CN116303841B

CN116303841B - Two-degree-of-freedom space-based multi-element map annotation optimization DDEGA-NM algorithm and system

Info

Publication number: CN116303841B
Application number: CN202211491090.XA
Authority: CN
Inventors: 邓吉秋; 郭志勇; 邱蓝; 李�荣; 吴军; 王飞龙
Original assignee: Central South University
Current assignee: Central South University
Priority date: 2022-11-25
Filing date: 2022-11-25
Publication date: 2025-09-09
Anticipated expiration: 2042-11-25
Also published as: CN116303841A

Abstract

The present invention discloses a DDEGA-NM method and system for multi-element map annotation optimization based on two-degree-of-freedom space. Under the premise that the appropriate number of candidate locations (N*M) and the minimum and maximum buffer distances have been determined, the initial population generation and random iteration process of the DDEGA algorithm are improved, thereby improving the quality of annotation configuration and the speed of algorithm convergence. Finally, a DDEGA optimization algorithm based on annotation configuration of multiple candidate locations in two-degree-of-freedom space is formed. The optimization is divided into two steps: the first step is to optimize the initial population by adding a local optimal permutation and combination method and controlling the range of random generation of the initial population to improve the quality of the initial population; the second step is to optimize the genetic process by changing the proportion of individuals generated by the differential algorithm and the genetic algorithm in the new population and increasing a portion of individuals randomly generated from high-quality candidate locations.

Description

Two-degree-of-freedom space-based multi-element map annotation optimization DDEGA-NM algorithm and system

Technical Field

The invention belongs to the field of algorithms for automatic configuration of multi-element annotation, and particularly relates to a multi-element map annotation optimization DDEGA-NM algorithm based on a two-degree-of-freedom space. Meanwhile, the invention also relates to a multi-element map annotation optimization DDEGA-NM system based on the two-degree-of-freedom space.

Background

At present, on the algorithm of multi-element annotation automatic configuration, the Discrete Differential Evolution and Genetic Algorithm (DDEGA) has better effect, takes the framework of the genetic algorithm as a main body, adopts a hybrid propagation mode of the differential evolution algorithm and the genetic algorithm to improve the optimization efficiency, and simultaneously increases the calculation operator of random gene positions, so that the two algorithms are complementary in priority, and simultaneously have the capability of eliminating the success or failure of the genetic algorithm and the intelligent memory and searching capability of the differential evolution algorithm group, and can adaptively adjust the searching direction to continuously optimize the algorithm result.

The DDEGA algorithm has a good effect on multi-element map annotation configuration, but also has the problems of long running time, low convergence speed and the like.

Disclosure of Invention

The invention aims to solve the defects in the prior art, and provides a two-degree-of-freedom space-based multi-element map annotation optimization DDEGA-NM algorithm and a system, which are used for improving the quality of an initial population by adding a local optimal arrangement and combination mode and controlling the range of random generation of the initial population, changing the proportion of individuals of a new population generated by a differential algorithm and a genetic algorithm, and adding a part of individuals generated by high-quality candidate positions randomly.

In order to achieve the above purpose, the present invention provides the following technical solutions:

A multi-element map annotation optimization DDEGA-NM algorithm based on two-degree-of-freedom space comprises the following steps:

s1, inputting experimental vector Map data Map0, wherein the Map comprises different layers of points, lines and planes, and each layer is provided with a plurality of elements;

S2, outputting a result of Outcome lists, wherein the result comprises a annotation candidate position Name, type, X and Y data information which are selected by each element, and Outcome = [ Name, type, X and Y ];

S3, defining a one-dimensional list Labels _1 and a two-dimensional list Labels _2, wherein Labels_1 and Labels _2 are used for representing all candidate position sets of all elements;

S4, determining the number of candidate positions and minimum and maximum buffer distance values of the marks from the elements based on a vector Map0, and generating candidate positions for the point, line and surface elements respectively to obtain a candidate position set;

s5, randomly generating an initial population by using an optimization DDEGA algorithm;

S6, acquiring each candidate position Name, type, X and Y-coordinate information one by adopting a circulation function, and scoring each candidate position by adopting a annotation gland, annotation ambiguity and annotation priority factors;

s7, generating each element, sequencing the elements according to the theoretical limit value of the candidate position from small to large, wherein the lower the grading value is, the higher the priority of the corresponding candidate position is;

S8, combining the candidate position theory limit values as an individual X (1) of the initial population, and ensuring that DDEGA-NM algorithm can keep the individual with the lowest score in each loop iteration;

s9, four scoring factors including annotation conflict, annotation gland, annotation ambiguity and annotation priority are adopted to evaluate the populations X (1) -X (100), and the minimum scoring value combination is directly added into the next generation population so as to ensure that the current optimal annotation candidate position combination is not lost;

s10, sorting the scoring values of the generated initial population X (1) -X (100), taking the individuals with the scores of the first 50% to participate in generating a next generation population, generating 30% and 60% of new individuals through a genetic algorithm and a differential evolution algorithm respectively, randomly selecting candidate position serial numbers [1- |N ] of each element to combine and generate 10% of individuals, and adding the individuals into the next generation new population to form new population individuals X (1) -X (100);

s11, judging whether the new populations X (1) -X (100) meet the iteration termination condition or not, wherein the method specifically comprises the following steps:

if the iteration condition is not met, the step S9 is entered, and the next loop iteration is started;

if the iteration termination condition is met, outputting X (n) with the smallest scoring value in the new populations X (1) -X (100);

S12, decoding the X (n) to obtain the determined candidate position coordinate information of each element, and storing Outcome, outcome = [ Name, type, X, Y ];

And S13, outputting a note configuration result Outcome after finishing.

Preferably, in step S3, the one-dimensional list Labels _1 and the two-dimensional list Labels _2 each include five fields, which are Name of the element, type of the element, X coordinate and Y coordinate of the candidate location center point of the annotation, and theoretical limit score F, label= [ Name, type, X, Y, F ].

Preferably, the determining the number of candidate positions and the maximum buffer distance value of the mark from the element in step S4 specifically includes:

setting different buffer distance parameters;

and (3) finding a point with the first derivative of 0 by adopting a curve fitting mode, wherein the point is a stable point of the curve, and further determining a proper maximum buffer distance value.

Preferably, the optimization DDEGA algorithm described in step S5 randomly generates the initial population:

1) Firstly, scoring the candidate positions generated by the candidate positions according to annotation gland, annotation ambiguity and annotation priority;

2) Then, the candidate positions of each element are ranked in order from small to large according to the scoring value;

3) Finally, generating the quality of the initial population based on the newly ordered candidate position optimization DDEGA.

Preferably, step S6 further includes:

The scoring value is expressed by F, the label= [ Name, type, X, Y, F ] data obtained each time are stored in a two-dimensional list Labels _1, and Labels _1 represents all candidate position information of all elements;

Namely:

Labels_1＝[Label_1-1,Label_1-2,Label_1-3,……,Label_1-N*M,Label_2-1,Label_2-2,

Label_2-3,……,Label_2-N*M,……,Label_S-1,Label_S-2,Label_S-3,……,Label_S-N*M];

Label _S-N*M represents the N.times.M candidate position coordinates and theoretical limit value scoring information set of the S-th element.

Preferably, the ordered result is denoted Labels _2:

Labels_2=[Label_1-[1],Label_1-[2],Label_1-[3],……,Label_1-[N*M],Label_2-[1],

Label_2-[2],Label_2-[3],……,Label_2-[N*M],……,Label_S-[1],Label_S-[2],Label_S-[3],……,Label_S-[N*M]];

Wherein Label _S-[N*M] represents the candidate set of positional information for theoretical limit value Fth row N.times.M in the S-th element.

Preferably, the first candidate position Label _S-[1] of each element in Labels _2 is called the theoretical optimal candidate position of the element, and the set (Label _1-[1],Label_2-[1],Label_3-[1],……,Label_S-[1]) is called the candidate position theoretical limit value combination V, and contains the optimal candidate position of each element under the condition of no annotation conflict, and the selection should be considered first in the annotation configuration process.

Preferably, in step S9, for the other 99 population individuals, the candidate positions are no longer randomly generated from the element numbers [1] - [ n×m ], but after the candidate position score value of Labels _2 is ordered, the candidate position selection interval is manually controlled, and the candidate position selection interval is randomly selected only in the candidate position numbers [1] - [ |n×m/3| ] (|n×m/3|represents the maximum integer not greater than n×m/3) of each element (Label ₁-label_S), so as to optimize the 99 population individuals.

A two-degree-of-freedom space-based multi-element map annotation optimization DDEGA-NM system, comprising:

a central processing unit, a memory;

The memory is a short-term memory or a persistent memory;

The central processor is configured to communicate with the memory, execute instruction operations in the memory on the two-degree-of-freedom space-based multi-element map annotation optimization DDEGA-NM system to perform a two-degree-of-freedom space-based multi-element map annotation optimization DDEGA-NM algorithm;

Further comprises:

a computer-readable storage medium comprising instructions that when executed on a computer cause the computer to perform a two-degree-of-freedom space-based multi-element map annotation optimization DDEGA-NM algorithm.

The invention has the technical effects and advantages that:

1. Optimizing an initial population, and improving the quality of the initial population by adding a local optimal arrangement and combination mode and controlling the random generation range of the initial population;

2. Optimizing the genetic process, changing the differential algorithm and the proportion of individuals in the new population generated by the genetic algorithm, and adding a part of individuals randomly generated by high-quality candidate positions;

3. By optimizing DDEGA algorithm with better effect in element map annotation configuration, the initial population quality and genetic iteration process are controlled, so that the time for searching high-quality annotation candidate position combination mode can be reduced, the efficiency of automatic annotation configuration can be improved, and annotation gland and annotation ambiguity can be reduced to a certain extent.

Drawings

FIG. 1 is a flowchart of an algorithm after DDEGA algorithm optimization in an embodiment of the present invention;

FIG. 2 is a flowchart of a DDEGA algorithm before optimization in accordance with an embodiment of the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, based on the embodiments of the invention, which are apparent to those of ordinary skill in the art without inventive faculty, are intended to be within the scope of the invention.

Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, based on the embodiments of the invention, which are apparent to those of ordinary skill in the art without inventive faculty, are intended to be within the scope of the invention.

The invention provides a multi-element map annotation optimization DDEGA-NM algorithm based on a two-degree-of-freedom space, which comprises input description, output description and annotation configuration processes. The detailed process of the technical scheme is described as follows:

(1) The input instruction input is experimental vector Map data Map0 of the patent of the invention, the Map comprises different layers of points, lines and planes, and each layer is provided with a plurality of elements;

(2) The output description is based on the technical method of the invention, realizes annotation configuration for each element of the map, and the output result is Outcome list, which contains the annotation candidate position Name, type, X and Y data information of each element selected, outcome = [ Name, type, X, Y ], as shown in the following table:

Table 1 outputs candidate location coordinate information Outcome

Field name	Meaning of field	Field type	Field description
				Name	Element name	TEXT	The name information of the element(s),
Type	Element type	TEXT	Is divided into point, polyline, polygon types
				X	Abscissa of the circle	DECIMAL	Determined candidate location center point abscissa information
Y	Ordinate of the ordinate	DECIMAL	Determined candidate location center point ordinate information

As shown in fig. 1, (3) a token configuration process;

(3-1) definition of parameters;

(3-1-1) defining a one-dimensional list Label. Label is used for representing the coordinate information of each candidate position of the element, and comprises five fields, namely Name of the element, type of the element, X coordinate and Y coordinate of a center point of a marked candidate position, theoretical limit value score F, label= [ Name, type, X, Y, F ];

(3-1-2) defining two-dimensional lists Labels _1 and Labels _2, each representing the entire candidate location set for all elements;

(3-2) generating candidate positions, namely determining the number (N.times.M) of candidate positions and minimum and maximum buffer distance values (b _min and b _max) of the marks from elements based on a vector Map0, wherein the specific method is as follows:

Setting different buffer distance parameters, setting the minimum buffer distance of the notes to 1/n of a text height, setting the maximum buffer distance to 2/n, 3/n, 4/n, and 4/n of a text height, and storing all lists List1 into List2 by (for iin range) circulation function, namely, calculating the corresponding note score theoretical limit value F of different minimum and maximum buffer distances (b _min and b _max) by List1, wherein b _min＝D/8,b_max = 2*D/n, 3*D/n, 4*D/n, and n is equal to n, and the minimum and maximum buffer distances are represented by List1, and List1 is represented by List _min,b_max and F List2＝[[D/n,2*D/n,F₁],[D/n,3*D/n,F₂],[D/n,4*D/n,F₃],……,[D/n,n*D/n,F_n-1]];

Determining a proper maximum buffer distance value, along with the increase of the maximum buffer distance, meaning that the selectable space range of the candidate position of the annotation is increased, the annotation with less collision of the gland is likely to occur, the corresponding theoretical limit value is gradually reduced, and when the maximum buffer distance is continuously increased, the theoretical limit value gradually shows a decreasing trend, so that a curve equation can be obtained by adopting a fitting curve mode, and then a point with a first derivative of 0 is found, wherein the point is a stable point of the curve, and the specific process is as follows:

Filling b _max from 2*D/n to n D/n values into an array Ld, wherein ld= [2*D/n,3*D/n,4*D/n, ], and n is D/n, and the next step is to obtain abscissa data of a fitting curve;

Filling theoretical limit values F corresponding to different buffer distances in a List2 into an array Fd, wherein Fd= [ F ₁,F₂,F₃,……,F_n-1 ] is used as ordinate data of a fitting curve in the next step;

Taking data in the array Ld as an abscissa and data in the array Fd as an ordinate, performing curve cubic polynomial function fitting, and solving an expression P1 of a cubic polynomial curve fitting equation, wherein P1=Ax ³+Bx² +Cx+D;

deriving the cubic polynomial P1 to obtain a first-order derivative function p2=3ax ² +2bx+c, and obtaining a solution x1 which is a positive number by setting the first-order derivative function to 0; if there are two positive solutions x1 and x2 (x 1< x 2), taking the smaller value of x1 as the solution, x representing the plateau under the curve;

Calculating the smallest absolute value of the difference between n and x1 in the array ld= [2*D/n,3*D/n, 4*D/n.. The number n is the appropriate maximum buffer distance b _max;

Then generating N.times.M candidate positions for the point, line and surface elements respectively to obtain a candidate position set Result [ ], which is specifically as follows:

Determining a reference position according to the position of a target element on the map, wherein the target element is an element needing to be added with a mark;

determining a candidate region according to the reference position and the buffer distance, wherein the buffer distance is a preset value, and the candidate region is a partial region on the map;

Dividing the candidate region;

Determining a plurality of candidate positions according to the division result of the candidate areas;

determining one candidate position in the plurality of candidate positions as a target position;

the annotation corresponding to the target element is arranged at the target location.

(3-3) Optimizing DDEGA the mode of randomly generating the initial population by the algorithm, namely firstly scoring the generated candidate positions in (3-2) by marking gland, marking ambiguity and marking priority, then sequencing the candidate positions of each element in the order from the scoring value to the big value again, and finally optimizing DDEGA based on the candidate positions after the new sequencing to generate the quality of the initial population;

(3-3-1) using a (for iin range) circulation function, obtaining each candidate position Name, type, X and Y-coordinate information in the list Result [ ] one by one with Label, grading each candidate position by using a annotation gland, annotation ambiguity and annotation priority factor, wherein the grading value is represented by F, the label= [ Name, type, X, Y, F ] data obtained each time are stored in a two-dimensional list Labels _1, and Labels _1 represents all candidate position information of all elements;

Namely:

Wherein Label _S-N*M represents the N.times.M candidate position coordinates and theoretical limit value scoring information set of the S-th element, enter (3-3-2)

(3-3-2) Ranking each element of Labels _1 generated in (3-3-1) in order of decreasing score value according to theoretical limit value F of candidate position, the lower the score value, the higher the priority of the corresponding candidate position. The ordered result is denoted Labels _2:

Labels_2=[Label_1-[1],Label_1-[2],Label_1-[3],……,Label_1-[N*M],Label_2-[1],

Wherein Label _S-[N*M] represents a set of candidate position information for theoretical limit value F row N.times.M in the S-th element, the first candidate position Label _S-[1] of each element in Labels _2 is called theoretical optimal candidate position of the element, the set formed (Label _1-[1],Label_2-[1],Label_3-[1],……,Label_S-[1]) is called candidate position theoretical limit value combination V, and the optimal candidate position of each element under no annotation conflict is included, and the selection should be considered firstly in the annotation configuration process;

(3-3-3) taking the combination V of the theoretical limit values of candidate positions as an individual X (1) of the initial population, ensuring that DDEGA-NM algorithm can keep the individual with the lowest score in each cycle iteration, for the other 99 population individuals, randomly generating candidate positions from element numbers [1] - [ N ] M, artificially controlling candidate position selection intervals only in candidate position numbers [1] - [ |N ] M/3| ] ] (|N is the largest integer not more than N|M/3) of each element (Label ₁-label_S) after being sequenced by Labels _2 candidate position scoring values, optimizing 99 population individuals to obtain initial population individuals X (2) -X (100), and forming initial populations Y1:X (1) -X (100);

(3-4) evaluating the population X (1) -X (100) by adopting four scoring factors of annotation conflict, annotation gland, annotation ambiguity and annotation priority, directly adding the minimum scoring value combination into the next generation population to ensure that the current optimal annotation candidate position combination is not lost, and entering (3-5);

(3-5) optimizing a genetic loop iteration process, namely sorting the scoring values of the initial populations X (1) -X (100) generated in the step (3-4), taking the individuals with the scores of the first 50% to participate in generating a next generation population, respectively generating 30% and 60% of new individuals through a Genetic Algorithm (GA) and a differential evolution algorithm (DDE), randomly selecting candidate position serial numbers [1- |N|M/3| ] of each element (Label ₁-label_S) to combine and generate 10% of individuals, adding the 10% of individuals into the next generation new population to form new population individuals X (1) -X (100), and entering (3-6);

(3-6) judging whether the new populations X (1) -X (100) meet the iteration termination condition, if not, entering (3-4) to start the next loop iteration, and if so, outputting X (n) with the minimum scoring value in the new populations X (1) -X (100) to enter (3-7);

(3-7) decoding X (n) to obtain the determined candidate position coordinate information of each element (Label ₁-label_S), storing Outcome, outcome = [ Name, type, X, Y ];

and (3-8) ending, and outputting a note configuration result Outcome.

According to the method, the quality of the initial population and the genetic iteration process are controlled by optimizing the DDEGA algorithm with a good effect in the element map annotation configuration, so that the time for searching a high-quality annotation candidate position combination mode can be reduced, the efficiency of automatic annotation configuration is improved, and annotation pressing and annotation ambiguity can be reduced to a certain extent.

Comparative example 1

As shown in fig. 2, the DDEGA algorithm steps are as follows:

(1) The initial population is generated by setting 8 candidate positions for each element, and then randomly selecting one candidate position around each element for combination to obtain a position set X (1), namely a group of solutions of the initial population, which are also called a group of chromosomes. In this way again, a set of Y (Y is set to 100 in the algorithm) positions X (1) -X (100), the initial population Y1, is obtained in total;

(2) Evaluating candidate positions by using four factors of annotation conflict, annotation gland, annotation ambiguity and annotation priority, so as to calculate fitness scores of all chromosomes, and then directly adding the chromosomes with the lowest scores into the next generation population;

(3) Sorting the scoring values of the initial population generated in the step (2), taking the individuals with the scores of the first 50 percent to participate in the next generation reproduction, and jointly generating a new population by a Genetic Algorithm (GA) and a differential evolution algorithm (DDE), wherein 70 percent of chromosomes are generated by the differential evolution algorithm, and 30 percent of chromosomes are generated by the genetic crossover mutation algorithm;

(4) And (3) performing loop iteration, namely outputting the optimal chromosome combination and decoding if the maximum iteration value set by iteration is reached, and entering the step (2) if the maximum iteration value is not reached, and continuing the loop iteration.

The DDEGA algorithm has a good effect on the multi-element map annotation configuration, but also has the problems of long running time and low convergence speed.

Compared with DDEGA algorithm, the method improves the quality of the initial population by adding a local optimal arrangement and combination mode and controlling the random generation range of the initial population;

secondly, changing the proportion of individuals in the new population generated by a difference algorithm and a genetic algorithm, and adding a part of individuals randomly generated by high-quality candidate positions;

Finally, by optimizing DDEGA algorithm with better effect in element map annotation configuration, the initial population quality and the genetic iteration process are controlled, so that the time for searching high-quality annotation candidate position combination mode can be reduced, the efficiency of automatic annotation configuration can be improved, and annotation pressing cover and annotation ambiguity can be reduced to a certain extent.

It should be noted that the foregoing description is only a preferred embodiment of the present invention, and although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood that modifications, equivalents, improvements and modifications to the technical solution described in the foregoing embodiments may occur to those skilled in the art, and all modifications, equivalents, and improvements are intended to be included within the spirit and principle of the present invention.

Claims

1. A DDEGA-NM method for optimizing multi-element map annotation based on two-degree-of-freedom space, characterized by comprising the following steps:

S1. Input experimental vector map data Map0. The map contains different layers of points, lines and surfaces, and each layer has multiple elements.

S2. Output the annotation configuration for each feature of the map. The output result is an Outcome list, which contains the Name, Type, X, and Y data information of the selected annotation candidate location for each feature. Outcome = [Name, Type, X, Y];

S3. Define a one-dimensional list Labels_1 and a two-dimensional list Labels_2. Both Labels_1 and Labels_2 are used to represent the set of all candidate locations of all elements.

S4. Determine the number of candidate locations and the minimum and maximum buffer distances between the annotation and the feature based on the vector map Map0, and then generate candidate locations for the point, line, and area features respectively to obtain a set of candidate locations;

S5. Optimize the DDEGA algorithm to randomly generate the initial population;

S6. Using a loop function, obtain the Name, Type, X and Y coordinate information of each candidate location one by one, and score each candidate location using annotation overlap, annotation ambiguity, and annotation priority factors;

S7. Generating each factor and sorting it in ascending order according to the theoretical limit value of the candidate position. The lower the score value, the higher the priority of the corresponding candidate position.

S8, taking the theoretical limit value combination of the candidate positions as an individual X(1) of the initial population, ensuring that the DDEGA-NM algorithm can retain the individual with the lowest score in each loop iteration;

S9, use the four scoring factors of annotation conflict, annotation overlap, annotation ambiguity, and annotation priority to evaluate the population X(1)-X(100), and directly add the minimum scoring value combination to the next generation population to ensure that the current optimal annotation candidate position combination will not be lost;

S10, sort the scores of the generated initial population X(1)-X(100), select the top 50% of the individuals to participate in the generation of the next generation population, generate 30% and 60% of the new individuals by genetic algorithm and differential evolution algorithm respectively, and then randomly select the candidate position number [1-|N*M/3|] of each element to combine and generate 10% of the individuals, and add them to the next generation new population to form the new population individuals X(1)-X(100);

S11. Determine whether the new population X(1)-X(100) meets the conditions for terminating the iteration, as follows:

If the iteration condition is not met, go to step S9 and start the next loop iteration;

If the termination condition is met, output the new population X(1)-X(100) with the smallest score value X(n);

S12. Decode X(n) to obtain the determined candidate position coordinate information of each element and store it in Outcome, where Outcome = [Name, Type, X, Y].

S13. End, output the annotation configuration result Outcome.

2. The DDEGA-NM method for multi-element map annotation optimization based on two-degree-of-freedom space according to claim 1 is characterized in that: in step S3, the one-dimensional list Labels_1 and the two-dimensional list Labels_2 each contain five fields, namely the name of the element Name, the type of the element Type, the X and Y coordinates of the center point of the annotation candidate location, and the theoretical limit value score F, and Label = [Name, Type, X, Y, F].

3. The DDEGA-NM method for optimizing multi-element map annotations based on a two-degree-of-freedom space according to claim 1, wherein the step S4 of determining the number of candidate locations and the maximum buffer distance between the annotation and the element is as follows:

Set different buffer distance parameters;

By fitting the curve, we find the point where the first-order derivative is 0. This point is the stationary point of the curve, and then we can determine the appropriate maximum buffer distance value.

4. The DDEGA-NM method for optimizing multi-element map annotation based on a two-degree-of-freedom space according to claim 1, wherein the optimized DDEGA algorithm in step S5 randomly generates an initial population:

1) First, the generated candidate locations are scored for annotation overlap, annotation ambiguity, and annotation priority;

2) Then re-sort the candidate locations of each element in ascending order of score;

3) Finally, the quality of the initial population generated by DDEGA is optimized based on the newly sorted candidate positions.

5. The DDEGA-NM method for optimizing multi-element map annotation based on two-degree-of-freedom space according to claim 1, wherein step S6 further comprises:

The score value is represented by F, and the Label = [Name, Type, X, Y, F] data obtained each time is stored in the two-dimensional list Labels_1, which represents all candidate location information of all elements;

Right now:

Labels_1＝[Label _1-1 , Label _1-2 , Label _1-3 ,...Label _1-N*M , Label _2-1 , Label _2-2 ,

Label _2-3 ,...Label _2-N*M ,..., Label _S-1 , Label _S-2 , Label _S-3 ,...Label _SN*M ];

Among them, Label _SN*M represents the N*Mth candidate position coordinates and theoretical limit value score information set of the Sth element.

6. The DDEGA-NM method for optimizing multi-element map annotation based on a two-degree-of-freedom space according to any one of claims 1 or 5, wherein the sorted results are represented by Labels_2:

Labels_2＝[Label _1-[1] , Label _1-[2] , Label _1-[3] ,...Label _1-[N*M] , Label _2-[1] ,

Label _2-[2] , Label _2-[3] ,...Label _2-[N*M] ,..., Label _S-[1] , Label _S-[2] , Label _S-[3] ,...Label _S-[N*M] ];

Among them, Label _S-[N*M] represents the set of candidate position information of the N*Mth theoretical limit value F in the Sth element.

7. A DDEGA-NM method for multi-element map annotation optimization based on two-degree-of-freedom space according to claim 6, characterized in that: the first candidate position Label _S-[1] of each element in Labels_2 is called the theoretical optimal candidate position of the element, and the set formed (Label _1-[1] , Label _2-[1] , Label _3-[1] , ..., Label _S-[1] ) is called the theoretical limit value combination V of candidate positions, which includes the optimal candidate positions of each element in the absence of annotation conflicts and should be considered first in the annotation configuration process.

8. The DDEGA-NM method for multi-element map annotation optimization based on two-degree-of-freedom space according to claim 1 is characterized in that: in step S9, for the other 99 population individuals, candidate positions are no longer randomly generated from the element numbers [1]-[N*M]. Instead, after the candidate position scores of Labels_2 are sorted, the candidate position selection interval is manually controlled, and only the candidate position numbers [1]-[|N _* _M /3|] (|N*M/3| represents the maximum integer not greater than N*M/3) of each element (Label 1 -label S) are randomly selected, thereby optimizing the 99 population individuals.

9. A DDEGA-NM system for multi-element map annotation optimization based on two-degree-of-freedom space, characterized by comprising:

CPU, memory;

The memory is a transient storage memory or a persistent storage memory;

The central processing unit is configured to communicate with the memory and execute instruction operations in the memory on the two-degree-of-freedom space-based multi-element map annotation optimization DDEGA-NM system to perform the method according to any one of claims 1 to 8;

Also includes:

A computer-readable storage medium comprising instructions, which, when executed on a computer, causes the computer to perform the method according to any one of claims 1 to 8.