A machine vision-based real-time monitoring method and system for ship anchoring

By using machine vision technology to dynamically partition and assess the real-time status of anchorages, the problem of data distortion in traditional monitoring methods is solved, enabling high-precision monitoring of ship anchoring safety and reducing the risk of safety accidents.

CN123118609APending Publication Date: 2026-11-17ZHEJIANG INT MARITIME COLLEGE
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611518215.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-29
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Traditional ship anchoring monitoring relies on sensors built into the anchor winch equipment, which lacks real-time visual data support from the sea surface, resulting in data distortion and an inability to accurately monitor the spatial movement of the anchor, which can easily lead to safety accidents.

Method used

By employing a machine vision-based approach, dynamic zoning of anchorages is achieved through the collection of anchor winch data and visual time-series data of anchoring on the sea surface. Combined with motion time-series characteristics and safety constraint data, hierarchical monitoring constraints are generated to realize real-time status judgment and standardized monitoring.

Benefits of technology

It improves the accuracy and efficiency of anchor movement monitoring, can identify local deviation anomalies, reduces the probability of safety accidents, and is suitable for routine anchoring monitoring of near-shore and ocean-going vessels.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123118609A_ABST
    Figure CN123118609A_ABST
Patent Text Reader

Abstract

This invention discloses a machine vision-based real-time monitoring method and system for ship anchoring, relating to the field of ship anchoring technology. The key technical solutions include the following steps: collecting anchor winch data and sea surface anchoring visual time-series data of the ship to be monitored; the anchor winch data includes anchor winch geometric parameters and anchoring safety constraint data; spatially clustering the anchor winch geometric parameters to obtain dynamic anchorage partitions corresponding to ship anchoring operations; obtaining the anchorage position offset amplitude and motion change trend based on the dynamic anchorage partitions, and obtaining motion time-series characteristics based on the position offset amplitude and motion change trend; extracting standardized monitoring constraint parameters based on hierarchical monitoring constraints; obtaining anchoring state adaptation loss data based on the spatial motion layout parameters of the real-time anchoring state judgment results and the standardized monitoring constraint parameters; and generating a real-time monitoring result set for ship anchoring based on the adaptation loss data. The effect is to reduce the probability of safety accidents such as anchor dragging and anchorage collisions.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of ship anchoring technology, and more specifically, to a machine vision-based method and system for real-time monitoring of ship anchoring. Background Technology

[0002] Traditional monitoring relies on sensors built into the anchor winch to collect basic operational data, lacking the data support of real-time visual observation of the sea surface. Environmental interference such as wind, waves, ship rolling, and obstruction by surrounding floating objects can distort the sensor data. Crew members cannot fully grasp the actual spatial movement of the anchor by relying on data from a single device, easily leading to missed detections of potential hazards such as anchor dragging and excessive anchor chain tension. Furthermore, current technology does not implement zoned management of the anchor operating space, failing to differentiate the varying risks of anchor deviation in near-ship and mid-to-far-ship areas. It also cannot integrate anchor winch geometric hardware parameters and visual temporal dynamic information for calculation, and cannot quantify the deviation between the actual anchor movement and safety constraints, potentially causing maritime safety accidents such as anchor damage, ship collisions, and vessel drift. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a method and system for real-time monitoring of ship anchoring based on machine vision.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A machine vision-based real-time monitoring method for ship anchoring includes the following steps:

[0006] Collect anchor winch data and visual time-series data of the ship to be monitored; the anchor winch data includes anchor winch geometric parameters and anchoring safety constraint data;

[0007] Spatial clustering of anchor winch geometric parameters yields dynamic anchorage partitions corresponding to ship anchoring operations;

[0008] The position offset amplitude and motion change trend of the anchor are obtained based on the dynamic zoning of the anchor, and the motion time sequence characteristics are obtained based on the position offset amplitude and motion change trend.

[0009] Based on motion time-series characteristics and anchoring safety constraint data, the hierarchical monitoring constraints of the ship's anchoring process are obtained; based on the hierarchical monitoring constraints and the visual time-series data of anchoring on the sea surface, the real-time status judgment result of the ship's anchoring is obtained.

[0010] Standardized monitoring constraint parameters are extracted based on hierarchical monitoring constraints; and the adaptation loss data of the anchoring state is obtained based on the spatial motion layout parameters and standardized monitoring constraint parameters according to the real-time anchoring state judgment results.

[0011] A real-time monitoring result set of the ship's anchoring is generated based on the adapted loss data.

[0012] Preferably, spatial clustering of the anchor winch geometry parameters is used to obtain dynamic anchorage partitions corresponding to ship anchoring operations, specifically including the following steps:

[0013] Based on the spatial motion attributes and operational range of ship anchoring operations, the geometric parameters of anchor winches are hierarchically classified and integrated to obtain classified parameter data.

[0014] The basic operating space of the anchorage is defined based on the classification parameter data. The basic operating space is then divided into zones based on the basic operating space and the dynamic motion characteristics of the ship's anchoring operation, resulting in the dynamic zone of the anchorage corresponding to the ship's anchoring operation.

[0015] Preferably, the positional offset amplitude and motion change trend of the anchor are obtained based on the dynamic zoning of the anchor, specifically including the following steps:

[0016] Lock the anchorage in the actual working space of the anchorage dynamic zone;

[0017] The positional offset of the anchor is obtained by comparing the actual working space position with the reference space range of the anchor dynamic zoning.

[0018] By tracking the spatial switching trajectory and displacement characteristics of anchors between different anchor dynamic zones, the trend of motion change can be obtained.

[0019] Preferably, the motion time sequence characteristics are obtained based on the position offset amplitude and motion change trend, specifically as follows:

[0020] By temporally correlating and fusing the positional shift magnitude at different times with the corresponding motion change trend over the time period, motion temporal characteristics are obtained.

[0021] Preferably, the hierarchical monitoring constraints for the ship's anchoring process are obtained based on motion timing characteristics and anchoring safety constraint data, specifically including the following steps:

[0022] Determine the basic operational constraint benchmarks for anchoring based on the dynamic zoning of anchorages;

[0023] The revised hierarchical benchmark is obtained by revising the coverage of the basic operation constraint benchmark based on the motion time sequence characteristics;

[0024] Based on the anchoring safety constraint data, the safety specifications of the modified hierarchical benchmark are defined to obtain the hierarchical monitoring constraints for the ship anchoring process.

[0025] Preferably, the real-time status judgment result of the ship's anchoring is obtained based on the layered monitoring constraints and the visual time-series data of anchoring on the sea surface, specifically including the following steps:

[0026] Based on the safety control standards corresponding to the hierarchical monitoring constraints, verify the actual operational status of the dynamic zoning of anchorages in the visual time series data of sea surface anchorage.

[0027] The comparison results are obtained by comparing the actual movement performance of the anchor with the matching difference of the constraint level under the actual operation state.

[0028] Based on the comparison results, normal operation and abnormal deviation conditions are distinguished to obtain real-time status judgment results.

[0029] Preferably, the standardized monitoring constraint parameters are extracted based on the hierarchical monitoring constraint conditions, specifically including the following steps:

[0030] The anchor position offset amplitude and motion change trend are matched with the control standards of each level of hierarchical monitoring constraints to obtain the matching results;

[0031] Based on the matching results, select safety constraints that fit the actual working conditions of anchoring.

[0032] Standardized monitoring constraint parameters adapted for anchorage detection are obtained based on the safety constraints.

[0033] Preferably, the spatial motion layout parameters of the anchoring real-time status judgment result are combined with the standardized monitoring constraint parameters to obtain the anchoring status adaptation loss data, specifically including the following steps:

[0034] The spatial motion layout parameters of the real-time anchoring status judgment results are compared with the standardized monitoring constraint parameters to obtain deviation data.

[0035] By integrating the deviation data, we obtain the adaptation loss data for the anchorage state.

[0036] Preferably, generating a real-time monitoring result set of the ship's anchoring based on the adaptation loss data specifically includes the following steps:

[0037] The degree of deviation of the ship's real-time anchoring status is determined based on the adaptation loss data;

[0038] Based on the dynamic change trend and timing characteristics of the anchorage dynamic zoning, the anchoring conditions with different degrees of deviation are classified and merged to obtain merged condition information.

[0039] A real-time monitoring result set, including anchorage safety level, operating condition status, and offset risk, is generated based on the merged operating condition information.

[0040] A machine vision-based real-time monitoring system for ship anchoring includes:

[0041] Data Acquisition Module: Acquires anchor winch data and visual time-series data of the vessel to be monitored; the anchor winch data includes anchor winch geometric parameters and anchoring safety constraint data;

[0042] Partitioning module: Spatial clustering of anchor winch geometric parameters is used to obtain dynamic anchorage partitions corresponding to ship anchoring operations;

[0043] The first processing module obtains the position offset amplitude and motion change trend of the anchorage based on the dynamic zoning of the anchorage, and obtains the motion time sequence characteristics based on the position offset amplitude and motion change trend;

[0044] The second processing module: Based on motion time series characteristics and anchoring safety constraint data, it obtains the layered monitoring constraints of the ship's anchoring process; based on the layered monitoring constraints and the visual time series data of anchoring on the sea surface, it obtains the real-time status judgment result of the ship's anchoring.

[0045] Extraction module: Extracts standardized monitoring constraint parameters based on hierarchical monitoring constraints; obtains the adaptation loss data of the anchoring state based on the spatial motion layout parameters of the real-time anchoring state judgment results and the standardized monitoring constraint parameters;

[0046] Generation module: Generates a set of real-time monitoring results of ship anchoring based on the adapted loss data.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] This invention uses spatial clustering based on anchor winch geometric parameters to divide anchor dynamic zones, enabling separate analysis of anchor movement states in different areas near, mid-, and far from ships. It captures subtle anomalies such as small local anchor displacements and generates hierarchical monitoring constraints by combining motion temporal characteristics with anchoring safety constraints. This yields real-time anchoring status judgment results and standardized monitoring constraint parameters. Hierarchical constraints achieve differentiated control benchmarks for different operating conditions. A standardized real-time monitoring result set is generated based on adapted loss data, facilitating crew identification of anchor displacement risks and early intervention, reducing the probability of safety accidents such as anchor dragging and anchor collisions. This application balances monitoring accuracy and operational efficiency, adapting to routine anchoring monitoring scenarios for various near-shore and ocean-going vessels. Attached Figure Description

[0049] Figure 1 This is a flowchart illustrating a machine vision-based real-time monitoring method for ship anchoring, as provided in an embodiment of the present invention.

[0050] Figure 2 This is a schematic diagram of a machine vision-based real-time monitoring system for ship anchoring, provided in an embodiment of the present invention. Detailed Implementation

[0051] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0052] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0053] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0054] Reference Figures 1-2 As shown.

[0055] The embodiments further illustrate the machine vision-based real-time monitoring method and system for ship anchoring proposed in this invention.

[0056] A machine vision-based real-time monitoring method for ship anchoring includes the following steps:

[0057] Collect anchor winch data and visual time-series data of anchoring on the sea surface of the vessel to be monitored; the anchor winch data includes anchor winch geometric parameters and anchoring safety constraint data;

[0058] Spatial clustering of anchor winch geometric parameters yields dynamic anchorage partitions corresponding to ship anchoring operations;

[0059] The position offset amplitude and motion change trend of the anchor are obtained based on the dynamic zoning of the anchor, and the motion time sequence characteristics are obtained based on the position offset amplitude and motion change trend.

[0060] Based on motion time-series characteristics and anchoring safety constraint data, the hierarchical monitoring constraints of the ship's anchoring process are obtained; based on the hierarchical monitoring constraints and the visual time-series data of anchoring on the sea surface, the real-time status judgment result of the ship's anchoring is obtained.

[0061] Standardized monitoring constraint parameters are extracted based on hierarchical monitoring constraints; and the adaptation loss data of the anchoring state is obtained based on the spatial motion layout parameters and standardized monitoring constraint parameters according to the real-time anchoring state judgment results.

[0062] A real-time monitoring result set of the ship's anchoring is generated based on the adapted loss data.

[0063] Spatial clustering of anchor winch geometry parameters yields dynamic anchorage partitions corresponding to ship anchoring operations, specifically including the following steps:

[0064] Based on the spatial motion attributes and operational range of ship anchoring operations, the geometric parameters of anchor winches are hierarchically classified and integrated to obtain classified parameter data.

[0065] The basic operating space of the anchorage is defined based on the classification parameter data. The basic operating space is then divided into zones based on the basic operating space and the dynamic motion characteristics of the ship's anchoring operation, resulting in the dynamic zone of the anchorage corresponding to the ship's anchoring operation.

[0066] The spatial motion attributes of ship anchoring operations include the longitudinal range of motion corresponding to the length of the anchor chain extension / retraction, the lateral sway amplitude corresponding to the rotation angle of the anchor winch drum, and the planar displacement caused by the underwater drag after the anchor claws enter the water. The ship's anchoring operation range includes the vertical operating height corresponding to the ship's rated draft, the planar operating radius corresponding to the ship's designed anchoring radius, and the boundary dimensions corresponding to the port-defined anchoring area. After extracting all anchor winch geometric parameters, the staff categorized and integrated them according to hierarchical logic. The first level of categorization dimension is the parameters related to spatial motion attributes, and the second level of categorization dimension is the parameters related to the operation range. The two levels of parameter data are then summarized and merged to form the categorized parameter data. For example, the geometric parameters of an anchor winch for a 5,000-ton bulk carrier include a drum diameter of 1.2 meters, a single anchor chain length of 27.5 meters, a maximum total length of anchor chain extended and retracted of 330 meters, a ship's draft of 8 meters, a designed anchoring radius of 120 meters, and a maximum drum rotation angle of ±35 degrees. First, the drum diameter, single anchor chain length, maximum total length of anchor chain extended and retracted, and drum rotation angle are classified into the spatial motion attribute level. Then, the ship's draft and designed anchoring radius are classified into the operating range level. After integrating the two sets of parameters, the complete classification parameter data of the ship is obtained.

[0067] The basic working space of the anchorage is defined based on the classification parameters. This basic working space is a three-dimensional space, with the three dimensions being vertical height, radial distance, and sway angle. Vertical height = draft + maximum anchor chain length × anchor chain vertical drop coefficient, where the anchor chain vertical drop coefficient ranges from 0 to 1, varying with sea surface wind and wave levels. The radial distance is equal to the ship's designed anchoring radius, and the sway angle is equal to the maximum rotation angle of the anchor winch drum. The combination of these three dimensions forms a closed three-dimensional space, which constitutes the basic working space of the anchorage. The dynamic motion characteristics of ship anchoring operations are introduced to further divide the basic working space. These dynamic motion characteristics include the low-speed motion of the anchor chain being lowered at a constant speed, the dragging and sliding characteristics after the anchor claws touch the seabed, the small-amplitude swaying characteristics of the ship driving the anchorage to reciprocate under the influence of wind and waves, and the stationary limiting characteristics after the anchor chain is fully released. The basic working space is divided according to the spatial motion amplitude threshold corresponding to the motion characteristics. Low-speed descent corresponds to the core cylindrical area inside the basic space, dragging and sliding corresponds to the annular strip area outside the core area, reciprocating small-amplitude swing corresponds to the transition buffer area outside the annular area, and static limiting corresponds to the outer boundary of the buffer area. The three-dimensional space is combined to form the dynamic zoning of anchorage for ship anchoring operations. For example, for a 5,000-ton bulk carrier, the classification parameters are: draft 8 meters, maximum anchor chain length 330 meters, and the vertical drop coefficient of the anchor chain is taken as 0.9 when the wind and waves are calm. Substituting into the formula, the vertical height = 8 + 330 × 0.9 = 305 meters. The radial distance in the plane is fixed at 120 meters, and the plane swing angle is ±35 degrees. Thus, a cylindrical basic working space with a defined length, width, and height is constructed. This allows for the setting of spatial segmentation thresholds. When the anchor chain is lowered at low speed, the radial offset should not exceed 15 meters, thus defining an inner core zone with a radius of 15 meters. The radial offset range of the anchor claws dragging underwater is 15 to 60 meters, defining an intermediate annular drag zone. The radial offset range of wind and waves is 60 to 100 meters, defining an outer buffer swing zone. After the anchor chain is fully released, the upper limit of the radial offset is 120 meters, defining the outermost static limit zone. These elements together constitute the complete dynamic zoning of the ship's anchorage.

[0068] The positional offset and movement trend of the anchor are obtained based on the dynamic zoning of the anchor, specifically including the following steps:

[0069] Lock the anchorage in the actual working space of the anchorage dynamic zone;

[0070] The positional offset of the anchor is obtained by comparing the actual working space position with the reference space range of the anchor dynamic zoning.

[0071] By tracking the spatial switching trajectory and displacement characteristics of anchors between different anchor dynamic zones, the trend of motion change can be obtained.

[0072] Based on the time-series visual data of sea surface anchoring acquired by machine vision, the three-dimensional coordinate information of the anchor entities on the sea surface and underwater is identified in real time. The three-dimensional coordinates include vertical height, radial distance, and sway angle. These three sets of coordinate values ​​together constitute the actual working space position of the anchor. The machine vision equipment continuously acquires images at fixed time intervals and outputs a set of three-dimensional coordinate data every 0.5 seconds, continuously recording the spatial position of the anchor at every moment throughout the entire anchoring process. For example, when the ship is anchoring for 10 seconds, the visual recognition obtains a vertical height of 120 meters, a radial distance of 32 meters, and a sway angle of 18 degrees for the anchor. The combination of these three sets of values ​​locks in the actual working space position of the anchor at that time point.

[0073] The reference space range for dynamic zoning of anchorages includes the standard vertical height, standard radial distance, and standard swing angle of the corresponding zone. Position offset amplitude is divided into vertical offset amplitude, radial offset amplitude, and angular offset amplitude. Vertical offset amplitude = actual vertical height of anchorage - standard vertical height of zone; radial offset amplitude = actual planar radial distance of anchorage - standard radial distance of zone; angular offset amplitude = actual planar swing angle of anchorage - standard swing angle of zone. These values ​​are combined to obtain the anchorage position offset amplitude data. For example, at the 10th second, the anchorage position data shows that the reference parameters for the inner core zone where the anchorage is located are a standard vertical height of 125 meters, a standard radial distance of 15 meters, and a standard swing angle of 10 degrees. The vertical offset amplitude is 120 - 125 = -5 meters, the radial offset amplitude is 32 - 15 = 17 meters, and the angular offset amplitude is 18 - 10 = 8 degrees. These three sets of calculated values ​​are integrated to obtain the anchorage position offset amplitude at that moment. The positive or negative sign of the value represents the direction of the anchorage offset, and the absolute value represents the magnitude of the offset.

[0074] The movement trajectory of the anchorage across zones is tracked to extract movement change trends. The actual working space position of the anchorage is continuously read at different time points, and the dynamic zone type of the anchorage at each point is compared. The entire trajectory of the anchorage switching from one zone to another is recorded, and the displacement change characteristics corresponding to each segment of the trajectory are extracted. Displacement change characteristics include total radial displacement per unit time, total vertical displacement per unit time, and total angular deflection per unit time. The unit time is set to 1 second. The increase or decrease of the values ​​of these three indicators corresponds to different movement change trends. If the total radial displacement per unit time continuously increases, it represents a spreading movement trend of the anchorage continuously sliding outward into the zone. If the total radial displacement per unit time continuously decreases, it represents a contracting movement trend of the anchorage retracting and converging inward into the zone. If the total angular deflection per unit time repeatedly alternates between positive and negative, it represents a reciprocating oscillating movement trend caused by wind and waves. If all three displacement indicators approach 0, it represents a static movement trend of the anchorage being stably stationary. For example, during the 10th to 30th second of the anchoring operation, the radial displacement of the anchor increased by 0.8 meters every second, continuously switching from the inner core zone to the middle annular drag zone. The total angle deflection per unit time remained stable within 1 degree, and the vertical height value did not fluctuate significantly. By integrating the trajectory of this segment with the displacement quantification index, it was determined that the anchor was in a trend of outward dragging and sliding during this stage.

[0075] The temporal characteristics of motion are obtained based on the magnitude of positional offset and the trend of motion change, specifically:

[0076] By temporally correlating and fusing the positional shift magnitude at different times with the corresponding motion change trend over the time period, motion temporal characteristics are obtained.

[0077] The positional offset amplitude includes vertical offset amplitude, radial offset amplitude, and angular offset amplitude. The motion change trend covers a continuous time interval. Each set of instantaneous positional offset amplitudes is matched to the 5-second time segment to which its acquisition time belongs, establishing a corresponding binding relationship between the time value and the motion trend of the segment. For example, during anchoring operations, 5 independent sets of positional offset amplitudes are generated at the 10th, 10.5th, 11th, 11.5th, and 12th seconds. These 5 moments all belong to the continuous time period from the 10th to the 15th second. The motion change trend corresponding to this time period is the outward dragging and sliding of the anchor. All 5 sets of instantaneous offset amplitudes are bound and matched with the dragging and sliding trend.

[0078] The vertical dimension fusion value is calculated as follows: Vertical dimension fusion value = Sum of vertical offset amplitudes at all times within the time period ÷ Total number of times within the time period; Vertical fusion value = Sum of vertical offset amplitudes at all times ÷ Total number of times within the time period. Radial dimension fusion value = Sum of radial offset amplitudes at all times within the time period ÷ Total number of times within the time period; Radial fusion value = Sum of radial offset amplitudes at all times ÷ Total number of times within the time period. Angular dimension fusion value = Sum of angular offset amplitudes at all times within the time period ÷ Total number of times within the time period; Angular fusion value = Sum of angular offset amplitudes at all times ÷ Total number of times within the time period. These three fusion values ​​are integrated with the corresponding motion change trends of the time period to form a basic temporal unit for a single time period. By concatenating these basic temporal units of all consecutive time periods in chronological order, the motion temporal characteristics are obtained.

[0079] Based on motion time-series characteristics and anchoring safety constraint data, the hierarchical monitoring constraints for the ship's anchoring process are obtained, specifically including the following steps:

[0080] Determine the basic operational constraint benchmarks for anchoring based on the dynamic zoning of anchorages;

[0081] The revised hierarchical benchmark is obtained by revising the coverage of the basic operation constraint benchmark based on the motion time sequence characteristics;

[0082] Based on the anchoring safety constraint data, the safety specifications of the modified hierarchical benchmark are defined to obtain the hierarchical monitoring constraints for the ship anchoring process.

[0083] The anchor dynamic zoning is pre-divided into multiple annular spatial intervals according to radial distance. Each zoning layer is matched with a fixed vertical height threshold and swing angle threshold. The built-in reference space boundary values ​​of each zoning layer are extracted, and all boundary values ​​are integrated to form the basic operational constraint benchmark for the anchoring operation space. The basic operational constraint benchmark includes the maximum allowable radial range, the upper and lower limits of vertical height, and the allowable swing angle range. The maximum allowable radial range = the radial distance value of the outermost dynamic zoning layer; the lower limit of vertical height = the standard vertical height value of the inner core zoning layer; the upper limit of vertical height = the standard vertical height value of the outermost zoning layer; the lower limit of swing angle = the minimum swing angle extreme value of the zoning layer; and the upper limit of swing angle = the maximum swing angle extreme value of the zoning layer. For example, the dynamic zoning of the anchorage is set with three layers of annular intervals. The upper limit of the radial distance of the inner zone is 15 meters, the upper limit of the radial distance of the middle zone is 32 meters, and the upper limit of the radial distance of the outer zone is 50 meters. The standard vertical height of the inner zone is 125 meters, and the standard vertical height of the outer zone is 200 meters. The extreme values ​​of the standard swing angle of the zone are -10 degrees and 10 degrees. After integrating the values, the basic operation constraint benchmark is obtained, where the maximum allowable radial range is 50 meters, the lower limit of the vertical height is 125 meters, the upper limit of the vertical height is 200 meters, the lower limit of the swing angle is -10 degrees, and the upper limit of the swing angle is 10 degrees. This set of values ​​constitutes a static benchmark without taking into account the actual movement state of the anchorage.

[0084] Motion timing features include vertical blending values, radial blending values, and angular blending values. Correction operations are divided into radial correction, vertical correction, and angular correction. The radially corrected boundary value = the basic datum radial boundary value + the radial blending value; the vertically corrected lower limit value = the basic datum vertical lower limit value - the vertical blending value; the vertically corrected upper limit value = the basic datum vertical upper limit value + the vertical blending value; the angularly corrected lower limit value = the basic datum angle lower limit value - the angular blending value; the angularly corrected upper limit value = the basic datum angle upper limit value + the angular blending value. For example, if the radial fusion value of the motion sequence characteristics at a certain time period is 19.4 meters, the vertical fusion value is -5 meters, and the angular fusion value is 8 degrees, the maximum boundary after radial correction is 50 + 19.4 = 69.4 meters, the lower limit after vertical correction is 125 - (-5) = 130 meters, the upper limit after vertical correction is 200 + (-5) = 195 meters, the lower limit after angular correction is -10 - 8 = -18 degrees, and the upper limit after angular correction is 10 + 8 = 18 degrees, thus integrating to form a correction level benchmark adapted to the real-time motion state of the anchor. This benchmark is different from the static foundation benchmark and can dynamically adjust the spatial coverage range according to the displacement trend of the anchor.

[0085] The system retrieves pre-set anchoring safety constraint data, defines safety regulations based on the revised tier benchmark, and ultimately outputs tiered monitoring constraints for the ship's anchoring process. The anchoring safety constraint data includes multiple safety thresholds, corresponding to mild warning, moderate warning, and severe danger levels. The revised boundary values ​​of each tier benchmark are compared one by one with the tiered thresholds in the anchoring safety constraint data to divide the space into tiered monitoring intervals. The interval division calculation formulas are: upper limit of mild warning interval = revised tier benchmark boundary value × 0.8; upper limit of moderate warning interval = revised tier benchmark boundary value × 0.9; lower limit of severe danger interval = revised tier benchmark boundary value × 0.9; upper limit of severe danger interval = revised tier benchmark boundary value × 1. Vertical and angular dimensions are divided into tiered intervals using the same proportional calculation logic. All dimensional tiered intervals are integrated with their corresponding safety warning levels to form tiered monitoring constraints for the ship's anchoring process. These constraints are directly used to determine in real time whether the anchorage operation space is within the safety regulations.

[0086] The real-time status assessment of a ship's anchoring is obtained based on the layered monitoring constraints and visual time-series data of anchoring on the sea surface. The specific steps include:

[0087] Based on the safety control standards corresponding to the hierarchical monitoring constraints, verify the actual operational status of the dynamic zoning of anchorages in the visual time series data of sea surface anchorage.

[0088] The comparison results are obtained by comparing the actual movement performance of the anchor with the matching difference of the constraint level under the actual operation state.

[0089] Based on the comparison results, normal operation and abnormal deviation conditions are distinguished to obtain real-time status judgment results.

[0090] The three observed values ​​of a single time-series anchor are compared with the boundary values ​​of each level of the hierarchical monitoring constraints to determine the current dynamic zoning level of the anchor at that time-series node. After all time-series nodes have completed zoning matching, they are integrated to form the actual operating status. Anchor radial zoning level identifier = 0 represents a safe zone; anchor radial zoning level identifier = 1 represents a mild warning zone; anchor radial zoning level identifier = 2 represents a moderate warning zone; anchor radial zoning level identifier = 3 represents a severe danger zone. When the real-time radial position is less than or equal to the upper limit of the mild warning zone, the radial zoning level identifier = 0; when the upper limit of the mild warning zone is less than or equal to the upper limit of the moderate warning zone, the radial zoning level identifier = 1; when the upper limit of the moderate warning zone is less than or equal to the maximum radial boundary of the corrected reference zone, the radial zoning level identifier = 2; when the real-time radial position is greater than or equal to the maximum radial boundary of the corrected reference zone, the radial zoning level identifier = 3. For example, within the stratified monitoring constraints, the upper limit of the radial light warning interval is 55.52 meters, the upper limit of the radial moderate warning interval is 62.46 meters, the maximum radial boundary of the correction benchmark is 69.4 meters, and the real-time radial position of the anchor is collected by a certain visual time-series node at 58 meters. Substituting this into the judgment logic, the radial partition level identifier of this node is 1, indicating that the anchor is currently in the dynamic partition corresponding to the light warning. The three-dimensional level identifiers of all collected nodes within this time-series data are assigned, and the actual operating status of the anchor's dynamic partition is obtained by summarizing all node identifier data.

[0091] The system reads the actual operating status output from the preceding sequence and performs a difference calculation to compare the actual movement performance of the anchor with the constraint level to obtain the comparison difference results. The comparison difference is divided into radial difference, vertical difference, and angular difference. The radial comparison difference value = real-time radial position of the anchor - standard radial boundary value of the current partition level; the vertical comparison difference value = real-time vertical height of the anchor - standard vertical boundary value of the current partition level; and the angular comparison difference value = real-time swing angle of the anchor - standard angular boundary value of the current partition level. For example, if the real-time radial position of the anchor at this time-series node is 58 meters, the current standard radial boundary value for the light warning zone is 55.52 meters, and the radial comparison difference is 58 - 55.52 = 2.48 meters, and the real-time vertical height of the same node is 142 meters, the standard lower vertical boundary for the light warning zone is 130 meters, and the vertical comparison difference is 142 - 130 = 12 meters, the real-time swing angle is 12 degrees, the standard upper angle boundary for the light warning zone is 18 degrees, and the angle comparison difference is 12 - 18 = -6 degrees, the three-dimensional sub-item difference values ​​are integrated into a single-node comparison array, and the comparison arrays of all time-series nodes are sequentially concatenated to form the comparison difference results of the time-series dimensions.

[0092] The system retrieves and compares the differences, using threshold-based logic to distinguish between normal operating conditions and abnormal offset conditions, thus outputting real-time status judgment results. Pre-set allowable difference thresholds for each dimension: radial allowable difference threshold = 5 meters, vertical allowable difference threshold = 20 meters, angular allowable difference threshold = 10 degrees. A radial condition indicator of 0 represents normal operation, while a radial condition indicator of 1 represents abnormal offset. When the absolute value of the radial comparison difference is less than or equal to the radial allowable difference threshold, the radial condition indicator is 0; when the absolute value of the radial comparison difference is greater than the radial allowable difference threshold, the radial condition indicator is 1. The vertical dimension condition indicator uses the same absolute value comparison calculation logic as the angular dimension condition indicator. The overall condition judgment rule for a single time-series node is: when all three indicators (radial, vertical, and angular) are equal to 0, it is judged as a normal operating condition; when any one indicator is equal to 1, it is judged as an abnormal offset condition. For example, if the radial comparison difference is 2.48 meters (absolute value = 2.48), the radial condition indicator is 0; if the vertical comparison difference is 12 meters (absolute value = 12), the vertical condition indicator is 0; and if the angle comparison difference is -6 degrees (absolute value = 6), the angle condition indicator is 0, and all three indicators are 0, this time node is determined to be in normal operating condition. However, if the radial comparison difference at a certain node is 6.2 meters (absolute value = 6.2), and the radial condition indicator is 1, even if the other two indicators are 0, this node is still determined to be in an abnormal offset condition. The condition determination conclusions of all time nodes are summarized in chronological order to obtain the real-time status judgment result of the anchoring cycle.

[0093] The extraction of standardized monitoring constraint parameters based on hierarchical monitoring constraints includes the following steps:

[0094] The anchor position offset amplitude and motion change trend are matched with the control standards of each level of hierarchical monitoring constraints to obtain the matching results;

[0095] Based on the matching results, select safety constraints that fit the actual working conditions of anchoring.

[0096] Standardized monitoring constraint parameters adapted for anchorage detection are obtained based on the safety constraints.

[0097] The anchor position offset amplitude and motion change trend are respectively matched with the control standards of each level of the hierarchical monitoring constraint conditions to generate a complete matching result. The anchor position offset amplitude is divided into radial offset amplitude, vertical offset amplitude, and angular offset amplitude. The motion change trend is divided into radial change rate, vertical change rate, and angular change rate. The control standards of each level of the hierarchical monitoring constraint conditions include independent thresholds for safety level, mild warning level, moderate warning level, and severe danger level. Each set of thresholds is configured with two sets of reference values: upper limit of offset amplitude and upper limit of motion rate. Radial position offset amplitude = real-time radial coordinate of anchor - reference radial coordinate of ship anchor point; vertical position offset amplitude = real-time vertical coordinate of anchor - reference vertical coordinate of ship anchor point; angular position offset amplitude = real-time swing angle of anchor - reference swing angle of ship anchor point. Radial motion change rate = change in radial coordinate per unit time ÷ 1 second; vertical motion change rate = change in vertical coordinate per unit time ÷ 1 second; angular motion change rate = change in swing angle per unit time ÷ 1 second. When the single-dimensional offset amplitude of the anchor is less than or equal to the upper limit of the offset amplitude of the corresponding control standard, and the single-dimensional motion change rate of the anchor is less than or equal to the upper limit of the rate of the corresponding control standard, it is determined that the dimension matches the current control standard. After all three dimensions match, a matching identifier for the corresponding level is generated. The matching identifiers of all dimensions and all levels are summarized and integrated to form a complete matching result. For example, the control standard for the mild warning level of layered monitoring constraints sets the upper limit of radial offset amplitude = 55.52 meters and the upper limit of radial motion change rate = 0.8 meters per second. At a certain moment, the real-time radial coordinate of the anchor is 42 meters, the reference radial coordinate of the ship's anchor point is 0 meters, and the radial position offset amplitude is 42-0=42 meters. Within 1 second during this period, the change in the radial coordinate of the anchor is 0.5 meters, and the radial motion change rate is 0.5÷1=0.5 meters per second. Since 42 meters ≤ 55.52 meters and 0.5 meters per second ≤ 0.8 meters per second, it is determined that the radial dimension matches the mild warning level control standard. The vertical dimension and the angular dimension use the same double threshold matching calculation logic. After the three-dimensional matching identifiers are integrated, the complete matching result corresponding to this time node is obtained.

[0098] Within the matching results, the hierarchical matching identifier corresponding to each time sequence node is traversed in chronological order. The total duration of consecutive occurrence of a single hierarchical identifier is counted. The hierarchical continuous duration = the number of time nodes in which the same hierarchical identifier exists consecutively × 1 second. The preset working condition fit judgment duration threshold is 30 seconds. When the continuous duration of a certain level is ≥30 seconds, it is determined that the constraint content corresponding to the control standard of that level fits the current actual anchoring movement condition. Otherwise, the constraint content is removed. For example, if the mild warning level matching indicator appears continuously for 46 time nodes, the continuous duration of the level is 46 × 1 = 46 seconds. Since 46 seconds ≥ 30 seconds, all boundary thresholds and warning triggering rules associated with the mild warning level are retained as safety constraints that conform to the working conditions. If the severe danger level matching indicator appears continuously for only 12 time nodes, the continuous duration of the level is 12 × 1 = 12 seconds. Since 12 seconds < 30 seconds, the constraints associated with the severe danger level are directly eliminated. All level control standards that pass the duration verification are integrated and summarized to form safety constraints that conform to the actual movement conditions of anchoring.

[0099] Based on the safety constraints, standardized conversion calculations are performed to output standardized monitoring constraint parameters adapted for anchoring status detection. The amplitude standardized parameter = offset amplitude threshold within the safety constraint ÷ reference offset value; the velocity standardized parameter = movement velocity threshold within the safety constraint ÷ reference velocity value. For example, if the radial offset amplitude threshold = 55.52 meters, the radial amplitude standardized parameter = 55.52 ÷ 100 = 0.5552; ​​if the radial movement velocity threshold = 0.8 meters per second, the radial velocity standardized parameter = 0.8 ÷ 1 = 0.8. The vertical and angular dimensions are standardized using the same conversion logic. All amplitude and velocity standardized parameters are stored uniformly as standardized monitoring constraint parameters adapted for anchoring status detection.

[0100] The spatial motion layout parameters from the real-time anchoring status assessment results are combined with standardized monitoring constraint parameters to obtain the anchoring status adaptation loss data. This process includes the following steps:

[0101] The spatial motion layout parameters of the real-time anchoring status judgment results are compared with the standardized monitoring constraint parameters to obtain deviation data.

[0102] By integrating the deviation data, we obtain the adaptation loss data for the anchorage state.

[0103] The spatial motion layout parameters include radial layout parameters, vertical layout parameters, and angular layout parameters. The standardized monitoring constraint parameters include radial constraint reference parameters, vertical constraint reference parameters, and angular constraint reference parameters. The radial single deviation value = radial layout parameter - radial constraint reference parameter; the vertical single deviation value = vertical layout parameter - vertical constraint reference parameter; and the angular single deviation value = angular layout parameter - angular constraint reference parameter. The three-dimensional single deviation values ​​of all time-series nodes are integrated to form complete deviation status data. For example, at a certain time-series node, the radial layout parameter is 0.62, the radial constraint reference parameter is 0.5552, and the radial single deviation value is 0.62-0.5552=0.0648. At the same node, the vertical layout parameter is 0.38, the vertical constraint reference parameter is 0.35, and the vertical single deviation value is 0.38-0.35=0.03. The angle layout parameter is 0.72, the angle constraint reference parameter is 0.8, and the angle single deviation value is 0.72-0.8=-0.08. The three sets of single deviation values ​​for this node are entered into the deviation dataset. The sub-deviation values ​​of the time-series nodes are continuously collected throughout the entire anchoring cycle, thereby forming deviation status data covering the entire operation cycle.

[0104] Pre-set three-dimensional fixed weight coefficients: radial weight coefficient = 0.5, vertical weight coefficient = 0.3, and angular weight coefficient = 0.2. The calculation formula for the adaptation loss value of a single time-series node is: Single node adaptation loss value = absolute value of radial single deviation value × radial weight coefficient + absolute value of vertical single deviation value × vertical weight coefficient + absolute value of angular single deviation value × angular weight coefficient. The absolute value of the deviation value is used to eliminate the calculation interference caused by positive and negative offset directions. The single node adaptation loss values ​​of all time-series nodes are arranged and summarized in chronological order to obtain the complete adaptation loss data of the anchored state. For example, the absolute value of the radial single deviation is 0.0648, the absolute value of the vertical single deviation is 0.03, the absolute value of the angular single deviation is 0.08, and the single-node adaptation loss is 0.0648×0.5+0.03×0.3+0.08×0.2=0.0574. This value is the single-node adaptation loss value corresponding to this time sequence node, and this data can be directly used for quantitative assessment of the safety level of anchoring operations.

[0105] The process of generating a real-time monitoring result set of the ship anchoring based on the adaptation loss data includes the following steps:

[0106] The degree of deviation of the ship's real-time anchoring status is determined based on the adaptation loss data;

[0107] Based on the dynamic change trend and timing characteristics of the anchorage dynamic zoning, the anchoring conditions with different degrees of deviation are classified and merged to obtain merged condition information.

[0108] A real-time monitoring result set, including anchorage safety level, operating condition status, and offset risk, is generated based on the merged operating condition information.

[0109] The degree of deviation corresponding to the real-time state of the ship at anchor is defined based on the range of adaptation loss values. Four consecutive, non-overlapping ranges of adaptation loss values ​​are pre-defined, corresponding to no deviation, slight deviation, moderate deviation, and severe deviation, respectively. An adaptation loss value less than or equal to 0.03 is considered no deviation; an adaptation loss value greater than 0.03 and less than or equal to 0.08 is considered slight deviation; an adaptation loss value greater than 0.08 and less than or equal to 0.15 is considered moderate deviation; and an adaptation loss value greater than 0.15 is considered severe deviation. For example, if the adaptation loss value of a certain time-series node is 0.0574, which is between 0.03 and 0.08, then the real-time state of the ship anchoring at that moment is defined as a slight deviation. If the adaptation loss value of another node is 0.18, then the real-time state of the ship anchoring at that moment is defined as a severe deviation. All time-series nodes are matched with their corresponding deviation degree labels and bound to the original adaptation loss values ​​for storage, forming a complete dataset with deviation labels.

[0110] The motion trend parameters and motion time sequence feature parameters corresponding to the dynamic zones of the anchorage are extracted. Combined with different deviation degree labels, the anchoring conditions throughout the entire cycle are classified and merged to output merged condition information. The dynamic zones of the anchorage are divided into near-ship zone, mid-distance zone, and far-ship zone. Motion trend parameters and motion time sequence feature parameters are collected for each zone. The motion trend parameters include the displacement growth rate and the swing growth rate. The motion time sequence feature parameter is the total duration of the same deviation degree label within the same zone. The zone duration = the number of consecutive time sequence nodes with the same deviation degree × 1 second. The preset stable condition judgment duration is equal to 20 seconds. When the duration of the zone corresponding to the same deviation degree within a zone is greater than or equal to 20 seconds, all time sequence nodes in that period are classified into the same type of merged condition unit. If the zone duration is less than 20 seconds, the time sequence nodes in that period are merged into the adjacent time-limited standard condition units. For example, if there are 32 consecutive time-series nodes indicating slight deviation in the far-ship zone, and the zone duration is 32 × 1 = 32 seconds, these 32 time-series nodes are independently divided into a slight deviation merging condition unit for the far-ship zone. If there are only 11 consecutive time-series nodes indicating severe deviation in the near-ship zone, and the zone duration is 11 × 1 = 11 seconds, these 11 time-series nodes are merged into the moderate deviation condition unit of the preceding mid-distance zone. After all zones have completed duration verification and unit merging, the zone position, deviation degree, duration, and motion acceleration value of all independent condition units are summarized to form complete merged condition information.

[0111] Extract the core quantitative indicators corresponding to each set of working condition units to generate a real-time monitoring result set including anchorage safety level, working condition status, and deviation risk. A pre-established mapping between deviation degree and anchorage safety level is implemented: no deviation corresponds to level 1 safety, slight deviation to level 2, moderate deviation to level 3, and severe deviation to level 4. The average adaptation loss value of a unit is calculated as the sum of the adaptation losses of all timing nodes within the unit divided by the total number of timing nodes in the unit. For example, the sum of the adaptation loss values ​​of the 32 time-series nodes of the "slight deviation" working condition unit in the "distant ship zone" is equal to 1.8368. The average adaptation loss value of a single unit is 1.8368 ÷ 32 = 0.0574. This working condition unit matches the Level 2 anchoring safety level. The working condition status is marked as "distant ship zone" for 32 seconds, and the deviation risk value is equal to 0.0574. The three items corresponding to each working condition unit are formed into a monitoring record. The monitoring of all working condition units throughout the entire cycle forms a real-time monitoring result set. The real-time monitoring result set is directly used for the visualization display and risk warning call of ship anchoring operations.

[0112] A machine vision-based real-time monitoring system for ship anchoring includes:

[0113] Data Acquisition Module: Acquires anchor winch data and visual time-series data of anchoring on the sea surface of the vessel to be monitored; the anchor winch data includes anchor winch geometric parameters and anchoring safety constraint data;

[0114] Partitioning module: Spatial clustering of anchor winch geometric parameters is used to obtain dynamic anchorage partitions corresponding to ship anchoring operations;

[0115] The first processing module obtains the position offset amplitude and motion change trend of the anchorage based on the dynamic zoning of the anchorage, and obtains the motion time sequence characteristics based on the position offset amplitude and motion change trend;

[0116] The second processing module: Based on motion time series characteristics and anchoring safety constraint data, it obtains the layered monitoring constraints of the ship's anchoring process; based on the layered monitoring constraints and the visual time series data of anchoring on the sea surface, it obtains the real-time status judgment result of the ship's anchoring.

[0117] Extraction module: Extracts standardized monitoring constraint parameters based on hierarchical monitoring constraints; obtains the adaptation loss data of the anchoring state based on the spatial motion layout parameters of the real-time anchoring state judgment results and the standardized monitoring constraint parameters;

[0118] Generation module: Generates a set of real-time monitoring results of ship anchoring based on the adapted loss data.

[0119] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A machine vision-based real-time monitoring method for ship anchoring, characterized in that, Includes the following steps: Collect anchor winch data and visual time-series data of the vessel to be monitored; the anchor winch data includes anchor winch geometric parameters and anchoring safety constraint data; Spatial clustering of anchor winch geometric parameters yields dynamic anchorage partitions corresponding to ship anchoring operations; The position offset amplitude and motion change trend of the anchor are obtained based on the dynamic zoning of the anchor, and the motion time sequence characteristics are obtained based on the position offset amplitude and motion change trend. Based on motion timing characteristics and anchoring safety constraint data, the hierarchical monitoring constraints of the ship's anchoring process are obtained; The real-time status judgment of the ship's anchoring is obtained based on the layered monitoring constraints and the visual time series data of the sea surface anchoring. Standardized monitoring constraint parameters are extracted based on hierarchical monitoring constraints; adaptation loss data of anchoring state is obtained based on the spatial motion layout parameters and standardized monitoring constraint parameters according to the real-time anchoring state judgment results. A real-time monitoring result set of the ship's anchoring is generated based on the adapted loss data.

2. The method for real-time monitoring of ship anchoring based on machine vision according to claim 1, characterized in that, Spatial clustering of anchor winch geometry parameters yields dynamic anchorage partitions corresponding to ship anchoring operations, specifically including the following steps: Based on the spatial motion attributes and operational range of ship anchoring operations, the geometric parameters of anchor winches are hierarchically classified and integrated to obtain classified parameter data. The basic operating space of the anchorage is defined based on the classification parameter data. The basic operating space is then divided into zones based on the basic operating space and the dynamic motion characteristics of the ship's anchoring operation, resulting in the dynamic zone of the anchorage corresponding to the ship's anchoring operation.

3. The method for real-time monitoring of ship anchoring based on machine vision according to claim 1, characterized in that, The positional offset and movement trend of the anchor are obtained based on the dynamic zoning of the anchor, specifically including the following steps: Lock the anchorage in the actual working space of the anchorage dynamic zone; The positional offset of the anchor is obtained by comparing the actual working space position with the reference space range of the anchor dynamic zoning. By tracking the spatial switching trajectory and displacement characteristics of anchors between different anchor dynamic zones, the trend of motion change can be obtained.

4. The method for real-time monitoring of ship anchoring based on machine vision according to claim 1, characterized in that, The temporal characteristics of motion are obtained based on the magnitude of positional offset and the trend of motion change, specifically: By temporally correlating and fusing the positional shift magnitude at different times with the corresponding motion change trend over the time period, motion temporal characteristics are obtained.

5. The method and system for real-time monitoring of ship anchoring based on machine vision according to claim 1, characterized in that, Based on motion time-series characteristics and anchoring safety constraint data, the hierarchical monitoring constraints for the ship's anchoring process are obtained, specifically including the following steps: Determine the basic operational constraint benchmarks for anchoring based on the dynamic zoning of anchorages; The revised hierarchical benchmark is obtained by revising the coverage of the basic operation constraint benchmark based on the motion time sequence characteristics; Based on the anchoring safety constraint data, the safety specifications of the modified hierarchical benchmark are defined to obtain the hierarchical monitoring constraints for the ship anchoring process.

6. The method for real-time monitoring of ship anchoring based on machine vision according to claim 5, characterized in that, The real-time status assessment of a ship's anchoring is obtained based on the layered monitoring constraints and visual time-series data of anchoring on the sea surface. The specific steps include: Based on the safety control standards corresponding to the hierarchical monitoring constraints, verify the actual operational status of the dynamic zoning of anchorages in the visual time series data of sea surface anchorage. The comparison results are obtained by comparing the actual movement performance of the anchor with the matching difference of the constraint level under the actual operation state. Based on the comparison results, normal operation and abnormal deviation conditions are distinguished to obtain real-time status judgment results.

7. The method for real-time monitoring of ship anchoring based on machine vision according to claim 6, characterized in that, The extraction of standardized monitoring constraint parameters based on hierarchical monitoring constraints includes the following steps: The anchor position offset amplitude and motion change trend are matched with the control standards of each level of hierarchical monitoring constraints to obtain the matching results; Based on the matching results, select safety constraints that fit the actual working conditions of anchoring. Standardized monitoring constraint parameters adapted for anchorage detection are obtained based on the safety constraints.

8. The method for real-time monitoring of ship anchoring based on machine vision according to claim 1, characterized in that, The spatial motion layout parameters from the real-time anchoring status assessment results are combined with standardized monitoring constraint parameters to obtain the anchoring status adaptation loss data. This process includes the following steps: The spatial motion layout parameters of the real-time anchoring status judgment results are compared with the standardized monitoring constraint parameters to obtain deviation data. By integrating the deviation data, we obtain the adaptation loss data for the anchorage state.

9. A method for real-time monitoring of ship anchoring based on machine vision according to claim 8, characterized in that, The process of generating a real-time monitoring result set of the ship anchoring based on the adaptation loss data includes the following steps: The degree of deviation of the ship's real-time anchoring status is determined based on the adaptation loss data; Based on the dynamic change trend and timing characteristics of the anchorage dynamic zoning, the anchoring conditions with different degrees of deviation are classified and merged to obtain merged condition information. A real-time monitoring result set, including anchorage safety level, operating condition status, and offset risk, is generated based on the merged operating condition information.

10. A machine vision-based real-time monitoring system for ship anchoring, applied to the machine vision-based real-time monitoring method for ship anchoring as described in any one of claims 1-9, characterized in that, include: Data Acquisition Module: Acquires anchor winch data and visual time-series data of the vessel to be monitored; the anchor winch data includes anchor winch geometric parameters and anchoring safety constraint data; Partitioning module: Spatial clustering of anchor winch geometric parameters is used to obtain dynamic anchorage partitions corresponding to ship anchoring operations; The first processing module obtains the position offset amplitude and motion change trend of the anchorage based on the dynamic zoning of the anchorage, and obtains the motion time sequence characteristics based on the position offset amplitude and motion change trend; The second processing module obtains the hierarchical monitoring constraints for the ship's anchoring process based on motion timing characteristics and anchoring safety constraint data. The real-time status judgment of the ship's anchoring is obtained based on the layered monitoring constraints and the visual time series data of the sea surface anchoring. Extraction module: Extracts standardized monitoring constraint parameters based on hierarchical monitoring constraints; obtains the adaptation loss data of the anchoring state based on the spatial motion layout parameters of the real-time anchoring state judgment results and the standardized monitoring constraint parameters; Generation module: Generates a set of real-time monitoring results of ship anchoring based on the adapted loss data.