CN104574457A - Shorter-path robot image drawing method - Google Patents

Abstract

The invention relates to an image drawing method, in particular to a shorter-path robot image drawing method, which is implemented according to the steps of previous preparation, contour drawing, starting point determination and degree control drawing. According to the shorter-path robot image drawing method, the operation efficiency is improved, and operation steps are simplified.

Description

A kind of robot graphics's method for drafting had compared with short path

Technical field

The present invention relates to a kind of image drawing method, particularly relate to a kind of robot graphics's method for drafting had compared with short path.

Background technology

Drawing is a very common task, and printer, plane plate all have this function.But its function is also only limitted to this.

In industrial technical field, when giving different end-of-arm toolings, the function of drawing can also expand to cut, mark, Linear cut etc.For realizing specific movement locus, this work at present main numerical control code that relies on realizes.For raster image, can't directly process.

In industrial robot field, movement locus completes mainly through the mode of pointwise teaching.When figure is comparatively complicated, the task of teaching is quite heavy.

Summary of the invention

The present invention mainly solves the deficiencies in the prior art, a kind of teaching task heavy when overcoming the plane complicated figure of robot drawing is provided, optimize the track route of end effector of robot simultaneously, reduce the redirect time between outline line, thus a kind of robot graphics's method for drafting had compared with short path of increasing work efficiency.

Above-mentioned technical matters of the present invention is mainly solved by following technical proposals:

There is the robot graphics's method for drafting compared with short path, carry out according to the following steps:

(1), early-stage preparations:

First this image is converted into gray-scale map, then by threshold process by this image binaryzation, extract foreground target, then, operated by mathematical morphology, delete isolated point, eliminate salt-pepper noise, afterwards, extract objective contour, the data point forming profile is simplified;

(2), profile is drawn:

First draw peripheral largest contours, then drawing all the other large profile, by drawing successively smoothly, finally drawing details profile;

The drawing order of details profile, is solved with traveling salesman problem;

For this reason, the center of gravity of each outline line is considered as city one by one, problem be find out one through all cities once and only shortest path once, unique difference is, after having accessed last city, does not need to get back to initial city;

Following solution procedure is carried out by ant group algorithm:

(1) n people worker ant is placed at random in m city, gets n=m here; Suppose sequential path 0,1 ..., the length of m-1} is L, the pheromones content by between each city of following formula initialization:

${\mathrm{\τ}}_{\mathrm{ij}}\left(0\right)=\left\{\begin{array}{cc}\frac{1}{L},& i,j\∈\{\mathrm{0,1},...,m-1\},i\≠j,\\ 0,& i=j\end{array}\right.$

(2) to a kth people worker ant, when selecting next city j by current location i, select probability is determined by following formula:

In formula, η _ijfor city j is relative to the visibility of city i, generally get wherein, d _ijfor the distance between city i and city j;

α, β, for representing the significance level of correlative, get α=1 here, β=10; represent the city list of current still end access;

The selection of city j is determined by roulette, and select probability is larger, and area shared on wheel disc is larger, and selected probability is also larger;

When k changes to n-1 by 0, when i changes to m-2 by 1, n people worker ant have accessed m city once and only once, every people worker ant forms a feasible solution respectively;

(3) according to the length of access path, n people worker ant is arranged from small to large; Suppose that this shortest path length of trying to achieve that circulates is L', if L' >=L, EOP (end of program), L' is overall bee-line; Otherwise, make L'=L, go to step (4);

(4) the pheromones content on path is upgraded by following formula:

${\mathrm{\τ}}_{\mathrm{ij}}(t+1)=(1-\mathrm{\ρ}){\mathrm{\τ}}_{\mathrm{ij}}\left(t\right)+\underset{r=0}{\overset{w-1}{\mathrm{\Σ}}}(w-r)\mathrm{\Δ}{\mathrm{\τ}}_{\mathrm{ij}}^r$

In formula, ρ is the volatility of pheromones, gets ρ=0.1 here; W is the ant quantity being used to lastest imformation element, gets w=n/2 here; by following formula value:

Wherein, C ^rrepresent the path that the people worker ant that grade is r is passed by; R=0 this near-optimal solution corresponding, r=1 represents this suboptimal solution, and the rest may be inferred for all the other; Q gets 1;

After completing above work, jump to step (2) circulation and perform;

According to above method, obtain the access order of outline line;

(3), determine to start to write a little:

Need to determine on outline line suitable to start to write a little, make redirect path short as far as possible;

First determine the bee-line of the first two profile, thus obtain these two profiles start to write a little, then find and second a little nearest point of starting to write on the 3rd profile, start to write a little as the 3rd, then find and the 3rd a little nearest point of starting to write on the 4th profile, start to write a little as the 4th, all the other the like, thus determine a relatively short path;

Or,

Each outline line is got a little at random, forms multiple different solution, choose optimum solution wherein as shortest path; For improving program speed of convergence, use heuristic search algorithm;

Ant group algorithm can not ensure all to try to achieve optimum solution at every turn, but the solution of gained is always very close to optimum solution;

(4), controlling extent drawing process:

First, at teaching pattern lower-pilot teach box, paintbrush is positioned to map sheet central authorities, nib touches paper; When host computer procedure performs, paintbrush is raised 10mm, and move to first profile start to write a little, paintbrush moves down 10mm to start to draw, and after Article 1 profile paints, 10mm raised by paintbrush, move to the drawing point of second profile, paintbrush moves down 10mm to start to draw Article 2 profile, and the rest may be inferred for all the other, until paint all profiles;

The end offset of the equal opposed robots's ruling pen of all data points;

The initial position of the corresponding ruling pen of figure central authorities;

By arranging drawing ratio, the actual size of institute's drawing shape can be changed, or reduce its original size with certain precision;

Steering order is sent to controller by netting twine, or expresses drawing for order with robot language, and download to controller in the mode of operation, the mode being called operation by teach box performs drawing tasks.

The invention provides a kind of robot graphics's method for drafting had compared with short path, improve operating efficiency, simplify the operation step.

Accompanying drawing explanation

Fig. 1 is example image pending in the present invention;

Fig. 2 is the plotting mode that in the present invention, a kind of path is longer;

Fig. 3 is the line of each outline line center of gravity after drawing path optimization in the present invention;

Fig. 4 gives a kind of shorter drawing path in the present invention.

Embodiment

Below by embodiment, and by reference to the accompanying drawings, technical scheme of the present invention is described in further detail.

Embodiment 1: as shown in Figure 1, Figure 2, Figure 3 and Figure 4, a kind of robot graphics's method for drafting had compared with short path, carries out according to the following steps:

(1), early-stage preparations:

First this image is converted into gray-scale map, then by threshold process by this image binaryzation, extract foreground target, then, operated by mathematical morphology, delete isolated point, eliminate salt-pepper noise, afterwards, extract objective contour, the data point forming profile is simplified;

(2), profile is drawn:

First draw peripheral largest contours, then drawing all the other large profile, by drawing successively smoothly, finally drawing details profile;

The drawing order of details profile, is solved with traveling salesman problem;

For this reason, the center of gravity of each outline line is considered as city one by one, problem be find out one through all cities once and only shortest path once, unique difference is, after having accessed last city, does not need to get back to initial city;

Following solution procedure is carried out by ant group algorithm:

(1) n people worker ant is placed at random in m city, gets n=m here; Suppose sequential path 0,1 ..., the length of m-1} is L, the pheromones content by between each city of following formula initialization:

${\mathrm{\τ}}_{\mathrm{ij}}\left(0\right)=\left\{\begin{array}{cc}\frac{1}{L},& i,j\∈\{\mathrm{0,1},...,m-1\},i\≠j,\\ 0,& i=j\end{array}\right.$

(2) to a kth people worker ant, when selecting next city j by current location i, select probability is determined by following formula:

In formula, η _ijfor city j is relative to the visibility of city i, generally get wherein, d _ijfor the distance between city i and city j;

α, β, for representing the significance level of correlative, get α=1 here, β=10; represent the city list of current still end access;

The selection of city j is determined by roulette, and select probability is larger, and area shared on wheel disc is larger, and selected probability is also larger;

When k changes to n-1 by 0, when i changes to m-2 by 1, n people worker ant have accessed m city once and only once, every people worker ant forms a feasible solution respectively;

(3) according to the length of access path, n people worker ant is arranged from small to large; Suppose that this shortest path length of trying to achieve that circulates is L', if L' >=L, EOP (end of program), L' is overall bee-line; Otherwise, make L'=L, go to step (4);

(4) the pheromones content on path is upgraded by following formula:

${\mathrm{\τ}}_{\mathrm{ij}}(t+1)=(1-\mathrm{\ρ}){\mathrm{\τ}}_{\mathrm{ij}}\left(t\right)+\underset{r=0}{\overset{w-1}{\mathrm{\Σ}}}(w-r)\mathrm{\Δ}{\mathrm{\τ}}_{\mathrm{ij}}^r$

In formula, ρ is the volatility of pheromones, gets ρ=0.1 here; W is the ant quantity being used to lastest imformation element, gets w=n/2 here; by following formula value:

Wherein, C ^rrepresent the path that the people worker ant that grade is r is passed by; R=0 this near-optimal solution corresponding, r=1 represents this suboptimal solution, and the rest may be inferred for all the other; Q gets 1;

After completing above work, jump to step (2) circulation and perform;

According to above method, obtain the access order of outline line;

(3), determine to start to write a little:

Need to determine on outline line suitable to start to write a little, make redirect path short as far as possible;

First determine the bee-line of the first two profile, thus obtain these two profiles start to write a little, then find and second a little nearest point of starting to write on the 3rd profile, start to write a little as the 3rd, then find and the 3rd a little nearest point of starting to write on the 4th profile, start to write a little as the 4th, all the other the like, thus determine a relatively short path;

Or,

Each outline line is got a little at random, forms multiple different solution, choose optimum solution wherein as shortest path; For improving program speed of convergence, use heuristic search algorithm;

Ant group algorithm can not ensure all to try to achieve optimum solution at every turn, but the solution of gained is always very close to optimum solution;

(4), controlling extent drawing process:

First, at teaching pattern lower-pilot teach box, paintbrush is positioned to map sheet central authorities, nib touches paper; When host computer procedure performs, paintbrush is raised 10mm, and move to first profile start to write a little, paintbrush moves down 10mm to start to draw, and after Article 1 profile paints, 10mm raised by paintbrush, move to the drawing point of second profile, paintbrush moves down 10mm to start to draw Article 2 profile, and the rest may be inferred for all the other, until paint all profiles;

The end offset of the equal opposed robots's ruling pen of all data points;

The initial position of the corresponding ruling pen of figure central authorities;

By arranging drawing ratio, the actual size of institute's drawing shape can be changed, or reduce its original size with certain precision;

Steering order is sent to controller by netting twine, or expresses drawing for order with robot language, and download to controller in the mode of operation, the mode being called operation by teach box performs drawing tasks.

In Fig. 4, give one of them solving result.As seen from the figure, the redirect route between profile obviously shortens, but due to data point be discrete, and have passed through and delete, therefore, what obtain is a relatively short path, instead of theoretic shortest path.

Claims (1)

1. there is the robot graphics's method for drafting compared with short path, it is characterized in that carrying out according to the following steps:

(1), early-stage preparations:

First this image is converted into gray-scale map, then by threshold process by this image binaryzation, extract foreground target, then, operated by mathematical morphology, delete isolated point, eliminate salt-pepper noise, afterwards, extract objective contour, the data point forming profile is simplified;

(2), profile is drawn:

First draw peripheral largest contours, then drawing all the other large profile, by drawing successively smoothly, finally drawing details profile;

The drawing order of details profile, is solved with traveling salesman problem;

For this reason, the center of gravity of each outline line is considered as city one by one, problem be find out one through all cities once and only shortest path once, unique difference is, after having accessed last city, does not need to get back to initial city;

Following solution procedure is carried out by ant group algorithm:

(1) n people worker ant is placed at random in m city, gets n=m here; Suppose sequential path 0,1 ..., the length of m-1} is L, the pheromones content by between each city of following formula initialization:

${\mathrm{\τ}}_{\mathrm{ij}}\left(0\right)=\left\{\begin{array}{cc}\frac{1}{L},& i,j\∈\{\mathrm{0,1},...,m-1\},i\≠j,\\ 0,& i=j\end{array}\right.$

(2) to a kth people worker ant, when selecting next city j by current location i, select probability is determined by following formula:

In formula, η _ijfor city j is relative to the visibility of city i, generally get wherein, d _ijfor the distance between city i and city j;

α, β, for representing the significance level of correlative, get α=1 here, β=10; represent the city list of current still end access;

The selection of city j is determined by roulette, and select probability is larger, and area shared on wheel disc is larger, and selected probability is also larger;

When k changes to n-1 by 0, when i changes to m-2 by 1, n people worker ant have accessed m city once and only once, every people worker ant forms a feasible solution respectively;

(3) according to the length of access path, n people worker ant is arranged from small to large; Suppose that this shortest path length of trying to achieve that circulates is L', if L' >=L, EOP (end of program), L' is overall bee-line; Otherwise, make L'=L, go to step (4);

(4) the pheromones content on path is upgraded by following formula:

${\mathrm{\τ}}_{\mathrm{ij}}(t+1)=(1-\mathrm{\ρ}){\mathrm{\τ}}_{\mathrm{ij}}\left(t\right)+\underset{r=0}{\overset{w-1}{\mathrm{\Σ}}}(w-r){\mathrm{\Δ\τ}}_{\mathrm{ij}}^r$

In formula, ρ is the volatility of pheromones, gets ρ=0.1 here; W is the ant quantity being used to lastest imformation element, gets w=n/2 here; by following formula value:

Wherein, C ^rrepresent the path that the people worker ant that grade is r is passed by; R=0 this near-optimal solution corresponding, r=1 represents this suboptimal solution, and the rest may be inferred for all the other; Q gets 1;

After completing above work, jump to step (2) circulation and perform;

According to above method, obtain the access order of outline line;

(3), determine to start to write a little:

Need to determine on outline line suitable to start to write a little, make redirect path short as far as possible;

First determine the bee-line of the first two profile, thus obtain these two profiles start to write a little, then find and second a little nearest point of starting to write on the 3rd profile, start to write a little as the 3rd, then find and the 3rd a little nearest point of starting to write on the 4th profile, start to write a little as the 4th, all the other the like, thus determine a relatively short path;

Or,

Each outline line is got a little at random, forms multiple different solution, choose optimum solution wherein as shortest path; For improving program speed of convergence, use heuristic search algorithm;

Ant group algorithm can not ensure all to try to achieve optimum solution at every turn, but the solution of gained is always very close to optimum solution;

(4), controlling extent drawing process:

First, at teaching pattern lower-pilot teach box, paintbrush is positioned to map sheet central authorities, nib touches paper; When host computer procedure performs, paintbrush is raised 10mm, and move to first profile start to write a little, paintbrush moves down 10mm to start to draw, and after Article 1 profile paints, 10mm raised by paintbrush, move to the drawing point of second profile, paintbrush moves down 10mm to start to draw Article 2 profile, and the rest may be inferred for all the other, until paint all profiles;

The end offset of the equal opposed robots's ruling pen of all data points;

The initial position of the corresponding ruling pen of figure central authorities;

By arranging drawing ratio, the actual size of institute's drawing shape can be changed, or reduce its original size with certain precision;

Steering order is sent to controller by netting twine, or expresses drawing for order with robot language, and download to controller in the mode of operation, the mode being called operation by teach box performs drawing tasks.