A sandbox agent felling analysis system based on a distributed queue drive

Abstract

This invention relates to the field of sandbox agent technology, specifically to a sandbox agent-based logging analysis system driven by a distributed queue. The system includes: a boundary repair module, which acquires logging boundary information based on the logging area and performs boundary validity verification to determine the logging baseline boundary; and a sandbox queue creation module, which determines an image analysis queue and a monitoring analysis queue based on the logging baseline boundary and updates the monitoring analysis queue based on boundary crossing tendencies. The image analysis queue acquires on-site image data to determine the logging progress area and performs advanced analysis to determine the logging operation area. The monitoring analysis queue determines the approximate area based on the logging operation boundary and the logging baseline boundary and predicts boundary crossing tendencies. This system achieves multi-project data isolation, automatic boundary repair, early warning of boundary crossing risks, and dynamic allocation of monitoring resources, improving the automation level of forestry logging monitoring.

Description

Technical Field

[0001] This invention relates to the field of sandbox agent technology, and more specifically to a sandbox agent harvesting analysis system based on distributed queue-driven processing. Background Technology

[0002] With the development of digital forestry supervision and spatial data processing technologies, the monitoring of logging operations is gradually shifting from manual on-site verification to automatic identification of remote sensing images, analysis of operation trajectories, and early warning of boundary risks. Existing monitoring systems typically employ centralized task scheduling and unified data storage. In scenarios where multiple logging projects are monitored concurrently, image data, boundary data, operation progress data, and risk assessment data from different projects are easily intertwined, leading to task scheduling conflicts, chaotic data access, and difficulties in anomaly detection.

[0003] While existing technologies can achieve asynchronous processing of interactive requests and allocation of computing power, they still have shortcomings for timber harvesting operations. They lack a customized regulatory architecture that combines sandboxes and distributed queues, cannot automatically repair topological defects at harvesting boundaries, cannot build image analysis queues and regulatory analysis queues based on the number of trees in the harvesting area, cannot identify harvesting operation areas by combining the progress of operations, and cannot dynamically update the regulatory queues by combining the tendency to cross boundaries and the remaining time before the boundary approaches. This results in a mismatch between monitoring resources and operational risks, and a lag in the identification of risks of crossing boundaries.

[0004] Chinese Patent Publication No. CN121326535A discloses an automated AI agent's interaction with the environment and resource management method and system. The method includes: intercepting and taking over all interaction requests between the agent and the environment; storing pending interaction requests, sorting and managing the requests according to a preset or dynamic scheduling strategy to achieve asynchronous processing of the requests; packaging trajectory fragments before the call occurs into a partial trajectory sample and sending it to an experience replay pool for use by the trainer; and being responsible for actually executing tool call requests in the interaction requests, monitoring the call frequency, time consumption, and resource consumption of different tools, and allocating and adjusting computing resources for different types of requests. Simultaneously, based on the analysis of historical tool call data and call behavior, a smoothing estimation method is used to predict the agent's usage needs for the environmental sandbox and pre-start the sandbox.

[0005] Therefore, the following problems exist in the existing technology: Existing technologies do not consider using image analysis queues and regulatory analysis queues to achieve data isolation for logging monitoring operations, nor do they consider conducting advance analysis of operations through image analysis queues after repairing topological defects at the logging baseline boundary, and then adaptively updating the regulatory analysis queue based on the tendency to cross the boundary and the remaining time for boundary approach. This results in data crosstalk during concurrent monitoring of multiple projects, risks of delayed judgment due to boundary crossing, and monitoring scheduling that cannot keep up with the dynamic changes in logging operation risks. Summary of the Invention

[0006] To address these issues, this invention provides a sandbox-based intelligent agent logging analysis system driven by a distributed queue. This system overcomes the technical problems of existing intelligent agent monitoring architectures that lack image analysis queues and monitoring analysis queues, are unable to automatically repair the topological integrity of logging operation baseline boundaries, cannot dynamically adjust image acquisition based on the progress of the operation, and cannot update the monitoring analysis queue based on the risk of exceeding boundaries and the remaining time before the boundary approaches.

[0007] To achieve the above objectives, this invention provides a sandbox-based intelligent agent harvesting analysis system driven by a distributed queue, comprising: The boundary repair module obtains logging boundary information based on the logging area, performs boundary validity verification based on the logging boundary information, and determines the logging benchmark boundary. The sandbox queue creation module determines the image analysis queue and the monitoring analysis queue based on the logging baseline boundary and the number of trees in the corresponding logging area; and determines and updates the monitoring analysis queue based on the tendency to cross the boundary, so as to monitor the logging situation in real time. The image analysis queue includes several image analysis units. Each image analysis unit is used to acquire on-site image data of the corresponding logging analysis area within the current image acquisition cycle to determine the logging progress area; and to perform advanced analysis based on the logging progress area and the logging benchmark boundary to identify the logging operation area based on the supplementary images acquired in advance. The monitoring analysis queue includes several monitoring analysis units. Each monitoring analysis unit is used to determine the approach area based on the logging operation boundary and the logging benchmark boundary of the logging operation area in order to predict the tendency to cross the boundary.

[0008] Furthermore, the boundary repair module performs a closure state analysis based on the logging boundary information to determine whether there are any breakpoints or gaps, and generates complete line segments for the breakpoints and gaps to determine the logging baseline boundary.

[0009] Furthermore, the boundary repair module is also used to filter and delete invalid holes in the logging baseline boundary.

[0010] Furthermore, the sandbox queue creation module divides the logging area into several logging analysis areas and several logging baseline boundary analysis segments based on the number of trees in the logging area, in order to determine the image analysis queue and the monitoring analysis queue.

[0011] Furthermore, the image analysis unit distinguishes between bare areas that have been logged and unlogged areas based on the spectral characteristics of the acquired field image data, in order to determine the logging progress area.

[0012] Furthermore, the image analysis unit determines the operation progress rate based on the current image acquisition cycle and the logging progress area corresponding to the previous image acquisition cycle, and determines that the logging operation is in an advanced state based on the judgment result that the operation progress rate is greater than the preset rate value.

[0013] Furthermore, in response to the logging operation being in an advanced state, the image analysis unit calculates the minimum spatial distance between the outer boundary of the logging progress area and the logging reference boundary.

[0014] Furthermore, the image analysis unit determines the supplementary images acquired in advance based on the estimated approach time to identify the logging operation area, wherein the estimated approach time is calculated based on the minimum spatial distance and the operation advancement rate.

[0015] Furthermore, the monitoring and analysis unit determines the tendency to cross the boundary based on the changing trend of the approximation area reduction.

[0016] Furthermore, the sandbox queue creation module updates the monitoring and analysis queue based on the boundary crossing tendency and the remaining time required for the logging operation boundary to reach the logging benchmark boundary, so as to continuously conduct real-time approach monitoring of the logging operation boundary.

[0017] Compared with existing technologies, the beneficial effects of this invention are that it achieves data isolation and distributed queue-driven operation for multiple logging projects through a boundary repair module and a sandbox queue creation module. It automatically repairs boundary topology defects, interprets logging progress based on images, proactively identifies boundary crossing risks, and dynamically allocates monitoring resources. This solves the problems of data crosstalk, boundary misjudgment, and resource misallocation in concurrent monitoring of multiple projects, thereby improving the automation level and resource utilization efficiency of forestry logging supervision.

[0018] Furthermore, this invention detects whether there are any breaks or gaps in the logging boundary through closed-state analysis, and generates complete line segments for the locations of the breaks and gaps. This achieves automatic repair of the connection state between the beginning and end of the logging boundary, avoiding abnormalities in the correlation analysis between the logging progress and the boundary due to the disconnection between the beginning and end of the boundary. This ensures the integrity and reliability of the logging benchmark boundary and provides an accurate basis for subsequent spatial calculations.

[0019] Furthermore, this invention automatically filters and deletes tiny closed holes inside the logging baseline boundary to prevent broken holes from interfering with subsequent area calculations and boundary crossing analysis. At the same time, it persistently stores the repaired boundary results to ensure that the calculation basis for subsequent image interpretation, spatial comparison, and approximate area measurement remains reliable and consistent, thereby improving the usability of boundary data.

[0020] Furthermore, this invention divides the logging analysis area and the logging baseline boundary analysis segment according to the number of trees in the logging area, and allocates image analysis units and monitoring analysis units accordingly. This realizes the construction of a distributed queue that matches the monitoring tasks with the distribution of trees, improves the utilization efficiency of monitoring resources, and provides an isolated and orderly task scheduling basis for concurrent monitoring of multiple projects.

[0021] Furthermore, this invention automatically distinguishes between logged and unlogged areas by utilizing the spectral characteristics of on-site image data, thereby determining the logging progress area and realizing the automated interpretation of logging operation progress. This provides a stable and objective on-site time-series data source for subsequent boundary risk pre-analysis, improving the real-time performance and accuracy of monitoring.

[0022] Furthermore, this invention calculates the operation progress rate by using the current image acquisition cycle and the logging progress area of ​​the previous cycle, and compares it with a preset rate value to determine whether it is in an advanced state. This enables automatic identification of whether the logging operation progress speed exceeds the normal level, and can promptly detect the risk of overspeeding operations, providing triggering conditions for subsequent advanced analysis and monitoring of resource allocation.

[0023] Furthermore, this invention obtains the operation progress rate in the logging progress area compared to the previous image acquisition cycle, and determines that the logging operation is in an advanced state based on the judgment result that the operation progress rate is higher than the preset rate value. This realizes the automatic identification of whether the operation equipment's advancement speed exceeds the normal level. By analyzing the minimum spatial distance between the outer boundary of the logging progress area and the logging benchmark boundary, the monitoring attention is focused on active operation areas with high risk of boundary crossing, providing basic distance parameters for subsequent calculation of the expected approach time and assessment of monitoring accuracy.

[0024] Furthermore, this invention calculates the ratio of the minimum spatial distance to the operation advancement rate to obtain the expected approach time, and determines the need for image analysis unit enhancement based on the expected approach time. This enables proactive enhancement of image acquisition before the logging operation reaches the baseline boundary, matching the logging operation area with the actual operation advancement situation, and overcoming the problem of judgment lag caused by the mismatch between image data and operation risk rhythm under a fixed image acquisition cycle.

[0025] Furthermore, this invention dynamically quantifies and assesses the risk of boundary crossing by analyzing the changing trend of the reduction in the approximate area between the logging operation boundary and the logging benchmark boundary. This allows the tendency to cross boundaries to be judged based on the actual changing trend of the operation boundary's advancement, and enables dynamic quantification and assessment of the tendency to cross boundaries based on the actual changing trend of the operation boundary's advancement, providing a clear risk basis for monitoring resource adjustments.

[0026] Furthermore, this invention updates the monitoring and analysis unit by updating the boundary risk and the remaining time required for the logging operation boundary to reach the logging benchmark boundary, ensuring that boundary crossing events can be confirmed in a timely manner. At the same time, it achieves adaptive matching between monitoring resources and dynamically changing boundary crossing risks, overcoming the problem of resource waste and insufficient monitoring coexisting under fixed monitoring frequency. Attached Figure Description

[0027] Figure 1 This is a module connection diagram of the sandbox intelligent agent logging analysis system based on distributed queue driving according to an embodiment of the present invention; Figure 2 This is a logic diagram for determining the location of a break or gap in an embodiment of the present invention; Figure 3 A flowchart illustrating the workflow for determining the logging progress zone in an embodiment of the present invention; Figure 4 A flowchart illustrating the process of determining the logging operation area for an embodiment of the present invention. Detailed Implementation

[0028] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments. It should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0029] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0030] As an explanation, this embodiment is used to achieve isolated intelligent supervision in scenarios with multiple logging plots operating concurrently. It achieves data isolation for multiple projects through sandbox agents and distributed queues, automatically repairs defects in logging permit boundaries, periodically interprets logging operation progress based on image analysis units, proactively identifies risks of speeding operations, and dynamically updates the distributed queue in conjunction with boundary crossing tendencies. This solves the problems of data crosstalk, misjudgment caused by boundary defects, and mismatch between monitoring resources and risks when monitoring multiple logging projects in parallel, thereby improving the automation level and resource utilization efficiency of forest logging area supervision.

[0031] Please see Figure 1 The diagram shown is a module connection diagram of a sandbox intelligent agent logging analysis system based on a distributed queue-driven approach according to an embodiment of the present invention. The sandbox intelligent agent logging analysis system based on a distributed queue-driven approach according to an embodiment of the present invention includes: The boundary repair module obtains logging boundary information based on the logging area, performs boundary validity verification based on the logging boundary information, and determines the logging benchmark boundary. The sandbox queue creation module determines the image analysis queue and the monitoring analysis queue based on the logging baseline boundary and the number of trees in the corresponding logging area; and determines and updates the monitoring analysis queue based on the tendency to cross the boundary, so as to monitor the logging situation in real time. The image analysis queue includes several image analysis units. Each image analysis unit is used to acquire on-site image data of the corresponding logging analysis area within the current image acquisition cycle to determine the logging progress area; and to perform advanced analysis based on the logging progress area and the logging benchmark boundary to identify the logging operation area based on the supplementary images acquired in advance. The monitoring analysis queue includes several monitoring analysis units. Each monitoring analysis unit is used to determine the approach area based on the logging operation boundary and the logging benchmark boundary of the logging operation area in order to predict the tendency to cross the boundary.

[0032] As an explanation, the boundary repair module verifies the validity of the logging boundary information, completing gaps and eliminating invalid holes to obtain the logging baseline boundary. The sandbox queue creation module reads the logging baseline boundary and the number of trees in the logging area, divides the logging analysis area and the logging baseline boundary analysis segment, allocates the image analysis queue and the monitoring analysis queue, and enters cyclic monitoring after completing the construction of the sandbox agent. The image analysis unit of the image analysis queue processes the on-site images, obtains the logging progress area, and further identifies the logging operation area, providing it to the monitoring analysis queue; the monitoring analysis unit of the monitoring analysis queue calculates the approach area, determines the tendency to cross the boundary, and sends it back to the sandbox queue creation module. The sandbox queue creation module updates the monitoring analysis queue by combining the tendency to cross the boundary and the remaining time for the logging operation boundary to reach the logging baseline boundary, realizing real-time approach monitoring. The process executes cyclically, and monitoring terminates upon receiving a logging operation completion signal.

[0033] Specifically, the boundary repair module obtains logging boundary information based on the logging area, performs boundary validity verification based on the logging boundary information, and determines the logging benchmark boundary.

[0034] In this embodiment, the boundary repair module is used to obtain logging boundary information of the logging area. This logging boundary information includes the name of the logging area and its initial boundary. The initial boundary of the logging area is its outer contour line, used to describe the permitted logging area.

[0035] In this embodiment, the boundary repair module includes a spatial geometry calculation unit, which performs boundary validity verification based on the logging boundary information to determine whether the logging boundary information meets the validity requirements. The logging baseline boundary is a valid polygon obtained after breakpoint repair and invalid hole removal, used for subsequent approximation area calculation and boundary violation tendency calculation. The validity check includes breakpoint detection and invalid hole detection; if there are breakpoints or gaps in the logging boundary information, repair processing is performed; if there are invalid holes, screening and removal processing is performed.

[0036] Please see Figure 2 As shown, it is a logic determination diagram for determining whether there is a break or gap location in an embodiment of the present invention.

[0037] Specifically, the spatial geometry calculation unit performs a closure state analysis based on the logging boundary information to determine whether there are any breakpoints or gaps, and generates complete line segments for the breakpoints and gaps to determine the logging baseline boundary.

[0038] In this embodiment, if the initial boundary is closed, it is determined as the logging baseline boundary, and invalid hole detection is performed. If the initial boundary is not closed, the coordinates corresponding to the unclosed endpoints of the initial boundary are extracted, and the spatial distance between the two endpoints is calculated.

[0039] In this embodiment, the spatial distance is compared with a preset spatial distance. If the spatial distance between two points is less than or equal to the preset spatial distance, a break or gap is identified. If the spatial distance between two points is greater than the preset spatial distance, a boundary anomaly warning and a boundary verification warning are issued, and the conditions for boundary completion are not met to avoid deviation from the actual logging boundary.

[0040] In this embodiment, the preset spatial distance is the maximum distance allowed between two adjacent sampling points based on relevant surveying and mapping benchmarks during the process of determining geographical boundary data in forestry logging operations.

[0041] In this embodiment, the boundary repair module connects the two endpoints of the unclosed position of the initial boundary to generate a complete line segment for the breakpoint gap position. Based on the complete line segment and the initial boundary, the logging benchmark boundary is generated.

[0042] Specifically, the boundary repair module is also used to filter and delete invalid holes in the logging baseline boundary.

[0043] In this embodiment, invalid holes are screened at the logging baseline boundary. All closed loops within the logging baseline boundary are extracted, and the area of ​​each loop is calculated. Loops with an area smaller than the preset hole area are identified as invalid holes and deleted. The preset hole area is determined based on the average crown area of ​​trees in the forest area where the logging area is located.

[0044] It is understandable that invalid holes are tiny closed loops with an area smaller than the preset hole area inside the logging baseline boundary. They have no practical significance for subsequent approximation area calculations and will cause interference.

[0045] Specifically, the sandbox queue creation module determines the image analysis queue and the monitoring analysis queue based on the logging baseline boundary and the number of trees in the corresponding logging area; and determines and updates the monitoring analysis queue based on the tendency to cross the boundary, so as to monitor the logging situation in real time.

[0046] Specifically, the sandbox queue creation module divides several logging analysis areas and several logging baseline boundary analysis segments based on the number of trees in the logging area to determine the image analysis queue and the monitoring analysis queue.

[0047] In this embodiment, the sandbox queue creation module reads the repaired logging baseline boundary and simultaneously obtains the tree distribution points within the logging area. These tree distribution points are derived from forest resource data statistics and are used to represent the distribution of trees within the logging permit area.

[0048] In this embodiment, the logging area corresponding to the logging benchmark boundary is divided into several unit areas. The number of trees in each unit area is counted based on the tree distribution points. The division method of the unit areas can be determined according to the actual terrain, forest stand distribution or accessibility of the logging area to ensure the convenience and accuracy of subsequent statistics and grouping.

[0049] In this embodiment, the trees in a unit area are grouped according to the number of trees in that unit area. Several adjacent unit areas are grouped together and accumulated. When the accumulated number of trees in the group reaches or is very close to the preset tree number threshold, the area corresponding to the group is divided into a logging analysis area. The preset tree number threshold can be set according to logging operation capacity, regulatory resource allocation or transportation scheduling capacity to balance the workload of each logging analysis area and avoid too many or too few trees in a single area.

[0050] Understandably, when it is impossible to ensure that the number of trees in each logging analysis area is exactly the same, the allowable deviation is determined based on the similarity standard and the preset tree number threshold. The deviation is determined according to the fluctuation range that the same batch of workers, logging equipment, transportation conditions and supervision forces can bear in a balanced manner, so that the logging analysis volume corresponding to logging analysis areas with different areas is comparable. When dividing, the cumulative number of trees in adjacent unit areas is compared with the threshold. If it falls within the allowable deviation range, it is considered to be extremely similar. If it exceeds the range, it is corrected by swapping adjacent unit areas, fine-tuning their corresponding logging benchmark boundary segments, or merging the excess part into an adjacent logging analysis area that still has the capacity to bear it, so as to ensure that the supervision boundary is clear and the operation intensity is similar.

[0051] In this embodiment, since each logging analysis area must correspond to a logging benchmark boundary during the division process, it is also necessary to determine whether the boundary corresponding to the current group is a continuous part of the logging benchmark boundary when grouping. If it is not, the combination method or boundary correspondence of adjacent unit areas is adjusted until each logging analysis area corresponds to a logging benchmark boundary.

[0052] In this embodiment, the impact analysis queue includes several image analysis units, and the monitoring analysis queue includes several monitoring analysis units. The number of image analysis units and monitoring analysis units is determined based on the number of logging analysis areas that have been divided. Each logging analysis area corresponds to one impact analysis unit and one monitoring analysis unit to execute the corresponding image analysis task and monitoring analysis task. Once all units in the image analysis queue and monitoring analysis queue have been allocated, the image analysis queue and monitoring analysis queue are constructed, the sandbox agent is created, and the subsequent cyclic monitoring process can be started.

[0053] Understandably, the logging analysis area is a closed vector region, used as the spatial processing range for image analysis units and regulatory analysis units to perform image acquisition, spectral interpretation, identification of logging progress areas, and analysis of boundary crossing tendencies. The logging baseline boundary analysis segment is the outline segment of the logging baseline boundary corresponding to the logging analysis area, used for subsequent approximation area calculation and boundary crossing tendency assessment, realizing segmented risk monitoring of the boundary.

[0054] Please see Figure 3 As shown, it is a flowchart illustrating the workflow for determining the logging progress area in an embodiment of the present invention.

[0055] Specifically, each image analysis unit is used to acquire on-site image data of the corresponding logging analysis area within the current image acquisition period to determine the logging progress area; and to perform advanced analysis based on the logging progress area and the logging benchmark boundary to identify the logging operation area based on supplementary images acquired in advance.

[0056] In this embodiment, the image analysis unit is used to acquire on-site image data of its corresponding logging analysis area. The on-site image data is remote sensing imagery that has undergone georegistration processing, establishing a coordinate mapping relationship between the imagery and planar geographic locations. The on-site image data requires radiometric correction to eliminate acquisition distortion and bias caused during the shooting process, ensuring that each logging analysis area on the imagery can be accurately matched to its actual geographical location.

[0057] In this embodiment, the image acquisition cycle is a pre-set fixed time interval used to control the interval between two adjacent image acquisitions. The image acquisition cycle is set based on the image data interpretation time and the conventional logging rate, and is used to periodically track and record the progress of logging operations.

[0058] Specifically, the image analysis unit distinguishes between bare areas that have been logged and unlogged areas based on the spectral characteristics of the acquired field image data, in order to determine the logging progress area.

[0059] In this embodiment, the image analysis unit distinguishes between bare areas that have been logged and unlogged areas based on the differences in image grayscale values ​​of the logging analysis area. Bare areas are areas where the surface soil is directly exposed after logging, while unlogged areas are areas that still retain their original forest cover. There are significant differences in image grayscale values ​​between the two types of areas.

[0060] In this embodiment, the image analysis unit extracts the grayscale value sets of the confirmed exposed areas and the grayscale value sets of the unharvested areas from the historical image data of the logging area. It then calculates the distribution range of each set of grayscale values, using the midpoint between the lower limit of the grayscale value distribution range for exposed areas and the upper limit for the grayscale value distribution range for unharvested areas as the grayscale value classification boundary. When the grayscale value of a region in the on-site image data is greater than the grayscale value classification boundary, it is identified as an exposed area; when the grayscale value of the target area is less than or equal to the grayscale value classification boundary, it is identified as an unharvested area. The exposed areas are then identified as logging progress areas.

[0061] Specifically, the image analysis unit determines the operation progress rate based on the current image acquisition cycle and the logging progress area corresponding to the previous image acquisition cycle, and determines that the logging operation is in an advanced state based on the judgment result that the operation progress rate is greater than the preset rate value.

[0062] In this embodiment, the image analysis unit extracts the logging progress area from the current image acquisition cycle and the logging progress area from the previous image acquisition cycle. The logging progress areas from the two cycles are spatially superimposed, and the ratio of the outward expansion distance of the outer boundary of the logging progress area between the two cycles to the duration of the image acquisition cycle is calculated to determine the operation progress rate. This rate quantifies the average advancement speed of the logging operation boundary within the current image acquisition cycle, providing a rate parameter for subsequent determination of whether the operation is ahead of schedule and for estimating the expected approach time. The outer boundary of the logging progress area is the contour segment of the logging progress area that is closest to the logging baseline boundary analysis segment.

[0063] In this embodiment, the image analysis unit compares the operation progress rate with a preset rate value. If the operation progress rate is greater than the preset rate value, it determines that the current logging operation is in an advanced state.

[0064] If the operation progress rate is less than or equal to the preset rate value, it is determined that the current logging operation is not in an advanced state, and subsequent analysis will not be performed, but continuous monitoring must be maintained.

[0065] Understandably, the advanced status indicates that the current logging operation is proceeding at a rate that exceeds the normal level. At the current rate of progress, there is a possibility that the operation may reach or exceed the logging baseline boundary before the next image acquisition cycle arrives. Therefore, an advanced analysis process needs to be initiated to assess whether there is an enhanced need for on-site image analysis.

[0066] In this embodiment, the preset rate value is determined based on the operational progress rate of several image acquisition cycles completed since the start of the current logging project. Specifically, the operational progress rate of all historical image acquisition cycles since the project started is acquired, and the preset rate value is determined based on the average operational progress rate of several consecutive image acquisition cycles prior to the current image acquisition cycle, in order to quantify the normal progress level.

[0067] Specifically, in response to the logging operation being in an advanced state, the image analysis unit calculates the minimum spatial distance between the outer boundary of the logging progress area and the logging reference boundary.

[0068] In this embodiment, in response to the current logging operation being in an advanced state, the outer boundary of the logging progress area is extracted and the distance between it and the logging baseline boundary analysis segment is calculated. The minimum value among all distances is taken as the minimum spatial distance, which is used to quantify the shortest spatial interval between the outer edge of the current logging operation and the permitted boundary.

[0069] Please see Figure 4 As shown, it is a flowchart illustrating the process of determining the logging operation area in an embodiment of the present invention.

[0070] Specifically, the image analysis unit determines the supplementary images acquired in advance based on the estimated approach time to identify the logging operation area, wherein the estimated approach time is calculated based on the minimum spatial distance and the operation advancement rate.

[0071] In this embodiment, the image analysis unit determines the estimated approach time by the ratio of the minimum spatial distance to the operation advancement rate. This time is used to represent the remaining time required for the outer boundary of the logging progress area to advance from the current position to the logging benchmark boundary analysis section, provided that the current advancement rate remains unchanged. This is used to assess whether there is a risk of boundary crossing within the current image acquisition cycle.

[0072] In this embodiment, the expected approach time is compared with the current image acquisition cycle time. If the expected approach time is less than the current image acquisition cycle time, it means that the logging operation will reach the logging baseline boundary analysis segment before the next image acquisition cycle arrives, and it is determined that there is an enhanced demand for on-site image analysis.

[0073] In this embodiment, if the expected approach time is greater than or equal to the current image acquisition cycle time, it means that the outer edge of the logging operation will not touch the logging baseline boundary analysis segment before the next image acquisition cycle arrives. At this time, it is determined that there is no need for enhancement of on-site image analysis, and the normal image analysis is maintained.

[0074] In this embodiment, in response to the determination of the need for enhancement of on-site image analysis, the image analysis unit is used to determine the image analysis enhancement parameters for subsequent image acquisition, so as to determine the supplementary images acquired in advance, and then identify the logging operation area.

[0075] Among them, the image analysis enhancement parameters acquired in subsequent imagery include the geographic contour line of the enhanced target area and the resolution configuration of the supplementary imagery acquisition.

[0076] Specifically, the image analysis unit delineates the target enhanced geographic range based on the spatial relationship between the outer boundary of the logging progress area and the logging benchmark boundary.

[0077] In this embodiment, the boundary segment of the outer boundary of the logging progress area is extracted, and the distance from the logging benchmark boundary analysis segment does not exceed the warning distance determined by the remaining duration of the current image acquisition cycle and the operation progress rate. The area covered by the above boundary segment and buffered outward by a preset buffer width is used as the target enhanced geographic range.

[0078] In this embodiment, the preset buffer width is determined based on the product of the operation progress rate within the current image acquisition cycle and the remaining duration of the current image acquisition cycle, so that the buffer width can cover the spatial range of the logging operation that can be pushed outward to the maximum extent possible during the remaining period of the current image acquisition cycle, thus avoiding omissions in the geographical range of the enhanced acquisition.

[0079] In this embodiment, the resolution accuracy is configured to reduce the ground sampling distance of the supplementary image within the target augmentation geographic range to a level sufficient to distinguish subtle relative displacements between the logging edge and the logging baseline boundary, and this level is determined based on the ground sampling distance of the conventional image, the capabilities of the supplementary image platform, and the requirements for boundary disturbance identification.

[0080] Specifically, the image acquisition cycle is updated based on the expected approach time to shorten the image acquisition interval in high-risk areas and adapt to the risk scenarios where logging operations rapidly approach the baseline boundary.

[0081] In this embodiment, the updated image acquisition cycle is determined based on the expected approach time and the preset number of monitoring times. As the logging operation continuously approaches the logging baseline boundary, the expected approach time continues to shorten, and the updated image acquisition cycle duration also continues to decrease. This shortens the image acquisition interval, increases the frequency of image acquisition for the target augmented geographical area, and improves the monitoring accuracy of the correlation between the logging operation and the baseline boundary.

[0082] The preset number of monitoring sessions is a fixed number determined by the image acquisition platform before the system is activated, based on the platform's ability to revisit the enhanced geographic range of the target, the latency of a single image processing session, and the cost of boundary correction. During a single update, the system will not recalculate based on the expected approach time. The updated image acquisition cycle will be obtained by dividing the expected approach time by the fixed number of sessions. This ensures that the updated image acquisition cycle is shortened only as the expected approach time decreases, allowing sufficient time for observation and processing before the outer edge of the logging reaches the logging baseline boundary. At the same time, it avoids resource waste caused by excessively dense data collection or oversights in boundary judgment.

[0083] The estimated approach time is the approach time predicted based on the previous available logging operation boundary; the remaining time used for subsequent monitoring and analysis queue updates is the current approach time recalculated based on the latest confirmed logging operation boundary. The two correspond to the prediction stage and the confirmation stage, respectively, and are connected sequentially in the same cycle.

[0084] In this embodiment, the image analysis unit acquires supplementary images according to the image analysis enhancement parameters and the updated image acquisition cycle.

[0085] In this embodiment, after acquiring the supplementary image, the image analysis unit performs the same interpretation processing on the supplementary image as on the aforementioned image, including radiometric correction and georegistration, and then uses the corresponding logging progress area in the supplementary image as the logging operation area, and the boundary outline of the logging operation area as the logging operation boundary.

[0086] Understandably, the logging operation area refers to the logged area identified from supplementary imagery within the target augmented geographic range, representing a newly extended portion of the previously confirmed logging progress area. When the newly extended area is spatially contiguous with the previously confirmed logging progress area, it is merged into the current cycle's logging progress area. When they are not contiguous, they are marked as newly logged patches, retaining their adjacency with the previous cycle's area. The logging operation area is used by the monitoring and analysis unit to calculate the approximation area in subsequent cycles, thereby obtaining boundary advancement data in a timely manner before the logging operation reaches the logging baseline boundary.

[0087] Specifically, the monitoring and analysis unit acquires the logging operation area and its boundary from the image analysis unit, and determines the approximation area in conjunction with the logging baseline boundary to predict boundary crossing tendencies. Specifically, when the logging operation boundary is entirely inside the logging baseline boundary and the two do not intersect, the area of ​​the region enclosed between the logging operation boundary and the logging baseline boundary is calculated as the approximation area. When there are intersections, partial boundary crossings, or multiple disconnected enclosed areas, the areas of each inward enclosed area are combined into the approximation area, and the crossed segments are marked separately and treated as boundary crossings, no longer included in the forward approximation area. The smaller the approximation area, the closer the outer edge of the logging operation is to the logging baseline boundary; the larger the approximation area, the farther the outer edge of the logging operation is from the logging baseline boundary.

[0088] Specifically, the monitoring and analysis unit determines the tendency to cross the boundary based on the changing trend of the shrinkage of the approximation area.

[0089] In this embodiment, the monitoring and analysis unit calculates several approximate areas based on the real-time updated logging operation area, arranges them in the order of collection, and forms an approximate area time series.

[0090] In this embodiment, the monitoring and analysis unit calculates the trend of reduction based on the approximation area time series of no less than three periods, and judges the trend by the difference in reduction between adjacent periods or the fitted slope. When the reduction shows a continuous increasing trend, it is determined that the outer edge of the logging operation is accelerating towards the logging benchmark boundary, and the tendency to cross the boundary is identified as a high risk of accelerated approach, i.e., the first risk of crossing the boundary. When the reduction shows a continuous decreasing trend, stable fluctuation, or the absolute value of the slope does not exceed the stable bandwidth jointly determined by the stability of the operation and the image registration error, the tendency to cross the boundary is determined as a low risk, i.e., the second risk of crossing the boundary.

[0091] Specifically, the sandbox queue creation module updates the monitoring and analysis queue based on the boundary crossing tendency and the remaining time required for the logging operation boundary to reach the logging benchmark boundary, in order to continuously conduct real-time approach monitoring of the logging operation boundary. The remaining time is recalculated based on the latest confirmed logging operation boundary.

[0092] In this embodiment, the sandbox queue creation module updates the monitoring and analysis queue based on the remaining time required for the logging operation boundary to reach the logging baseline boundary in response to the first boundary crossing risk. When the second boundary crossing risk is detected, if the expected approach time is still less than the current image acquisition cycle, the target augmentation geographic range and supplementary image acquisition are retained, but the number of monitoring and analysis units is not increased.

[0093] In this embodiment, the monitoring interval of the regulatory analysis task is synchronously checked with the updated image acquisition cycle based on the remaining time, so that the remaining time contains at least two complete image acquisition cycles; one of which is used to acquire and interpret supplementary images, and the other is used to check the boundary position and trigger an alarm before the logging operation touches the logging benchmark boundary.

[0094] In this embodiment, when there are fewer than two complete image acquisition cycles available in the remaining time, the task concurrency required to complete two consecutive boundary verifications is calculated. Based on the number of boundary monitoring segments that can be divided within the same logging analysis area, the principle that segments handled by each monitoring analysis unit do not overlap, and supplementary image sharing permissions, the number of monitoring analysis units is increased or the current monitoring analysis task is split. The newly added monitoring analysis unit is only responsible for the assigned boundary segments and does not repeatedly calculate the approximation area that has been confirmed by other monitoring analysis units. After the alarm is triggered, the queue is restored to the state before the risk is resolved.

[0095] In this embodiment, real-time approach monitoring is continuously carried out on the logging operation boundary. As the logging operation continues to advance, the monitoring and analysis unit calculates the approach area by obtaining the logging benchmark boundary analysis segment corresponding to the logging analysis area, and determines the tendency to cross the boundary based on the changing trend of the reduction of the approach area, so as to realize the continuous assessment of the risk of crossing the boundary.

[0096] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A sandbox-based intelligent agent logging analysis system driven by a distributed queue, characterized in that, include: The boundary repair module obtains logging boundary information based on the logging area, performs boundary validity verification based on the logging boundary information, and determines the logging benchmark boundary. The sandbox queue creation module determines the image analysis queue and the monitoring analysis queue based on the logging baseline boundary and the number of trees in the corresponding logging area; and determines and updates the monitoring analysis queue based on the tendency to cross the boundary, so as to monitor the logging situation in real time. The image analysis queue includes several image analysis units. Each image analysis unit is used to acquire on-site image data of the corresponding logging analysis area within the current image acquisition cycle to determine the logging progress area; and to perform advanced analysis based on the logging progress area and the logging benchmark boundary to identify the logging operation area based on the supplementary images acquired in advance. The monitoring analysis queue includes several monitoring analysis units. Each monitoring analysis unit is used to determine the approach area based on the logging operation boundary and the logging benchmark boundary of the logging operation area in order to predict the tendency to cross the boundary.

2. The sandbox intelligent agent logging analysis system based on distributed queue driving according to claim 1, characterized in that, The boundary repair module performs a closure state analysis based on the logging boundary information to determine whether there are any breakpoints or gaps, and generates complete line segments for the breakpoints and gaps to determine the logging baseline boundary.

3. The sandbox intelligent agent logging analysis system based on distributed queue driving according to claim 2, characterized in that, The boundary repair module is also used to filter and delete invalid holes in the logging baseline boundary.

4. The sandbox intelligent agent logging analysis system based on distributed queue driving according to claim 3, characterized in that, The sandbox queue creation module divides the logging area into several logging analysis areas and several logging baseline boundary analysis segments based on the number of trees in the logging area, in order to determine the image analysis queue and the monitoring analysis queue.

5. The sandbox intelligent agent logging analysis system based on distributed queue driving according to claim 4, characterized in that, The image analysis unit distinguishes between bare areas that have been logged and unlogged areas based on the spectral characteristics of the acquired field image data, in order to determine the logging progress area.

6. The sandbox intelligent agent logging analysis system based on distributed queue driving according to claim 5, characterized in that, The image analysis unit determines the operation progress rate based on the current image acquisition cycle and the logging progress area corresponding to the previous image acquisition cycle, and determines that the logging operation is in an advanced state based on the judgment result that the operation progress rate is greater than the preset rate value.

7. The sandbox intelligent agent logging analysis system based on distributed queue driving according to claim 6, characterized in that, The image analysis unit, in response to the logging operation being in an advanced state, calculates the minimum spatial distance between the outer boundary of the logging progress area and the logging reference boundary.

8. The sandbox intelligent agent logging analysis system based on distributed queue driving according to claim 7, characterized in that, The image analysis unit determines the supplementary images acquired in advance based on the estimated approach time to identify the logging operation area, wherein the estimated approach time is calculated based on the minimum spatial distance and the operation advancement rate.

9. The sandbox intelligent agent logging analysis system based on distributed queue driving according to claim 8, characterized in that, The monitoring and analysis unit determines the tendency to cross the boundary based on the changing trend of the reduction in the approximate area.

10. The sandbox intelligent agent logging analysis system based on distributed queue driving according to claim 9, characterized in that, The sandbox queue creation module updates the monitoring and analysis queue based on the boundary crossing tendency and the remaining time required for the logging operation boundary to reach the logging benchmark boundary, so as to continuously conduct real-time approach monitoring of the logging operation boundary.

Images