阅读说明：本技术 一种基于标准球阵的零件特征机器人快速视觉定位方法 (Part feature robot rapid visual positioning method based on standard ball array ) 是由 李文龙 田亚明 王刚 陈颖茂 于 2020-05-13 设计创作，主要内容包括：本发明属于自动化测量相关技术领域,其公开了一种基于标准球阵的零件特征机器人快速视觉定位方法,所述零件特征机器人快速视觉定位方法基于在位置1处扫描标准球阵及零件计算得到的零件特征在标准球阵局部坐标系下的坐标和通过扫描位置2处的标准球阵创建的局部坐标系,可以快速定位零件特征在空间中的坐标,并指导机器人完成零件的安装和拆卸；其中,基于得到的变换矩阵将位置1处的零件特征在标准球阵点云局部坐标系下的坐标通过坐标变换计算得到位置2处的零件特征在机器人基坐标系下的坐标,由此完成快速视觉定位。本发明的适用性较强,且极大地提高了精度,适用范围广。(The invention belongs to the technical field related to automatic measurement, and discloses a part feature robot rapid visual positioning method based on a standard ball array, which is based on the coordinate of a part feature obtained by scanning the standard ball array at a position 1 and calculating the part in a standard ball array local coordinate system and the local coordinate system created by scanning the standard ball array at a position 2, can rapidly position the coordinate of the part feature in the space and guide a robot to complete the installation and the disassembly of the part; and calculating the coordinates of the part feature at the position 1 under the standard spherical array point cloud local coordinate system through coordinate transformation based on the obtained transformation matrix to obtain the coordinates of the part feature at the position 2 under the robot base coordinate system, thereby completing the quick visual positioning. The invention has strong applicability, greatly improves the precision and has wide application range.)

1. A part feature robot rapid visual positioning method based on a standard ball array is characterized by comprising the following steps:

(1) a robot positioned at a position 1 around a part drives a scanner to respectively scan and measure a standard ball original point cloud 1 at least containing three standard balls with different diameters and an original point cloud of the part to be positioned;

(2) performing point cloud segmentation on the standard sphere original point cloud 1 to obtain a standard sphere array point cloud;

(3) constructing a local coordinate system based on the standard spherical array point cloud obtained by segmentation and calculating to obtain a transformation matrix from the local coordinate system to a measurement coordinate system;

(4) performing point cloud segmentation on the original point cloud of the part to be positioned to obtain a part measurement point cloud;

(5) matching the part measurement point cloud serving as a test model with a part standard model serving as a reference model, and calculating to obtain a part characteristic pose;

(6) calculating the coordinates of the part features under a standard spherical array point cloud local coordinate system based on a coordinate transformation matrix group consisting of the obtained transformation matrices;

(7) the robot at the position 2 drives the scanner to scan and measure a standard ball original point cloud 2 at least containing three standard balls with different diameters;

(8) performing point cloud segmentation on the standard sphere original point cloud 2 to obtain a standard sphere array point cloud 2;

(9) constructing a local coordinate system based on the standard spherical point cloud 2 and calculating a transformation matrix from the local coordinate system to a corresponding measurement coordinate system;

(10) and (3) calculating the coordinates of the part feature at the position 1 under the standard spherical array point cloud local coordinate system through coordinate transformation based on the obtained transformation matrix to obtain the coordinates of the part feature at the position 2 under the robot base coordinate system, thereby completing the rapid visual positioning.

2. The method for quickly visually positioning the part feature robot based on the standard ball array as claimed in claim 1, wherein: the standard ball array model of the standard ball original point cloud 1 and the standard ball original point cloud 2 at least containing three standard balls with different diameters is the same solid model, and the relative position of the solid model and the solid model of the part original point cloud to be positioned in the space is kept unchanged.

3. The method for quickly visually positioning the part feature robot based on the standard ball array as claimed in claim 1, wherein: and performing point cloud segmentation on the standard ball original point cloud and the part original point cloud by adopting random sampling consistency and an Euclidean clustering mode.

4. The method for quickly visually positioning the part feature robot based on the standard ball array as claimed in claim 1, wherein: in the step (3) and the step (9), the diameter and the spherical center coordinate of the standard sphere in the segmented point cloud of the standard spherical array are calculated through least square fitting, and the standard ball arrays are arranged in ascending or descending order according to the diameter, the three standard balls which are arranged most at the front are respectively numbered as 1, 2 and 3, taking the center of the No. 1 standard ball as an origin, respectively calculating the distance vectors from the origin to the centers of the No. 2 standard ball and the No. 3 standard ball, taking the distance vector between the centers of the No. 1 standard ball and the No. 2 standard ball as an X axis, taking the distance vector between the centers of the No. 1 standard ball and the No. 3 standard ball as a Y axis, calculating the Z axis according to the right-hand rule, recalculating the Y axis based on the X axis and the Z axis according to the right-hand rule, processing the X, Y, Z axis vector by adopting standardization and Schmidt orthogonalization, and further creating a local coordinate system, and calculating a transformation matrix from the local coordinate system to the measurement coordinate system by combining the coordinates of the sphere center of the No. 1 standard sphere in the measurement coordinate system.

5. The method for quickly visually positioning the part feature robot based on the standard ball array as claimed in claim 1, wherein: and (5) matching by adopting an ADF algorithm, and calculating the characteristic pose of the part by solving an inverse transformation matrix based on the pose information in the created part standard model.

6. The method for quickly visually positioning the part feature robot based on the standard ball array as claimed in claim 1, wherein: and (6) and (10) respectively calculating the coordinates of the part features under the standard spherical array local coordinate system and the robot base coordinate system under the condition that the relative positions of the standard spherical array local coordinate system at the position 1 and the position 2 and the part features in the space are unchanged.

7. The robot rapid visual positioning method for part features based on standard ball arrays as claimed in any one of claims 1 to 6, characterized in that: the robot is a six-degree-of-freedom industrial robot, and the scanner is a grating type binocular area array scanner.

8. The robot rapid visual positioning method for part features based on standard ball arrays as claimed in any one of claims 1 to 6, characterized in that: keeping the relative positions of the standard spherical array solid model and the part characteristic solid model in the space unchanged, and moving the standard spherical array solid model and the part characteristic solid model from the position 1 to the position 2; secondly, controlling the robot at the position 2 to drive the scanner to scan and measure the standard spherical array in multiple postures, recording the pose of the robot corresponding to each position, and calculating the hand-eye relationship based on the AX (X) -XB (X-ray) theoretical model, namely a transformation matrix from a measurement coordinate system to a robot tail end coordinate system and recording the transformation matrix as the transformation matrix

9. The robot rapid visual positioning method for part features based on standard ball arrays as claimed in any one of claims 1 to 6, characterized in that: in step (6), firstly, a transformation matrix from the terminal coordinate system to the base coordinate system when the robot 1 at the position 1 measures the standard spherical array is calculated based on the Pose Pose1 of the robot 1, and the transformation matrix is recorded as

Secondly, calculating a transformation matrix from the terminal coordinate system to the base coordinate system when the robot 1 at the position 1 measures the characteristics of the part based on the Pose Pose2 of the robot 1, and recording the transformation matrix as

Finally, based on the coordinate transformation matrix group at the position 1, calculating the coordinate of the part feature under the local coordinate system of the standard spherical array point cloud intoWherein the hand-eye matrixTherefore, it is

Technical Field

The invention belongs to the technical field related to automatic measurement, and particularly relates to a part feature robot rapid visual positioning method based on a standard ball array.

Background

The determination of the feature pose of the part is an important premise for positioning, mounting and dismounting the part, and in the field of intelligent manufacturing and equipment using a robot as a main carrier, a sensor and operation software are integrated, the visual positioning replaces the traditional manual naked eye identification positioning, and the pose of the part feature in the space is obtained through analysis and calculation, so that the determination of the feature pose of the part becomes one of the research directions in the field of intelligent manufacturing and automatic measurement at present.

The service conditions of the parts are often restricted by the surrounding environment, but when the parts are in service in extremely severe environments such as high temperature, high dust and the like, the mounting and dismounting of the parts become very difficult. The traditional part mounting and dismounting adopts an artificial naked eye identification positioning method, the method depends on manual operation, the efficiency is low, the randomness is high, the subjective consciousness is strong, the positioning is inaccurate, the mounting and dismounting of the part are unreliable, and safety accidents are easy to happen. Accordingly, there is a technical need in the art to develop a fast visual positioning method for a part feature robot based on a standard ball array, which has better positioning accuracy.

Disclosure of Invention

Aiming at the defects or improvement requirements of the prior art, the invention provides a part feature robot rapid visual positioning method based on a standard ball array. The non-contact measurement is that a six-degree-of-freedom industrial robot drives a grating type binocular area array scanner to complete automatic measurement, standard balls in a standard ball array are all standard matte ceramic balls, coordinates of part characteristics under a standard ball array local coordinate system and a local coordinate system created by scanning the standard ball array at a position 2 are obtained through calculation based on scanning of the standard ball array at the position 1 and parts, the coordinates of the part characteristics in the space can be rapidly located, and the robot is guided to complete installation and disassembly of the parts. Meanwhile, the method utilizes a six-degree-of-freedom industrial robot to drive a grating type binocular area array scanner to scan and measure the standard spherical array and the part under the environment-friendly condition to determine the space coordinate of the part characteristic in the local coordinate system of the standard spherical array, and the position and the attitude of the part characteristic in the space can be rapidly positioned at other positions in the severe environment only by scanning the standard spherical array and establishing the local coordinate system in the same way.

In order to achieve the aim, the invention provides a part feature robot rapid visual positioning method based on a standard ball array, which comprises the following steps:

(1) a robot positioned at a position 1 around a part drives a scanner to respectively scan and measure a standard ball original point cloud 1 at least containing three standard balls with different diameters and an original point cloud of the part to be positioned;

(2) performing point cloud segmentation on the standard sphere original point cloud 1 to obtain a standard sphere array point cloud;

(3) constructing a local coordinate system based on the standard spherical array point cloud obtained by segmentation and calculating to obtain a transformation matrix from the local coordinate system to a measurement coordinate system;

(4) performing point cloud segmentation on the original point cloud of the part to be positioned to obtain a part measurement point cloud;

(5) matching the part measurement point cloud serving as a test model with a part standard model serving as a reference model, and calculating to obtain a part characteristic pose;

(6) calculating the coordinates of the part features under a standard spherical array point cloud local coordinate system based on a coordinate transformation matrix group consisting of the obtained transformation matrices;

(7) the robot at the position 2 drives the scanner to scan and measure a standard ball original point cloud 2 at least containing three standard balls with different diameters;

(8) performing point cloud segmentation on the standard sphere original point cloud 2 to obtain a standard sphere array point cloud 2;

(9) constructing a local coordinate system based on the standard spherical point cloud 2 and calculating a transformation matrix from the local coordinate system to a corresponding measurement coordinate system;

(10) and (3) calculating the coordinates of the part feature at the position 1 under the standard spherical array point cloud local coordinate system through coordinate transformation based on the obtained transformation matrix to obtain the coordinates of the part feature at the position 2 under the robot base coordinate system, thereby completing the rapid visual positioning.

Furthermore, the standard ball array model of the standard ball original point cloud 1 and the standard ball original point cloud 2 at least containing three standard balls with different diameters is the same solid model, and the relative position of the solid model and the solid model of the part original point cloud to be positioned in the space is kept unchanged.

Further, point cloud segmentation is performed on the standard ball original point cloud and the part original point cloud in a random sampling consistency and Euclidean clustering mode.

Further, in the step (3) and the step (9), the diameter and the spherical center coordinates of the standard sphere in the segmented point cloud of the standard spherical array are calculated through least square fitting, and the standard ball arrays are arranged in ascending or descending order according to the diameter, the three standard balls which are arranged most at the front are respectively numbered as 1, 2 and 3, taking the center of the No. 1 standard ball as an origin, respectively calculating the distance vectors from the origin to the centers of the No. 2 standard ball and the No. 3 standard ball, taking the distance vector between the centers of the No. 1 standard ball and the No. 2 standard ball as an X axis, taking the distance vector between the centers of the No. 1 standard ball and the No. 3 standard ball as a Y axis, calculating the Z axis according to the right-hand rule, recalculating the Y axis based on the X axis and the Z axis according to the right-hand rule, processing the X, Y, Z axis vector by adopting standardization and Schmidt orthogonalization, and further creating a local coordinate system, and calculating a transformation matrix from the local coordinate system to the measurement coordinate system by combining the coordinates of the sphere center of the No. 1 standard sphere in the measurement coordinate system.

And further, matching by adopting an ADF algorithm in the step (5), and calculating the characteristic pose of the part by solving an inverse transformation matrix based on the pose information in the created part standard model.

Further, in the step (6) and the step (10), under the condition that the relative positions of the standard spherical array local coordinate system at the position 1 and the position 2 and the part feature in the space are kept unchanged, the coordinates of the part feature under the standard spherical array local coordinate system and the robot-based coordinate system are respectively calculated.

Further, the robot is a six-degree-of-freedom industrial robot, and the scanner is a grating type binocular area array scanner.

Further, the relative positions of the standard ball array solid model and the part feature solid model in the space are kept unchanged, and the standard ball array solid model and the part feature solid model are moved from the position 1 to the position 2; secondly, controlling the robot at the position 2 to drive the scanner to scan and measure the standard spherical array in multiple postures, recording the pose of the robot corresponding to each position, and calculating the hand-eye relationship based on the AX (X) -XB (X-ray) theoretical model, namely a transformation matrix from a measurement coordinate system to a robot tail end coordinate system and recording the transformation matrix as the transformation matrix

Further, in step (6), first, a transformation matrix from the end coordinate system to the base coordinate system when the robot 1 at the position 1 measures the standard spherical matrix is calculated based on the Pose1 of the robot 1, and is recorded as

Secondly, calculating a transformation matrix from the terminal coordinate system to the base coordinate system when the robot 1 at the position 1 measures the characteristics of the part based on the Pose Pose2 of the robot 1, and recording the transformation matrix as

Finally, based on the coordinate transformation matrix group at the position 1, calculating the coordinate of the part feature under the local coordinate system of the standard spherical array point cloud intoWherein the hand-eye matrixTherefore, it is

Generally, compared with the prior art, the technical scheme of the invention has the following advantages that:

1. the invention utilizes the six-degree-of-freedom industrial robot to drive the grating type binocular area array scanner to scan the standard ball array and the parts in an environment friendly way so as to determine the position relation between the parts and the standard ball array, and then only the standard ball is scanned at the severe part service environment, so that the rapid visual positioning of the part characteristics can be realized.

2. The invention can realize point cloud segmentation and standard spherical array coordinate system creation only by carrying out single scanning measurement on the standard spherical array, does not need point cloud splicing, and has easy implementation and high measurement efficiency.

3. The invention realizes the part characteristic positioning based on the standard ball array comprising three standard balls with different diameters, and the method for establishing the local coordinate system is simple and flexible and has accurate positioning.

4. The invention adopts ADF algorithm to match point cloud, and calculates the pose of the part feature under the measurement coordinate system based on the matching transformation matrix, and has high calculation efficiency and wide application range.

Drawings

FIG. 1 is a schematic flow chart of a method for robot fast visual positioning of part features based on a standard ball array according to a preferred embodiment of the present invention;

FIG. 2 is a schematic diagram of a system for performing fast feature location of a part based on a standard ball grid array according to an embodiment of the present invention; wherein { B } is a robot base coordinate system, { E } is a robot tip coordinate system, { T } is a measurement coordinate system, { W }₁Is a standard spherical array local coordinate system, W₂The method comprises the following steps of (1) taking a local coordinate system of a part;

FIG. 3 is a schematic diagram of a three-dimensional feature of a standard ball array including three standard balls provided by the present invention;

FIG. 4 is a schematic diagram of three-dimensional features of a part to be positioned according to the present invention;

FIG. 5 is a schematic diagram of a system position 1 for performing fast part feature location based on a standard ball grid array according to an embodiment of the present invention;

fig. 6 is a schematic diagram of a system position 2 for performing fast part feature location based on a standard ball array according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

Referring to fig. 1, fig. 2, fig. 3, fig. 4, fig. 5 and fig. 6, the method for quickly positioning part features by a robot based on a standard spherical array according to the present invention is suitable for quickly positioning most part features, and uses a three-dimensional robot measurement technique to quickly position the part features based on an environmentally friendly standard spherical array; the pose of the part features under the local coordinate system of the standard ball array is analyzed and calculated through the standard ball array at a certain position and the scanning and measuring result of the part features, and based on the premise that the pose is not changed, the rapid positioning of the part features can be realized only by scanning and measuring the standard ball array at other positions. The method mainly comprises the following steps:

firstly, a robot located at a position 1 around a part drives a scanner to scan and measure a standard ball original point cloud 1 at least containing three standard balls with different diameters and an original point cloud of the part to be positioned respectively.

Specifically, the robot 1 is controlled at the position 1 to drive the scanner 1 to scan and measure a standard spherical array in multiple postures, the pose of the robot corresponding to each position is recorded, and the hand-eye relationship is calculated based on an AX (X-XB) theoretical model, namely a transformation matrix from a measurement coordinate system to a robot tail end coordinate system is recorded asIn this embodiment, location 1 is located in a part-friendly working environment.

Secondly, controlling the robot 1 at the position 1 to drive the scanner 1 to scan and measure the standard spherical Array in a single frame, recording the current Pose of the robot 1 as Pose1, and recording the measurement point cloud as Array 1;

and finally, controlling the robot 1 at the position 1 to drive the scanner 1 to scan and measure the characteristics of the part in a single way, recording the current Pose of the robot 1 as Pose2, and recording the measurement point cloud as originPart.

And step two, performing point cloud segmentation on the standard ball original point cloud 1 to obtain a standard ball array point cloud.

Specifically, the measurement point cloud Array1 is segmented by a standard ball point cloud segmentation module to obtain a standard ball Array point cloud containing three standard balls, and the standard ball Array point cloud is marked as Spheres 1.

And step three, constructing a local coordinate system based on the standard spherical array point cloud obtained by segmentation and calculating to obtain a transformation matrix from the local coordinate system to the measurement coordinate system.

Specifically, firstly, the diameters and the spherical center coordinates of three standard Spheres in the standard spherical array point cloud Spheres1 are respectively calculated by adopting least square fitting;

secondly, arranging the three standard balls in a descending order according to the diameter, and numbering the standard balls as 1, 2 and 3 respectively;

finally, taking the center of the No. 1 standard ball as an origin, respectively calculating the distance vectors from the origin to the centers of the No. 2 standard ball and the No. 3 standard ball, taking the distance vector from the center of the No. 1 standard ball to the center of the No. 2 standard ball as an X axis, taking the distance vector from the center of the No. 1 standard ball to the center of the No. 3 standard ball as a Y axis, calculating the Z axis according to right-hand rules, and recalculating the Y axis based on the X axis and the Z axis so as to prevent the X axis and the Y axis from being collinear according to the right-hand rules; finally, processing the X, Y, Z axial vector by adopting standardization and Schmidt orthogonalization to further create a local coordinate system, and calculating a transformation matrix from the local coordinate system to the measurement coordinate system by combining the coordinates of the sphere center of the No. 1 standard sphere under the measurement coordinate system, and recording the transformation matrix as

And step four, performing point cloud segmentation on the original point cloud of the part to be positioned to obtain a part measurement point cloud.

Specifically, a part feature point cloud segmentation module is used for segmenting an original measurement point cloud OriginPart of a part to be positioned to obtain a part feature point cloud which is marked as FeaturePart.

And fifthly, matching the part measurement point cloud serving as a test model with a part standard model serving as a reference model, and calculating to obtain the part characteristic pose.

Specifically, the part measurement point cloud FeaturePart is used as a test model, a part standard stl model is used as a reference model, ADF matching is carried out, and the coordinates of the part characteristics under a measurement coordinate system are obtained through analysis and calculation and are recorded as

And sixthly, calculating the coordinates of the part features under the local coordinate system of the standard spherical array point cloud based on a coordinate transformation matrix group consisting of the obtained transformation matrices.

Specifically, as shown in fig. 5, first, a transformation matrix from the end coordinate system to the base coordinate system when the robot 1 at the position 1 measures the standard spherical array is calculated based on the Pose1 of the robot 1, and is denoted as

Secondly, calculating a transformation matrix from the terminal coordinate system to the base coordinate system when the robot 1 at the position 1 measures the characteristics of the part based on the Pose Pose2 of the robot 1, and recording the transformation matrix as

Finally, based on the coordinate transformation matrix group at the position 1, calculating the coordinate of the part feature under the local coordinate system of the standard spherical array point cloud intoWherein the hand-eye matrixTherefore, it is

And step seven, the robot at the position 2 drives the scanner to scan and measure the standard ball original point cloud 2 at least containing three standard balls with different diameters.

Specifically, firstly, the relative positions of the standard ball array solid model and the part feature solid model in the space are kept unchanged, and the standard ball array solid model and the part feature solid model are moved from a position 1 to a position 2; in this embodiment, location 2 is located in the harsh environment in which the part is in service.

Secondly, controlling the robot 2 at the position 2 to drive the scanner 2 to scan and measure the standard spherical array in multiple postures, recording the pose of the robot corresponding to each position, and calculating the hand-eye relationship based on the AX (X-XB) theoretical model, namely a transformation matrix from a measurement coordinate system to a robot tail end coordinate system is recorded as

And finally, controlling the robot 2 at the position 2 to drive the scanner 2 to scan and measure the standard spherical Array in a single frame, recording the current Pose of the robot 2 as Pose, and recording the measured point cloud as Array.

And step eight, performing point cloud segmentation on the standard ball original point cloud 2 to obtain a standard ball array point cloud 2.

Specifically, the measurement point cloud Array is segmented by a standard ball point cloud segmentation module to obtain a standard ball Array point cloud containing three standard balls, and the standard ball Array point cloud is marked as Spheres 2.

And step nine, constructing a local coordinate system based on the standard ball point cloud 2 and calculating a transformation matrix from the local coordinate system to a corresponding measurement coordinate system.

Specifically, the same operation as the third step is adopted for the standard spherical array point clouds Spheres2, and finally, a transformation matrix from the standard spherical array local coordinate system to the measurement coordinate system is obtained through analysis and calculation and is marked as

And step ten, calculating the coordinates of the part features at the position 1 under the local coordinate system of the standard spherical array point cloud based on the obtained transformation matrix through coordinate transformation to obtain the coordinates of the part features at the position 2 under the robot base coordinate system, thereby completing the rapid visual positioning.

In particular, as shown in figure 6,firstly, calculating a transformation matrix from the terminal coordinate system to the base coordinate system when the robot 2 at the position 2 measures the standard spherical array based on the Pose Pose of the robot 2, and recording the transformation matrix as

Secondly, note the coordinates of the part feature at position 2 in the standard spherical array local coordinate system as

Finally, based on the coordinate transformation matrix group at the position 2, calculating the coordinates of the part features under the robot 2-based coordinate system intoWherein the hand-eye matrixSince the relative positions of the standard spherical array solid model and the part feature solid model in the space are kept unchanged, namely the coordinates of the part features at the position 1 and the position 2 are the same under the local coordinate system of the standard spherical array, the coordinates are the sameThe part feature positioning is completed.

It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.