Unmanned aerial vehicle inspection method and system based on unmanned aerial vehicle nest

Abstract

The application discloses a UAV inspection method and system based on a UAV nest, relates to the technical field of UAV inspection, and specifically comprises the following steps: analyzing data collection behaviors triggered by a UAV in an inspection path, generating corresponding data collection codes, constructing a behavior intention sequence according to a triggering sequence of the data collection codes, mapping the behavior intention sequence with an inspection task code and an inspection path progress, and when the UAV nest outputs an inspection completion determination, identifying a case where the UAV triggers the data collection behaviors and the UAV nest determines that the inspection is completed through differences between the behavior intention sequence and the inspection path progress. The application solves the problem that the UAV nest determines the task completion based on the inspection path and ignores the lag of the data collection behaviors of the UAV, which leads to the early end of the task, realizes dynamic correction and end control of the task determination based on the inspection behavior state, and thus improves the inspection data integrity and the task execution accuracy.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) inspection technology, and specifically to a UAV inspection method and system based on UAV nests. Background Technology

[0002] Drone inspection based on drone nests refers to a technical solution that uses drone nests, equipped with drone parking, automatic charging, communication connectivity, and control management functions, as ground infrastructure to centrally manage and schedule drones, thereby achieving automated execution of inspection tasks. In this technical system, the drone nest typically serves as the core node of the system, establishing communication connections with the backend management platform and the drones to issue inspection tasks and transmit inspection data back. Generally, after the backend management platform generates an inspection task, the task information is transmitted to the drone nest via the communication network. The nest then performs status checks on the currently parked drones and selects the appropriate drone to perform the inspection task based on the task requirements. Subsequently, the drone takes off under the control of the nest and inspects the target area according to a preset route or planned path. During the inspection, it collects image data or other sensor information and transmits the relevant data to the nest or backend platform via the communication link. After completing the inspection task, the drone returns to the nest and lands, entering a charging or standby state. In the above process, the system usually also manages the inspection task execution status, the drone operation status, and the data transmission process, thus forming an inspection process that covers task generation, task scheduling, drone take-off and landing control, inspection execution, and data processing.

[0003] The existing technology has the following shortcomings: In drone inspection methods based on drone nests, when a drone triggers data collection at inspection points during its inspection mission, and the drone nest determines that the drone has completed its inspection mission based on the inspection progress, a situation arises where the drone is still collecting critical inspection data while the drone nest has ended its task management. This is because the drone nest typically determines the task based on the completion status of the inspection path and does not perceive the lag in the data collection behavior triggered by the drone at the inspection point. This leads to a desynchronization between the inspection behavior and the task determination. Existing technologies cannot correct the task completion determination based on the inspection behavior status when the drone triggers data collection and the drone nest determines that the inspection is complete. This results in subsequent data collected by the drone not being included in the current inspection mission, leading to missing critical inspection data and incomplete inspection results.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a drone inspection method and system based on drone nests to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a drone inspection method based on drone nests, specifically including the following steps: S1. Analyze the data collection behavior triggered by the UAV in the inspection path, generate the corresponding data collection code, and construct the behavior intention sequence according to the triggering order of the data collection code. Map the behavior intention sequence with the inspection task code and the inspection path progress. When the UAV nest outputs the inspection completion judgment, identify the situation where the UAV triggers the data collection behavior and the UAV nest judges the inspection to be completed by the difference between the behavior intention sequence and the inspection path progress. S2. Perform mapping and transformation processing on the behavioral intent sequence to generate a state image set, and divide the state image set into an untriggered interval, a process interval, and a completed interval according to the execution process of the data acquisition encoding, so as to determine the inspection behavior status. S3. Perform reverse processing on the status image set and the inspection task code. When there is an untriggered interval or an in-process interval in the status image set, reverse the inspection completion judgment to keep the inspection task code in the execution state and continuously include the subsequent data collection code into the inspection task code. S4. Combine the interval transition process of the state mirror set to dynamically adjust the timing of the end of the inspection task coding. When the entire state mirror set enters the completion interval, the inspection task coding end processing is executed.

[0007] Preferably, S1 specifically includes the following steps: S101. Analyze the data collection behavior triggered by the UAV in the inspection path, extract the trigger signal identifier, the corresponding inspection path progress and the trigger time sequence for each data collection behavior, and separate and record the data collection behavior according to the trigger signal identifier. S102. Construct coding fields based on the inspection task code, inspection path progress and trigger time sequence, concatenate the coding fields in a preset order to generate data collection codes, and sort the data collection codes according to the trigger time sequence. S103. Construct a behavioral intent sequence based on the sorted data collection codes, and map each data collection code in the behavioral intent sequence to the inspection task code and the inspection path progress according to the inspection path progress. S104. When the UAV nest outputs the inspection completion judgment, the inspection path progress corresponding to each data acquisition code in the behavior intention sequence is compared with the inspection path progress item by item. When there is a data acquisition code that has not reached the inspection path progress, the situation where the UAV triggers the data acquisition behavior and the UAV nest determines that the inspection is completed is identified.

[0008] Preferably, S103 is as follows: The data collection codes are arranged sequentially according to their trigger time, and their positions are marked according to the progress of the inspection path. Data collection codes with continuous inspection path progress are sequentially connected to construct a sequence of behavioral intent. Based on the behavioral intent sequence, the data collection codes in the behavioral intent sequence are divided into intervals according to the progress of the inspection path, and the data collection codes are matched with the inspection task codes according to the progress of the inspection path to form a correspondence between the data collection codes and the inspection task codes. By combining the inspection path progress, each data collection code in the behavioral intent sequence is mapped item by item, and each data collection code in the behavioral intent sequence is mapped to a specific position in the inspection path progress, thus forming a corresponding mapping between the behavioral intent sequence, the inspection task code, and the inspection path progress.

[0009] Preferably, S104 is as follows: When the drone nest output inspection completion determination is made, the inspection path progress corresponding to each data collection code in the behavior intention sequence is extracted, and the inspection path progress corresponding to the drone nest output inspection completion determination is obtained. For each data collection code in the behavioral intent sequence, the progress of the inspection path is compared with the progress of the inspection path. The difference between the progress of the inspection path corresponding to each data collection code and the progress of the inspection path is calculated and the difference result is recorded. Based on the difference results, each data collection code is filtered. When there is a data collection code whose inspection path progress is less than the inspection path progress, the corresponding data collection code is determined as the data collection code that has not reached the inspection path progress, and the situation where the drone triggers the data collection behavior and the drone nest determines that the inspection is completed is identified.

[0010] Preferably, S2 specifically includes the following steps: S201. Perform mapping transformation processing on the behavioral intent sequence, extract each data acquisition code according to the execution order of the data acquisition code, rearrange the positions of each data acquisition code according to the execution order, convert the rearranged data acquisition code into the corresponding execution mark, and form a mapping result that corresponds one-to-one with the behavioral intent sequence. S202. Based on the mapping results, the execution markers corresponding to each data acquisition code are combined according to the execution order, and the execution markers are continuously merged to generate a state mirror set, so that each execution marker in the state mirror set maintains a corresponding relationship with the data acquisition code. S203. Based on the execution process of the data acquisition encoding, the state image set is traversed and analyzed. The occurrence order and change boundary of each execution mark are identified according to the execution order. When no execution mark appears in a continuous segment, it is divided into an untriggered interval. When the execution mark appears continuously without forming a termination change, it is divided into a progress interval. When the execution mark has a start and end corresponding change, it is divided into a completed interval. The inspection behavior status is determined according to the distribution position of the untriggered interval, progress interval and completed interval in the execution order.

[0011] Preferably, S202 specifically refers to: Based on the mapping results, the execution tags corresponding to each data acquisition code are read sequentially according to the execution order, and the execution tags are arranged item by item according to the execution order. The execution tags are then connected and combined in an adjacent order to form a continuous sequence of execution tags. For the execution mark sequence, adjacent execution marks are compared one by one according to the execution order. When adjacent execution marks are consistent, they are merged. When adjacent execution marks change, they are segmented. The merged execution mark sequence is constructed into a state mirror set, and each execution mark in the state mirror set is kept in correspondence with the data acquisition code according to the execution order.

[0012] Preferably, S3 specifically includes the following steps: S301. Perform reverse processing on the status image set and the inspection task code. Traverse the status image set item by item according to the execution order of the data acquisition code, and align and compare each interval in the status image set with the execution process corresponding to the inspection task code. Identify the untriggered intervals or intervals in the status image set based on the alignment and comparison results. S302. In the case that there is an untriggered interval or an in-process interval in the state mirror set, the inspection completion judgment is reversed, the inspection completion judgment is changed from the completed state to the execution state, and the inspection task code corresponding to the inspection completion judgment is kept in the execution state. S303. Under the condition that the inspection task code remains in the execution state, the subsequent data acquisition codes are received and processed according to the execution order of the data acquisition codes, and the subsequent data acquisition codes are sequentially assigned to the inspection task code, so that the subsequent data acquisition codes are continuously assigned to the inspection task code.

[0013] Preferably, S302 is as follows: If there are untriggered or in-process intervals in the status image set, the status image set is scanned item by item according to the execution order of the data acquisition codes to identify the data acquisition code range corresponding to the untriggered or in-process intervals, and the data acquisition code range is aligned with the execution process corresponding to the inspection task code. Based on the alignment results between the data collection coding range and the execution process corresponding to the inspection task coding, the inspection completion judgment is reversed and adjusted. The inspection completion judgment is changed from the completed state to the execution state, and the status of the inspection completion judgment corresponding to the inspection task coding is updated to the execution state, so that the inspection completion judgment corresponding to the inspection task coding remains in the execution state.

[0014] Preferably, S4 is as follows: By combining the interval transition process of the state mirror set, the state mirror set is traversed item by item according to the execution order of data acquisition encoding, the transition status of each interval in the state mirror set in the execution order is identified, and the transition status of each interval to the completed interval is determined. The timing of the end of the inspection task code is dynamically adjusted based on the interval transfer process. If there are cases where the inspection task code has not fully entered the completed interval, the timing of the end of the inspection task code is delayed to keep the inspection task code in the execution state. When the entire state image set has entered the completion interval, the inspection task code is processed to end, and the inspection task code is changed from the execution state to the completion state.

[0015] Preferably, the drone inspection system based on the drone nest includes a behavior recognition mapping module, a state modeling mapping module, a judgment back-inference correction module, and an end control execution module; The behavior recognition and mapping module parses the data collection behavior triggered by the UAV in the inspection path, generates the corresponding data collection code, and constructs the behavior intent sequence based on the triggering order of the data collection code. The behavior intent sequence is mapped with the inspection task code and the inspection path progress. When the UAV's nest outputs the inspection completion judgment, the difference between the behavior intent sequence and the inspection path progress is used to identify the situation where the UAV triggers the data collection behavior and the UAV nest judges the inspection to be completed. The state modeling and mapping module performs mapping and transformation processing on the behavioral intent sequence, generates a state image set, and divides the state image set into non-triggered interval, in progress interval and completed interval according to the execution process of data acquisition and encoding, so as to determine the inspection behavior state. The judgment and reverse correction module performs reverse processing on the state image set and the inspection task code. When there is an untriggered interval or an in-process interval in the state image set, the inspection completion judgment is reversed to keep the inspection task code in the execution state and the subsequent data collection code is continuously included in the inspection task code. The control execution module ends, and the timing of the end of the inspection task code is dynamically adjusted in conjunction with the interval transition process of the state image set. The inspection task code end processing is executed when all state image sets enter the completion interval.

[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention transforms the data collection behavior of UAVs during inspection from discrete events into continuously analyzable behavioral states by introducing behavioral intent sequences and state mirror sets. Furthermore, it models the correlation between these behavioral states and inspection task coding and inspection path progress, thus breaking through the traditional technical model that relies solely on the completion status of the inspection path for task determination. When the UAV's nest outputs the inspection completion determination, the difference recognition mechanism between the behavioral intent sequence and the inspection path progress accurately detects situations where the UAV has not yet completed data collection. This effectively identifies the asynchrony between the inspection behavior state and task determination, avoiding premature end of the inspection task due to delayed data collection, and fundamentally solving the technical defect of key inspection data not being included in the task results.

[0017] 2. When an inconsistency is identified between the inspection behavior status and the inspection completion determination, this invention employs a reverse processing mechanism between the state mirror set and the inspection task code to reversely adjust the inspection completion determination, restoring and maintaining the inspection task code in its execution state. Simultaneously, it continuously receives and incorporates subsequent data acquisition codes, thereby establishing a traceable and correctable dynamic control process for task determination. This processing method ensures that all data acquisition codes generated during the inspection process are incorporated into the inspection task code for unified management and processing, avoiding data omissions caused by premature task termination, improving the integrity and continuity of inspection data, and making the inspection results more accurately reflect the actual execution status of the UAV.

[0018] 3. In the final stage of the inspection task, this invention dynamically adjusts the termination time of the inspection task code by combining the interval transition process of the state mirror set. The task termination condition is changed from a single path completion judgment to a completion interval judgment mechanism based on behavioral state. The inspection task code termination process is only executed when all state mirror sets have entered the completion interval, thus achieving strict consistency between the termination time of the inspection task and the actual completion state of the inspection behavior. This dynamic adjustment mechanism not only improves the accuracy of the inspection task termination judgment but also enhances the adaptability of the entire UAV nest inspection system to complex inspection scenarios. It transforms the inspection process from static judgment to dynamic perception and control, comprehensively improving the reliability, accuracy, and system stability of UAV inspection operations. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0020] Figure 1 This is a schematic diagram of the process of the present invention.

[0021] Figure 2 This is a schematic diagram of the modules of the present invention. Detailed Implementation

[0022] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0023] This invention provides, for example Figure 1 The drone inspection method based on drone nests shown includes the following steps: S1. Analyze the data collection behavior triggered by the UAV in the inspection path, generate the corresponding data collection code, and construct the behavior intention sequence according to the triggering order of the data collection code. Map the behavior intention sequence with the inspection task code and the inspection path progress. When the UAV nest outputs the inspection completion judgment, identify the situation where the UAV triggers the data collection behavior and the UAV nest judges the inspection to be completed by the difference between the behavior intention sequence and the inspection path progress. In this embodiment, S1 specifically includes the following steps: S101. Analyze the data collection behavior triggered by the UAV in the inspection path, extract the trigger signal identifier, the corresponding inspection path progress and the trigger time sequence for each data collection behavior, and separate and record the data collection behavior according to the trigger signal identifier. During UAV inspection missions, data acquisition is typically triggered by onboard sensor signals, such as camera shutter trigger signals, infrared sensor sampling signals, or laser scanning trigger signals. Analyzing this data acquisition can be done by reading the event log or sensor trigger queue in the UAV flight control system, capturing each trigger signal, and extracting the trigger signal identifier, inspection path progress, and trigger time sequence. The inspection path progress can be obtained by projecting the UAV's real-time positioning coordinates onto a preset inspection path trajectory. For example, on a 1000-meter-long inspection path, when the UAV is at 300 meters, the corresponding inspection path progress is 0.3. The trigger time sequence can be achieved by sorting by timestamps. For instance, if three data acquisition events are collected with timestamps of 10:00:01, 10:00:03, and 10:00:05, the trigger time sequence increases sequentially. Based on this, the data acquisition events can be recorded separately according to the trigger signal identifier. A key-value mapping structure can be used to store different types of trigger signals in independent record sets. For example, in an actual inspection experiment, the UAV triggers image acquisition signals at path progress points of 0.25, 0.52, and 0.78, and infrared acquisition signals at path progress points of 0.30 and 0.55, respectively, during its flight along the route. The image acquisition behavior can be recorded as [(0.25, t1), (0.52, t2), (0.78, t3)], and the infrared acquisition behavior can be recorded as [(0.30, t4), (0.55, t5)], where t1 to t5 represent the corresponding time sequence. This enables independent parsing and structured storage of different data acquisition behaviors, providing basic data support for subsequent data acquisition encoding generation.

[0024] Unmanned aerial vehicles (UAVs) are flight vehicles that perform inspection missions, capable of spatial movement and triggering data acquisition actions according to a preset inspection path. The inspection path is a pre-planned spatial trajectory, which can consist of multiple consecutive path points, used to define the UAV's inspection range and movement sequence. Data acquisition actions refer to the specific acquisition actions triggered by sensors during the UAV's inspection process, such as image acquisition or environmental parameter acquisition. Trigger signal identifiers are used to distinguish different types of data acquisition actions and can be composed of sensor type or trigger command identifiers. The corresponding inspection path progress characterizes the position of the data acquisition action within the inspection path and can be represented by the path length ratio or path point sequence number. The trigger time sequence describes the sequential relationship of each data acquisition action and is usually obtained by timestamp sorting. Separate recording refers to storing different types of data acquisition actions separately according to the trigger signal identifiers, enabling each type of data acquisition action to be independently retrieved and analyzed in subsequent processing. Through the synergistic effect of the above elements, the discrete data acquisition actions generated during UAV inspections can be transformed into a structured data set with a clear time sequence and spatial location, thus providing a foundation for subsequent action intent sequence construction and inspection judgment.

[0025] S102. Construct coding fields based on the inspection task code, inspection path progress and trigger time sequence, concatenate the coding fields in a preset order to generate data collection codes, and sort the data collection codes according to the trigger time sequence. The inspection task code, inspection path progress, and trigger time sequence can be combined using structured data processing to form a data acquisition code with a unified format. Specifically, the inspection task code can first be encoded with a fixed length, for example, using the string identifier "TASK20260101A" to represent a specific inspection task. Then, the inspection path progress can be numerically normalized, representing the proportion of the drone's current position within the inspection path as a value between 0.000 and 1.000. For example, when the drone is at 420 meters of the total path length (1000 meters), the corresponding inspection path progress is 0.420. Finally, the trigger time sequence can be serialized by sorting timestamps and assigning them incremental numbers. For example, triggering data acquisition at 10:15:02, 10:15:05, and 10:15:09 corresponds to trigger timestamps of 001, 002, and 003, respectively. After extracting each field, the inspection task code, inspection path progress, and trigger time sequence are used as encoding fields and concatenated according to a preset order. For example, data acquisition codes are generated in the order of "inspection task code - inspection path progress - trigger time sequence". In an actual inspection test, the inspection task code for a certain transmission line is "TASK20260101A". The drone triggers data acquisition behavior at path progress positions of 0.128, 0.563, and 0.907 respectively, with corresponding trigger time sequences of 001, 002, and 003. This generates data acquisition codes "TASK20260101A-0.128-001", "TASK20260101A-0.563-002", and "TASK20260101A-0.907-003". All data acquisition codes are then sorted according to the trigger time sequence to obtain a time-ordered and position-resolvable encoding sequence.

[0026] The inspection task code distinguishes different inspection tasks and serves as a global identifier for data collection codes. It is typically composed of a task number, date information, or area identifier. The inspection path progress indicates the spatial location of the data collection behavior within the inspection path, which can be quantified by path length ratio or path node number. The trigger time sequence describes the chronological order of data collection behaviors, which can be generated as an increasing sequence value by sorting by timestamps. The coding field is a structured data item formed by independently extracting the inspection task code, inspection path progress, and trigger time sequence. The preset order limits the arrangement of the coding fields to ensure a unified format across different data collection codes. The data collection code is a combined identifier formed by concatenating multiple coding fields in a preset order, used to uniquely identify a single data collection behavior. By uniformly encoding the inspection task code, inspection path progress, and trigger time sequence, time and spatial information can be expressed in the same identifier. This allows data collection behaviors to be sorted, compared, and mapped according to unified rules during subsequent processing, thus providing a stable data foundation for constructing behavioral intent sequences.

[0027] S103. Construct a behavioral intent sequence based on the sorted data collection codes, and map each data collection code in the behavioral intent sequence to the inspection task code and the inspection path progress according to the inspection path progress. S104. When the UAV nest outputs the inspection completion judgment, the inspection path progress corresponding to each data acquisition code in the behavior intention sequence is compared with the inspection path progress item by item. When there is a data acquisition code that has not reached the inspection path progress, the situation where the UAV triggers the data acquisition behavior and the UAV nest determines that the inspection is completed is identified.

[0028] In this embodiment, S103 specifically refers to: The data collection codes are arranged sequentially according to their trigger time, and their positions are marked according to the progress of the inspection path. Data collection codes with continuous inspection path progress are sequentially connected to construct a sequence of behavioral intent. Based on the behavioral intent sequence, the data collection codes in the behavioral intent sequence are divided into intervals according to the progress of the inspection path, and the data collection codes are matched with the inspection task codes according to the progress of the inspection path to form a correspondence between the data collection codes and the inspection task codes. By combining the inspection path progress, each data collection code in the behavioral intent sequence is mapped item by item, and each data collection code in the behavioral intent sequence is mapped to a specific position in the inspection path progress, thus forming a corresponding mapping between the behavioral intent sequence, the inspection task code, and the inspection path progress.

[0029] When processing the sorted data acquisition codes, a linear sequence can first be formed according to the trigger time order, and the location of each data acquisition code can be marked based on the inspection path progress. Specifically, the inspection path progress field in each data acquisition code can be extracted as an independent variable and arranged sequentially according to the trigger time order to form an ordered list. Then, using the actual spatial trajectory of the inspection path as a reference, each inspection path progress is mapped to a specific segment location on the path. Continuity can be determined by the trend of inspection path progress changes between adjacent data acquisition codes. When the inspection path progress corresponding to adjacent data acquisition codes remains monotonically increasing and the direction of change is consistent, it is determined as a data acquisition code with continuous inspection path progress, and then sequentially connected according to the trigger time order to construct a behavioral intent sequence. This processing method transforms discretely triggered data acquisition codes into a sequence structure with a clear temporal order and spatial trajectory continuity, enabling subsequent processing to perform stable analysis based on the sequence structure.

[0030] Once the behavioral intent sequence is formed, it can be divided into intervals based on the progress of the inspection path, and the data collection codes and inspection task codes can be matched accordingly. Specifically, the behavioral intent sequence can be segmented according to the range of changes in the inspection path progress, grouping data collection codes with similar progress into the same interval, and maintaining the same trigger time order within each interval. Interval division can be achieved by comparing the differences in inspection path progress between adjacent data collection codes; when the trend of difference changes in a turning point or becomes discontinuous, this serves as the boundary for interval division. Subsequently, each data collection code is bound to an inspection task code, ensuring that each data collection code carries a corresponding inspection task code identifier, thus forming a stable correspondence between data collection codes and inspection task codes, and ensuring that the behavioral intent sequence has a consistent affiliation in the task dimension.

[0031] After completing the interval division and corresponding matching, the behavioral intent sequence can be mapped item by item in conjunction with the inspection path progress. Specifically, the inspection path can be represented as an ordered structure composed of continuous path segments. By performing position conversion on the inspection path progress, each data acquisition code in the behavioral intent sequence is mapped to a specific location within the corresponding path segment. During the mapping process, a path projection method can be used to map the inspection path progress to a relative position on the path trajectory and further determine its position index within the path structure, thereby achieving a one-to-one correspondence between "data acquisition code—inspection path position." This item-by-item mapping process ensures that each data acquisition code has a unique spatial location, enabling the behavioral intent sequence to fully reflect the actual behavioral trajectory of the UAV on the inspection path and forming a corresponding mapping structure between the behavioral intent sequence, the inspection task code, and the inspection path progress.

[0032] Inspection path progress describes the relative position of the UAV on the inspection path and is a fundamental parameter connecting time series and spatial position. Its value can be derived from path projection calculation or path node index conversion. Position calibration is the process of mapping the inspection path progress in the data acquisition code to the actual spatial position of the inspection path, giving the data acquisition behavior spatial positioning attributes. Data acquisition codes with continuous inspection path progress refer to a set of data acquisition codes whose inspection path progress changes monotonically and does not reverse direction when arranged in the order of trigger time. Behavioral intent sequence is an ordered set of multiple data acquisition codes arranged in the order of trigger time and forming a continuous trajectory on the inspection path. Interval division is based on the characteristics of the change in inspection path progress to classify the behavior. The process of segmenting the intent sequence is used to identify inspection behaviors at different stages; the inspection task code is used to identify the inspection task as a whole, so that the data collection code has a unified affiliation in the task dimension; the correspondence between the data collection code and the inspection task code is used to ensure that the data collection behavior can be correctly classified in subsequent processing; item-by-item mapping refers to performing spatial location correspondence processing on each data collection code in the behavior intent sequence; the specific position in the inspection path progress represents the actual spatial landing point of the data collection behavior in the inspection path; the correspondence mapping between the behavior intent sequence, the inspection task code, and the inspection path progress is used to construct a unified three-dimensional association structure of time, space, and task, so that the data collection behavior can be accurately described and support subsequent judgment processing.

[0033] In this embodiment, S104 specifically refers to: When the drone nest output inspection completion determination is made, the inspection path progress corresponding to each data collection code in the behavior intention sequence is extracted, and the inspection path progress corresponding to the drone nest output inspection completion determination is obtained. For each data collection code in the behavioral intent sequence, the progress of the inspection path is compared with the progress of the inspection path. The difference between the progress of the inspection path corresponding to each data collection code and the progress of the inspection path is calculated and the difference result is recorded. Based on the difference results, each data collection code is filtered. When there is a data collection code whose inspection path progress is less than the inspection path progress, the corresponding data collection code is determined as the data collection code that has not reached the inspection path progress, and the situation where the drone triggers the data collection behavior and the drone nest determines that the inspection is completed is identified.

[0034] After receiving the inspection status information transmitted back by the UAV, the UAV's homing system can generate an inspection completion determination based on the inspection path endpoint trigger flag or path coverage completion flag, and simultaneously extract the corresponding inspection path progress record value. This inspection path progress record value can be obtained from the inspection path progress field carried by the UAV during its last location report, or it can be directly determined from the path progress flag corresponding to the inspection path endpoint. Simultaneously, the system iterates through the behavioral intent sequence, reading each data acquisition code one by one, and parsing the inspection path progress field contained in each data acquisition code. The parsing results are written to a temporary storage set. This temporary storage set can adopt a sequential storage structure, ensuring a one-to-one correspondence between the data acquisition code and the corresponding inspection path progress, thus completing the unified data preparation for the behavioral intent sequence and the inspection completion determination.

[0035] After extracting the inspection path progress, a step-by-step comparison can be performed between the inspection path progress corresponding to each data acquisition code in the behavioral intent sequence and the actual inspection path progress. Specifically, a sequential traversal approach can be used to scan the temporary storage set. Each time the inspection path progress corresponding to a data acquisition code is read, it is compared with the inspection path progress recorded by the UAV's nest, and the offset between the two is written as the difference result into the corresponding data record entry. The difference result can be represented by a signed offset identifier to distinguish whether the path progress is before or after the inspection completion position. All difference results can be uniformly stored in a difference result set, maintaining the same order as the original data acquisition codes, thus forming a complete "data acquisition code—inspection path progress—difference result" association structure.

[0036] After obtaining the set of difference results, the data acquisition codes can be filtered based on these results. Specifically, the set of difference results can be iterated again, and a directional judgment can be performed on each difference result. When the difference result indicates that the inspection path progress corresponding to the data acquisition code is ahead of the inspection path progress recorded by the UAV's nest, the data acquisition code is marked as a data acquisition code that has not reached the inspection path progress and written into the incomplete set. This incomplete set can be stored as an independent data set to record all data acquisition codes that have not reached the inspection path progress. In an actual inspection experiment, some data acquisition codes in the behavioral intent sequence had their inspection path progress stuck in the middle of the path, while the UAV nest had already generated an inspection completion judgment based on the path endpoint. Through filtering, these data acquisition codes can be reliably identified, thus forming a set of data acquisition codes that have not reached the inspection path progress.

[0037] When a data acquisition code exists in the incomplete set, it can be determined that the UAV triggered a data acquisition action and the UAV nest determined the inspection was completed. The UAV nest is used for unified management of UAV inspection tasks. Its output inspection completion determination indicates that the inspection path coverage has reached the termination condition. The inspection path progress corresponding to each data acquisition code describes the spatial location where the data acquisition action occurred. The difference calculation reflects the offset relationship between the inspection path progress in the action intent sequence and the inspection path progress corresponding to the inspection completion determination. The difference result is used to characterize the offset direction and offset state. The existence of a data acquisition code with an inspection path progress less than the inspection path progress indicates that the data acquisition action occurred before the inspection completion position. Data acquisition codes that have not reached the inspection path progress indicate data items that are still in the incomplete acquisition state. When the UAV triggers a data acquisition action and the UAV nest determines the inspection is completed, it indicates that there is an inconsistency between the inspection action state and the task determination state. By constructing the difference result set and the incomplete set, the stable identification and data representation of this inconsistent state can be achieved, thus providing a reliable basis for subsequent task determination correction.

[0038] S2. Perform mapping and transformation processing on the behavioral intent sequence to generate a state image set, and divide the state image set into an untriggered interval, a process interval, and a completed interval according to the execution process of the data acquisition encoding, so as to determine the inspection behavior status. In this embodiment, S2 specifically includes the following steps: S201. Perform mapping transformation processing on the behavioral intent sequence, extract each data acquisition code according to the execution order of the data acquisition code, rearrange the positions of each data acquisition code according to the execution order, convert the rearranged data acquisition code into the corresponding execution mark, and form a mapping result that corresponds one-to-one with the behavioral intent sequence. When performing mapping transformation on a behavioral intent sequence, the structured conversion from data acquisition codes to execution markers can be achieved by parsing and reconstructing the execution order of the data acquisition codes. Specifically, the data acquisition codes are first read sequentially from the behavioral intent sequence according to their existing order. The execution order of the data acquisition codes is then determined based on the trigger time sequence or sequence position identifier contained within them. Next, the positions of the data acquisition codes are rearranged according to the execution order to ensure consistency with the sequence arrangement. After rearranging the positions, execution attributes, such as the acquisition trigger state or execution stage identifier, can be extracted for each data acquisition code. These attributes are then mapped to corresponding execution markers, which can use discrete identifiers to represent different execution stages. Taking a real-world inspection experiment as an example, during the inspection of a transmission line, the behavioral intent sequence contains multiple data acquisition codes in sequence. These data acquisition codes are distributed at different locations along the path according to their time sequence. By sorting their trigger times, a stable execution order can be obtained. Then, the data acquisition codes are rearranged according to the execution order, and an execution mark is generated based on the acquisition status corresponding to each data acquisition code. Finally, a mapping result corresponding to each item in the behavioral intent sequence is formed, so that each data acquisition code in the behavioral intent sequence has a clear execution mark expression, thus providing a foundation for the subsequent construction of the state mirror set.

[0039] The behavioral intent sequence is an ordered structure composed of multiple data acquisition codes arranged in execution order, used to describe the continuous data acquisition behaviors that occur during UAV inspections. Mapping and transformation processing converts the data acquisition codes in the behavioral intent sequence into another structural expression that can be used for state analysis. The execution order of the data acquisition codes characterizes the sequential execution relationship of each code during the inspection process and can be determined by trigger time or sequence position. Position rearrangement adjusts the arrangement of the data acquisition codes in the sequence according to the execution order, ensuring the sequence structure remains consistent with the execution order. The corresponding execution marker is an abstract representation of the execution state of the data acquisition code, used to describe the execution stage of the data acquisition behavior. The one-to-one mapping result with the behavioral intent sequence means that each data acquisition code corresponds to an execution marker, and the mapping relationship maintains a consistent order. This mapping result transforms the behavioral intent sequence from a single code sequence into a sequence structure containing execution state information, thus providing a foundation for subsequent state division and inspection behavior state determination.

[0040] S202. Based on the mapping results, the execution markers corresponding to each data acquisition code are combined according to the execution order, and the execution markers are continuously merged to generate a state mirror set, so that each execution marker in the state mirror set maintains a corresponding relationship with the data acquisition code. S203. Based on the execution process of the data acquisition encoding, the state image set is traversed and analyzed. The occurrence order and change boundary of each execution mark are identified according to the execution order. When no execution mark appears in a continuous segment, it is divided into an untriggered interval. When the execution mark appears continuously without forming a termination change, it is divided into a progress interval. When the execution mark has a start and end corresponding change, it is divided into a completed interval. The inspection behavior status is determined according to the distribution position of the untriggered interval, progress interval and completed interval in the execution order.

[0041] When processing the state image set, the execution sequence of the state image set can be traversed and analyzed based on the execution process of data acquisition and encoding. Each execution marker in the state image set is read item by item through continuous scanning, and the order of appearance of the execution markers and the change positions between adjacent execution markers are recorded during the reading process. Specifically, starting from the beginning of the state image set, execution markers are read sequentially, and the current execution marker is compared with the previous execution marker. When an execution marker appears for the first time, it is recorded as the starting position; when an execution marker changes, the change position is recorded as the boundary position, thus forming a segment sequence structure composed of multiple continuous segments. In actual inspection testing, the behavioral intent sequence is processed to form a state image set, where the execution markers exhibit a continuous change process of "not appearing—continuously appearing—change ending" during the execution process. Through traversal analysis, segments where execution markers do not appear consecutively can be identified as untriggered intervals; segments where execution markers appear continuously without forming a termination change can be identified as progress intervals; and segments where execution markers form a closed change from appearance to end can be identified as completed intervals. After the intervals are divided, further analysis can be performed based on the position of each interval in the execution order. When the completed interval is at the end of the execution process and the proceeding interval no longer continues, the corresponding execution status is determined as the inspection behavior completed status. When the proceeding interval is still at the end of the continuation position, the execution status is determined as the inspection behavior proceeding status, thereby realizing the structured determination of the inspection behavior status.

[0042] The execution process of data acquisition encoding describes the execution order of data acquisition encoding in the behavioral intent sequence and is the basic sequence basis for the entire traversal analysis. Traversal analysis is a process of sequentially reading and continuously comparing the state mirror set. The appearance order and change boundaries of each execution marker are used to characterize the starting position, duration, and change position of the execution marker in the execution process. The absence of an execution marker in a continuous segment indicates that no corresponding execution marker was detected in a segment of a continuous execution process, and such segments constitute a non-triggered interval. The non-triggered interval is used to represent the execution stage where data acquisition behavior has not yet occurred. The continuous appearance of execution markers without a termination change indicates that the execution markers remain within the continuous segment. In the state, this type of segment constitutes a progress interval; the progress interval is used to indicate that the data acquisition behavior is in the continuous execution phase; the corresponding changes in the start and end of the execution mark indicate that the execution mark has formed a complete appearance and end process in the execution process, and this type of segment constitutes a completion interval; the completion interval is used to indicate that the data acquisition behavior has formed a closed execution process; the distribution position in the execution sequence is used to describe the arrangement of each interval in the entire execution process; the inspection behavior state is the execution state determined based on the distribution position of the untriggered interval, the progress interval, and the completion interval in the execution sequence. Through the joint analysis of the segment structure and distribution position, the stable identification and expression of the inspection behavior state can be achieved.

[0043] In this embodiment, S202 specifically refers to: Based on the mapping results, the execution tags corresponding to each data acquisition code are read sequentially according to the execution order, and the execution tags are arranged item by item according to the execution order. The execution tags are then connected and combined in an adjacent order to form a continuous sequence of execution tags. For the execution mark sequence, adjacent execution marks are compared one by one according to the execution order. When adjacent execution marks are consistent, they are merged. When adjacent execution marks change, they are segmented. The merged execution mark sequence is constructed into a state mirror set, and each execution mark in the state mirror set is kept in correspondence with the data acquisition code according to the execution order.

[0044] The construction of the execution marker sequence can be achieved by sequentially traversing the mapping results. During the traversal, the execution markers corresponding to each data acquisition code are read item by item according to the execution order, and the read results are written sequentially into a sequential storage structure to form a continuously arranged sequence of execution markers. In specific implementation, a linear storage structure can be used to carry the execution markers, maintaining the execution order during the writing process, so that the execution markers form a one-dimensional continuous arrangement in the storage space. At the same time, the association index between each execution marker and the corresponding data acquisition code is recorded during the writing process, so that the execution marker sequence and the data acquisition code maintain a synchronous mapping relationship. In actual inspection tests, the behavioral intent sequence forms a set of execution markers after mapping and transformation. After sequential traversal and writing, a continuously arranged sequence of execution markers according to the execution order can be obtained, providing a basic data structure for subsequent adjacency relationship analysis.

[0045] When processing the execution mark sequence, adjacent execution marks can be compared item by item using a sequential scanning method, and a segment structure can be constructed synchronously during the scanning process. Specifically, the sequence is read starting from the beginning, with the current execution mark used as the initial mark for the segment. Adjacent execution marks are then read sequentially. When an adjacent execution mark matches the current one, it is added to the current segment and merged, while the segment boundary is updated. When an adjacent execution mark changes, the current segment is encapsulated into an independent segment structure and written into the segment set. The changed execution mark is then used as the new initial mark for the segment, and the scanning process continues. Through this continuous scanning and dynamic segment construction method, the execution mark sequence can be divided into multiple segment structures with continuous attributes, thus achieving unified execution of merging and segmentation processing.

[0046] After segmentation, all segment structures can be organized according to execution order to form a state mirror set, maintaining the sequential correspondence between execution markers and data acquisition codes within the state mirror set. A continuously arranged sequence of execution markers refers to a set of execution markers arranged without interruption in the storage structure according to execution order. Item-by-item comparison is the process of continuously reading and comparing the states of adjacent elements in the execution marker sequence. Consistency between adjacent execution markers indicates that continuously read execution markers maintain the same state attributes. Merging is the process of aggregating consecutive and consistent execution markers into a single segment. Changes in adjacent execution markers indicate a switch in execution state within the sequence. Segmentation is the process of dividing the execution marker sequence using the change position as the boundary. The merged execution marker sequence is a structured sequence composed of multiple continuous segments. The state mirror set is a set structure formed by arranging these segments according to execution order, used to reflect the continuous change trajectory of execution markers during the execution process. By constructing the state mirror set, execution markers can be transformed from a point-by-point sequence into a segmented expression, making the change process of execution state clearly presented structurally, thus providing a stable data foundation for subsequent interval segmentation and inspection behavior state determination.

[0047] S3. Perform reverse processing on the status image set and the inspection task code. When there is an untriggered interval or an in-process interval in the status image set, reverse the inspection completion judgment to keep the inspection task code in the execution state and continuously include the subsequent data collection code into the inspection task code. In this embodiment, S3 specifically includes the following steps: S301. Perform reverse processing on the status image set and the inspection task code. Traverse the status image set item by item according to the execution order of the data acquisition code, and align and compare each interval in the status image set with the execution process corresponding to the inspection task code. Identify the untriggered intervals or intervals in the status image set based on the alignment and comparison results. When performing reverse processing on the state image set and the inspection task code, the state image set can be traversed item by item according to the execution order of the data acquisition code. During the traversal, an execution process sequence corresponding to the inspection task code is constructed, aligning the position of each interval in the state image set in the execution order with the execution process position of the inspection task code. Specifically, a sequential scanning method can be used, reading each interval sequentially from the beginning of the state image set and synchronously mapping the read interval position with the process position of the inspection task code during execution. The correspondence between interval position and execution process is recorded during the mapping process. When an interval position and execution process are inconsistent, the inconsistent position is marked as an alignment offset point, and the difference segment is divided using this offset point as the boundary. In actual inspection experiments, there may be untriggered intervals or ongoing intervals in the state image set that extend into the later stages of the execution process, while the inspection task code has already reached its execution end position. Sequential alignment comparison can identify the range of inconsistencies between the intervals and the execution process, thereby determining the specific position of the untriggered or ongoing intervals in the execution order, providing a basis for subsequent judgment.

[0048] The state mirror set and the inspection task code represent the inspection behavior status and task execution process, respectively. The alignment comparison between the two is used to achieve reverse verification between the behavior status and the task status. The reverse calculation process is the process of comparing the execution process of the inspection task code with the state mirror set. The alignment comparison is a process of matching the position of each interval in the state mirror set in the execution order with the corresponding execution process position of the inspection task code. The untriggered intervals or in-process intervals in the state mirror set indicate that there are still data collection behaviors that have not started or completed in the execution process. The untriggered intervals are used to represent the segments where data collection behaviors have not yet been triggered, and the in-process intervals are used to represent the segments where data collection behaviors are in the continuous execution process. The alignment comparison can identify the offset relationship between the state mirror set and the inspection task code. When there are untriggered intervals or in-process intervals that extend to the execution end position corresponding to the inspection task code, it can be determined that the execution status of the inspection task code is inconsistent with the actual inspection behavior status, thus providing a basis for the reverse adjustment of the subsequent inspection completion determination.

[0049] S302. In the case that there is an untriggered interval or an in-process interval in the state mirror set, the inspection completion judgment is reversed, the inspection completion judgment is changed from the completed state to the execution state, and the inspection task code corresponding to the inspection completion judgment is kept in the execution state. S303. Under the condition that the inspection task code remains in the execution state, the subsequent data acquisition codes are received and processed according to the execution order of the data acquisition codes, and the subsequent data acquisition codes are sequentially assigned to the inspection task code, so that the subsequent data acquisition codes are continuously assigned to the inspection task code.

[0050] While the inspection task code remains in execution, the continuously transmitted data stream from the UAV can be monitored and processed. Subsequent data acquisition codes can be received and processed according to their execution order. Specifically, a continuous receiving entry point can be set within the execution process corresponding to the inspection task code. When the UAV triggers a new data acquisition action during the inspection, the generated data acquisition code is written to the receiving buffer in real time. The data acquisition codes in the receiving buffer are then arranged sequentially according to the execution order information contained within them, ensuring that subsequent data acquisition codes are corrected in order before entering the processing flow. In an actual inspection experiment, the UAV continued to trigger data acquisition actions at the end of the inspection path. Subsequent data acquisition codes were continuously transmitted back via wireless link. After being written to the buffer through the receiving entry point and arranged in execution order, the consistency of the subsequent data acquisition codes' order within the execution process is guaranteed, thus providing ordered input for processing.

[0051] After receiving and processing subsequent data acquisition codes, they can be sequentially assigned to the inspection task codes according to their execution order. Specifically, an extension space can be reserved in the storage structure corresponding to the inspection task codes. Upon receiving a new data acquisition code, it is inserted into the end of the sequence or the corresponding position of the inspection task code according to the execution order, and the execution process identifier of the inspection task code is updated, making the newly assigned data acquisition code part of the inspection task code. Simultaneously, receiving and assignment processing continues while the inspection task code remains in the execution state, ensuring that subsequent data acquisition codes form a continuously expanding structure within the inspection task code. In actual inspection, once the inspection completion judgment is reverse-adjusted to the execution state, subsequent data acquisition codes are continuously received and written to the storage structure corresponding to the inspection task code, thus ensuring that all subsequent data acquisition codes are included in the scope of the inspection task code and preventing the data acquisition behavior from becoming detached from the inspection task code.

[0052] In this embodiment, S302 specifically refers to: If there are untriggered or in-process intervals in the status image set, the status image set is scanned item by item according to the execution order of the data acquisition codes to identify the data acquisition code range corresponding to the untriggered or in-process intervals, and the data acquisition code range is aligned with the execution process corresponding to the inspection task code. Based on the alignment results between the data collection coding range and the execution process corresponding to the inspection task coding, the inspection completion judgment is reversed and adjusted. The inspection completion judgment is changed from the completed state to the execution state, and the status of the inspection completion judgment corresponding to the inspection task coding is updated to the execution state, so that the inspection completion judgment corresponding to the inspection task coding remains in the execution state.

[0053] When the state mirror set contains untriggered or ongoing intervals, continuous scanning can be performed on the state mirror set according to the execution order of data acquisition codes. During the scanning process, the start and end positions of each interval are identified, and the corresponding data acquisition code range is extracted based on the interval boundaries. Specifically, the state mirror set can be sequentially traversed, marking consecutive segments belonging to untriggered or ongoing intervals as a single interval. During traversal, the data acquisition codes corresponding to the start and end positions of each interval are recorded, forming a continuous set of data acquisition code ranges. Subsequently, this set of data acquisition code ranges is mapped to the execution process corresponding to the inspection task code. A positional correspondence is established through the execution order, allowing the data acquisition code range to obtain the corresponding segment position within the execution process. In actual inspection experiments, some ongoing intervals in the state mirror set extend to the end of the execution process, while the execution process corresponding to the inspection task code has ended. Continuous scanning and boundary identification can determine the data acquisition code range corresponding to this ongoing interval, and its coverage area can be marked within the execution process.

[0054] After aligning the data acquisition coding range with the execution process corresponding to the inspection task coding, a reverse adjustment process can be performed on the inspection completion judgment based on the alignment result. Specifically, the state storage location corresponding to the inspection completion judgment can be accessed, the original completion status identifier can be replaced, and the execution status identifier can be written to the same location. Simultaneously, the status identifier corresponding to the inspection task coding is updated to the execution status, ensuring consistency between the inspection completion judgment and the inspection task coding at the state level. During execution, reverse adjustment can be performed through state overwriting. That is, when it is detected that the data acquisition coding range overwrites the execution end position, a state rollback operation is performed on the inspection completion judgment, causing the inspection task coding to re-enter the execution state. In actual inspection processes, when the inspection completion judgment has output a completion status but the corresponding data acquisition coding range still exists, the reverse adjustment process triggered by the alignment result can restore the inspection completion judgment to the execution state.

[0055] The presence of untriggered or in-process intervals in the state mirror set indicates that there are still data acquisition behaviors that have not started or completed in the execution process. The data acquisition coding range corresponding to the untriggered or in-process intervals is a continuous set of data acquisition codes obtained by identifying the interval boundaries. The inspection completion judgment is an identifier used to mark the end status of the inspection task. The reverse adjustment process is a process of rolling back and replacing the inspection completion judgment status. Adjusting from the completion status to the execution status means replacing the status identifier corresponding to the inspection completion judgment with the execution identifier. By aligning the state mirror set with the execution process corresponding to the inspection task code, the offset between the execution process and the actual inspection behavior can be identified. When the data acquisition coding range corresponding to the untriggered or in-process interval covers the execution end position, the reverse adjustment process restores the inspection task code to the execution status, thereby ensuring that the inspection task continues to maintain a valid state in the execution process and providing conditions for subsequent data acquisition codes to be continuously incorporated into the inspection task code.

[0056] S4. Combine the interval transition process of the state mirror set to dynamically adjust the timing of the end of the inspection task coding. When the entire state mirror set enters the completion interval, the inspection task coding end processing is executed.

[0057] In this embodiment, S4 specifically refers to: By combining the interval transition process of the state mirror set, the state mirror set is traversed item by item according to the execution order of data acquisition encoding, the transition status of each interval in the state mirror set in the execution order is identified, and the transition status of each interval to the completed interval is determined. The timing of the end of the inspection task code is dynamically adjusted based on the interval transfer process. If there are cases where the inspection task code has not fully entered the completed interval, the timing of the end of the inspection task code is delayed to keep the inspection task code in the execution state. When the entire state image set has entered the completion interval, the inspection task code is processed to end, and the inspection task code is changed from the execution state to the completion state.

[0058] When combining the interval transition process with the state mirror set, a continuously updated record can be established for the state mirror set to continuously track the state changes of each interval in the execution sequence. In the specific implementation, a state record bit can be set for each interval in the state mirror set, and the state mirror set can be updated and scanned after each data acquisition and encoding input. By comparing the interval states before and after the update, the change from an untriggered interval or a progress interval to a completed interval can be identified, and the change position is recorded as an interval transition identifier, thus forming a continuous description of the interval transition process. In actual inspection experiments, as data acquisition and encoding are continuously input, some progress intervals gradually transform into completed intervals. Through continuous updates and state comparison, each interval transition can be accurately identified, making the interval transition process continuously traceable.

[0059] When dynamically adjusting the termination time of inspection task codes based on the interval transfer process, this can be achieved by setting an termination control flag corresponding to the inspection task code. During execution, the termination control flag can be checked after each interval transfer identification. If there are still untriggered or in progress intervals in the state mirror set, the termination control flag is maintained, keeping the inspection task code running and preventing it from entering the termination process. In actual inspection, when some intervals have completed transfers but others have not, continuously checking the termination control flag can prevent premature termination of the inspection task code, thus ensuring consistency between the inspection behavior state and the task execution state.

[0060] When the entire state image set has entered the completion interval, the inspection task coding termination process can be triggered by performing an integrity check on the state image set. Specifically, a full scan of the state image set can be performed. If the scan result indicates that there are no untriggered or in-process intervals, this state is used as the criterion for determining that the entire set has entered the completion interval, and the inspection task coding termination process is triggered. In actual inspection, after the last in-process interval completes its transition to the completion interval, a full scan can confirm that the entire state image set has entered the completion interval, thus initiating the termination process.

[0061] When the inspection task code ends, a status overwrite operation can be performed on the status storage location corresponding to the inspection task code, replacing the execution status identifier with the completion status identifier. Simultaneously, the data acquisition code receiving entry point is closed, and the process of transferring data acquisition codes to the inspection task code is terminated, preventing the inspection task code from receiving new data acquisition codes. In actual execution, after the end-of-process is triggered, the data stream receiving and transfer operations are synchronously closed, thus forming a clear task termination boundary.

[0062] The interval transition process of the state mirror set is used to describe the change trajectory of untriggered intervals, in-process intervals, and completed intervals in the execution order. The interval transition identifier is used to record the position where the interval state changes. The end time of the inspection task code is the trigger node when the inspection task transitions from the execution state to the completed state. Dynamic adjustment is the process of updating the end control identifier by continuously monitoring the interval transition process. All entering the completed interval indicates that there are no untriggered intervals or in-process intervals in the state mirror set. The inspection task code end processing is the process of overwriting the execution state of the inspection task code and terminating the data acquisition code ingress. By continuously tracking the interval transition process and dynamically adjusting the end control identifier, the end time of the inspection task code and the actual inspection behavior completion state can be synchronized, thereby ensuring the integrity and consistency of the inspection task.

[0063] like Figure 2 The drone inspection system based on drone nests shown includes a behavior recognition mapping module, a state modeling mapping module, a judgment back-inference correction module, and an end control execution module. The behavior recognition and mapping module parses the data collection behavior triggered by the UAV in the inspection path, generates the corresponding data collection code, and constructs the behavior intent sequence based on the triggering order of the data collection code. The behavior intent sequence is mapped with the inspection task code and the inspection path progress. When the UAV's nest outputs the inspection completion judgment, the difference between the behavior intent sequence and the inspection path progress is used to identify the situation where the UAV triggers the data collection behavior and the UAV nest judges the inspection to be completed. The state modeling and mapping module performs mapping and transformation processing on the behavioral intent sequence, generates a state image set, and divides the state image set into non-triggered interval, in progress interval and completed interval according to the execution process of data acquisition and encoding, so as to determine the inspection behavior state. The judgment and reverse correction module performs reverse processing on the state image set and the inspection task code. When there is an untriggered interval or an in-process interval in the state image set, the inspection completion judgment is reversed to keep the inspection task code in the execution state and the subsequent data collection code is continuously included in the inspection task code. The control execution module ends, and the timing of the end of the inspection task code is dynamically adjusted in conjunction with the interval transition process of the state image set. The inspection task code end processing is executed when all state image sets enter the completion interval.

[0064] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0065] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0066] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0067] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0068] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0069] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0070] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A drone inspection method based on drone nests, characterized in that, Specifically, the following steps are included: S1. Analyze the data collection behavior triggered by the UAV in the inspection path, generate the corresponding data collection code, and construct the behavior intention sequence according to the triggering order of the data collection code. Map the behavior intention sequence with the inspection task code and the inspection path progress. When the UAV nest outputs the inspection completion judgment, identify the situation where the UAV triggers the data collection behavior and the UAV nest judges the inspection to be completed by the difference between the behavior intention sequence and the inspection path progress. S2. Perform mapping and transformation processing on the behavioral intent sequence to generate a state image set, and divide the state image set into an untriggered interval, a process interval, and a completed interval according to the execution process of the data acquisition encoding, so as to determine the inspection behavior status. S3. Perform reverse processing on the status image set and the inspection task code. When there is an untriggered interval or an in-process interval in the status image set, reverse the inspection completion judgment to keep the inspection task code in the execution state and continuously include the subsequent data collection code into the inspection task code. S4. Combine the interval transition process of the state mirror set to dynamically adjust the timing of the end of the inspection task coding. When the entire state mirror set enters the completion interval, the inspection task coding end processing is executed.

2. The UAV inspection method based on UAV nests according to claim 1, characterized in that, S1 specifically includes the following steps: S101. Analyze the data collection behavior triggered by the UAV in the inspection path, extract the trigger signal identifier, the corresponding inspection path progress and the trigger time sequence for each data collection behavior, and separate and record the data collection behavior according to the trigger signal identifier. S102. Construct coding fields based on the inspection task code, inspection path progress and trigger time sequence, concatenate the coding fields in a preset order to generate data collection codes, and sort the data collection codes according to the trigger time sequence. S103. Construct a behavioral intent sequence based on the sorted data collection codes, and map each data collection code in the behavioral intent sequence to the inspection task code and the inspection path progress according to the inspection path progress. S104. When the UAV nest outputs the inspection completion judgment, the inspection path progress corresponding to each data acquisition code in the behavior intention sequence is compared with the inspection path progress item by item. When there is a data acquisition code that has not reached the inspection path progress, the situation where the UAV triggers the data acquisition behavior and the UAV nest determines that the inspection is completed is identified.

3. The UAV inspection method based on UAV nests according to claim 2, characterized in that, S103 specifically refers to: The data collection codes are arranged sequentially according to their trigger time, and their positions are marked according to the progress of the inspection path. Data collection codes with continuous inspection path progress are sequentially connected to construct a sequence of behavioral intent. Based on the behavioral intent sequence, the data collection codes in the behavioral intent sequence are divided into intervals according to the progress of the inspection path, and the data collection codes are matched with the inspection task codes according to the progress of the inspection path to form a correspondence between the data collection codes and the inspection task codes. By combining the inspection path progress, each data collection code in the behavioral intent sequence is mapped item by item, and each data collection code in the behavioral intent sequence is mapped to a specific position in the inspection path progress, thus forming a corresponding mapping between the behavioral intent sequence, the inspection task code, and the inspection path progress.

4. The UAV inspection method based on UAV nests according to claim 2, characterized in that, S104 specifically refers to: When the drone nest output inspection completion determination is made, the inspection path progress corresponding to each data collection code in the behavior intention sequence is extracted, and the inspection path progress corresponding to the drone nest output inspection completion determination is obtained. For each data collection code in the behavioral intent sequence, the progress of the inspection path is compared with the progress of the inspection path. The difference between the progress of the inspection path corresponding to each data collection code and the progress of the inspection path is calculated and the difference result is recorded. Based on the difference results, each data collection code is filtered. When there is a data collection code whose inspection path progress is less than the inspection path progress, the corresponding data collection code is determined as the data collection code that has not reached the inspection path progress, and the situation where the drone triggers the data collection behavior and the drone nest determines that the inspection is completed is identified.

5. The UAV inspection method based on UAV nests according to claim 1, characterized in that, S2 specifically includes the following steps: S201. Perform mapping transformation processing on the behavioral intent sequence, extract each data acquisition code according to the execution order of the data acquisition code, rearrange the positions of each data acquisition code according to the execution order, convert the rearranged data acquisition code into the corresponding execution mark, and form a mapping result that corresponds one-to-one with the behavioral intent sequence. S202. Based on the mapping results, the execution markers corresponding to each data acquisition code are combined according to the execution order, and the execution markers are continuously merged to generate a state mirror set, so that each execution marker in the state mirror set maintains a corresponding relationship with the data acquisition code. S203. Based on the execution process of the data acquisition encoding, the state image set is traversed and analyzed. The occurrence order and change boundary of each execution mark are identified according to the execution order. When no execution mark appears in a continuous segment, it is divided into an untriggered interval. When the execution mark appears continuously without forming a termination change, it is divided into a progress interval. When the execution mark has a start and end corresponding change, it is divided into a completed interval. The inspection behavior status is determined according to the distribution position of the untriggered interval, progress interval and completed interval in the execution order.

6. The UAV inspection method based on UAV nests according to claim 5, characterized in that, S202 specifically refers to: Based on the mapping results, the execution tags corresponding to each data acquisition code are read sequentially according to the execution order, and the execution tags are arranged item by item according to the execution order. The execution tags are then connected and combined in an adjacent order to form a continuous sequence of execution tags. For the execution mark sequence, adjacent execution marks are compared one by one according to the execution order. When adjacent execution marks are consistent, they are merged. When adjacent execution marks change, they are segmented. The merged execution mark sequence is constructed into a state mirror set, and each execution mark in the state mirror set is kept in correspondence with the data acquisition code according to the execution order.

7. The UAV inspection method based on UAV nests according to claim 1, characterized in that, S3 specifically includes the following steps: S301. Perform reverse processing on the status image set and the inspection task code. Traverse the status image set item by item according to the execution order of the data acquisition code, and align and compare each interval in the status image set with the execution process corresponding to the inspection task code. Identify the untriggered intervals or intervals in the status image set based on the alignment and comparison results. S302. In the case that there is an untriggered interval or an in-process interval in the state mirror set, the inspection completion judgment is reversed, the inspection completion judgment is changed from the completed state to the execution state, and the inspection task code corresponding to the inspection completion judgment is kept in the execution state. S303. Under the condition that the inspection task code remains in the execution state, the subsequent data acquisition codes are received and processed according to the execution order of the data acquisition codes, and the subsequent data acquisition codes are sequentially assigned to the inspection task code, so that the subsequent data acquisition codes are continuously assigned to the inspection task code.

8. The UAV inspection method based on UAV nests according to claim 7, characterized in that, S302 specifically refers to: If there are untriggered or in-process intervals in the status image set, the status image set is scanned item by item according to the execution order of the data acquisition codes to identify the data acquisition code range corresponding to the untriggered or in-process intervals, and the data acquisition code range is aligned with the execution process corresponding to the inspection task code. Based on the alignment results between the data collection coding range and the execution process corresponding to the inspection task coding, the inspection completion judgment is reversed and adjusted. The inspection completion judgment is changed from the completed state to the execution state, and the status of the inspection completion judgment corresponding to the inspection task coding is updated to the execution state, so that the inspection completion judgment corresponding to the inspection task coding remains in the execution state.

9. The UAV inspection method based on UAV nests according to claim 1, characterized in that, S4 specifically refers to: By combining the interval transition process of the state mirror set, the state mirror set is traversed item by item according to the execution order of data acquisition encoding, the transition status of each interval in the state mirror set in the execution order is identified, and the transition status of each interval to the completed interval is determined. The timing of the end of the inspection task code is dynamically adjusted based on the interval transfer process. If there are cases where the inspection task code has not fully entered the completed interval, the timing of the end of the inspection task code is delayed to keep the inspection task code in the execution state. When the entire state image set has entered the completion interval, the inspection task code is processed to end, and the inspection task code is changed from the execution state to the completion state.

10. A drone inspection system based on drone nests, used to implement the drone inspection method based on drone nests as described in any one of claims 1-9, characterized in that, It includes a behavior recognition mapping module, a state modeling mapping module, a judgment back-calculation correction module, and a termination control execution module; The behavior recognition and mapping module parses the data collection behavior triggered by the UAV in the inspection path, generates the corresponding data collection code, and constructs the behavior intent sequence based on the triggering order of the data collection code. The behavior intent sequence is mapped with the inspection task code and the inspection path progress. When the UAV's nest outputs the inspection completion judgment, the difference between the behavior intent sequence and the inspection path progress is used to identify the situation where the UAV triggers the data collection behavior and the UAV nest judges the inspection to be completed. The state modeling and mapping module performs mapping and transformation processing on the behavioral intent sequence, generates a state image set, and divides the state image set into non-triggered interval, in progress interval and completed interval according to the execution process of data acquisition and encoding, so as to determine the inspection behavior state. The judgment and reverse correction module performs reverse processing on the state image set and the inspection task code. When there is an untriggered interval or an in-process interval in the state image set, the inspection completion judgment is reversed to keep the inspection task code in the execution state and the subsequent data collection code is continuously included in the inspection task code. The control execution module ends, and the timing of the end of the inspection task code is dynamically adjusted in conjunction with the interval transition process of the state image set. The inspection task code end processing is executed when all state image sets enter the completion interval.

Images