CN103706571A - Visual positioning sorting method - Google Patents

Abstract

The invention belongs to the field of mechanical automation, and in particular relates to a visual positioning sorting method based on a six-degree-of-freedom industrial robot. The visual positioning sorting method can be applied to the industries of logistics, assembly lines and the like. The method has the advantages that an image coordinate system of work pieces to be sorted can be accurately and quickly acquired through an image processing technology and a pattern recognition technology according to an image acquired by a camera; the image coordinate system is converted into coordinates under a robot coordinate system according to a coordinate conversion technology; the work pieces are placed at the designated positions through a manipulator according to the class by flexibly using the coordinates and work piece class information.

Description

A kind of vision locating, sorting method

Technical field

The invention belongs to mechanical automation field, be specifically related to a kind ofly based on Six-DOF industrial robot vision locating, sorting method, can be applicable to the industries such as logistics, assembling line.

Background technology

At present, industrial robot has been widely used in the fields such as automobile and auto parts and components manufacturing industry, mechanical processing industry, electric industry.The working method of the robot applying in industry now mostly adopts teaching to reproduce, and the working procedure storing by teach programming can repeat expection work, and the pose of robot arrival impact point is fixed like this, and each action all repeats.The working method limitation that this teaching is reproduced is very large, for some material putting position the unknowns in the situation that, the use of industrial robot is restricted, particularly for the workpiece being sticked together, industrial robot is difficult to judge accurately the type of location of workpiece coordinate, workpiece when operation, thereby cannot reach the object that workpiece is sorted.

In the sorting area the present invention relates to, workpiece to be sorted and workbench must have obvious gray scale difference.

Summary of the invention

In order to overcome the defect in background technology, the invention provides a kind of can a plurality of location of workpiece coordinates that are sticked together of Obtaining Accurate, the vision locating, sorting method of workpiece type.

Concrete technical scheme of the present invention is:

A vision locating, sorting method, is characterized in that, comprises the following steps:

1] obtain the coordinate figure of all workpiece to be sorted under image coordinate system; Its concrete steps are:

1.1] pictures taken, carries out pretreatment to picture, extracts the part interested of picture;

1.2] by the method for gray level threshold segmentation, by picture processing, be black and white picture, obtain the image of workpiece to be sorted;

1.3] in step 1.2] the image of workpiece to be sorted in, utilize minimum circumscribed circle method to extract and can comprise the circular image profile of all workpiece to be sorted, and obtain the center point coordinate of circular image profile;

1.4] according to step 1.3] in center point coordinate, select the external square area of minimum of circular image profile;

1.5] to step 1.4] in the external square area of minimum carry out whole scan, determine the coordinate of all workpiece to be sorted;

2] coordinate figure of all workpiece to be sorted under image coordinate system is transformed into the coordinate under robot coordinate system;

3] coordinate figure of the to be sorted workpiece of manipulator based under robot coordinate system, picks up workpiece to be sorted respectively, and workpiece to be sorted is put to assigned address.

Above-mentioned steps 1.5] concrete steps be:

1.5.1] set the square area that area is less than workpiece to be sorted, the external square of minimum is carried out to whole scan;

1.5.2] set gray threshold T, determine whether square area is workpiece to be detected; If square area gray threshold meets the requirement of threshold value T, think that square area is workpiece to be sorted and retains, otherwise give up;

1.5.3] respectively all point coordinates in the square area retaining are averaged, determine the coordinate of each workpiece to be sorted.

Above-mentioned steps 2] adopt the method for two-dimensional transformations to obtain the coordinate figure of workpiece to be sorted under robot coordinate system.

The invention has the beneficial effects as follows:

1, the method Ke Shi robot that the present invention adopts well conforms, can sort workpiece by autonomous classification band, carry out task operating, thereby expanded the application of robot in industries such as logistics, equipment production lines, under the flexibility and intelligentized prerequisite that realize robot, improve production efficiency and the reliability of robot.

2, the present invention adopts minimum circumscribed circle method and sets the square area that the length of side is less than workpiece to be detected, can be good at the workpiece to be sorted that identification is sticked together, and obtains accurately the coordinate of workpiece to be sorted.

Accompanying drawing explanation

Fig. 1 is particular flow sheet of the present invention.

The specific embodiment

The invention provides a kind of vision locating, sorting method that can obtain accurately and rapidly a plurality of location of workpiece coordinates to be sorted, workpiece type.

When employing is of the present invention, need image capture device, industrial computer, display and robot body to form.Image capture device adopts dimension industrial camera, industrial computer be take industrial control board as core, the CPU module forming with industrial control board and keyset spreads over standard computer expansion bus on passive backplane, each module is connected in passive backplane, thereby realize, by standard computer expansion bus and industrial control mainboard, carries out the function that exchanges data realizes corresponding module; The communication of industrial camera and industrial computer adopts ethernet communication, and display adopts regular display.

During work, the image that industrial camera collects, utilize that image processing techniques and mode identification technology are accurate, the coordinate figure of rapid extraction workpiece to be sorted under image coordinate system, and according to Coordinate Conversion technology, image coordinate is transformed into the coordinate under robot coordinate system, finally coordinate and workpiece classification information are sent to industrial computer flexibly, industrial computer control accurately picks up workpiece to be sorted.

Concrete grammar of the present invention is:

Step 1] utilize industrial camera to take pictures, the picture of taking is processed, and finally obtained the coordinate figure (X of all workpiece to be sorted under image coordinate system ₁, Y ₁);

This step is specifically:

Step 1.1] pictures taken, picture is carried out to pretreatment, extract the part interested of picture; The preprocess method that the present invention adopts in this step is pyramid image partition method, and real image is dwindled and obtains the interested part of picture;

Step 1.2] by the method for gray level threshold segmentation, by picture processing, be black and white picture, obtain the image of workpiece to be sorted;

Step 1.3] in step 1.2] the image of workpiece to be sorted in, utilize minimum circumscribed circle method to extract and can comprise the circular image profile of all workpiece to be sorted, and obtain the center point coordinate (X, Y) of circular image profile;

Step 1.4] according to step 1.3] in center point coordinate (X, Y), select the external square area of minimum of circular image profile;

Step 1.5] to step 1.4] in the external square area of minimum carry out whole scan, determine the coordinate of all workpiece to be sorted;

This step is specifically:

Step 1.5.1] set the square area that area is less than workpiece to be sorted, the external square of minimum is carried out to whole scan;

Step 1.5.2] set gray threshold T, determine whether square area is workpiece to be detected; If square area gray scale meets the requirement of threshold value T, think that square area is workpiece to be sorted and retains, otherwise give up;

Step 1.5.3] respectively the square area retaining is asked for to central point, determine the coordinate (X of each workpiece to be sorted ₁, Y ₁).

Step 2] adopt the method for two-dimensional transformations by the coordinate figure (X of all workpiece to be sorted under image coordinate system ₁, Y ₁) be transformed into the coordinate figure (X under robot coordinate system ₂, Y ₂).

The concrete calculation relational expression of this step is as follows:

$\left[\begin{array}{c}{X}_2\\ Y_2\end{array}\right]=\left(\begin{array}{cc}\cos \mathrm{\θ}& -\sin \\ \sin \mathrm{\θ}& \cos \mathrm{\θ}\end{array}\right)\left[\begin{array}{c}{X}_1\\ Y_1\end{array}\right]+\left[\begin{array}{c}{X}_0\\ Y_0\end{array}\right]$

Wherein:

[X ₁y ₁] ^tcoordinate for workpiece to be sorted under image coordinate system;

[X ₂y ₂] ^tfor workpiece to be sorted coordinate under robot coordinate system;

[X ₀y ₀] ^tfor robot coordinate system's initial point coordinate under image coordinate system.

θ is that robot coordinate ties up to directions X cosine value under image coordinate system;

Step 3] by the workpiece all to be sorted coordinate under robot coordinate system obtaining, utilize industrial computer control that workpiece to be sorted is put to assigned address.

Claims (3)

1. a vision locating, sorting method, is characterized in that, comprises the following steps:

1] obtain the coordinate figure of all workpiece to be sorted under image coordinate system;

1.1] pictures taken, carries out pretreatment to picture, extracts the part interested of picture;

1.2] by the method for gray level threshold segmentation, by picture processing, be black and white picture, obtain the image of workpiece to be sorted;

1.3] in step 1.2] the image of workpiece to be sorted in, utilize minimum circumscribed circle method to extract and can comprise the circular image profile of all workpiece to be sorted, and obtain the center point coordinate of circular image profile;

1.4] according to step 1.3] in center point coordinate, select the external square area of minimum of circular image profile;

1.5] to step 1.4] in the external square area of minimum carry out whole scan, determine the coordinate of all workpiece to be sorted;

2] coordinate figure of all workpiece to be sorted under image coordinate system is transformed into the coordinate under robot coordinate system;

3] coordinate figure of the to be sorted workpiece of manipulator based under robot coordinate system, picks up workpiece to be sorted respectively, and workpiece to be sorted is put to assigned address.

2. according to the vision locating, sorting method described in claims 1, it is characterized in that: described step 1.5] concrete steps be:

1.5.1] set the square area that area is less than workpiece to be sorted, the external square of minimum is carried out to whole scan;

1.5.2] set gray threshold T, determine whether square area is workpiece to be detected; If square area gray threshold meets the requirement of threshold value T, think that square area is workpiece to be sorted and retains, otherwise give up;

1.5.3] respectively all point coordinates in the square area retaining are averaged, determine the coordinate of each workpiece to be sorted.

3. according to the vision locating, sorting method described in claims 1 or 2, it is characterized in that: described step 2] adopt the method for two-dimensional transformations to obtain the coordinate figure of workpiece to be sorted under robot coordinate system.

Images