Deck height monitoring system and method for berthed ships based on laser radar

Abstract

The present invention relates to a deck height monitoring system and a monitoring method for berthing ships based on laser radar. The monitoring system comprises a point cloud acquisition subsystem and a data processing subsystem; the point cloud acquisition subsystem comprises a three-dimensional point cloud data acquisition module, and the three-dimensional point cloud data acquisition module is used to collect the original point cloud data of the entire ship in real time; the data processing subsystem comprises a data processing module, a data storage module and a data transmission module, and the data processing module is used to process the original point cloud data of the ship output by the three-dimensional point cloud data acquisition module in real time; the data storage module is used to save the ship point cloud data transmitted by the data processing module in real time; and the data transmission module is used to transmit the ship point cloud data transmitted by the data processing module. The present invention can accurately and in real time monitor the deck height of the ship, can improve the safety and stability of the shore power system when docking with the ship, and at the same time improve the efficiency of port management.

Description

Technical Field

[0001] The present invention relates to the field of smart port technology, and more specifically, to a deck height monitoring system and a monitoring method for a berthed ship based on a laser radar. Background Art

[0002] With the continuous development of port trade and maritime transportation, the demand for safe management and operational efficiency of ships during berthing and docking is increasing. When a ship is berthed, it is usually necessary to connect to the shore power system to supply power and reduce port pollution. However, the current shore power system often cannot monitor the height of the ship deck in real time when docking with the ship, which may lead to safety hazards during docking and unstable connection of cables. At present, some traditional ship berthing monitoring systems usually use technologies such as cameras, one-dimensional laser rangefinders and ultrasonic sensors for monitoring. However, these technologies are often limited by environmental conditions, limited data acquisition capabilities and high costs, resulting in low monitoring accuracy and real-time performance. The traditional ship monitoring system is limited by environmental conditions, limited data acquisition capabilities and high costs, resulting in low monitoring accuracy and real-time performance. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a deck height monitoring system and a monitoring method for a berthed ship based on laser radar, which can accurately and in real time monitor the deck height of the ship to improve the safety and stability when the shore power system is docked with the ship, while improving the efficiency of port management.

[0004] The technical solution adopted by the present invention to solve the technical problem is: constructing a deck height monitoring system for berthed ships based on laser radar, including a point cloud acquisition subsystem and a data processing subsystem;

[0005] The point cloud acquisition subsystem includes a three-dimensional point cloud data acquisition module, which is used to collect the original point cloud data of the entire ship in real time and send the collected original point cloud data of the ship to the data processing subsystem;

[0006] The data processing subsystem includes a data processing module, a data storage module and a data transmission module. The data processing module is used to process the original point cloud data of the ship output by the three-dimensional point cloud data acquisition module in real time, obtain real-time ship point cloud data and deck distance laser radar height information, and transmit the ship point cloud data and the deck distance laser radar height information to the data storage module and the data transmission module respectively; the data storage module is used to save the ship point cloud data and the deck distance laser radar height information transmitted by the data processing module in real time; the data transmission module is used to transmit the ship point cloud data and the deck distance laser radar height information transmitted by the data processing module to an external terminal.

[0007] According to the above scheme, the three-dimensional point cloud data acquisition module is a three-dimensional pan-tilt platform composed of a single-line laser radar and a servo motor, which realizes the collection function of ship point cloud data.

[0008] According to the above scheme, the data processing subsystem is composed of an industrial computer or an embedded device to realize point cloud data processing, data storage and data transmission functions.

[0009] The present invention also provides a method for monitoring the deck height of a berthed ship based on a laser radar, comprising the following steps:

[0010] S1. The point cloud acquisition subsystem sends the collected original point cloud data A to the data processing subsystem. The data processing subsystem removes shore-based fixed facilities from the original point cloud data A through a straight-through filter and only retains the point cloud data of the ship. Then, the radius filter is used to remove outliers and the voxel filter is used to downsample and reduce the point cloud density to obtain the processed point cloud data B.

[0011] S2, projecting the point cloud data B along the Z axis to the xoy plane, recording the projected point cloud data as point cloud data C, applying an edge detection algorithm to the point cloud data C to obtain a series of boundary points, extracting the obtained boundary points and individually restoring them to obtain point cloud data D of the outermost contour of the ship;

[0012] S3. After obtaining the ship boundary point cloud information, the RANSAC straight line fitting algorithm is used for the extracted outermost contour point cloud of the ship, i.e., point cloud data D, to extract the equation of the straight line closest to the scanning platform and the longest line, and obtain the characteristic straight line equation of the ship. The boundary point with the largest coordinate value on the characteristic straight line equation and in the positive direction of the X-axis is identified and recorded as point B (X b ,Y b ), extract the coordinates of point B; return to the three-dimensional point cloud data of the ship, find the point with the largest x-axis coordinate, extract the coordinates of this point, and record it as point C (X c ,Y c ,Z c );

[0013] S4. After obtaining the coordinates of key points B and C, the point cloud data B is subjected to through-filtering. The through-filtering method is as follows: based on the x-axis coordinate of point B, the point cloud data to the right of the x-axis coordinate of point B is segmented, and all the point cloud data below are filtered out based on 2 / 3 of the z-axis coordinate of point C, and the remaining point cloud data is saved, so that only the bow deck and part of the bow and side of the ship are left. The deck plane is fitted with the RANSAC algorithm to find the plane with the largest number of point clouds, which is the deck plane. Finally, the z-axis coordinates of the point clouds on the deck plane are averaged, and the average value obtained is the height of the deck.

[0014] According to the above scheme, in step S1, the method of through filtering is: setting a threshold range, filtering out all points outside the threshold range in the original point cloud data directly obtained from the point cloud acquisition system, and only retaining points within the threshold, specifically:

[0015] pass.setFilterFieldName("z"); Set the filter field to the z-axis direction;

[0016] pass.setFilterLimits(0.0,50); Set the filter range in the filter direction;

[0017] pass.setNegative(true); Set whether to keep the points within the range or filter out the points within the range. If the value in the brackets is true, the points within the range are kept.

[0018] The radius filtering method is as follows: take any point in the point cloud data as the center, give a radius size, draw a circle with the selected point as the center and the given value as the radius, and count the number of points within the circle according to the Euclidean distance between the points. When the number of points in the circle is greater than the set threshold, the point is retained, otherwise it is removed.

[0019] The voxel filtering method is: after obtaining the point cloud coordinate data, calculate the maximum value on the X, Y, and Z coordinate axes , , and minimum value , , , according to the maximum and minimum values ​​on each coordinate axis, find the length of each side of the minimum volume surrounding the point cloud data , , ;

[0020]

[0021] Given a voxel grid size of a, the X, Y, and Z coordinate axes are divided into P, Q, and R parts, resulting in a total of A grid of individual pixels, where Indicates rounding downward;

[0022]

[0023]

[0024] All voxel grids are numbered and recorded as , each data point Divide into voxel grids

[0025] Calculate the centroid of each voxel grid, denoted as , replace the coordinate information of all points in the voxel with the coordinates of the center of gravity of the voxel grid. If the center of gravity does not exist, replace the coordinates of all points in the voxel with the coordinates of the point closest to the center of gravity in terms of distance within the voxel;

[0026]

[0027] Where n is the number of data points in the voxel grid, is the data point.

[0028] According to the above scheme, in step S2, the algorithm for determining whether a point P is a boundary point is as follows: when the neighboring points near point P are all biased towards one side, point P is considered to be a boundary point of the point cloud; conversely, when the neighboring points near point P are evenly distributed around it, point P is an internal point.

[0029] According to the above scheme, in step S3, the ship's two-dimensional straight line feature extraction based on the RANSAC algorithm is: two points are randomly selected from the ship's point cloud , ,according to , Determine a straight line S based on the position of and determine the values ​​of the three parameters A, B, and C;

[0030] Assume that the equation of the straight line S to be fitted is:

[0031]

[0032] A and B are the components of the direction vector of the line, and C is the intercept of the line;

[0033] Count the number of points that fall on the straight line S determined by the A, B, and C parameters: Set the distance threshold to , calculate any point in the point cloud data Distance to line S ,when Distance to S Less than When, see is a point on S, where It is calculated by the following formula:

[0034]

[0035] statistics Less than The number of and record it, recorded as W;

[0036] Repeat the above steps N times, where N is the number of iterations of the RANSAC algorithm;

[0037] After the iteration is completed, the parameter straight line equation corresponding to the maximum W value is selected, and the straight line is the desired ship straight line equation.

[0038] According to the above scheme, in step S4, the method of fitting the plane is:

[0039] Randomly select three points from the ship point cloud , , according to , , The coordinates of the three points determine the values ​​of A, B, C, and D, thereby determining a plane L;

[0040] Let the equation of plane L be:

[0041]

[0042] A is the normal vector component of the plane on the X axis, B is the normal vector component of the plane on the Y axis, C is the normal vector component of the plane on the Z axis, and D is the offset of the plane;

[0043] Count the number of points that fall on the plane L determined by the A, B, C, and D parameters: Set the distance threshold to , calculate any point in the point cloud data Distance to plane L ,when Distance to L Less than When, see is a point on L, where It is calculated by the following formula:

[0044]

[0045] statistics Less than The number of and record it, recorded as Q;

[0046] Repeat the above steps N times, where N is the number of iterations of the RANSAC algorithm;

[0047] After the iteration is completed, the parametric plane equation corresponding to the maximum Q value is selected. The plane is the deck plane. The z-axis coordinates of all point clouds on the plane are averaged, and the average value obtained is the height H of the deck:

[0048] .

[0049] The laser radar-based deck height monitoring system and monitoring method for berthed ships of the present invention have the following beneficial effects:

[0050] 1. The invention has strong anti-environmental interference capability. The rotating pan-tilt composed of multi-line laser radar and servo motor can not only work normally during the day and night, but also effectively resist environmental interference, which greatly reduces the influence of environmental factors such as light and weather on monitoring accuracy, and ensures the stability and reliability of the system under various complex environmental conditions;

[0051] 2. The present invention can reduce system costs and improve data acquisition capabilities. The system uses a single-line laser radar combined with a servo motor to form a rotating pan-tilt platform. Compared with the traditional multi-line laser radar solution, the hardware cost is significantly reduced. At the same time, through the high-frequency scanning of the rotating pan-tilt platform, the ship point cloud data can be fully and quickly acquired, which significantly improves the coverage and accuracy of data collection;

[0052] 3. The real-time performance and accuracy of the present invention are greatly improved. Through high-frequency scanning and fast data processing algorithms, the point cloud data of the ship can be continuously acquired, and the data can be processed in real time to calculate the height of the ship deck. The plane fitting algorithm is combined with the RANSAC method, so that the monitoring system can quickly and accurately capture the height changes of the ship deck, ensuring the real-time and reliability of the data, thereby providing timely and effective height monitoring during the ship berthing process, and ensuring the safety and stability of the docking process;

[0053] 4. The efficient data processing and accurate height calculation of the present invention can greatly improve the ability to acquire ship point cloud data. Through the high-frequency scanning of the rotating pan-tilt head, the ship point cloud data can be acquired uninterruptedly, and then reliable data can be acquired in real time through algorithm processing, which improves the real-time and response speed of the monitoring system, enabling it to capture the ship's position and height changes in a timely manner, ensuring the safety and stability of the docking process. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0055] Figure 1 This is a schematic diagram of the ship boundary point cloud and feature points extracted by the present invention;

[0056] Figure 2 is a flow chart of the edge detection algorithm of the present invention;

[0057] Figure 3 is a schematic diagram of edge detection of the present invention;

[0058] Figure 4 This is a bird's-eye view of the shore power scene of the present invention;

[0059] Figure 5 It is a side view of the shore power field of the present invention;

[0060] Figure 6It is a flow chart of the deck height monitoring method of a berthed ship based on laser radar of the present invention. DETAILED DESCRIPTION

[0061] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0062] like Figure 1-6 As shown, the deck height monitoring system for berthed ships based on laser radar of the present invention includes a point cloud acquisition subsystem and a data processing subsystem.

[0063] The point cloud acquisition subsystem includes a three-dimensional point cloud data acquisition module, which is used to collect the original point cloud data of the entire ship in real time and send the collected original point cloud data of the ship to the data processing subsystem; the data processing subsystem includes a data processing module, a data storage module and a data transmission module, wherein: the data processing module is used to process the original point cloud data of the ship output by the three-dimensional point cloud data acquisition module in real time, obtain real-time ship point cloud data and deck distance laser radar height information, and transmit the ship point cloud data and the deck distance laser radar height information to the data storage module and the data transmission module respectively; the data storage module is used to save the ship point cloud data and the deck distance laser radar height information transmitted by the data processing module in real time; the data transmission module is used to transmit the ship point cloud data and the deck distance laser radar height information transmitted by the data processing module to an external terminal.

[0064] The point cloud data acquisition subsystem includes a three-dimensional point cloud data acquisition module, which is used to collect the original point cloud data of the entire ship in real time and send the collected original point cloud data of the ship to the data processing subsystem; the three-dimensional point cloud data acquisition module is a three-dimensional pan-tilt platform composed of a single-line laser radar and a servo motor. The three-dimensional point cloud data acquisition module realizes the collection function of ship point cloud data. The three-dimensional point cloud data acquisition module is arranged on the side of the mobile power connection platform close to the berth.

[0065] The data processing subsystem includes a data processing module, a data storage module and a data transmission module. The data processing subsystem is implemented by an industrial computer or an embedded device to realize the point cloud data processing, data storage and data transmission functions. The data processing module is used to process the original point cloud data of the ship output by the three-dimensional point cloud data acquisition module, obtain real-time ship point cloud data and deck distance laser radar height information, and transmit the ship point cloud data and deck distance laser radar height information to the data storage module and the data transmission module respectively. The data storage module is used to store the ship point cloud data and deck distance laser radar height information transmitted by the data processing module in real time, and the data transmission module is used to transmit the ship point cloud data and deck distance laser radar height information transmitted by the data processing module to an external terminal.

[0066] The present invention also provides a method for monitoring the deck height of a berthed ship based on a laser radar, comprising the following steps:

[0067] S1. The point cloud acquisition subsystem sends the collected original point cloud data A to the data processing subsystem. The data processing subsystem removes shore-side fixed facilities such as shore foundations from the original point cloud data A through direct filtering, retaining only the point cloud data of the ship, and then uses radius filtering to remove outliers and voxel filtering to downsample and reduce the point cloud density, thereby obtaining the processed point cloud data B.

[0068] The method of through filtering is: set a threshold range, filter out all points outside the threshold range in the original point cloud data directly obtained from the point cloud acquisition system, and only retain the points within the threshold range:

[0069] pass.setFilterFieldName("z"); Set the filter field to the z-axis direction;

[0070] pass.setFilterLimits(0.0, 50); Set the filter range in the filtering direction;

[0071] pass.setNegative(true); Set whether to keep the points in the range or filter out the points in the range. If the value in the brackets is true, the points in the range will be kept. In this case, only the point cloud data with Z-axis coordinate values ​​between 0-50 will be kept, and all the point cloud data with coordinate values ​​below 0 or above 50 will be deleted. After setting the appropriate range in the three directions of X, Y, and Z, you can cut and obtain complete ship point cloud data with few interfering objects.

[0072] The method of radius filtering is: first, take any point in the point cloud data as the center, give a radius size, draw a circle with the selected point as the center and the given value as the radius, then count the number of points within the circle according to the Euclidean distance between the points, and when the number of points in the circle is greater than the set threshold, the point is retained, otherwise it is removed. After radius filtering, some outliers or noise points will be removed, and the point cloud data of the ship will be more accurate.

[0073] The voxel filtering method is: after obtaining the point cloud coordinate data, calculate the maximum value on the X, Y, and Z coordinate axes , , and minimum value , , , and then find the length of each side of the minimum volume surrounding the point cloud data based on the maximum and minimum values ​​on each coordinate axis , , ;

[0074]

[0075] Given a voxel grid size of a, the X, Y, and Z coordinate axes can be divided into P, Q, and R parts, resulting in a total of A grid of individual pixels, where Indicates rounding downward.

[0076]

[0077]

[0078] All voxel grids are numbered and recorded as , each data point Divide into voxel grids

[0079] According to the above formula, the centroid of each voxel grid is calculated and recorded as , replace the coordinate information of all points in the voxel with the coordinates of the center of gravity of the voxel grid. When the center of gravity does not exist, replace the coordinates of all points in the voxel with the coordinates of the point closest to the center of gravity in terms of distance within the voxel;

[0080]

[0081] Where n is the number of data points in the voxel grid, is the data point.

[0082] Voxel filtering can greatly reduce the amount of point cloud data of a ship without losing key data and can also improve subsequent calculation efficiency.

[0083] S2, project the point cloud data B along the Z axis to the xoy plane (see the attached coordinate system for the method of choosing the coordinate system). Figure 5 ), the point cloud data after projection is recorded as point cloud data C, and an edge detection algorithm is applied to the point cloud data C to obtain a series of boundary points. The obtained boundary points are extracted and restored separately to obtain the point cloud data D of the outermost contour of the ship.

[0084] To determine whether a point P is a boundary point, the algorithm is as follows: when the neighboring points near point P are all biased to one side, point P is considered to be a boundary point of the point cloud; conversely, when the neighboring points near point P are evenly distributed around it, point P is an internal point. Specifically, first select a point P with the smallest y-axis value in the point cloud data C, use point P as the starting point for edge detection, use the kd tree to establish the topological relationship between point clouds, search for P's k neighboring points, arbitrarily select a neighboring point P1, use the vector PP1 as the reference, and calculate the rotation angle of all other neighboring points PP2, PP3, etc. in the clockwise direction, which are recorded as , then bubble sort the rotation angles, calculate the corresponding angle differences and select the maximum angle difference, and determine whether the maximum angle difference is greater than the maximum threshold. Here, the maximum threshold is set to 140°. If it is greater than the maximum threshold, mark the point as a boundary feature point and retain it, and then determine the next point; if it is less than the maximum threshold, determine the next point until all points are determined.

[0085] The angle formula between two adjacent connecting lines is calculated as:

[0086] .

[0087] S3. After obtaining the boundary point cloud information of the ship, the RANSAC straight line fitting algorithm is used for the extracted outermost contour point cloud of the ship, i.e., the point cloud data D, to extract the equation of the line closest to the scanning platform and the longest line, i.e., the characteristic straight line equation of the ship. Then, the boundary point with the largest coordinate value on the characteristic straight line equation and in the positive direction of the X-axis is found. This point is the intersection point between the bow and the side of the ship, recorded as point B (X b ,Y b ), extract the coordinates of point B. Return to the 3D point cloud data of the ship and find the point with the largest x-axis coordinate. This point is the coordinate of the tip of the bow. Extract the coordinates of this point and record it as point C (X c ,Y c ,Z c ).

[0088] Ship 2D straight line feature extraction based on RANSAC algorithm: Randomly select two points from the ship point cloud , ,according to , Determine a straight line S based on the position of and determine the values ​​of the three parameters A, B, and C;

[0089] Assume that the equation of the straight line S to be fitted is:

[0090]

[0091] A and B are the components of the direction vector of the line, and C is the intercept of the line.

[0092] Count the number of points that fall on the straight line S determined by the A, B, and C parameters: Set the distance threshold to , calculate any point in the point cloud data Distance to line S ,when Distance to S Less than When, see is a point on S, where It is calculated by the following formula:

[0093]

[0094] statistics Less than , and record it as W.

[0095] Repeat the above steps N times, where N is the number of iterations of the RANSAC algorithm.

[0096] After the iteration is completed, the parameter straight line equation corresponding to the maximum W value is selected, and the straight line is the desired ship straight line equation.

[0097] S4. After obtaining the coordinates of key points B and C, perform straight-through filtering on point cloud data B. The filtering method is: based on the x-axis coordinate of point B, all point cloud data on the right side of the x-axis coordinate of point B are filtered out, based on 2 / 3 of the z-axis coordinate of point C, all data below are filtered out, and the remaining point cloud data is saved, leaving only the bow deck and part of the bow and side of the ship. Finally, the RANSAC algorithm is used to fit the deck plane. The method of fitting the plane is similar to that of the straight line, except that the parameter equation and the calculation of the distance are different.

[0098] Let the plane equation be:

[0099]

[0100] Any point The distance to this plane is:

[0101]

[0102] Find the plane with the largest number of point clouds, which is the deck plane. Finally, take the average of the z-axis coordinates of the point clouds on the plane, and the average is the height H of the deck:

[0103] .

[0104] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A method for monitoring the deck height of a berthed ship based on laser radar, characterized in that: A deck height monitoring system for berthed ships based on laser radar is adopted. The deck height monitoring system for berthed ships based on laser radar includes a point cloud acquisition subsystem and a data processing subsystem; The point cloud acquisition subsystem includes a three-dimensional point cloud data acquisition module, which is used to collect the original point cloud data of the entire ship in real time and send the collected original point cloud data of the ship to the data processing subsystem; The data processing subsystem includes a data processing module, a data storage module and a data transmission module. The data processing module is used to process the original point cloud data of the ship output by the three-dimensional point cloud data acquisition module in real time, obtain real-time ship point cloud data and height information of the deck from the laser radar, and transmit the ship point cloud data and the height information of the deck from the laser radar to the data storage module and the data transmission module respectively; the data storage module is used to store the ship point cloud data and the height information of the deck from the laser radar transmitted by the data processing module in real time; the data transmission module is used to transmit the ship point cloud data and the height information of the deck from the laser radar transmitted by the data processing module to an external terminal; The laser radar-based deck height monitoring method for berthing ships comprises the following steps: S1. The point cloud acquisition subsystem sends the collected original point cloud data A to the data processing subsystem. The data processing subsystem removes shore-based fixed facilities from the original point cloud data A through a straight-through filter and only retains the point cloud data of the ship. Then, the radius filter is used to remove outliers and the voxel filter is used to downsample and reduce the point cloud density to obtain the processed point cloud data B. S2, projecting the point cloud data B along the Z axis to the xoy plane, recording the projected point cloud data as point cloud data C, applying an edge detection algorithm to the point cloud data C to obtain a series of boundary points, extracting the obtained boundary points and individually restoring them to obtain point cloud data D of the outermost contour of the ship; S3. After obtaining the ship boundary point cloud information, the RANSAC straight line fitting algorithm is used for the extracted outermost contour point cloud of the ship, i.e., point cloud data D, to extract the equation of the straight line closest to the scanning platform and the longest line, and obtain the characteristic straight line equation of the ship. The boundary point with the largest coordinate value on the characteristic straight line equation and in the positive direction of the X-axis is identified and recorded as point B (X b ,Y b ), extract the coordinates of point B; return to the three-dimensional point cloud data of the ship, find the point with the largest x-axis coordinate, extract the coordinates of this point, and record it as point C (X c ,Y c ,Z c ); S4. After obtaining the coordinates of key points B and C, the point cloud data B is subjected to through-filtering. The through-filtering method is as follows: based on the x-axis coordinate of point B, the point cloud data to the right of the x-axis coordinate of point B is segmented, and all the point cloud data below are filtered out based on 2 / 3 of the z-axis coordinate of point C, and the remaining point cloud data is saved, so that only the bow deck and part of the bow and side of the ship are left. The deck plane is fitted with the RANSAC algorithm to find the plane with the largest number of point clouds, which is the deck plane. Finally, the z-axis coordinates of the point clouds on the deck plane are averaged, and the average value obtained is the height of the deck.

2. The method for monitoring the deck height of a berthed ship based on laser radar according to claim 1, characterized in that: In step S1, the method of through filtering is: setting a threshold range, filtering out all points outside the threshold range in the original point cloud data directly obtained from the point cloud acquisition system, and only retaining points within the threshold range, specifically: pass.setFilterFieldName("z"); Set the filter field to the z-axis direction; pass.setFilterLimits(0.0,50); Set the filter range in the filter direction; pass.setNegative(true); Set whether to keep the points within the range or filter out the points within the range. If the value in the brackets is true, the points within the range are kept. The radius filtering method is as follows: take any point in the point cloud data as the center, give a radius size, draw a circle with the selected point as the center and the given value as the radius, and count the number of points within the circle according to the Euclidean distance between the points. When the number of points in the circle is greater than the set threshold, the point is retained, otherwise it is removed. The voxel filtering method is: after obtaining the point cloud coordinate data, calculate the maximum value on the X, Y, and Z coordinate axes , , and minimum value , , , according to the maximum and minimum values ​​on each coordinate axis, find the length of each side of the minimum volume surrounding the point cloud data , , ; Given a voxel grid size of a, the X, Y, and Z coordinate axes are divided into P, Q, and R parts, resulting in a total of A grid of individual pixels, where Indicates rounding downward; All voxel grids are numbered and recorded as , each data point Divide into voxel grids Calculate the centroid of each voxel grid, denoted as , replace the coordinate information of all points in the voxel with the coordinates of the center of gravity of the voxel grid. If the center of gravity does not exist, replace the coordinates of all points in the voxel with the coordinates of the point closest to the center of gravity in terms of distance within the voxel; Where n is the number of data points in the voxel grid, is the data point.

3. The method for monitoring the deck height of a berthed ship based on laser radar according to claim 1, characterized in that: In step S2, the algorithm for determining whether a point P is a boundary point is as follows: when the neighboring points near point P are all biased towards one side, point P is considered to be a boundary point of the point cloud; conversely, when the neighboring points near point P are evenly distributed around it, point P is an internal point.

4. The method for monitoring the deck height of a berthed ship based on laser radar according to claim 1, characterized in that: In the step S3, Ship 2D straight line feature extraction based on RANSAC algorithm: Randomly select two points from the ship point cloud , ,according to , Determine a straight line S based on the position of and determine the values ​​of the three parameters A, B, and C; Assume that the equation of the straight line S to be fitted is: A and B are the components of the direction vector of the line, and C is the intercept of the line; Count the number of points that fall on the straight line S determined by the A, B, and C parameters: Set the distance threshold to , calculate any point in the point cloud data Distance to line S ,when Distance to S Less than When, see is a point on S, where It is calculated by the following formula: statistics Less than The number of and record it, recorded as W; Repeat the above steps N times, where N is the number of iterations of the RANSAC algorithm; After the iteration is completed, the parameter straight line equation corresponding to the maximum W value is selected, and the straight line is the desired ship straight line equation.

5. The method for monitoring the deck height of a berthed ship based on laser radar according to claim 1, characterized in that: In step S4, the method of fitting the plane is: Randomly select three points from the ship point cloud , , according to , , The coordinates of the three points determine the values ​​of A, B, C, and D, thereby determining a plane L; Let the equation of plane L be: A is the normal vector component of the plane on the X axis, B is the normal vector component of the plane on the Y axis, C is the normal vector component of the plane on the Z axis, and D is the offset of the plane; Count the number of points that fall on the plane L determined by the A, B, C, and D parameters: Set the distance threshold to , calculate any point in the point cloud data Distance to plane L ,when Distance to L Less than When, see is a point on L, where It is calculated by the following formula: statistics Less than The number of and record it, recorded as Q; Repeat the above steps N times, where N is the number of iterations of the RANSAC algorithm; After the iteration is completed, the parametric plane equation corresponding to the maximum Q value is selected. The plane is the deck plane. The z-axis coordinates of all point clouds on the plane are averaged, and the average value obtained is the height H of the deck: 。 6. The method for monitoring the deck height of a berthed ship based on laser radar according to claim 1, characterized in that: The three-dimensional point cloud data acquisition module is a three-dimensional pan-tilt platform composed of a single-line laser radar and a servo motor, which realizes the collection function of ship point cloud data.

7. The method for monitoring the deck height of a berthed ship based on laser radar according to claim 1, characterized in that: The data processing subsystem is implemented by an industrial computer or an embedded device to realize point cloud data processing, data storage and data transmission functions.

Images