CN102156710A - Plant identification method based on cloud model and TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method - Google Patents
Plant identification method based on cloud model and TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method Download PDFInfo
- Publication number
- CN102156710A CN102156710A CN 201110048981 CN201110048981A CN102156710A CN 102156710 A CN102156710 A CN 102156710A CN 201110048981 CN201110048981 CN 201110048981 CN 201110048981 A CN201110048981 A CN 201110048981A CN 102156710 A CN102156710 A CN 102156710A
- Authority
- CN
- China
- Prior art keywords
- plant
- cloud model
- trapezoidal
- value
- membership degree
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 62
- 238000004364 calculation method Methods 0.000 claims abstract description 31
- 238000011156 evaluation Methods 0.000 claims abstract description 30
- 241000196324 Embryophyta Species 0.000 claims description 117
- 239000011159 matrix material Substances 0.000 claims description 29
- 238000012821 model calculation Methods 0.000 claims description 4
- 230000001174 ascending effect Effects 0.000 claims description 3
- 238000010835 comparative analysis Methods 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims 1
- 238000012854 evaluation process Methods 0.000 abstract description 2
- 230000000052 comparative effect Effects 0.000 abstract 2
- 238000012163 sequencing technique Methods 0.000 abstract 1
- 235000017166 Bambusa arundinacea Nutrition 0.000 description 27
- 235000017491 Bambusa tulda Nutrition 0.000 description 26
- 235000015334 Phyllostachys viridis Nutrition 0.000 description 26
- 241001330002 Bambuseae Species 0.000 description 22
- 239000011425 bamboo Substances 0.000 description 22
- 230000006870 function Effects 0.000 description 8
- 241000209128 Bambusa Species 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000005054 agglomeration Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 235000013305 food Nutrition 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 241000215338 unidentified plant Species 0.000 description 2
- 244000271437 Bambusa arundinacea Species 0.000 description 1
- 101100494448 Caenorhabditis elegans cab-1 gene Proteins 0.000 description 1
- 235000015366 Dendrocalamus farinosus Nutrition 0.000 description 1
- 241001654577 Dendrocalamus farinosus Species 0.000 description 1
- 241000764238 Isis Species 0.000 description 1
- 241000216509 Schizostachyum Species 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a plant identification method based on a cloud model and a TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method. The method comprises the following steps: constructing a plant shape feature specimen database; utilizing a trapezium-cloud model to compare the shape features of a plant to be identified with the plant shape feature specimen database to acquire a comparative membership between the plant to be identified and the shape feature specimen database, thus completing the primary identification of the plant to be identified; when a plurality of identification results exist, utilizing a normal cloud model to carry out accurate matching calculation on the retrieval results so as to acquire a comparative accurate membership between the plant to be identified and the shape feature specimen database; and comprehensively evaluating the membership by utilizing the TOPSIS method to identify the plant. The method can comprehensively evaluate the final identification result by adopting the TOPSIS method, and can completely, reasonably and accurately carry out advantage and disadvantage sequencing according to certain evaluation indexes, so that the evaluation process is clear, and the evaluation result is objective.
Description
Technical Field
The invention belongs to the technical field of plant identification, and particularly relates to a plant identification method based on a cloud model and a TOPSIS method.
Background
Agricultural databases play an important role in development and utilization of agricultural information resources, and a large number of agricultural databases are built by agricultural organizations at home and abroad currently, wherein the four agricultural databases of AGRIS, IFIS, AGRI-COLA and CABI, namely an agricultural system database (AGRIS) of the food and agriculture organization of the United nations, an international food information database (IFIS), an agricultural online access database (AGRI-COLA) of the United states department of agriculture and an international agricultural biological center database (CAB 1), are the most famous agricultural database systems at home and abroad. A representative plant database in China is as follows: china national agricultural science data sharing center (www.agridata.cn), China digital plant specimen library (http:// www.cvh.org.cn), China plant database (www.plant.csdb.cn), China plant science net (www.chinaplant.org), China biodiversity information system (bd.
The databases contain abundant professional knowledge and store a large amount of agricultural related scientific and technical information, and because the databases are high in speciality and can only be inquired in a keyword retrieval mode, the operation difficulty of an operator on the databases is increased, the use efficiency of the databases is reduced, and the resource waste is caused. It mainly shows that: (1) the requirement for operators is high. Operators using these databases must be able to know the database search interface and grasp the search strategy by using a computer. Meanwhile, the search concepts and search approaches, such as topics, keywords, organizations, full texts, title names, etc., must be understood and mastered. However, users of the plant database include not only field experts and agronomy personnel with strong professional knowledge, but also agricultural people and non-agricultural science and technology workers, and most of the personnel have difficulty in accurately inputting search terms according to requirements; (2) the keyword retrieval mode requires that the retrieval questions must be strictly input according to a specified format, and the retrieval results can be obtained only when the combination is completely matched, and the retrieval mode which is consistent with the retrieval question mark in the aspect of word is difficult to realize that the retrieval results meeting the requirements of users are retrieved in content and concept, so that the recall ratio and precision ratio of the retrieval results are low.
In order to solve the above problems, some researchers use digital image processing technology to realize plant identification, and then query the database according to the identification result, thereby obtaining related professional knowledge.
However, the following problems still exist in the current technology: (1) the current image acquisition is basically realized by a scanner, the background is simple, the image segmentation and description are relatively simple, and the complex background is difficult to process; (2) image databases used by different researchers are different, so that the performance of the recognition effect is difficult to compare, and the image databases are difficult to use in practice; (3) the current image classification system can only process several to hundreds of plants, has a small range, and is difficult to realize large-scale and multi-class plant identification. However, some researchers have proposed:
there are literature reports titled: "uncertainty in knowledge representation" (the author of this article is sudeo et, published in 2000, china engineering science, vol. 2, No. 10, pages 73-79). This article discloses that a cloud model is an uncertainty transformation model between a certain qualitative concept expressed in natural language values and its quantitative representation, which includes:
(1) definition of cloud
Let U be a quantitative discourse domain represented by an accurate numerical value, C be a qualitative concept on U, if the quantitative value is
Is e.g. U, and
is a random implementation of the qualitative concept C,degree of certainty for C
∈[O,1]Is a random number with a tendency to stabilize::U→[O，1]，
,
. Then
The distribution over the universe of discourse is called the cloud model, simply called the cloud, eachReferred to as a cloud droplet.
(2) Digital features of the cloud
Expectations for digital features of cloud model
Entropy of
And entropy
And (4) characterization, which reflects the overall characteristics of the qualitative concept C.
The expectation of the cloud drop in the discourse domain space distribution is that the point which can represent the qualitative concept most and reflects the cloud gravity center of the cloud drop group of the concept.
The method is an uncertainty measure of qualitative concepts, is jointly determined by randomness and ambiguity of the concepts, and reveals the relevance of the ambiguity and the randomness. Entropy of the entropy
On one hand, the method is a measure of randomness of a qualitative concept, and reflects the dispersion degree of cloud droplets capable of representing the qualitative concept; on the other hand, the method is also a measure of qualitative concept ambiguity, and reflects the value range of cloud droplets which can be accepted by the concept in the discourse space. By using
The digital feature reflects both ambiguity and randomness and also reflects the relevance between the ambiguity and the randomness.
Is a measure of the uncertainty in entropy, reflecting the degree of uncertainty agglomerations, i.e., cloud droplet agglomerations, at all points representing the concept in the domain space. The size indirectly represents the degree of dispersion and thickness of the cloud, and is determined by randomness and ambiguity of entropy.
The numerical characteristic of the cloud is that the whole cloud consisting of thousands of cloud droplets is delineated by three numerical values, and the ambiguity and the randomness in the language values of the qualitative representation are completely integrated. Different types of clouds, such as normal cloud model, trapezoidal cloud model, semi-cloud model, etc., are formed due to different specific implementation methods. The cloud model comprises a normal cloud model and a trapezoidal cloud model.
There are also books reported under the book name "Multiple Attribute Decision Making: methods and Application "(the authors of the present Application: C.L. Hwang and Yoon K., published in Berlin spring Press 1981). This publication discloses a method of sorting according to the closeness of a finite number of evaluation objects to an idealized target, discrete approximation to Ideal Solution sorting (Technique for Order Preference by Similarity to Ideal Solution, abbreviated TOPSIS). The principle is that based on a normalized original data matrix, an optimal scheme and a worst scheme in a limited scheme are found out to form a space, an object to be evaluated can be regarded as a point on the space, and accordingly, the distance (common Euclidean distance, also called Euclidean distance) between the point and the optimal scheme and the worst scheme can be obtained, so that the relative proximity degree of the object and the optimal scheme is obtained, and the goodness and the badness of the scheme can be evaluated.
Disclosure of Invention
In view of the problems and deficiencies of the prior art, the present invention provides a plant identification method based on a cloud model and a toposis method, which can conveniently and quickly retrieve a plant to be detected from a plurality of plants in the same group or a database storing a large number of plant samples to realize plant identification.
In order to solve the problems, the invention adopts the following technical scheme:
a plant identification method based on a cloud model and a TOPSIS method is characterized in that the method firstly constructs a shape characteristic sample database of a plant; then, comparing the appearance characteristics of the plant to be detected with the appearance characteristic sample database of the plant by using a trapezoidal cloud model to obtain the membership degree of the plant to be detected compared with the appearance characteristic sample database, thereby realizing the primary identification of the plant to be detected; when a plurality of identification results are obtained, performing accurate matching calculation on the retrieval result by using a normal cloud model to obtain the accurate membership degree of the detected plant compared with the appearance characteristic standard database; and finally, comprehensively evaluating the membership degree by using a TOPSIS method to identify the plants, wherein the method comprises the following specific steps of:
(1) constructing a shape characteristic sample database of the plant;
(2) calculating the membership degree of the plant to be detected by utilizing the trapezoidal cloud model, and preliminarily identifying the plant to be detected;
(3) judging whether the membership degree of the tested plant is less than or equal to 1, if the membership degree of the tested plant is less than or equal to 1, turning to the step (5), and if the membership degree of the tested plant is less than or equal to 1, turning to the step (4);
(4) carrying out normal cloud model calculation on the specimen with the membership degree of 1 and the plant to be detected by using a normal cloud model to obtain the accurate membership degree of the plant to be detected;
(5) and comprehensively evaluating the membership degree by using a TOPSIS method to identify the plant to be detected.
Constructing a database of appearance characteristic samples of plants in the step (1), wherein the database of appearance characteristics of plants is stored as the appearance characteristics of plants which can be used for identifying the types of the plants, and the appearance characteristic values are numerical data.
Calculating the membership degree of the plant to be detected by using the trapezoidal cloud model in the step (2) to preliminarily identify the plant to be detected, wherein the operation steps are as follows:
(21) determining the expected interval of the trapezoidal cloud model
: determining an expected interval of the trapezoidal cloud model according to the appearance characteristic value range in the appearance characteristic sample library of the plant
Wherein
Is the lower limit value of the appearance characteristic,
the upper limit value of the appearance characteristic value;
(22) computing trapezoidal cloud model entropy: expanding the value range of the appearance characteristics of the plants by utilizing the ascending cloud and the descending cloud of the trapezoidal cloud model, wherein the expansion interval is set to be [ -3 ]
,+3]；
(23) And describing the appearance characteristics of the plant to be detected by utilizing the digital characteristics of the trapezoidal cloud model: determining a trapezoidal cloud expectation curve equation through the trapezoidal cloud expectation and the entropy:
wherein,
is the membership degree of the x point of the trapezoidal cloud,
in order to have a desired lower limit of the interval,
in order to have an upper limit of the desired interval,
entropy of trapezoidal clouds;
(24) calculating the membership degree of the tested plant by utilizing a trapezoidal cloud model: respectively carrying out comparative analysis on all appearance characteristics of the tested plant and appearance characteristic values in the specimen library by utilizing a trapezoidal cloud model to obtain each appearance characteristic of the tested plant
。
Performing normal cloud model calculation on the specimen with the membership degree of 1 and the plant to be detected by using the normal cloud model in the step (4) to obtain the accurate membership degree of the plant to be detected, wherein the operation steps are as follows:
(41) determining the expected value of the normal cloud model: determining the expected value of the normal cloud model according to the appearance characteristic values in the appearance characteristic specimen library of the plants
，
The intermediate value of the appearance characteristic interval value;
(42) calculating the entropy of the Normal cloud model
Entropy of
The calculation formula of (A) is as follows:；
(43) the normal cloud expectation curve equation can be determined through the normal cloud expectation and the entropy, and the curve equation is as follows:
wherein,
is the membership degree of the x point of the trapezoidal cloud,
in order to be the desired value,
entropy of trapezoidal clouds;
(44) comparing and analyzing all the appearance characteristics of the tested plant with the specimen of which the membership degree of the tested plant is 1 calculated in the step (2) to obtain the appearance characteristics of each plant
。
The TOPSIS method is utilized to comprehensively evaluate the membership degree to identify the tested plant in the step (5), and the operation steps are as follows:
(51) and establishing an evaluation matrix F of comprehensive evaluation of membership: if only one item or no item of the membership value calculated in the step (2) is equal to 1, evaluating a matrix F as the membership matrix obtained in the step (2); otherwise, evaluating the matrix F as the result obtained by the calculation in the step (3);
(52) determining an ideal point of the evaluation matrix F, wherein the calculation formula of the ideal point is as follows:
wherein,
is an ideal point set, and is characterized in that,
for the value of the jth entry of the ith evaluated solution,
is to find the largest set of objective function numbers,
solving a function set of a minimum target;
(53) and determining the worst point of the evaluation matrix F, wherein the calculation formula of the worst point is as follows:
wherein,is the set of the worst points, and is,
for the value of the jth entry of the ith evaluated solution,
is to find the largest set of objective function numbers,
solving a function set of a minimum target;
(54) and calculating the distance from each evaluated scheme to the ideal point in the evaluation matrix F, wherein the calculation formula is as follows:
wherein,
for the distance of the ith evaluated solution to the ideal point,for the value of the jth entry of the ith evaluated solution,
the j item is the value of an ideal point, and n is the number of evaluated schemes;
(55) and calculating the distance from each evaluated scheme to the worst point in the evaluation matrix F, wherein the calculation formula is as follows:
wherein,
for the distance of the ith evaluated solution to the worst point,
for the value of the jth entry of the ith evaluated solution,
the j item value is the worst point, and n is the number of evaluated schemes;
(56) and calculating the relative closeness of each evaluated scheme in the evaluation matrix F to the ideal point, wherein the calculation formula is as follows:
wherein,
for the relative proximity of the ith evaluated solution to the ideal point,
for the distance of the ith evaluated solution to the worst point,the distance from the ith evaluated scheme to the optimal point is defined, and n is the number of the evaluated schemes;
(57) and relative closeness of each evaluated scheme to an ideal point according to the evaluation matrix F
And (5) sorting the good and the bad of each case.
The above step (22)The entropy of the trapezoid cloud model is calculated
The extension range of the model is 20% of the lower limit value of the appearance characteristics, and then the entropy of the trapezoidal cloud model is obtained
The calculation formula is as follows: entropy of the entropy
=(×0.2)/3。
Compared with the prior art, the plant identification method based on the cloud model and the TOPSIS method has the following effects:
(1) the method identifies the unknown plant by using the extrinsic characteristic value of the plant, so that the problems that the requirement on the professional knowledge of an operator is high and the retrieval result is not matched with the requirement when a keyword is used for retrieving the plant database are effectively solved, the retrieval range is expanded, and the retrieval precision is improved;
(2) the method utilizes the cloud model to carry out digital description on the appearance characteristic information of the plant, realizes the uncertain conversion between the qualitative and quantitative appearance characteristic information of the plant, simultaneously, respectively utilizes different digital characteristics of the trapezoidal cloud model and the normal cloud model to describe the appearance characteristic of the plant, expands the retrieval range and improves the identification effect under the condition of ensuring the identification precision;
(3) the method adopts TOPSIS to comprehensively evaluate the final identification result, can comprehensively, reasonably and accurately rank the advantages and disadvantages of certain evaluation indexes, and has clear evaluation process and objective evaluation result.
Drawings
FIG. 1 is a flow chart of a plant identification method based on the cloud model and TOPSIS method of the present invention;
FIG. 2 is a schematic diagram of a trapezoidal cloud model;
FIG. 3 is a schematic diagram of a normal cloud model;
FIG. 4 is a database of bamboo appearance feature samples;
FIG. 5 is a table of expected value intervals of a trapezoidal cloud model (with bamboo names omitted);
FIG. 6 is a table of shape feature entropy values;
FIG. 7 is a membership degree corresponding to the appearance characteristics of the measured bamboo and the sample database based on a ladder table cloud model;
FIG. 8 is normal cloud model expected values
Table;
FIG. 9 is entropy of normal cloud model
A table of values;
FIG. 10 is a membership degree corresponding to the appearance characteristics of the measured bamboo and the sample database based on the normal cloud model.
Detailed Description
The invention is described in further detail below with reference to the figures and specific embodiments.
This example uses an unknown plant, bamboo, as the identifier.
Referring to fig. 1, the plant identification method based on the cloud model and the TOPSIS method of the present invention includes the following steps:
(1) constructing a shape characteristic sample database of the plant
The sample database is stored as plant appearance characteristics which can be used for distinguishing plant types, the appearance characteristic value is numerical data, and 6 numerical parameters of the bamboo, such as the stalk height, the diameter, the internode length, the stalk wall thickness, the leaf length, the leaf width and the like, are selected as appearance characteristic parameters; then 15 common bamboos are selected from the world bamboos and vines, and a database of appearance characteristic sample of the bamboos is established. See fig. 4.
(2) Calculating the membership degree of the tested plant by utilizing the trapezoidal cloud model to realize the primary identification of the tested plant, wherein the operation steps are as follows:
(21) determining the expected interval of the trapezoidal cloud model
: determining expected value interval according to the data of the bamboo appearance characteristic sample database table 1For example, the expected interval takes the value of the interval [1 ~ 5 ]]I.e., expressed as,=1，=5, the conversion result is shown in fig. 5.
(22) Computing trapezoidal cloud model entropy
Expanding the value range of the appearance characteristics of the plants by utilizing the ascending cloud and the descending cloud of the trapezoidal cloud model, wherein the expansion interval is set to be [ -3 ]
,+3
]Extended range of topographical features20% of the lower limit value, so the trapezoidal cloud model entropy
Calculating formula: entropy of the entropy
=(
X 0.2)/3. Respectively calculating the entropy of the appearance characteristics of each plant to be detected
The results are shown in FIG. 6.
(23) And describing the appearance characteristics of the plant to be detected by utilizing the digital characteristics of the trapezoidal cloud model: determining a trapezoidal cloud expectation curve equation through the trapezoidal cloud expectation and the entropy:
wherein,the membership degree of a trapezoidal cloud x point;
For example, trapezoidal cloud model description is carried out on a first item in a bamboo appearance characteristic sample database, the height of a bamboo stalk is 1-5 m, and the numerical characteristics of the bamboo appearance characteristic described by the trapezoidal cloud model are as follows:
the schematic diagram of the trapezoidal cloud model can be referred to fig. 2, wherein an X axis X in fig. 2 represents the height of a bamboo stalk, and the unit is meter; y-axis
Expressed as degree of membership
(ii) a Wherein,
in order to have a desired lower limit of the interval,
in order to have an upper limit of the desired interval,
is the entropy of the trapezoidal cloud.
(24) Calculating the membership degree of the tested plant by utilizing a trapezoidal cloud model: according to the calculated expected value interval
And entropy
And respectively calculating the membership degree of each appearance characteristic of the detected bamboo by using the digital characteristics of the detected plant described by the trapezoidal cloud, as shown in figure 7.
(3) Judging whether the membership degree of the tested plant is less than or equal to 1, if the membership degree of the tested plant is less than or equal to 1, turning to the step (5), and if the membership degree of the tested plant is less than or equal to 1, turning to the step (4);
(4) the method comprises the following steps of (1) utilizing a normal cloud model to calculate the normal cloud model of a specimen with the membership degree of 1 and a tested plant, obtaining that the membership degrees of the tested bamboo compared with three kinds of bamboos, namely 8, 11 and 14, are all 1 through the step (2), and identifying the tested bamboo with the three kinds of bamboos, namely 8, 11 and 14 by utilizing the normal cloud model, wherein the method comprises the following specific steps:
(41) determining expected values for a normal cloud model
Determining the expected value of the normal cloud model according to the appearance characteristic values in the appearance characteristic specimen library of the plants
，
The expected value of the normal cloud model is obtained as the middle value of the appearance feature interval value, as shown in fig. 8.
(42) Entropy of a computational normal cloud model
The calculation formula is as follows:according to the formula [ -3
,+3
]The entropy of the elements in between exceeds 99.73 percent
The calculation results are shown in fig. 9.
(43) The normal cloud expectation curve equation can be determined through the normal cloud expectation and the entropy, and the curve equation is as follows:
wherein,
the membership degree of a trapezoidal cloud x point;
is a desired value;
entropy of trapezoidal clouds;
for example, the height of the bamboo is 1-5 m, the appearance characteristics of the bamboo are described by trapezoidal clouds, and the numerical characteristics are as follows:
the schematic diagram of the normal cloud model can be referred to fig. 3, wherein an X axis X in fig. 3 represents the culm height of bamboo, and the unit is meter; y-axisExpressed as degree of membership
(ii) a Wherein,
in order to be the desired value,
is the entropy of the trapezoidal cloud.
(44) Comparing and analyzing all the appearance characteristics of the tested plant with the specimen of which the membership degree of the tested plant is 1 calculated in the step (1) to obtain the appearance characteristics of each plant
。
Based on the calculated expected value
And entropyAnd respectively calculating the membership degree of each appearance characteristic of the bamboo to be detected by utilizing the digital characteristics described by the trapezoidal clouds, wherein the result is shown in figure 10.
(5) Comprehensively evaluating the membership degree by using a TOPSIS method to identify the tested plant, and specifically comprising the following steps of:
(51) and determining an evaluation matrix F for comprehensive evaluation of membership
If the membership degree calculated in the step (2) has 8, 11 and 14 items which are all 1, the evaluation matrix F is the calculation result of the step (3), namely
(52) Determining an ideal point of the evaluation matrix F, wherein the calculation formula of the ideal point is as follows:
wherein, the point set is an ideal point set,
is the value of the jth item of the ith evaluated scenario, J is the largest set of objective function numbers,
the function set of the minimum target is solved, and the calculation result is as follows:
(53) and determining the worst point of the evaluation matrix F, wherein the calculation formula of the worst point is as follows:
wherein,is the set of the worst points, and is,
is the value of the jth item of the ith evaluated scenario, J is the largest set of objective function numbers,
is the set of functions for solving the minimum objective, which is calculated if:
(54) and calculating the distance from each evaluated scheme to the ideal point in the evaluation matrix F, wherein the calculation formula is as follows:
wherein,
for the distance of the ith evaluated solution to the ideal point,
for the value of the jth entry of the ith scheme,
the j is the value of the ideal point, n is the number of evaluated schemes, and the calculation result is as follows:
(55) and calculating the distance from each evaluated scheme to the worst point in the evaluation matrix F, wherein the calculation formula is as follows:
wherein,
for the distance of the ith evaluated solution to the worst point,
for the value of the jth entry of the ith scheme,
the j item value is the worst point, n is the number of evaluated schemes, and the calculation result is as follows:
(56) and calculating the relative closeness of each evaluated scheme in the evaluation matrix F to the ideal point, wherein the calculation formula is as follows:
wherein,
relative proximity to ideal points for the ith evaluated solution;
for the distance of the ith evaluated solution to the worst point,
the distance from the ith evaluated scheme to the optimal point is defined, n is the number of evaluated schemes, and the calculation result is as follows:
said relative proximity
Is a value between 0 and 1, for an "ideal point"
In other words, their relative proximity
Is 1, to the "worst point"
In other words, their relative proximity
Is 0, and thus relative proximity
The larger, the closer the point illustrating the scheme is to the most ideal point, the better; on the contrary, relative proximity
The smaller the points illustrating the scheme, the closer to the worst point, the worse.
(57) And relative closeness of each evaluated scheme to an ideal point according to the evaluation matrix F
And (5) sorting the good and the bad of each case.
Relative proximity of step (56)
It is known that the identification of the bamboo to be tested is most likely to be bamboo 14 (Schizostachyum brachycanum), followed by bamboo 8 (Dendrocalamus farinosus) and bamboo 11 (giganticochloa atroviolacea).
The method of the present invention is not limited to the examples described in the specific embodiments, and other embodiments derived from the technical solutions of the present invention by those skilled in the art also belong to the technical innovation scope of the present invention.
Claims (6)
1. A plant identification method based on a cloud model and a TOPSIS method is characterized in that the method firstly constructs a shape characteristic sample database of a plant; then, comparing the appearance characteristics of the plant to be detected with the appearance characteristic sample database of the plant by using a trapezoidal cloud model to obtain the membership degree of the plant to be detected compared with the appearance characteristic sample database, thereby realizing the primary identification of the plant to be detected; when a plurality of identification results are obtained, performing accurate matching calculation on the retrieval result by using a normal cloud model to obtain the accurate membership degree of the detected plant compared with the appearance characteristic standard database; and finally, comprehensively evaluating the membership degree by using a TOPSIS method to identify the plants, wherein the method comprises the following specific steps of:
(1) constructing a shape characteristic sample database of the plant;
(2) calculating the membership degree of the plant to be detected by utilizing the trapezoidal cloud model, and preliminarily identifying the plant to be detected;
(3) judging whether the membership degree of the tested plant is less than or equal to 1, if the membership degree of the tested plant is less than or equal to 1, turning to the step (5), and if the membership degree of the tested plant is less than or equal to 1, turning to the step (4);
(4) carrying out normal cloud model calculation on the specimen with the membership degree of 1 and the plant to be detected by using a normal cloud model to obtain the accurate membership degree of the plant to be detected;
(5) and comprehensively evaluating the membership degree by using a TOPSIS method to identify the plant to be detected.
2. The method for identifying plants based on cloud model and TOPSIS method as claimed in claim 1, wherein said step (1) is to construct a database of appearance characteristic samples of plants, which is stored as appearance characteristics of plants for identifying plant species, and the appearance characteristic values are numerical data.
3. The method for identifying plants based on cloud model and TOPSIS method as claimed in claim 1, wherein the step (2) of calculating the membership degree of the plant to be detected by using the trapezoidal cloud model to preliminarily identify the plant to be detected comprises the following steps:
(21) determining the expected interval of the trapezoidal cloud model
: determining an expected interval of the trapezoidal cloud model according to the appearance characteristic value range in the appearance characteristic sample library of the plant
Wherein
Is the lower limit value of the appearance characteristic,
the upper limit value of the appearance characteristic value;
(22) computing trapezoidal cloud model entropy
: expanding the value range of the appearance characteristics of the plants by utilizing the ascending cloud and the descending cloud of the trapezoidal cloud model, wherein the expansion interval is set to be [ -3 ],+3
]；
(23) And describing the appearance characteristics of the plant to be detected by utilizing the digital characteristics of the trapezoidal cloud model: determining a trapezoidal cloud expectation curve equation through the trapezoidal cloud expectation and the entropy:
wherein,
is the membership degree of the x point of the trapezoidal cloud,
period of time ofThe lower limit of the inspection interval is reached,
in order to have an upper limit of the desired interval,
entropy of trapezoidal clouds;
(24) calculating the membership degree of the tested plant by utilizing a trapezoidal cloud model: respectively carrying out comparative analysis on all appearance characteristics of the tested plant and appearance characteristic values in the specimen library by utilizing a trapezoidal cloud model to obtain each appearance characteristic of the tested plant
。
4. The method for identifying plants based on cloud model and TOPSIS method according to claim 1, characterized in that said normal cloud model is used in step (4) to perform normal cloud model calculation on the specimen with membership degree of 1 and the plant to be tested to obtain the accurate membership degree of the plant to be tested, and the operation steps are as follows:
(41) determining the expected value of the normal cloud model: determining the expected value of the normal cloud model according to the appearance characteristic values in the appearance characteristic specimen library of the plants，The intermediate value of the appearance characteristic interval value;
(42) calculating the entropy of the Normal cloud model
Entropy of
The calculation formula of (A) is as follows:；
(43) the normal cloud expectation curve equation can be determined through the normal cloud expectation and the entropy, and the curve equation is as follows:
wherein,
is the membership degree of the x point of the trapezoidal cloud,
in order to be the desired value,
entropy of trapezoidal clouds;
(44) comparing and analyzing all the appearance characteristics of the tested plant with the specimen of which the membership degree of the tested plant is 1 calculated in the step (2) to obtain the appearance characteristics of each plant
。
5. The method for identifying plants based on cloud model and TOPSIS method as claimed in claim 1, wherein the TOPSIS method is used in the step (5) to comprehensively evaluate the membership degree to identify the tested plants, and the operation steps are as follows:
(51) and establishing an evaluation matrix F of comprehensive evaluation of membership: if only one item or no item of the membership value calculated in the step (2) is equal to 1, evaluating a matrix F as the membership matrix obtained in the step (2); otherwise, evaluating the matrix F as the result obtained by the calculation in the step (3);
(52) determining an ideal point of the evaluation matrix F, wherein the calculation formula of the ideal point is as follows:
wherein,
is an ideal point set, and is characterized in that,
for the value of the jth entry of the ith evaluated solution,is to find the largest set of objective function numbers,
solving a function set of a minimum target;
(53) and determining the worst point of the evaluation matrix F, wherein the calculation formula of the worst point is as follows:
wherein,
is the set of the worst points, and is,
for the value of the jth entry of the ith evaluated solution,
is to find the largest set of objective function numbers,
solving a function set of a minimum target;
(54) and calculating the distance from each evaluated scheme to the ideal point in the evaluation matrix F, wherein the calculation formula is as follows:
wherein,
for the distance of the ith evaluated solution to the ideal point,
for the value of the jth entry of the ith evaluated solution,
the j item is the value of an ideal point, and n is the number of evaluated schemes;
(55) and calculating the distance from each evaluated scheme to the worst point in the evaluation matrix F, wherein the calculation formula is as follows:
wherein,
for the distance of the ith evaluated solution to the worst point,for the value of the jth entry of the ith evaluated solution,
the j item value is the worst point, and n is the number of evaluated schemes;
(56) and calculating the relative closeness of each evaluated scheme in the evaluation matrix F to the ideal point, wherein the calculation formula is as follows:
wherein,
for the relative proximity of the ith evaluated solution to the ideal point,
for the distance of the ith evaluated solution to the worst point,
the distance from the ith evaluated scheme to the optimal point is defined, and n is the number of the evaluated schemes;
(57) and relative closeness of each evaluated scheme to an ideal point according to the evaluation matrix F
And (5) sorting the good and the bad of each case.
6. The method for identifying plants based on cloud model and TOPSIS method as claimed in claim 5, wherein said step (22) of calculating trapezoidal cloud model entropy
The extension range of the model is 20% of the lower limit value of the appearance characteristics, and then the entropy of the trapezoidal cloud model is obtained
The calculation formula is as follows: entropy of the entropy=(
×0.2)/3。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110048981 CN102156710A (en) | 2011-03-02 | 2011-03-02 | Plant identification method based on cloud model and TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110048981 CN102156710A (en) | 2011-03-02 | 2011-03-02 | Plant identification method based on cloud model and TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102156710A true CN102156710A (en) | 2011-08-17 |
Family
ID=44438209
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201110048981 Pending CN102156710A (en) | 2011-03-02 | 2011-03-02 | Plant identification method based on cloud model and TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102156710A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740635A (en) * | 2016-02-03 | 2016-07-06 | 王永林 | Cloud ideal solution evaluation method for transformer electromagnetic design scheme |
| CN109874584A (en) * | 2019-03-19 | 2019-06-14 | 广州辰轩农业科技有限公司 | A kind of fruit tree growing way monitoring system based on deep learning convolutional neural networks |
| CN112580493A (en) * | 2020-12-16 | 2021-03-30 | 广东省林业科学研究院 | Plant identification method, device and equipment based on unmanned aerial vehicle remote sensing and storage medium |
| CN113627735A (en) * | 2021-07-16 | 2021-11-09 | 宁夏建设投资集团有限公司 | Early warning method and system for safety risk of engineering construction project |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101344929A (en) * | 2008-08-20 | 2009-01-14 | 盛秀英 | Garden tree classification method |
| EP2133822A1 (en) * | 2008-06-13 | 2009-12-16 | University College Cork | A method of stem taper, volume and product breakout prediction |
| CN101697167A (en) * | 2009-10-30 | 2010-04-21 | 邱建林 | Clustering-decision tree based selection method of fine corn seeds |
| US20100332475A1 (en) * | 2009-06-25 | 2010-12-30 | University Of Tennessee Research Foundation | Method and apparatus for predicting object properties and events using similarity-based information retrieval and modeling |
-
2011
- 2011-03-02 CN CN 201110048981 patent/CN102156710A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2133822A1 (en) * | 2008-06-13 | 2009-12-16 | University College Cork | A method of stem taper, volume and product breakout prediction |
| CN101344929A (en) * | 2008-08-20 | 2009-01-14 | 盛秀英 | Garden tree classification method |
| US20100332475A1 (en) * | 2009-06-25 | 2010-12-30 | University Of Tennessee Research Foundation | Method and apparatus for predicting object properties and events using similarity-based information retrieval and modeling |
| CN101697167A (en) * | 2009-10-30 | 2010-04-21 | 邱建林 | Clustering-decision tree based selection method of fine corn seeds |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105740635A (en) * | 2016-02-03 | 2016-07-06 | 王永林 | Cloud ideal solution evaluation method for transformer electromagnetic design scheme |
| CN105740635B (en) * | 2016-02-03 | 2018-01-30 | 中原工学院 | A kind of cloud ideal solution evaluation method of transformer electromagnetic design scheme |
| CN109874584A (en) * | 2019-03-19 | 2019-06-14 | 广州辰轩农业科技有限公司 | A kind of fruit tree growing way monitoring system based on deep learning convolutional neural networks |
| CN112580493A (en) * | 2020-12-16 | 2021-03-30 | 广东省林业科学研究院 | Plant identification method, device and equipment based on unmanned aerial vehicle remote sensing and storage medium |
| CN112580493B (en) * | 2020-12-16 | 2021-11-09 | 广东省林业科学研究院 | Plant identification method, device and equipment based on unmanned aerial vehicle remote sensing and storage medium |
| CN113627735A (en) * | 2021-07-16 | 2021-11-09 | 宁夏建设投资集团有限公司 | Early warning method and system for safety risk of engineering construction project |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Truong et al. | Housing price prediction via improved machine learning techniques | |
| CN107577688B (en) | Original article influence analysis system based on media information acquisition | |
| Huang et al. | Revealing density-based clustering structure from the core-connected tree of a network | |
| Clarke et al. | Clustering in non-parametric multivariate analyses | |
| CN107016068A (en) | Knowledge mapping construction method and device | |
| Gustafsson et al. | Comparison and validation of community structures in complex networks | |
| CN102663447B (en) | Cross-media searching method based on discrimination correlation analysis | |
| Zhou et al. | Relevance feature mapping for content-based multimedia information retrieval | |
| CN110717806B (en) | Product information pushing method, device, equipment and storage medium | |
| Novaliendry et al. | The optimized K-means clustering algorithms to analyzed the budget revenue expenditure in Padang | |
| CN106611016B (en) | A kind of image search method based on decomposable word packet model | |
| CN110347701A (en) | A kind of target type identification method of entity-oriented retrieval and inquisition | |
| CN111126865A (en) | Technology maturity judging method and system based on scientific and technological big data | |
| CN109359135A (en) | A kind of Algorithm for Similarity Search in Time Series based on segmentation weight | |
| CN102156710A (en) | Plant identification method based on cloud model and TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method | |
| CN109583712B (en) | Data index analysis method and device and storage medium | |
| JP2013257668A (en) | Interest analysis method, interest analyzer and program of the same | |
| Kobayashi et al. | Modeling collective anticipation and response on Wikipedia | |
| CN117114105B (en) | Target object recommendation method and system based on scientific research big data information | |
| Ni et al. | The analysis and research of clustering algorithm based on PCA | |
| Hassan et al. | Sampling technique selection framework for knowledge discovery | |
| Wang et al. | ApproxCCA: An approximate correlation analysis algorithm for multidimensional data streams | |
| Zheng | Individualized Recommendation Method of Multimedia Network Teaching Resources Based on Classification Algorithm in a Smart University | |
| Barret et al. | Predicting the Environment of a Neighborhood: A Use Case for France. | |
| CN113868493A (en) | Visual chart recommendation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20110817 |