CN113019763A

CN113019763A - Spraying robot track planning method based on grid projection algorithm

Info

Publication number: CN113019763A
Application number: CN202011454927.4A
Authority: CN
Inventors: 禹鑫燚; 程兆赢; 张毅凯; 郑坤; 欧林林
Original assignee: Zhejiang University of Technology ZJUT
Current assignee: Zhejiang University of Technology ZJUT
Priority date: 2020-12-10
Filing date: 2020-12-10
Publication date: 2021-06-25
Anticipated expiration: 2040-12-10
Also published as: CN113019763B

Abstract

A spraying robot track planning method based on a grid projection algorithm comprises the following steps: scanning by a laser sensor to obtain a point cloud model of a sprayed workpiece; after the point cloud slicing direction is determined, cutting the point cloud model of the spraying workpiece by using a series of parallel and equidistant cutting planes; providing a grid projection algorithm to solve the section contour point of each cutting plane on the workpiece point cloud model; translating the section contour point along the direction of the normal vector of the section contour point by a distance of the spraying height of a spray gun to obtain a spraying track point; and connecting the spraying track points in sequence in a straight line form to obtain a complete spraying track. The invention can plan the spraying track of workpieces in any shapes, and the proposed grid projection algorithm can quickly obtain the section contour points, thereby greatly improving the efficiency of the spraying track planning.

Description

Spraying robot track planning method based on grid projection algorithm

Technical Field

The invention relates to the field of automatic spraying, in particular to planning of tracks of spraying robots for the surfaces of various workpieces in the spraying industry.

Background

More and more industries gradually change to automation, and robots, which are products of the combination of information industry and mechanical industry, have the characteristics of full automation and high adaptability, and are being widely applied to the fields of automobiles, furniture, plastics, electronic products and the like. In the painting field, the extensive use of painting robots frees workers from toxic and harmful environments and greatly improves productivity. The trajectory planning technique of the painting robot is being widely studied as an important factor affecting the quality of painted products.

At present, the track planning technology of a spraying robot is not mature, and the problems that the efficiency is not high, the spraying robot is not suitable for spraying workpieces with free-form surfaces, the paint waste is easy to cause, the uniformity of a coating film on the surface of the workpiece is poor and the like exist. In the paper, Chen et al, used a piece-wise modeling of CAD model of a painted workpiece, and used a bounding box algorithm to plan the painting track on each piece of panel, and finally connected the painting tracks on each piece of panel. The slicing method cannot be applied to free-curved surface type spraying workpieces and the whole spraying process is too complex. In a path planning method, a device and a system of a spray robot and a storage medium [ P ], CN 110181516A (Chen nationwide, Yuxing Yu, Wang administrative), Chen nationwide and the like propose a method for segmenting a three-dimensional point cloud model of a spray workpiece by using a series of tangent planes, determining each spray data point of each tangent plane and obtaining a corresponding spray track according to the constraint points.

Disclosure of Invention

The invention aims to overcome the defects in the prior art and provides a spraying robot path planning method based on a grid projection algorithm, which is suitable for spraying workpieces in various shapes and has the characteristic of high efficiency.

The invention is realized by the following technical scheme: firstly, scanning by using a laser sensor to obtain a point cloud model of a sprayed workpiece; then determining the direction of the point cloud slice and cutting the point cloud model of the workpiece along the direction of the point cloud slice by using a series of parallel and equidistant cutting planes; then, solving the section contour points of the workpiece point cloud model on the tangent plane by using the proposed grid projection algorithm; then, translating the obtained section contour point along the direction of the normal vector of the section contour point by a distance of the spraying height of a spray gun to obtain a spraying track point; and finally, connecting the spraying track points point by point in a straight line mode to obtain a complete spraying path.

A spraying robot track planning method based on a grid projection algorithm comprises the following specific steps:

step 1: and scanning by a laser sensor to obtain a point cloud model of the sprayed workpiece.

Step 2: and determining the direction of the point cloud slice. The workpiece point cloud model is a pile of three-dimensional data points, the mass center and the characteristic vector of the workpiece point cloud model are obtained through calculation, the directions of the three characteristic vectors of the workpiece point cloud model are used as XYZ axes of a coordinate system of the workpiece point cloud model, the mass center of the workpiece point cloud model is used as the origin of the coordinate system of the workpiece point cloud model, and the coordinate system is called as a characteristic vector coordinate system. And selecting a direction vertical to the dominant feature vector of the workpiece point cloud model as the direction of the point cloud slice.

And step 3: the contour points of the workpiece point cloud model on the tangent plane are obtained by a grid projection algorithm, and the steps are as follows:

step 3-1: intercepting a point cloud band between adjacent tangent planes as a research object under a feature vector coordinate system;

step 3-2: assuming that the direction perpendicular to the Z axis is the direction of the point cloud slice, the entire mesh can just cover the point cloud band by projecting N × N meshes along the direction of the point cloud slice, and as shown in fig. 2, the intersection points of the meshes on the tangent plane are the mesh points. The grid points are section contour points;

step 3-3: under the previous assumption, the Y-coordinate and Z-coordinate values of the grid points are determined. However, given the complexity of the point cloud data distribution in the workpiece point cloud model, the same point cloud data as its Y and Z coordinate values may not exist for each grid point. With each grid point as the center, a square with a side length λ is created, as shown in fig. 3, where λ is a very small value;

step 3-4: and for each square, selecting k points with the maximum X coordinate value in the projection range of the point cloud slice direction, and taking the average value of the X coordinate values of the k points as the X coordinate value of the grid point. If the projection range of the square has no point cloud data, the value of lambda is continuously expanded until the projection range of the square of each grid point has point cloud data. Thus, the three-dimensional coordinates of all the grid points, namely the three-dimensional coordinates of all the section contour points, can be obtained.

And 4, step 4: and translating the section contour point along the direction of the normal vector of the section contour point by the distance of the spraying height of the spray gun to obtain a spraying track point. And finally, connecting the spraying track points point by point in a straight line form to form a complete spraying track.

The invention has the advantages that: the spraying track planning method can plan the spraying track of a workpiece in any shape, and the provided grid projection algorithm can quickly obtain the section contour points, so that the spraying track planning efficiency is greatly improved.

Drawings

FIG. 1 is a flow chart of the method of the present invention.

Fig. 2 is a schematic diagram of a grid projection algorithm.

Fig. 3 is a square corresponding to each grid point.

Fig. 4 is a spray coated workpiece.

FIG. 5 is a point cloud model of a painted workpiece.

Fig. 6a to 6b are schematic diagrams of coordinate system transformation of the workpiece point cloud model, wherein fig. 6a is an original coordinate system of the workpiece point cloud model, and fig. 6b is a feature vector coordinate system of the workpiece point cloud model.

Fig. 7 is a point cloud slice object map.

Fig. 8a to 8b are sectional contour points on a tangent plane, where fig. 8a is a thumbnail of the sectional contour point on the tangent plane and fig. 8b is an enlarged view of the sectional contour point on the tangent plane.

Fig. 9 is a schematic illustration of the painted trace points.

Fig. 10 is a schematic view of a spray trajectory.

Detailed description of the invention

In order to clearly understand the objects, features and advantages of the present invention, the following detailed description of the embodiments of the present invention will be made with reference to the accompanying drawings.

The invention relates to a spraying robot path planning method based on a grid projection algorithm, which comprises the following specific processes:

step 1: the sprayed workpiece is a free-form surface type workpiece as shown in fig. 4. A point cloud model of the sprayed workpiece is obtained by scanning with a laser sensor, as shown in fig. 5.

Step 2: the orientation of the point cloud slice is determined. The original coordinate system of the workpiece point cloud model is shown in fig. 6a), and the axes of red, green and blue represent the XYZ axes of the coordinate system, respectively. Since the workpiece point cloud model is a stack of three-dimensional data points, the centroid and the feature vector of the workpiece point cloud model are obtained through calculation, the three feature vector directions of the workpiece point cloud model are taken as the XYZ axes of the workpiece point cloud model coordinate system, the centroid of the workpiece point cloud model is taken as the origin of the workpiece point cloud model coordinate system, and the coordinate system is called as a feature vector coordinate system, as shown in fig. 6 b). Under a feature vector coordinate system, the lengths of the sprayed workpiece in XYZ axes are respectively 106.1mm, 248.7mm and 691.1mm, wherein the maximum value in the X axis direction is 24.3mm, the minimum value in the X axis direction is-81.8 mm, the maximum value in the Y axis direction is 127.0mm, the minimum value in the Y axis direction is-121.7 mm, the maximum value in the Z axis direction is 305.8mm, and the minimum value in the Z axis direction is-385.3 mm. And transforming the workpiece point cloud model from the original coordinate system to the characteristic vector coordinate system, and selecting the direction vertical to the dominant characteristic vector of the workpiece point cloud model as the direction of point cloud slicing, namely the direction vertical to the Z axis in the characteristic vector coordinate system is the direction of point cloud slicing. The following steps are performed in the feature vector coordinate system.

And step 3: the workpiece point cloud model was cut along the direction of the point cloud slices with a series of parallel and equidistant cutting planes, as shown in fig. 7, with a distance of 120mm between adjacent slices.

And 4, step 4: the proposed mesh projection algorithm is used to obtain the contour points of the workpiece point cloud model on the tangent plane, as shown in fig. 8.

And 5: the spray trajectory points are obtained by translating the sectional contour points in the direction of their normal vector by the distance of the spray height of the spray gun, which is 220mm, as shown in fig. 9.

Step 6: and connecting the spraying track points point by point in a linear mode to obtain a complete spraying track. As shown in fig. 10.

Claims

1. A spraying robot track planning method based on a grid projection algorithm comprises the following steps:

step 1: scanning by using a laser sensor to obtain a point cloud model of a sprayed workpiece;

step 2: determining the direction of the point cloud slice; the workpiece point cloud model is a pile of three-dimensional data points, the mass center and the characteristic vector of the workpiece point cloud model are obtained through calculation, the directions of the three characteristic vectors of the workpiece point cloud model are taken as XYZ axes of a coordinate system of the workpiece point cloud model, the mass center of the workpiece point cloud model is taken as the origin of the coordinate system of the workpiece point cloud model, and the coordinate system of the workpiece point cloud model is called as a characteristic vector coordinate system; selecting a direction vertical to the dominant feature vector of the workpiece point cloud model as a direction of point cloud slicing; after the point cloud slicing direction is determined, cutting the workpiece point cloud model along the point cloud slicing direction by using a series of parallel and equidistant cutting planes;

and step 3: acquiring a section contour point of the workpiece point cloud model on each cutting plane by using a proposed grid projection algorithm; the steps of the grid projection algorithm are as follows:

step 3-2: assuming that the direction of the point cloud slice is vertical to the Z axis, projecting the point cloud slice by using a N-N grid along the direction of the point cloud slice, so that the whole grid just can cover a point cloud band, and taking the intersection point of the grid on the two slices as a grid point; the grid points are section contour points;

step 3-3: under the previous assumption, the Y-coordinate and Z-coordinate values of the grid points are determined; however, given the complexity of the point cloud data distribution in the workpiece point cloud model, the same point cloud data as its Y and Z coordinate values may not exist for each grid point; establishing a square with the side length of lambda by taking each grid point as a center, wherein the lambda is a very small value;

step 3-4: for each square, selecting k points with the maximum X coordinate values in the point cloud slice direction, and taking the average value of the X coordinate values of the k points as the X coordinate value of the grid point; if the projection range of the square has no point cloud data, continuously expanding the value of lambda until the square projection range of each grid point has point cloud data; thus, the three-dimensional coordinate values of all the grid points can be obtained, namely the three-dimensional coordinate values of all the section contour points are obtained;

and 4, step 4: translating the section contour point along the direction of a normal vector of the section contour point by a distance of the spraying height of a spray gun to obtain a spraying track point; and connecting the spraying track points point by point in a straight line form to form a complete spraying track.