A method for extracting plant leaf shape features based on point cloud

Abstract

This invention discloses a method for extracting plant leaf shape features based on point clouds, including the following steps: Step 1. Reading the three-dimensional point cloud data of the plant leaf; Step 2. Calculating the midpoint coordinates of the leaf point cloud extracted in Step 1; Step 3. Calculating the moment of inertia of the leaf point cloud; Step 4. Calculating the minimum and maximum moments of inertia of the leaf point cloud; Step 5. Calculating the longest height of the leaf; Step 6. Calculating the shortest width of the leaf; Step 7. Determining the bending direction of the leaf; Step 8. Calculating the perimeter of the leaf edge; Step 9. Calculating the surface area of ​​the leaf. Through the design of steps 1-9, this method can directly obtain the leaf shape feature effect during use, and has the characteristics of fast data processing speed and good feature extraction effect.

Description

Technical Field

[0001] This invention belongs to the field of information technology and relates to pattern recognition and image processing technology, specifically to a method for extracting plant leaf shape features based on point clouds. Background Technology

[0002] Plant identification is a method of determining the species and family of a plant based on its phenotypic characteristics. It is of great significance in botany, agriculture, and horticulture. Traditional plant identification is done manually, with botanists identifying plant species by leaf shape, leaf margins, leaf base, and veins, which is time-consuming, labor-intensive, and inefficient. With the development of technology, researchers in the field of plant identification have proposed an intelligent identification approach—leaf recognition technology. This technology can use computers to identify plant leaves, thereby achieving the purpose of plant leaf identification. However, the above method is still in the theoretical verification stage, and no feasible means have been designed to implement the above technology.

[0003] The main leaf recognition technologies currently available include the "Plant Leaf Geometric Modeling Method Based on Point Cloud Features" disclosed in the invention patent "201810420070.0". This method mainly extracts the geometric features of the leaves by using Principal Component Analysis (PCA) to extract the principal direction and normal vector of the leaves, thus solving the problem that "traditional point cloud-based plant leaf modeling methods mostly use general point cloud processing methods, without considering the characteristics of leaf morphology, resulting in low computational efficiency". However, for specific leaf shape extraction, this method cannot achieve the effect of directly and quickly extracting leaf shape features from specific leaves.

[0004] Therefore, there is an urgent need to design a new method for extracting plant leaf shape features to achieve rapid extraction of leaf shape features, thereby solving the problems existing in the above-mentioned technologies and facilitating plant identification and botanical, agricultural and horticultural research in practice. Summary of the Invention

[0005] To address the aforementioned problems, this invention aims to provide a method for extracting plant leaf shape features based on point clouds. Through the design of steps Step 1-Step 9, this method can directly obtain leaf shape feature effects during use, and has the characteristics of fast data processing speed and good feature extraction effect.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for extracting plant leaf shape features based on point clouds, including the following steps:

[0008] Step 1. Read the 3D point cloud data of the plant leaves;

[0009] Step 2. Calculate the coordinates of the midpoint of the leaf point cloud extracted in Step 1;

[0010] Step 3. Calculate the moment of inertia of the leaf point cloud;

[0011] Step 4. Calculate the minimum and maximum moments of inertia of the leaf point cloud;

[0012] Step 5. Calculate the maximum height of the blade;

[0013] Step 6. Calculate the shortest blade width;

[0014] Step 7. Determine the direction of the blade's bending;

[0015] Step 8. Calculate the edge perimeter of the blade;

[0016] Step 9. Calculate the surface area of ​​the blade.

[0017] Preferably, the process of reading the three-dimensional point cloud data of plant leaves described in Step 1 includes:

[0018] Step 101. First, convert the point cloud data from external storage into a TXT text file format that retains only the XYZ coordinate values ​​of the point cloud;

[0019] Step 102. Then store the point cloud data from the external TXT text file into memory in the form of a singly linked list;

[0020] The process of calculating the midpoint coordinates of the blade point cloud described in Step 2 includes: the midpoint coordinates of the point cloud data are the average value of the coordinates of each component of the point cloud, and then the point cloud data is translated according to the midpoint coordinates so that the midpoint is translated to the origin of the coordinate system.

[0021] Preferably, the calculation process of the rotational inertia of the blade point cloud described in Step 3 includes:

[0022] Step 301. For a point cloud, the moment of inertia I is equal to the mass m of the point mass multiplied by the perpendicular distance r between the point mass and the axis of rotation:

[0023] I = mr;

[0024] Step 302. For n point clouds, assuming the mass of each point is 1 unit, calculate the average moment of inertia of the leaf point cloud about the three coordinate axes:

[0025] I x =∑sqrt(y 2 +z 2 ) / n

[0026] I y =∑sqrt(x 2+z 2 ) / n

[0027] I z =∑sqrt(x 2 +y 2 ) / n.

[0028] Preferably, the calculation process of the minimum and maximum moment of inertia of the blade point cloud in Step 4 includes: based on the moment of inertia around the three coordinate axes, the coordinate axis corresponding to the minimum value is taken as the coordinate axis closest to the main axis direction of the blade, and the coordinate axis corresponding to the maximum value is taken as the coordinate axis that the blade needs to rotate first.

[0029] The calculation process for the maximum blade height described in Step 5 includes:

[0030] Step 501. First, rotate the leaf point cloud around the Z-axis. This will always bring the leaf to its minimum height. Record the rotation angle A around the Z-axis corresponding to the minimum height. z ;

[0031] Step 502. Then rotate the initial point cloud around the Z-axis by A. z If the angle is +90°, the principal axis of the point cloud is parallel to the Y-axis. At this point, the height difference of the point cloud is the maximum height H of the blade.

[0032] Preferably, the calculation process for the shortest blade width described in Step 6 includes:

[0033] Rotating the point cloud of the blade around the X-axis will always bring the blade to its minimum height. Record the rotation angle A around the X-axis corresponding to the minimum height. x Then rotate the point cloud around the X-axis again by A. x The width of the point cloud obtained is the shortest width W of the leaf.

[0034] Preferably, the process for determining the blade bending direction in step 7 includes...

[0035] Let the coordinates of the four points of the minimum bounding rectangle of the blade be (X... l Y l ), (X t Y t ), (X r Y r ), (X b Y b ):

[0036] If |X b -(X r -X l ) / 2|<|X t -(X r -X l) / 2|, the leaves are curved upwards;

[0037] If |X b -(X r -X l ) / 2|>=|X t -(X r -X l ) / 2|, the blade is bent downwards.

[0038] Preferably, the calculation process for the blade edge perimeter described in Step 8 includes:

[0039] Step 801. First, based on the blade's main axis and bending direction, determine the storage method for the cylindrical coordinate values ​​when converting the blade point cloud coordinates from rectangular coordinates to cylindrical coordinates;

[0040] Step 802. After determining the storage method, convert the blade point cloud coordinates from rectangular coordinates to cylindrical coordinates;

[0041] Step 803. After converting the leaf point cloud coordinates from rectangular coordinates to cylindrical coordinates, fill in any missing coordinate points;

[0042] Step 804. Calculate the edge length of the blade in cylindrical coordinates.

[0043] Preferably, the process of converting the blade point cloud coordinates from rectangular coordinates to cylindrical coordinates as described in Step 802 includes:

[0044] (1) Calculate the minimum value Ymin at the bottom of the blade, subtract Ymin from all Y components of the blade point cloud, and shift the bottom of the blade to the origin of the coordinate system.

[0045] (2) Based on the calculated blade bending direction, the point cloud is translated along the Z-axis in the opposite direction of the bending by half the width of the blade, so that the bending is directed towards the Y coordinate, which is consistent with the cylindrical direction of the central axis on the Y-axis.

[0046] (3) Based on the conversion formula from rectangular coordinates to cylindrical coordinates, convert the rectangular coordinates (x, y, z) to cylindrical coordinates. Then, deposit middle;

[0047] When the radius r corresponds to When the value is a many-to-one relationship, take the average of x, y, and z as the mean. The coordinates of the point pointed to by the member pp in the table.

[0048] Preferably, the process of filling in the missing coordinate points described in Step 803 includes:

[0049] (1) Calculate the maximum angle of the blade boundary in cylindrical coordinates. With minimum angle

[0050] (2) For a certain height, the cylindrical coordinates of the blade boundary are less than or greater than In this case, use the actual non-empty boundary angle point coordinates to fill in the gaps. or Boundary point;

[0051] (3) For the empty spaces in the middle, the average method is used to fill them, and the neighborhood size is 5×5:

[0052]

[0053]

[0054]

[0055] In the formula: m is the number of non-empty elements in R[h+u][b+v].

[0056] Preferably, the calculation process for the edge length of the blade described in step 804 includes:

[0057] (1) In cylindrical coordinates, the angle range for each height has been standardized, and the coordinates of the blade edge points are obtained. and

[0058]

[0059] Where H is the blade height;

[0060] (2) Smooth the edge points, then calculate the distance between each height according to the distance formula, and finally calculate the sum:

[0061]

[0062] This allows us to obtain the edge length of the blade;

[0063] Step 9 describes the calculation process for the blade surface area, which includes...

[0064] In cylindrical coordinates, adjacent heights and adjacent angles naturally form a quadrilateral, which can be approximated as two planar triangles for area calculation, and finally the sum is calculated:

[0065]

[0066] The beneficial effects of this invention are: This invention discloses a method for extracting plant leaf shape features based on point clouds. Compared with the prior art, the improvement of this invention lies in:

[0067] 1. This invention designs a method for extracting plant leaf shape features based on point clouds. The method includes the following steps: Step 1. Reading the three-dimensional point cloud data of the plant leaf; Step 2. Calculating the midpoint coordinates of the leaf point cloud extracted in Step 1; Step 3. Calculating the moment of inertia of the leaf point cloud; Step 4. Calculating the minimum and maximum moments of inertia of the leaf point cloud; Step 5. Calculating the longest height of the leaf; Step 6. Calculating the shortest width of the leaf; Step 7. Determining the bending direction of the leaf; Step 8. Calculating the perimeter of the leaf edge; Step 9. Calculating the surface area of ​​the leaf. In use, this method can directly obtain the leaf shape feature effect, and has the advantages of fast data processing speed and good feature extraction effect.

[0068] 2. This method uses the moment of inertia to calculate the initial rotational coordinate axis of the blade, which can quickly calculate the height and width of the blade;

[0069] 3. This method proposes a new method to determine the bending direction of a blade based on the position of the bottom or top x-value of the four points of the minimum bounding rectangle of the blade and the horizontal center of the bounding rectangle. This method can effectively and quickly determine the bending direction of the blade.

[0070] 4. This method converts the rectangular coordinates of the blade to cylindrical coordinates for calculation, which can quickly obtain the adjacency relationship of the point cloud, thus facilitating the calculation of the blade area;

[0071] 5. This method can quickly obtain the coordinates of the blade edge points based on cylindrical coordinates, which facilitates the calculation of the edge length and enables rapid extraction of blade shape features. Attached Figure Description

[0072] Figure 1 This is a three-view diagram of the point cloud of the blade of the present invention.

[0073] Figure 2 This is a diagram showing the minimum height position of the blade rotating around the Z-axis in this invention.

[0074] Figure 3 This is a diagram showing the process of the blade rotating around the Z-axis in this invention.

[0075] Figure 4 This is a diagram of the point cloud of the blade of the present invention rotated 90 degrees around the Z-axis.

[0076] Figure 5 This is a diagram showing the minimum height position of the blade point cloud rotating around the X-axis according to the present invention.

[0077] Figure 6 This is a diagram showing the process of the blade rotating around the X-axis in this invention.

[0078] Figure 7 This is a schematic diagram of the bending direction of the blade in this invention.

[0079] Figure 8 This is a diagram showing the conversion of the rectangular coordinates of the blades in this invention to cylindrical coordinates.

[0080] Figure 9 This is a top-view projection of the blade at different heights after the rectangular coordinates of the present invention have been converted to cylindrical coordinates.

[0081] Figure 10 This is a coordinate projection diagram of the blade rotor of the present invention.

[0082] Figure 11 This is an edge extraction image of the blade of the present invention.

[0083] Figure 12 This is a smoothed edge diagram of the blade of the present invention.

[0084] Figure 13 This is a flowchart illustrating the extraction process of plant leaf shape features based on point clouds according to the present invention. Detailed Implementation

[0085] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0086] Example 1: Refer to Appendix Figure 1-13 The method shown is a point cloud-based method for extracting plant leaf shape features, including the following steps:

[0087] Step 1. Read the 3D point cloud data of the plant leaves.

[0088] The point cloud data processed in this embodiment is in TXT text file format, which only retains the XYZ coordinate values ​​of the point cloud. If other file formats are used, they can be converted to text files using relevant software. Storing the point cloud data from the external TXT text file into memory can speed up the point cloud data processing. This embodiment uses a singly linked list to store the point cloud in memory, and the structure type (global type) of each node is as follows:

[0089] typedef struct PointXYZ

[0090] {float X, Y, Z;

[0091] struct PointXYZ*NextP;

[0092] Point List;

[0093] Step 2. Calculate the midpoint coordinates of the point cloud data extracted in Step 1.

[0094] The midpoint coordinates of the point cloud data are the average of the coordinates of each component of the point cloud. Then, the point cloud data is translated based on the midpoint coordinates to move the midpoint to the origin of the coordinate system, which facilitates the subsequent rotation transformation of the blades.

[0095] Step 3. Calculate the moment of inertia of the point cloud.

[0096] Step 301. Moment of inertia is a measure of the inertia of a rigid body when it rotates about an axis. For a blade point cloud (particle), the moment of inertia I is equal to the mass m of the particle multiplied by the perpendicular distance r between the particle and the axis of rotation.

[0097] I = mr

[0098] Step 302. For n point clouds, assuming the mass of each point is 1 unit, the average moment of inertia of the leaf point cloud about the three coordinate axes can be calculated:

[0099] I x =∑sqrt(y 2 +z 2 ) / n

[0100] I y =∑sqrt(x 2 +z 2 ) / n

[0101] I z =∑sqrt(x 2 +y 2 ) / n;

[0102] Step 4. Calculate the minimum and maximum moments of inertia of the leaf point cloud.

[0103] Based on the moment of inertia about the three coordinate axes, the coordinate axis corresponding to the minimum value is taken as the coordinate axis closest to the main axis direction of the blade (such as the Y axis), and the coordinate axis corresponding to the maximum value is taken as the coordinate axis that the blade needs to rotate first (such as the Z axis).

[0104] Step 5. Calculate the maximum height of the blade.

[0105] Step 501. Rotate the blade point cloud around the Z-axis; this will always bring the blade to its minimum height. Figure 2-3 Record the rotation angle A around the Z-axis corresponding to the minimum height. z ;

[0106] Step 502. Then rotate the initial point cloud around the Z-axis by A. z +90°, then the principal axis of the point cloud is parallel to the Y-axis, such as... Figure 4 At this point, the height difference of the point cloud is the maximum height H of the leaf.

[0107] Step 6. Calculate the shortest blade width

[0108] Rotating the point cloud of the blade around the X-axis will always bring the blade to its minimum height. Record the rotation angle A around the X-axis corresponding to the minimum height. x ,like Figure 5-6 Then rotate the point cloud around the X-axis again by A. x Then the width of the point cloud is the shortest width W of the leaf.

[0109] Step 7. Direction of leaf bending

[0110] The blades all have a certain curvature, which can be used to determine the direction of curvature: (1) When the blade bends upwards, the bottom x-value of the minimum bounding rectangle is near the horizontal center of the bounding rectangle; (2) When the blade bends downwards, the top x-value of the minimum bounding rectangle is near the horizontal center of the bounding rectangle, such as Figure 7 As shown;

[0111] Let the coordinates of the four points of the minimum bounding rectangle of the blade be (X... l ,Y l ), (X t 、,Y t ), (X r ,Y r ), (X b ,Y b ):

[0112] If |X b -(X r -X l ) / 2|<|X t -(X r -X l ) / 2|The blades are curved upwards;

[0113] If |X b -(X r -X l ) / 2|>=|X t -(X r -X l ) / 2|The blades are bent downwards;

[0114] Step 8. Perimeter of the leaf edge

[0115] Step 801. First, based on the blade's principal axis and bending direction, determine the storage method for the cylindrical coordinate values ​​when converting the blade point cloud coordinates from rectangular coordinates to cylindrical coordinates.

[0116] (1) The blade shape is similar to a small curved surface. Based on the blade's principal axis and bending direction, it can be represented in cylindrical coordinates. The conversion formula from rectangular coordinates to cylindrical coordinates is as follows:

[0117]

[0118] h = y

[0119]

[0120] Among them, the angle of the cylindrical coordinates angle The increment can generally be set to 1°;

[0121] When converting rectangular coordinates to cylindrical coordinates, an adjacency list is used to store the cylindrical coordinate values.

[0122] (2) Convert the rectangular coordinates (x, y, z) to cylindrical coordinates. Subsequently, there may also exist radial radius r and The value represents a one-to-many relationship, and this information needs to be saved for later processing.

[0123] (3) An adjacency list is used to store multiple points with different radii corresponding to the same height angle. The node type is:

[0124] typedef struct HL

[0125] {PointList*pp; / / Pointer to the coordinates of points in the point cloud}

[0126] short r; / / The perpendicular distance of the point from the Y-axis, which is also the radius of the cylinder.

[0127] struct HL*NextHL; / / Points to the next point at the same height and angle.

[0128] HLList;

[0129] Store each elevation angle point in a two-dimensional pointer array, initially empty:

[0130] HLList*R[H]

[360] ={NULL};

[0131] Where H is a macro defined based on the actual blade size;

[0132] Step 802. After determining the storage method, convert the leaf point cloud coordinates from rectangular coordinates to cylindrical coordinates.

[0133] (1) Calculate the minimum Y value Ymin at the bottom of the blade, subtract Ymin from all Y components of the blade point cloud, and shift the bottom of the blade to the origin of the coordinate system, such as... Figure 8 (a) and Figure 8 (b);

[0134] (2) Based on the calculated blade bending direction, such as Figure 9 The point cloud is translated along the Z-axis in the opposite direction of the curvature by half the width of the blade, as shown below. Figure 9 This causes the bend to point towards the Y coordinate, aligning with the direction of the cylinder whose central axis lies on the Y-axis.

[0135] (3) Based on the conversion formula from rectangular coordinates to cylindrical coordinates, convert the rectangular coordinates (x, y, z) to cylindrical coordinates. Then, deposit In the middle. When the radius r corresponds to When the value is a many-to-one relationship, take the average of x, y, and z as the mean. The coordinates of the point pointed to by member pp in the file;

[0136] Step 803. After converting the leaf point cloud coordinates from rectangular coordinates to cylindrical coordinates, fill in any missing coordinate points.

[0137] (1) Calculate the maximum angle of the blade boundary in cylindrical coordinates. With minimum angle

[0138] (2) For a certain height, the cylindrical coordinates of the blade boundary are less than or greater than In some cases (e.g., leaf tip and leaf base, such as...) Figure 9 (), using actual non-empty boundary angle point coordinates to fill in the blanks. or Boundary point;

[0139] (3) For the empty spaces in the middle, the average method is used to fill them, and the neighborhood size is 5×5:

[0140]

[0141]

[0142]

[0143] In the formula: m is the number of non-empty elements in R[h+u][b+v], such as Figure 10 As shown;

[0144] Step 804. Calculate the edge length of the blade.

[0145] In cylindrical coordinates, since the angle range for each height is standardized, the coordinates of the blade edge points can be obtained. and

[0146] (H is the blade height), smooth the edge points, then calculate the distance between each height according to the distance formula, and finally calculate the sum:

[0147]

[0148] This leads to the edge length of the blade, such as Figure 11-12 As shown;

[0149] Step 9. Calculate the surface area of ​​the blade.

[0150] In cylindrical coordinates, adjacent heights and adjacent angles naturally form a quadrilateral, which can be approximated as two planar triangles for area calculation, and finally the sum is calculated:

[0151]

[0152] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for extracting features of plant leaf shape based on point cloud, characterized in that: Comprising the steps of Step 1. reading three-dimensional point cloud data of a plant leaf; Step 2. calculating the midpoint coordinates of the leaf point cloud extracted in Step 1; Step 3. calculating the moment of inertia of the leaf point cloud; Step 4. calculating the minimum and maximum moments of inertia of the leaf point cloud; Step 5. calculating the longest height of the leaf; Step 501. First, rotate the blade point cloud around the Z axis, which will always make the blade have the minimum height, and record the rotation angle A around the Z axis corresponding to the minimum height z ; Step 502. Rotating the initial point cloud around the Z axis by A again z + 90 o If the main axis direction of the point cloud is parallel to the Y axis, the height difference of the point cloud is the longest height H of the blade. Step 6. calculating the shortest width of the leaf; Rotating the blade point cloud around the X axis, always makes the blade at the minimum height, record the minimum height corresponding to the rotation angle A around the X axis x , and then re-rotating the point cloud around the X axis A x , the width of the resulting point cloud is the shortest width W of the blade; Step 7. determining the bending direction of the leaf; Step 8. calculating the edge circumference of the leaf; Step 801. first, according to the main axis of the leaf and the bending direction, determine the storage mode of the cylindrical coordinate value of the leaf point cloud coordinate when the coordinate is converted from rectangular coordinate to cylindrical coordinate; Step 802. after determining the storage mode, convert the leaf point cloud coordinate from rectangular coordinate to cylindrical coordinate; Step 803. after converting the leaf point cloud coordinate from rectangular coordinate to cylindrical coordinate, supplement the missing points; Step 804. calculate the edge length of the leaf under cylindrical coordinate; (1) In cylindrical coordinates, the angle range of each height has been unified to obtain the coordinates of the blade edge points and ， Wherein, H is the height of the leaf; (2) smoothing the edge points, then calculating the distance between each height according to the distance formula, and finally calculating the total sum: ； Thus obtaining the edge length of the leaf; Step 9. calculating the curved surface area of the leaf; Under cylindrical coordinate, the adjacent height and adjacent angle naturally form a four-sided surface block, which can be approximately divided into two plane triangles to calculate the area, and finally the total sum is calculated: 。 2. The method according to claim 1, wherein the reading process of the three-dimensional point cloud data of the plant leaf in Step 1 comprises: Step 101. first, convert the point cloud data format in external storage to TXT text file format which only retains the XYZ coordinate values of the point cloud; Step 102. then store the point cloud data in the TXT text file in external storage in the form of a single linked list in the internal storage; The process of calculating the midpoint coordinates of the leaf point cloud in Step 2 comprises: the midpoint coordinates in the point cloud data are the average values of each component coordinate of the point cloud, and then the point cloud data is translated according to the midpoint coordinates to make the midpoint move to the coordinate origin. The process of calculating the moment of inertia of the leaf point cloud in Step 3 comprises:

3. The method of claim 1, wherein the method comprises: Step 301. for the point cloud, the moment of inertia I is equal to the mass m of the particle multiplied by the vertical distance r between the particle and the rotation axis: I = mr; Step 302. for n point clouds, assuming that the mass of each point is 1, the average moments of inertia of the leaf point cloud around the three coordinate axes are calculated:

4. The method according to claim 1, wherein the process of calculating the minimum and maximum moments of inertia of the leaf point cloud in Step 4 comprises: according to the moments of inertia around the three coordinate axes, the coordinate axis corresponding to the minimum value is taken as the coordinate axis closest to the main axis direction of the leaf, and the coordinate axis corresponding to the maximum value is taken as the coordinate axis that needs to be rotated first. I x =∑sqrt(y 2 +z 2 ) / n I y =∑sqrt(x 2 +z 2 ) / n I z =∑sqrt(x 2 +y 2 ) / n. The process of determining the bending direction of the leaf in Step 7 comprises The process of converting the leaf point cloud coordinate from rectangular coordinate to cylindrical coordinate in Step 802 comprises 5. The method of claim 1, wherein: ​ Let the four point coordinates of the minimum circumscribed rectangle of the blade be (X l , Y l ), (X t , Y t ), (X r , Y r ), (X b , Y b ): If |X b -(X r -X l ) / 2| < |X t -(X r -X l ) / 2|, the blade bends upward; If |X b -(X r -X l ) / 2| >= |X t -(X r -X l ) / 2|, the blade bends downward.

6. The method of claim 5, wherein: ​ (1) Calculate the minimum value Y of the blade bottom Ymin, and subtract Ymin from the Y component of the blade point cloud to move the blade bottom to the coordinate origin; (2) According to the calculated bending direction of the blade, the point cloud is translated along the Z axis towards the opposite direction of the bending by half the width of the blade, so that the bending direction is consistent with the cylindrical direction of the center axis on the Y axis; (3) According to the conversion formula from rectangular coordinates to cylindrical coordinates, rectangular coordinates (x, y, z) are converted into cylindrical coordinates (h, r) and then stored in R[h][ ] and R[r][ ]; When the value of radius r corresponds to (h, ) a many-to-one relationship, the average value of x, y, z is taken as the point coordinate to which the member pp in R[h][ ] points.

7. The method of claim 5, wherein the method further comprises: The process of filling the missing points in the coordinate points in step Step803 includes (1) Calculate the maximum angle of the blade boundary in cylindrical coordinates max and the minimum angle min; (2) For a certain height, if the cylindrical coordinate of the blade boundary is less than max or greater than min, the boundary angle point coordinates that are not actually empty are used to complete to max or min boundary points; (3) For the middle missing points, the average method is used for filling, and the neighborhood size is 5x5: ； ； ； In the formula: m is the number of R[h +u][b+v] that is not empty.

Images