CN-121982195-A - Non-standard revolving body structural member modeling method based on laser scanning point cloud

CN121982195ACN 121982195 ACN121982195 ACN 121982195ACN-121982195-A

Abstract

The invention relates to the technical field of reverse modeling, and particularly discloses a non-standard revolving body structural member modeling method based on laser scanning point cloud, which comprises the following steps of S1, clamping a workpiece on a positioner; the method comprises the steps of clamping a 3D camera on a mechanical arm and calibrating the mechanical arm, S2 calibrating the camera and a position shifter at a photographing teaching position, S3 scanning a workpiece, collecting point cloud data of a corresponding surface area, S4 sequentially rotating the position shifter according to a set angle and surface continuity, sequentially collecting the point cloud data of the corresponding surface area, S5 aligning and registering two groups of point cloud data collected sequentially, S6 performing de-duplication optimization and fitting on the aligned and registered point cloud data, and S7 obtaining complete point cloud of the workpiece and completing modeling. The invention is designed aiming at the reverse modeling technology and the specificity of the nonstandard revolving body structural member, and can effectively meet the high-precision splicing modeling technology requirement of the nonstandard revolving body structural member.

Inventors

ZHUO LEI
CHEN YI
Liang Enpei
LIU ZIHAO

Assignees

东方电气集团数字科技有限公司

Dates

Publication Date: 20260505
Application Date: 20251224

Claims (10)

1. The modeling method for the nonstandard revolving body structural member based on the laser scanning point cloud is characterized by comprising the following steps of: S1, clamping a workpiece on a positioner; clamping the 3D camera on the mechanical arm, and performing hand-eye calibration treatment; S2, moving the mechanical arm to a set photographing teaching position; calibrating the rotating axis of the positioner and a coordinate system of the 3D camera; S3, scanning a workpiece clamped by the positioner at the current position, and collecting point cloud data of a corresponding surface area; s4, sequentially rotating the positioner according to the set rotation angle and the continuity of the surface of the workpiece; After the workpiece rotates in place at the present time, scanning the workpiece clamped by the positioner, and collecting point cloud data of a corresponding surface area; completing the acquisition of point cloud data of all the areas needing to be acquired on the surface of the workpiece according to the areas needing to be acquired on the surface of the workpiece; s5, aligning and registering two groups of adjacent point cloud data according to the sequence of the point cloud data acquired in the workpiece rotation sequence; s6, performing de-duplication optimization treatment on the aligned and registered point cloud data, and fitting the overlapping parts; and S7, sequentially completing alignment registration and duplicate removal fitting until all groups of point cloud data are aligned, obtaining the complete point cloud of the workpiece, and completing modeling.
2. The modeling method for the nonstandard revolving body structural member based on the laser scanning point cloud according to claim 1, wherein in the step S1, the hand-eye calibration processing adopts a mode that eyes are on hands so as to realize the fusion of a 3D camera coordinate system and a world coordinate system of the working operation of a mechanical arm.
3. The modeling method for the nonstandard revolving body structural member based on the laser scanning point cloud according to claim 2, wherein the hand-eye calibration processing satisfies the following relation: AX=XB; wherein A is a space transformation homogeneous matrix of the 3D camera in adjacent two motions; b is a transformation homogeneous matrix of the tail end coordinate system of the mechanical arm in two adjacent movements; X is a hand-eye matrix to be solved; wherein, the solving process of A satisfies the following relation: A1=H c2 *H c1 -1 ; A2=H c3 *H c2 -1 ; Wherein A1 is a space transformation homogeneous matrix of the 3D camera in the first movement; a2 is a space transformation homogeneous matrix of the 3D camera in the second motion; H c1 、H c2 、H c3 is a calibration picture external parameter matrix of the 3D camera in calibration; The solving process of B satisfies the following relation: B1=Hc2-1*Hc1; B2=Hc3-1*Hc2; B1 is a transformation homogeneous matrix of the tail end coordinate system of the mechanical arm in the first movement; b1 is a transformation homogeneous matrix of the tail end coordinate system of the mechanical arm in the second movement; hg1 is a coordinate system description matrix of the mechanical arm coordinate system corresponding to the calibration picture H c1 in the base coordinate system; Hg2 is a coordinate system description matrix of the mechanical arm coordinate system corresponding to the calibration picture H c2 in the base coordinate system; hg3 is a coordinate system description matrix of the mechanical arm coordinate system corresponding to the calibration picture H c3 in the base coordinate system.
4. The modeling method for the nonstandard revolving body structural member based on the laser scanning point cloud according to claim 1, wherein in the step S2, the shooting teaching position is right above a workpiece, and the 3D camera scans the workpiece in a mode of being vertical to the ground downwards.
5. The non-standard revolving body structural member modeling method based on the laser scanning point cloud as claimed in claim 1, wherein in the step S2, calibration of a rotating axis of a positioner and a 3D camera coordinate system is carried out by teaching calibration of the rotating axis of the positioner by taking the 3D camera coordinate system as a reference.
6. The modeling method for the nonstandard revolving body structural member based on the laser scanning point cloud, which is characterized by calibrating teaching of the rotation axis of the positioner, and meeting the following relation: = ; In the formula, A transformation matrix for the rotating axis of the positioner; In a coordinate system for the axis A normal vector of directions; In a coordinate system for the axis A normal vector of directions; In a coordinate system for the axis A normal vector of directions; K=1- ; M= - + 。
7. The modeling method for non-standard revolution solid structural members based on laser scanning point clouds according to claim 1, wherein in S4, the rotation angle of the positioner is 90 °.
8. The modeling method of the nonstandard revolving body structural member based on the laser scanning point cloud according to claim 1, wherein in S3 and S4, preprocessing of removing invalid points, setting ROI (region of interest) areas, stripping outliers and normal vector estimation is performed on collected point cloud data.
9. The modeling method of a nonstandard revolving body structural member based on laser scanning point cloud according to claim 1, wherein in S5, the alignment registration processing of two adjacent sets of point cloud data is to register the overlapping parts of the two sets of point cloud data, the overlapping parts are obtained by the distance between the two sets of point cloud data, and the points in the second set of point cloud data exist in the field range of the points in the first set of point cloud data, and are regarded as the overlapping parts; registering the two groups of point cloud data of the overlapping part by using an ICP algorithm, and obtaining a transformation matrix; And applying the transformation matrix to two adjacent complete sets of point cloud data, and enabling the two sets of point cloud data to be matched for the second time after rotation.
10. The modeling method of the nonstandard revolving body structural member based on the laser scanning point cloud according to claim 9, wherein the ICP algorithm performs registration to obtain a transformation matrix satisfying the following relation: ; In the formula, A transformation matrix which is the overlapping part of two adjacent groups of point cloud data; an ending sequence number of a point that is a synthesized partial point cloud data; a starting sequence number of a point which is a part of the point cloud data; is a rotation matrix; is a translation matrix; An ith point in the first set of point cloud data; is the i-th point in the second set of point cloud data.

Description

Non-standard revolving body structural member modeling method based on laser scanning point cloud Technical Field The invention relates to the technical field of reverse modeling, in particular to a non-standard revolving body structural member modeling method based on laser scanning point cloud. Background With the development of automation and intelligent technologies, the automatic operation of the mechanical arm can be realized by modeling and designing the workpiece and then by a track point automatic path planning algorithm, and the method has wide application in the technical fields of spraying, polishing, welding and the like. Modeling of a workpiece is divided into reverse modeling and forward modeling. The forward modeling is a process from drawing design to finished product production, the reverse modeling is reverse modeling, and the forward modeling is a process of reconstructing a mathematical model of an existing real object and re-designing and re-creating the existing product. Currently, reverse modeling techniques are relatively mature for fixed-attitude or standardized structural members, enabling higher-precision surface splice modeling. However, for nonstandard revolving body structural members, due to the technical characteristics of irregular shape, variable size and the like, a plurality of technical challenges are faced in the aspect of high-precision surface splicing of reverse modeling, and the conventional modeling method is often influenced by factors such as mechanical system errors and the like, so that point cloud data are lost or inaccurate, and further the precision of subsequent reverse modeling and automatic operation of a mechanical arm is influenced. Therefore, the development of the high-precision splicing modeling method capable of effectively aiming at the nonstandard revolving body structural member has important practical significance. Disclosure of Invention Aiming at the characteristics of the reverse modeling technology and the nonstandard revolving body structural member and the defects of the prior art, the invention provides a high-precision splicing modeling method which is based on laser scanning point cloud and can effectively cope with the nonstandard revolving body structural member. The technical aim of the invention is achieved by the following technical scheme, namely a non-standard revolving body structural member modeling method based on laser scanning point cloud, which comprises the following steps: S1, clamping a workpiece on a positioner; clamping the 3D camera on the mechanical arm, and performing hand-eye calibration treatment; S2, moving the mechanical arm to a set photographing teaching position; calibrating the rotating axis of the positioner and a coordinate system of the 3D camera; S3, scanning a workpiece clamped by the positioner at the current position, and collecting point cloud data of a corresponding surface area; s4, sequentially rotating the positioner according to the set rotation angle and the continuity of the surface of the workpiece; After the workpiece rotates in place at the present time, scanning the workpiece clamped by the positioner, and collecting point cloud data of a corresponding surface area; completing the acquisition of point cloud data of all the areas needing to be acquired on the surface of the workpiece according to the areas needing to be acquired on the surface of the workpiece; s5, aligning and registering two groups of adjacent point cloud data according to the sequence of the point cloud data acquired in the workpiece rotation sequence; s6, performing de-duplication optimization treatment on the aligned and registered point cloud data, and fitting the overlapping parts; and S7, sequentially completing alignment registration and duplicate removal fitting until all groups of point cloud data are aligned, obtaining the complete point cloud of the workpiece, and completing modeling. Further, in S1, the hand-eye calibration process uses a manner that the eyes are on the hands, so as to achieve fusion of the 3D camera coordinate system and the world coordinate system of the robot arm working and running. Further, the hand-eye calibration process satisfies the following relation: AX=XB; wherein A is a space transformation homogeneous matrix of the 3D camera in adjacent two motions; b is a transformation homogeneous matrix of the tail end coordinate system of the mechanical arm in two adjacent movements; X is a hand-eye matrix to be solved; wherein, the solving process of A satisfies the following relation: A1=Hc2*Hc1-1; A2=Hc3*Hc2-1; Wherein A1 is a space transformation homogeneous matrix of the 3D camera in the first movement; a2 is a space transformation homogeneous matrix of the 3D camera in the second motion; H c1、Hc2、Hc3 is a calibration picture external parameter matrix of the 3D camera in calibration; The solving process of B satisfies the following relation: B1=Hc2-1*Hc1; B2=Hc3-1*Hc2; B1 is a transform