CN106570512A - Forest subcompartment classification method - Google Patents

Forest subcompartment classification method Download PDF

Info

Publication number
CN106570512A
CN106570512A CN201510653565.4A CN201510653565A CN106570512A CN 106570512 A CN106570512 A CN 106570512A CN 201510653565 A CN201510653565 A CN 201510653565A CN 106570512 A CN106570512 A CN 106570512A
Authority
CN
China
Prior art keywords
classification
forest
classes
cluster
subcompartment
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
Application number
CN201510653565.4A
Other languages
Chinese (zh)
Inventor
冯仲科
马力
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Forestry University
Original Assignee
Beijing Forestry University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Forestry University filed Critical Beijing Forestry University
Priority to CN201510653565.4A priority Critical patent/CN106570512A/en
Publication of CN106570512A publication Critical patent/CN106570512A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Sorting Of Articles (AREA)

Abstract

The invention relates to a forest subcompartment classification method. The method comprises the steps: enabling a clustering analysis idea and method to be applied in the subcompartment division research; directly comparing the qualities in dominant tree species (groups) with the consistent operation type and standing type or between samples; searching the growth dominant factors (standing conditions) of a forest subcompartment, and enabling the growth dominant factors (standing conditions) to serve as clustering factors, wherein the growth dominant factors (standing conditions) comprise the topographic conditions (slope, exposure and slope position) and soil conditions (soil thickness and humus thickness); employing the shortest distance clustering algorithm (shown in the description), wherein dkr, dpk and dqk are the distances between different clusters; taking the forest subcompartment as the classification object, and dividing all to-be-clustered forest subcompartment data in different classification modes or according to different classification numbers (two classes, three classes, four classes,..., or more classes); selecting the classification mode or classification number which enables the number of forest subcompartments in different classes are close to each other through combining with the factors that the internal difference in the same cluster is small and the difference between different clusters is remarkable; describing the features of forest subcompartments in each class, and guiding the actual production.

Description

A kind of forest bottom class sorting technique
First, technical field
The present invention relates to a kind of forest bottom class sorting technique, the sorting technique of particularly a kind of dominant tree (group) forest bottom class consistent for working group, site type.
2nd, technical background
Up to the present, the classification Main Basiss forest inventory investigation of forest bottom class and monitoring, and conventional model modeling method, still without effective, ripe method.In the past transmission method can cause larger error：
1. the error that control time causes.Fact-finding process is lasting long-drawn-out, and investigator causes the elementary errors of thickness of soil according to before and after investigation order.
2. observation position is forbidden the error for causing.Due to landform, the impact of understory shrub, it is impossible to humus thickness is accurately measured, so as to produce elementary errors.
3. the error that Silvicultural Measures difference causes.Conventional sorting methods can cause Classification Change using every wooden dipping factor as class condition, seeds difference different because of Silvicultural Measures.
Therefore, in forest bottom class categorizing process, many troubles and error are inevitably brought using conventional sorting methods.
3rd, the content of the invention
For many troubles for overcoming existing sorting technique inevitably to bring in forest bottom class categorizing process again and errors, it is an object of the invention to provide a kind of rational cluster analyses thought and its using method.
The object of the present invention is achieved like this：
1. by cluster analyses thought and approach application into forest bottom class Research on partition, directly compare working group, in the consistent dominant tree of site type (group) or between sample, compare the distance between property (cluster key element) of each forest bottom class growth-dominated, property close (distance is near) is classified as into a class, nature difference larger (distance remote) is classified as inhomogeneity, take appropriate mode classification or classification number, so that same intra-cluster variants are as little as possible, and difference is as notable as possible between different clusters, and forest bottom class number contained by different clusters is substantially approximate, and then propose suitable Silvicultural Measures suggestion according to classification.
2. clustering method is as follows：1. it is sample (object of classification) to choose the consistent dominant tree (group) of working group, site type；2. the data such as forest bottom class's growth-dominated factor (land occupation condition), including orographic condition (gradient, slope aspect, slope position), edaphic condition (thickness of soil, humus thickness) are investigated；3. standard deviation Standardization Act is adopted, above-mentioned cluster key element is standardized；4. the Euclidean distance between key element is clustered after normalized is processed；5. according to the distance of cluster key element Euclidean distance, take beeline clustering procedure that all forest bottom classes to be clustered are carried out into clustering processing, and compare different classifications mode or classification number (being divided to two classes, three classes, four classes etc.) gained cluster result, the substantially close mode classification of forest bottom class number or classification number in different clusters are chosen as far as possible；6. forest bottom class feature in every kind of classification is described, actual production is and guided.
This invention has advantages below：
1. forest bottom class somatomedin using orographic condition, edaphic condition as division, it is to avoid because of the error that Silvicultural Measures are caused；
2. in terms of classification, cluster analyses acquired results are so that belonging to the distance between any inhomogeneous two data objects is all higher than in same class the distance between two data objects；
4th, specific embodiment：
Forest bottom class sorting technique is different from conventional sorting methods, has made significant improvement, specifically：
1) cluster analyses thought and approach application are into forest bottom class Research on partition, directly compare in working group, site type consistent dominant tree (group) or between sample, count the property of each forest bottom class growth-dominated and as cluster key element, after taking standard deviation Standardization Act to process individual cluster key element, calculate its related Euclidean distance, using beeline clustering procedure (Wherein dkr、dpk、dqkThe distance between respectively different clusters) all forest bottom classes to be clustered are divided, so that property is similar, (distance is near) is classified as a class, nature difference larger (distance remote) is classified as inhomogeneity, to reach same intra-cluster variants very little, significant difference between difference cluster, and the substantially close cluster result of different classes of middle forest bottom class number, forest bottom class feature in every kind of classification is described after the completion of classification, and then proposes that suitable Silvicultural Measures suggestion and guides actual production according to classification.
2) forest bottom class sorting technique：Forest bottom class sorting technique is as follows：1. it is sample (object of classification) to choose the consistent dominant tree (group) of working group, site type；2. forest bottom class's growth-dominated factor (land occupation condition), including orographic condition (gradient, slope aspect, slope position), edaphic condition (thickness of soil, humus thickness) etc. are investigated as cluster key element；3. standard deviation standardized algorithm is adopted, above-mentioned cluster key element is standardized；4. the Euclidean distance between cluster key element after normalized is processed, as subsequent classification foundation；5. according to the distance of above-mentioned cluster key element, beeline clustering procedure is taken to be divided all forest bottom classes to be clustered so that same intra-cluster variants very little, significant difference between difference cluster；6. compare different classifications mode or classification number (being divided to two classes, three classes, four classes etc.) gained cluster result, the substantially close mode classification of forest bottom class number or classification number in different clusters are chosen as far as possible；7. final classification mode or classification number are determined, and confirms cluster result；8. forest bottom class feature in every kind of classification is described, actual production is and guided.
3) formula for using in method：
Standard deviation standardization：Wherein xi'jFor the variate-value after standardization, xijFor real variable value,For initial data meansigma methodss, sjFor initial data standard deviation；
Euclidean distance：Wherein dijFor the Euclidean distance after calculating, xikFor the observed value of k-th index of i-th sample, xjkFor the observed value of k-th index of j-th sample；
Beeline clustering procedure：Wherein dkr、dpk、dqkThe distance between respectively different clusters.

Claims (2)

1. a kind of forest bottom class sorting technique, is characterized in that：By cluster analyses thought and approach application to forest bottom class Research on partition In, directly compare in working group, site type consistent dominant tree (group) or between sample, each forest bottom class growth master The property led, is classified as a class by property is akin, nature difference it is larger be classified as inhomogeneity, Different Forest bottom class is divided into Inhomogeneity, not Shi get same intra-cluster variants very little, significant difference between difference cluster, so propose according to classification it is mutually suitable Silvicultural Measures suggestion.
2. a kind of forest bottom class sorting technique realized described in claim 1, is characterized in that：1. working group, on the spot class are chosen The consistent dominant tree of type (group) is sample (object of classification)；2. forest bottom class's growth-dominated factor (land occupation condition) is investigated, Including orographic condition (gradient, slope aspect, slope position), edaphic condition (thickness of soil, humus thickness) etc. as cluster key element； 3. standard deviation standardization is adopted ( x i j ′ = x i j - x ‾ j s j ( i = 1 , 2 , ... , m ; j = 1 , 2 , ... , n ) , Wherein x 'ijFor the variate-value after standardization, xijFor real variable value,For initial data meansigma methodss, sjFor initial data standard deviation) algorithm enters rower to above-mentioned cluster key element Quasi-ization process；4. the Euclidean distance between cluster key element after normalized is processed Wherein dijFor the Euclidean distance after calculating, xikFor the observed value of k-th index of i-th sample, xjk For the observed value of k-th index of j-th sample), as subsequent classification foundation；5. the distance according to above-mentioned cluster key element is remote Closely, take beeline clustering procedure (Wherein dkr、dpk、dqkRespectively different clusters Between distance) all forest bottom classes to be clustered are divided so that same intra-cluster variants very little, it is poor between difference cluster It is different notable；6. compare different classifications mode or classification number (being divided to two classes, three classes, four classes etc.) gained cluster result, select as far as possible Take the substantially close mode classification of forest bottom class number or classification number in different clusters；7. final classification mode or classification are determined Number, and confirm cluster result；8. forest bottom class feature in every kind of classification is described, actual production is and guided.
CN201510653565.4A 2015-10-12 2015-10-12 Forest subcompartment classification method Pending CN106570512A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510653565.4A CN106570512A (en) 2015-10-12 2015-10-12 Forest subcompartment classification method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510653565.4A CN106570512A (en) 2015-10-12 2015-10-12 Forest subcompartment classification method

Publications (1)

Publication Number Publication Date
CN106570512A true CN106570512A (en) 2017-04-19

Family

ID=58506718

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510653565.4A Pending CN106570512A (en) 2015-10-12 2015-10-12 Forest subcompartment classification method

Country Status (1)

Country Link
CN (1) CN106570512A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108921194A (en) * 2018-05-31 2018-11-30 河海大学 A kind of adaptive clustering method of landform slope position classification
CN110020961A (en) * 2019-01-18 2019-07-16 北京林业大学 Chinese main arbor species plot/bottom class's sub-index grinds the technical method built
CN110197343A (en) * 2019-06-12 2019-09-03 中国林业科学研究院资源信息研究所 A kind of matching process and system of Forest management types and bottom class
CN111079221A (en) * 2019-12-25 2020-04-28 北京林业大学 Road slope land standing type dividing method and device

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101398317A (en) * 2008-10-28 2009-04-01 崔国发 Method for testing conifer forest community situation
US8396293B1 (en) * 2009-12-22 2013-03-12 Hrl Laboratories, Llc Recognizing geometrically salient objects from segmented point clouds using strip grid histograms
CN103268613A (en) * 2013-05-28 2013-08-28 北京林业大学 Method for detecting sub-compartment forest resource by remote sensing and geographical information system technology

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101398317A (en) * 2008-10-28 2009-04-01 崔国发 Method for testing conifer forest community situation
US8396293B1 (en) * 2009-12-22 2013-03-12 Hrl Laboratories, Llc Recognizing geometrically salient objects from segmented point clouds using strip grid histograms
CN103268613A (en) * 2013-05-28 2013-08-28 北京林业大学 Method for detecting sub-compartment forest resource by remote sensing and geographical information system technology

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
张晶晶: "渭北黄土高原刺槐林健康评价研究", 《中国优秀硕士学位论文全文数据库 农业科技辑》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108921194A (en) * 2018-05-31 2018-11-30 河海大学 A kind of adaptive clustering method of landform slope position classification
CN108921194B (en) * 2018-05-31 2021-10-08 河海大学 Terrain slope classification self-adaptive clustering method
CN110020961A (en) * 2019-01-18 2019-07-16 北京林业大学 Chinese main arbor species plot/bottom class's sub-index grinds the technical method built
CN110197343A (en) * 2019-06-12 2019-09-03 中国林业科学研究院资源信息研究所 A kind of matching process and system of Forest management types and bottom class
CN110197343B (en) * 2019-06-12 2021-08-20 中国林业科学研究院资源信息研究所 Forest management type and shift matching method and system
CN111079221A (en) * 2019-12-25 2020-04-28 北京林业大学 Road slope land standing type dividing method and device

Similar Documents

Publication Publication Date Title
CN106197424B (en) The unmanned plane during flying state identification method of telemetry driving
Zheng et al. Remote sensing and machine learning in crop phenotyping and management, with an emphasis on applications in strawberry farming
CN105630743B (en) A kind of system of selection of spectrum wave number
CN106570512A (en) Forest subcompartment classification method
CN105740759B (en) Semilate rice information decision tree classification approach based on feature extraction in multi-temporal data
CN107766883A (en) A kind of optimization random forest classification method and system based on weighted decision tree
CN104102929A (en) Hyperspectral remote sensing data classification method based on deep learning
CN109522924A (en) A kind of broad-leaf forest wood recognition method based on single photo
Ismail et al. Discriminating the early stages of Sirex noctilio infestation using classification tree ensembles and shortwave infrared bands
CN103839078B (en) A kind of hyperspectral image classification method based on Active Learning
CN105117734B (en) Corn seed classification hyperspectral imagery recognition methods based on model online updating
CN109800712B (en) Vehicle detection counting method and device based on deep convolutional neural network
CN108090628A (en) A kind of grain feelings security detection and analysis method based on PSO-LSSVM algorithms
An et al. Real-time monitoring method of strawberry fruit growth state based on YOLO improved model
CN111241939A (en) Rice yield estimation method based on unmanned aerial vehicle digital image
CN109165672A (en) A kind of Ensemble classifier method based on incremental learning
CN110263735A (en) A method of tree species classification being carried out to artificial forest high-spectral data using Three dimensional convolution neural network
CN110210433A (en) A kind of container number detection and recognition methods based on deep learning
CN106570537A (en) Random forest model selection method based on confusion matrix
CN107403205A (en) A kind of RFID warehouses package plain location method based on random forest
Cuong et al. Improved YOLO object detection algorithm to detect ripe pineapple phase
CN109063750A (en) SAR target classification method based on CNN and SVM decision fusion
Krongchai et al. Application of multiple self‐organizing maps for classification of soil samples in Thailand according to their geographic origins
CN114662621B (en) Agricultural machinery working area calculation method and system based on machine learning
Araújo et al. Appropriate search techniques to estimate Weibull function parameters in a Pinus spp. plantation

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20170419

WD01 Invention patent application deemed withdrawn after publication