retransmission request is sent to the cloud and the reason for the failure is recorded. When a unified time base is established, the acquisition end performs no less than N time synchronization interactions with the cloud at the beginning of the session, where N is preferably 5. Each interaction records the request sending time and response receiving time, and reads the reference timestamp carried by the cloud in the response. The acquisition end calculates the round-trip delay for each interaction and selects the one with the smallest round-trip delay from the N interactions as the time mapping basis for this session. It calculates the offset of the local clock relative to the cloud reference time and solidifies the offset as a session time mapping parameter. Subsequently, the acquisition end writes a correction timestamp under the unified time base to each lidar point cloud frame, inertial measurement sequence, and image frame. The correction timestamp is obtained by subtracting the offset from the locally recorded timestamp. After the observation stream is formed, the acquisition end divides the observation stream into windows according to the length of the observation slice time window, and integrates multi-source observations within the same time window. The observation stream segments are aggregated into a single observation stream segment, where the observation segment time window length is 2 seconds in one implementation. For each observation stream segment, the acquisition end serializes the observation records into a byte sequence according to a fixed field order, preferably sensor type, calibration timestamp, payload length, and payload data. The acquisition end calculates an observation summary for the byte sequence and uses the observation summary, along with the time window start and end, tile number, device identifier, and session identifier, to constitute the input elements of an atomic event. The observation summary refers to the hash value calculated from the serialized original observation byte sequence within the time window, used to generate an event identifier. When generating the event identifier, the acquisition end concatenates the device identifier, session identifier, time window identifier, tile number, and observation summary in that order to form a byte string to be identified, and performs a hash operation to obtain the event identifier. The event identifier calculation formula is as follows: , Where k represents the time window index. This represents the event identifier for the k-th time window, where H represents the hash function, preferably SHA-256. This indicates the device identifier issued by the cloud. This indicates a session identifier generated and distributed from the cloud. Indicates the time window identifier. This indicates the tile identifier within the set of tiles in the region. This represents a summary of the observation stream calculated from the byte sequence. The acquisition end writes the event identifier into the header of the observation stream segment and simultaneously establishes a unified identifier index table locally. The index table includes the event identifier, session identifier, time window start and end, tile number, observation summary, and local file offset address. When the acquisition end performs uploading or resuming of interrupted transmission of the observation stream segment, the acquisition end carries the event identifier and segment sequence number in the metadata of each uploaded segment and records the confirmation result of the corresponding segment in the cloud receipt. This enables the acquisition end to locate the specific observation stream segment and its segment boundary based on the unified identifier index table and perform subsequent transmission. Event identifiers are generated by combining device identifiers, session identifiers, time window identifiers, tile identifiers, and observation summaries, and a unified identifier index table is established. This ensures that any observation stream segment and its blocks can be uniquely located and reused during uploading, retransmission, and retransmission, avoiding the problems of duplicate uploading and mismatched merging caused by the lack of unique anchor points in existing technologies.

[0023] S2. The specific steps for calculating the metric digest to generate the evidence chain and performing consistency verification in the cloud are as follows: The acquisition end serializes the observation stream segments, uses the serialized byte sequence as the object to be digested, calculates the metric digest for the time window using a preset digest algorithm, and concatenates the metric digest with the metadata of the time window to form a message to be signed. The metric digest refers to the hash value calculated after re-serializing the header and data body of the complete observation stream segment, which is used for digital signature and building evidence chain.

[0024] Specifically, the preset digest algorithm is preferably SHA-256. It uses a byte sequence as input to calculate the metric digest for the time window. Specifically, the acquisition end first appends padding bits and the original length information to the end of the byte sequence to meet the requirement of 512-bit grouping. Then, the padded message is divided into several groups of 512 bits each, and each group is split into 32-bit words, generating 64 message words according to the message expansion rules specified by SHA-256. The acquisition end then uses the 8 initial chain values ​​and 64 round constants specified by SHA-256 to perform 64 rounds of compression iteration operations on each group. In each round, temporary values ​​are calculated and working variables are updated according to the specified selection function, majority function, and cyclic shift operation. After grouping, the working variables are accumulated back to the chain value. When all grouping is completed, the acquisition end concatenates the final 8 32-bit chain values ​​in a fixed order to form a 256-bit output, which is then saved as a metric digest in hexadecimal form.

[0025] The acquisition terminal calls the device's private key to perform a digital signature on the message to be signed, obtaining a signature value. The device's private key is generated internally within the Trusted Execution Environment (TEE) during the device's initial registration and stored as a key object in the protected storage area of ​​the TEE. A chain of evidence nodes is then constructed. The TEE refers to a secure execution area deployed within the edge computing unit of the acquisition terminal, divided by the processor's hardware security mechanism. The secure execution area and the normal operating area are hardware isolated in terms of instruction execution and storage access. The TEE includes a protected storage area and a controlled calling interface. The device's private key is generated internally within the TEE and stored in the protected storage area. The normal operating area can only submit messages to be signed and obtain signature values ​​through the controlled calling interface. The normal operating area does not obtain the plaintext content of the device's private key. The chain of evidence nodes includes an event identifier, a metric digest, a signature value, a chain pointer value of the previous evidence node, a time window sequence number, and a parameter package version number. The chain pointer value is taken from the chain digest of the previous evidence node.

[0026] Specifically, the process of calling the device's private key to perform a digital signature on the message to be signed involves the trusted execution environment at the acquisition end receiving a request to generate a key pair from the normal operating area. This request carries the device identifier. The trusted execution environment internally generates an SM2 key pair and stores the device's private key as a persistent key object in the protected storage area of ​​the trusted execution environment, while simultaneously writing a unique key alias for this key object. The trusted execution environment then exports the corresponding device public key to the normal operating area, which uploads the device public key along with the device identifier to the device registration table in the cloud to complete the binding. During the acquisition process, when a signature needs to be generated for the evidence node in the k-th time window, the acquisition end's normal operating area... The line area first constructs the message to be signed. Specifically, it performs deterministic encoding according to a fixed field order, using the TLV rule, where each field contains a type code, length, and value. Then, the normal execution area calculates the SM3 digest of the TLV byte sequence to obtain the message digest value, and submits the device identifier, key alias, message digest value, and algorithm identifier as a signature request to the signature interface of the Trusted Execution Environment (TEE). After receiving the request, the TEE first loads the corresponding key object from the protected storage area according to the key alias, verifies that the key object is in an available state and consistent with the device identifier, and then calls the SM2 signature primitive within the TEE to perform signature calculation on the message digest value to obtain the signature result.

[0027] After writing the chained evidence nodes into the local evidence cache table and establishing a correspondence with the observation stream fragments, the upload is performed. During the upload, the chained evidence nodes, the block data of the observation stream fragment, and the block sequence number are used as metadata for the same upload transaction. After receiving the data, the cloud first reads the public key and algorithm identifier of the corresponding device according to the device registration table, and verifies the signature value. After successful verification, the cloud performs chained consistency verification on the chained evidence nodes. Specifically, the cloud queries the previous evidence node using the chain pointer value as an index, and verifies that the previous evidence node exists and the session identifier is consistent. The cloud verifies that the time window sequence number is incremented and that the start and end of the time window meet the non-overlapping rule. The cloud then verifies the received observations. The stream segment is reserialized and a metric digest is calculated using the same fixed field order as the acquisition end. The metric digest is verified to be completely consistent with the metric digest within the chain of evidence nodes. When all the above verification and consistency checks pass, the cloud writes the chain of evidence node into the event ledger. The ledger is saved using an append-only method. When any check fails, the chain of evidence node and its corresponding observation data are written into the isolation area in blocks. At the same time, a retransmission task is generated and returned to the acquisition end. The retransmission task includes an event identifier, a failure reason code, and a retransmission deadline. Based on the retransmission task, the corresponding block is located from the local evidence cache table and the block cache and retransmission is performed.

[0028] S3. The specific steps for generating acquisition segments based on trajectory uncertainty estimation and quality gating are as follows: Under a unified time reference, tight-combination real-time localization and mapping (RTD) processing is performed on the lidar point cloud and inertial measurement data within the k-th time window to obtain the residual index and pose covariance matrix for that time window. Based on these, a quality score is calculated and gating is performed. Specifically, the acquisition end performs TRD processing within the time window... Keyframes for the lidar are selected according to fixed rules. Indicates the start timestamp. The end timestamp is indicated. The keyframe selection rule is that the translation increment of adjacent keyframes is not less than 0.3 meters or the heading angle increment is not less than 3 degrees. If either condition is met, it is determined as a keyframe. The acquisition end constructs a sliding window state quantity with keyframes as nodes. The state quantity includes the pose, velocity and inertial bias corresponding to the keyframe. The acquisition end pre-integrates the high-frequency sequence of IMU data within the time window according to the keyframe timestamp to generate inertial constraints between adjacent keyframes. The inertial pre-integration result is used as the first type of constraint factor for compact combination optimization. Distortion removal and voxel downsampling are performed on the point cloud of each keyframe. Distortion removal adopts the method of IMU data pre-integration attitude interpolation combined with point timestamp compensation to uniformly transform each point in the point cloud to the keyframe time. Based on the local sub-map accumulated in the previous time window, scanning matching is performed on the point cloud of the current keyframe. The scanning matching adopts a point-to-surface ICP algorithm with a fixed number of iterations, where the number of iterations is preferably 10. The registration error is constructed as the second type of constraint factor for compact combination optimization. The two types of factors are jointly solved in the same nonlinear least squares objective function, and the state variables are iteratively updated to complete the joint solution. The acquisition end uses the residuals of all factors at the end of the optimization to form a residual set, and calculates the residual index of this time window. The joint optimization objective formula is: , in Let represent the state vector to be solved, such as pose or velocity; let i represent the i-th constraint factor, such as IMU data or LiDAR data; let I represent the total number of constraint factors; and let argmin represent the variable corresponding to the minimum value. Represents the residual calculation function for the i-th constraint factor; The pose covariance matrix is ​​obtained by calculating the information matrix and inverting it based on the Jacobian matrix and weight matrix. The pose covariance matrix is ​​taken from the covariance sub-block of the pose at the end of the keyframe of the time window. The acquisition end inputs the residual index and the pose covariance matrix into the quality scoring function to calculate the quality score, and compares it with the gating threshold. The quality score calculation formula is as follows: , in This represents the quality score for the k-th time window, with values ​​ranging from 0 to 1. Represents an exponential function. Represents the residual index. Let represent the pose covariance matrix of the k-th time window. The trace of the covariance matrix is ​​used to compress multidimensional uncertainty into a single scalar. and These are the corresponding weight coefficients, which are distributed by the cloud in the parameter package and fixed within the session, such as the preferred option. Take 0.5, Take 0.1; When the quality score is greater than or equal to the gating threshold, the time window is encapsulated as a qualified acquisition segment and enters the incremental patching process. When the quality score is less than the gating threshold, a re-acquisition instruction is generated and sent back to the acquisition end for execution. The re-acquisition instruction includes a time window identifier, tile number, suggested re-acquisition path segment, suggested speed limit, and suggested sampling configuration increment, so that the acquisition end can re-acquire the same spatial unit and form a replaceable time window segment.

[0029] Specifically, the threshold value is preferably set to 0.8. First, multiple data collections are performed under different environments, and the quality score for each time window is calculated and compared with the actual positioning accuracy and map accuracy. Statistical analysis shows that when the quality score reaches 0.80, the positioning accuracy and map accuracy of the data meet the expected requirements, effectively avoiding data distortion caused by excessive errors. To ensure the stability and efficiency of the system, after multiple rounds of testing and optimization, the threshold value is finally set to 0.80 as the standard for judging whether the data is qualified. This threshold can both ensure data quality and adapt to different collection environments, avoiding excessive re-collection requests due to an overly strict threshold. The acquisition end uses a tight combination of LiDAR and IMU data for real-time positioning and mapping, outputs residual index and pose covariance matrix, and calculates quality score based on these. By using a gating threshold, time window segments are judged as qualified or re-acquisition commands are triggered, so that quality control is changed from manual screening after the fact to real-time gating during acquisition, avoiding the problem of low-quality segments being mixed in and causing rework of the overall results in existing technologies.

[0030] The specific steps for generating and rolling back version differential patches and packaging the delivery session results in S4 are as follows: When a regional data update request is received, the system first identifies the version differences between the currently updated region and the patch based on a predefined tile partitioning scheme. Then, the cloud determines the incremental patch range and generates a version differential patch based on each valid fragment in the collected regional tile set and the corresponding patch index. The differential patch is generated by comparing the differences between the target region and historical versions to calculate the added, updated, and deleted content, thus forming an incremental patch package. When merging patches in the cloud, the system first checks for version conflicts. If a version conflict is found during the merging process, a conflict detection mechanism is triggered. Specifically, based on the patch index and version control information, it determines whether the target segment has been modified or updated by other patches. If a conflict is found, a conflict package is generated, which includes the conflict content, conflict source, resolution strategy, etc., and is output through a rollback pointer. After completing the incremental merge and handling version conflicts, the cloud encapsulates the session results and generates a delivery signature to ensure the integrity and immutability of the data. Specifically, a unique signature identifier is generated by applying an encryption algorithm to the merged final dataset. This signature is generated based on the hash value of the data and includes metadata such as timestamps and version numbers to verify the correctness and compliance of the data delivery. The delivery signature is then packaged and output along with the data to ensure that the data has not been tampered with during transmission, storage, or delivery.

[0031] like Figure 2The graph shows a comparison of the effects of cloud computing-based surveying and mapping geographic information data acquisition methods, equipment, and media. The horizontal axis of the graph displays the indicators, and the vertical axis displays the performance value of each indicator, i.e., the strength of the technology's performance on that indicator. As can be seen from the graph, the present invention shows higher performance than the existing technology in all indicators. This indicates that the present invention has advantages in data acquisition, merging, version control, and delivery, especially in trajectory reliability and differential patching success rate, which are improved compared with the existing technology. Example 2

[0032] Based on the above embodiment 1, the cloud computing-based surveying and mapping geographic information data acquisition method in surveying and mapping acquisition scenarios where GNSS is unavailable is specifically as follows: Taking the collection of geographic information data for urban underground utility tunnels as an example; Step 1: When creating the pipeline gallery data acquisition task, the cloud receives the device registration information reported by the acquisition terminal via the wired network. The cloud generates a device identifier based on the device serial number and registration time, and associates it with the device registration table. When the acquisition terminal initiates a data acquisition start request, the cloud generates a session identifier and creates a session record. The cloud projects the task coverage area onto the local coordinate system according to the pipeline gallery design drawings, divides the grid according to the 20-meter tile side length, and generates a tile set and its spatial boundary. The cloud encapsulates the tile set and sensor sampling configuration into a parameter packet and sends it down via the wired network. After receiving the parameter packet, the acquisition terminal verifies its integrity. If the verification is successful, it writes the device identifier, session identifier, tile set, and sampling configuration into the local session header file, and then enters the unified time reference establishment process: the acquisition terminal performs N time synchronization interactions with the cloud via the wired network, selects the minimum round-trip delay to calculate the clock offset, and solidifies it as a session time mapping parameter to provide a unified time reference for subsequent multi-source data.

[0033] Step 2: The acquisition end serializes the generated multi-source observations under a unified time reference. For the k-th time window, the acquisition end concatenates the LiDAR point cloud and the panoramic image byte stream according to a fixed field order to form a byte sequence to be digested. Subsequently, the acquisition end uses the SHA-256 algorithm to calculate the metric digest of the sequence. It then calls the device private key pre-installed in the security chip for signing. This private key is generated and permanently stored inside the Trusted Execution Environment (TEE) when the device is deployed in the corridor. The corresponding public key has been pre-registered to the cloud. The acquisition end encodes the metric digest, time window identifier, and other metadata into TLV format and submits it to the TEE to generate an SM2 signature value. Using the signature result, the acquisition end constructs a chain of evidence nodes and writes the node into the local evidence cache table, binding it with the observation data. During upload, the evidence node and the observation data are sent together in blocks. After receiving the data, the cloud first verifies the authenticity of the signature using the bound device public key. Then, the cloud recalculates the metric digest of the received data and compares it with the record in the evidence node. After all verifications are successful, the evidence node is added to the event ledger. If any step fails, the batch of data is transferred to the isolation area, and a retransmission command is immediately generated and sent to the acquisition end to trigger the location and retransmission from the local cache.

[0034] Step 3: Based on a unified time reference, perform tight-combined SLAM processing on the lidar point cloud and IMU data within the k-th time window. Select keyframes from the distortion-free point cloud according to the rule of translation greater than 0.3 meters or heading change greater than 3 degrees. In sliding window optimization, fuse the IMU data pre-integration factor and the point-to-surface ICP algorithm scanning matching factor to jointly solve for pose, velocity, and bias. After optimization, calculate the covariance matrix of the pose of the final keyframe and the comprehensive residual index of all factors. Calculate the quality score using weight coefficients distributed from the cloud. Considering the repetitive nature of the structure within the utility tunnel and the tendency to accumulate errors, after extensive field testing and calibration, strictly set the gate threshold to 0.80. When the quality score is greater than or equal to 0.80, the data in that time window is considered qualified, and its pose result will be used to update the local sub-map and encapsulate the segment for subsequent processes. If the quality score is less than 0.80, a specific resampling instruction is generated.

[0035] Step 4: When the cloud receives a regional data update request for a specific corridor segment, it first locates the set of tiles to be updated according to the tile division scheme, using session identifier, time window sequence number, and SLAM quality score as markers. The cloud compares the latest collected data with historical versions, calculates the added, updated, or deleted content through point cloud registration and feature comparison, and generates a binary differential patch package. During the merging process, the cloud performs strict conflict detection. When multiple qualified versions of the same tile are detected, the conflict resolution mechanism prioritizes the data segment with the higher quality score and smaller pose covariance. If the quality scores are close, the decision is made based on the collection timestamp and sensor data integrity, and the information of the overwritten version is recorded in the conflict log. After all incremental merging is completed, the cloud encapsulates the final full-element dataset of the utility tunnel.

[0036] This embodiment provides an electronic device, which includes a processor and a memory. The memory is communicatively connected to the processor and stores program instructions that can be executed by the processor. The processor is used to execute the program instructions to implement all the steps of the surveying and mapping geographic information data acquisition method described in the above method embodiment.

[0037] The processor can be a central processing unit, or other general-purpose processors, digital signal processors, application-specific integrated circuits, field-programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., with data processing capabilities; the general-purpose processor can be a microprocessor or any conventional processor.

[0038] The memory may include various media capable of storing program instructions, such as read-only memory, random access memory, magnetic disk or optical disk; the memory and the processor may be integrated on the same chip, or they may be set separately and connected by bus communication.

[0039] This embodiment provides a storage medium storing computer-readable instructions. When these computer-readable instructions are executed by one or more processors, they implement all the steps of the surveying and mapping geographic information data acquisition method described in the above method embodiment.

[0040] The storage medium may be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk or solid-state drive, etc.; the computer-readable storage medium may also be various combinations of the above-mentioned storage media.

[0041] The embodiments of the present invention described above are subject to modification and change of method by those skilled in the art without departing from the embodiments and broader aspects of the present invention. The appended claims are intended to include all such modifications and changes of method that do not depart from the present invention.

Claims

1. A cloud computing-based method for collecting surveying and mapping geographic information data, characterized in that, include: Establish an event log and collection identifiers, acquire the observation stream, and generate event identifiers based on the collection identifiers and observation streams; A metric digest is calculated on the observation stream. A signature is generated using the metric digest and the event identifier as payloads. The metric digest, signature, and event identifier are encapsulated into a chain of evidence nodes. A consistency check is performed on the chain of evidence nodes. If the check passes, the data is written into the event ledger. The observation stream is processed to construct a location, resulting in a residual index and a pose covariance matrix. The residual index and pose covariance matrix are then weighted and fused to obtain a quality score. When the quality score is not less than the preset gating threshold, the observation stream is encapsulated into a qualified acquisition segment. Using the event log as an index, the qualified acquisition segments are differentially merged to generate a version patch.

2. The surveying and mapping geographic information data acquisition method according to claim 1, characterized in that, The generated event identifier includes: The acquisition end divides the observation stream into windows according to a preset time window length, aggregating multi-source observations within the same time window into a single observation stream segment. For each observation stream segment, the acquisition end serializes the observation records into a byte sequence according to a fixed field order. The acquisition end calculates an observation summary for the byte sequence and uses the observation summary, along with the time window start and end points, tile number, device identifier, and session identifier, to constitute the input elements of an atomic event. The input elements are concatenated in sequence to form a byte string to be identified, and a hash operation is performed to obtain the event identifier. The event identifier calculation formula is as follows: ， Where k represents the time window index. This represents the event identifier for the k-th time window, where H represents the hash function. This indicates the device identifier issued by the cloud. This indicates a session identifier generated and distributed from the cloud. Indicates the time window identifier. This indicates the tile identifier within the set of tiles in the region. This represents the observation stream summary calculated from the byte sequence.

3. The surveying and mapping geographic information data acquisition method according to claim 2, characterized in that, The generated event identifiers also include the step of establishing a unified identifier index: The acquisition end writes the event identifier into the header of the observation stream segment and simultaneously establishes a unified identifier index table locally. The index table includes the event identifier, session identifier, time window start and end, tile number, observation summary, and local file offset address. When the acquisition end performs uploading or resuming of interrupted transmission of the observation stream segment, the acquisition end carries the event identifier and segment sequence number in the metadata of each uploaded segment and records the confirmation result of the corresponding segment in the cloud receipt, enabling the acquisition end to locate the specific observation stream segment and its segment boundary based on the unified identifier index table.

4. The surveying and mapping geographic information data acquisition method according to claim 1, characterized in that, The metric summary calculated for the observed flow includes: The observation stream segment is serialized, and the serialized byte sequence is used as the object to be digested. A preset digest algorithm is used to calculate the metric digest of the time window, and the metric digest is concatenated with the metadata of the time window to form a message to be signed. The metric digest refers to the hash value calculated after re-serializing the header and data body of the complete observation stream segment.

5. The surveying and mapping geographic information data acquisition method according to claim 1, characterized in that, The consistency check performed on the chain of evidence nodes includes: After writing the chained evidence nodes into the local evidence cache table and establishing a correspondence with the observation stream fragments, the upload is performed. During the upload, the chained evidence nodes, the block data of the observation stream fragments, and the block sequence number are used as metadata for the same upload transaction. After receiving the data, the cloud first reads the public key and algorithm identifier of the corresponding device according to the device registration table and verifies the signature value. After the verification is successful, the chained consistency check is performed on the chained evidence nodes. The previous evidence node is queried using the chain pointer value as an index, and the existence of the previous evidence node and the consistency of the session identifier are verified. The cloud verifies that the time window sequence number is incremented and that the start and end of the time window meet the non-overlapping rule. The cloud re-serializes the received observation stream fragments according to the same fixed field order as the acquisition end and calculates the metric digest, verifying that the metric digest is completely consistent with the metric digest in the chained evidence nodes.

6. The surveying and mapping geographic information data acquisition method according to claim 5, characterized in that, The aforementioned write event log includes: When all the verification and consistency checks pass, the cloud writes the chain evidence node into the event ledger. The ledger is saved using an append-only method. When any check fails, the chain evidence node and its corresponding observation data are written into the isolation area in blocks. At the same time, a retransmission task is generated and returned to the acquisition end. Based on the retransmission task, the corresponding block is located from the local evidence cache table and the block cache and retransmission is performed.

7. The surveying and mapping geographic information data acquisition method according to claim 1, characterized in that, The calculation process for the quality score includes: The acquisition unit inputs the residual index and pose covariance matrix into the quality scoring function to calculate the quality score, and compares it with the gating threshold. The quality score calculation formula is as follows: ， in This represents the quality score for the k-th time window, with values ​​ranging from 0 to 1. Represents an exponential function. Represents the residual index. Let represent the pose covariance matrix of the k-th time window. The trace of the covariance matrix is ​​represented. and These are the weight coefficients for the corresponding items.

8. The surveying and mapping geographic information data acquisition method according to claim 1, characterized in that, The comparison between the quality score and the gate threshold includes: When the quality score is greater than or equal to the gate threshold, the time window is encapsulated as a qualified acquisition segment and enters the incremental patching process. When the quality score is less than the gate threshold, a re-acquisition instruction is generated and sent back to the acquisition end for execution, so that the acquisition end can re-acquire the same spatial unit and form a replaceable time window segment.

9. An electronic device, characterized in that, include: processor; as well as A memory storing program instructions that, when executed by the processor, cause the electronic device to perform the method according to any one of claims 1-8.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium storing computer-readable instructions thereon, which, when executed by one or more processors, implement the method as described in any one of claims 1-8.

Images