About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms

Anatolii Kuzmin; Leonid Grekov; Nataliia Kuzmina; Oleksii Petrov

doi:doi:10.11648/j.ajrs.20221002.11

| Peer-Reviewed

About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms

Anatolii Kuzmin, Leonid Grekov, Nataliia Kuzmina, Oleksii Petrov

Published in American Journal of Remote Sensing (Volume 10, Issue 2)

Received: 10 August 2022 Accepted: 29 August 2022 Published: 5 September 2022

Views: Downloads:

Download PDF

Share This Article

Twitter
Linked In
Facebook

Abstract

Applied problems of studying the earth's surface using satellite images of remote sensing of the Earth are considered for the study of forest, agricultural, water and other natural resources, where clustering and classification algorithms are instrumental research methods. It is noted that the most well-known procedures for classifying and segmenting multispectral space images in GIS systems, such as ArcGIS, ERDAS, ENVI, are built-in. The need to expand the toolkit for a more efficient solution of applied problems of this class is noted. New universal clustering and classification algorithms based on a unified approach are proposed. Both methods belong to grid-type algorithms, and at the first stage of their work they group points of a set of n - dimensional vectors into grid cells, each cell saves only the numbers of points belonging to it and is characterized by a unique code. The vector grid spacing is a parameter of the method and is set by the user using a single integer value. At the next stage, the clustering algorithm combines the cells and the points belonging to them into clusters using the cell neighborhood principle. In this case, the algorithm does not attach the next cell to the cluster in the case when its density is less than the specified value. The classification algorithm refers the points of the cell of the main set to the class to which the cell with the same code of the training set belongs. The algorithms can be used to process large data sets of large spatial dimensions, including satellite images. Clustering and classification algorithms do not require a preliminary specification of the number of clusters and information about the nature of the distribution of points in the input set.

Published in	American Journal of Remote Sensing (Volume 10, Issue 2)
DOI	10.11648/j.ajrs.20221002.11
Page(s)	30-38
Creative Commons	This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.
Copyright	Copyright © The Author(s), 2022. Published by Science Publishing Group

Keywords

Remote Sensing Methods, Images Segmentation, Clustering Algorithms, Classification Algorithms, Grid Methods, Neighborhood Relation, Cell Density

References

[1]	A. Kuzmin, L. Grekov, О. Petrov, О. Medvedenko (2017) Computational procedures for thematic processing of satellite images in the interest of monitoring agricultural resources (part 1). Environmental safety and nature management, № 1-2 (23), 70-78.
[2]	А. Kuzmin, L. Grekov, N. Kuzmina, О. Petrov, О. Medvedenko (2020) Computational procedures for thematic processing of satellite images in the interest of monitoring agricultural resources (part 2). Environmental safety and nature management, № 1 (33), 87–93. https://doi.org/10.32347/2411-4049.2020.1.87-94.
[3]	Tou J. T, Gonzalez R. C. Pattern recognition principles. Boston, MA, USA: Addison-Wesley Publ. Company, 1974. 395 p.
[4]	ArcGIS Desktop https://desktop.arcgis.com/ru/arcmap/10.3/main/get-started/arcgis-tutorials.htm.
[5]	ERDAS ER Mapper. Hexagon Geospatial. https://www.hexagongeospatial.com/brochure-pages/erdas-ermapper-professional-benefit-brochure.
[6]	ENVI— Environment for Visualizing Images. Harris Geospatial Solutions https://www.l3harrisgeospatial.com/docs/using_envi_home.html.
[7]	N. Abramov, D. Makarov, А. Talalayev, V. Fralenko Modern methods of intellectual data processing of remote sensing data. Software systems: Theory and applications vol. 9, № 4 (39), p. 417-442.
[8]	Scheinberg K. An efficient implementation of an active set method for svms // J. Mach. Learn. Res. — 2006. — Vol. 7. — Pp. 2237-2257.
[9]	Yizong Cheng Mean Shift, Mode Seeking, and Clustering // IEEE Transactions on Pattern Analysis and Machine Intelligence. — IEEE, 1995. — August (vol. 17, rel. 8).— doi: 10.1109/34.400568.
[10]	Ester, Martin; Kriegel, Hans-Peter; Sander, Jörg; Xu, Xiaowei (1996). У Simoudis, Evangelos; Han, Jiawei; Fayyad, Usama M. A density-based algorithm for discovering clusters in large spatial databases with noise Proceedings of the Second International Conference on Knowledge Discovery and Data Mining (KDD-96). AAAI Press. p. 226 –231. ISBN 1-57735-004-9.
[11]	Ester, M. A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise / M. Ester, H.-P. Kriegel, J. Sander, X. Xu // In Proc. ACM SIGMOD Int. Conf. on Management of Data, Portland, OR, 1996. –P. 226-231.
[12]	Sarmah S., Bhattacharyya D. K. (2012) A grid-density based technique for finding clusters in satellite image. Pattern Recognition Letters, V. 33, 589-604.
[13]	I. Pestunov, Y. Siniavsky (2012) Clustering Algorithm in Satellite Image Segmentation Problems. Bulletin of the Kemerovo State University №4 (52) vol. 2, 110-125.
[14]	I. Pestunov, S. Rylov. «A Method for Constructing an Ensemble of Grid Hierarchical Clustering Algorithms for Satellite Image Segmentation», Regional problems of remote sensing of the Earth, Materials of the international scientific conference, Siberian Federal University, Krasnoyarsk, 2014, p. 215-223.
[15]	Y. Kulikova, I. Pestunov, Y. Siniavsky. «Nonparametric Clustering Algorithm for Processing Large Data Arrays», Proceedings of the 14th scientific conference "Mathematical methods for pattern recognition", MAKS Press, М., 2009, p. 149-152.
[16]	Agrawal, R. Automatic Subspace Clustering of High Dimensional Data for Data Mining Applications / R. Agrawal, J. Gehrke, D. Gunopulos, P. Raghavan // In Proc. ACM SIGMOD Int. Conf. on Management of Data, Seattle, Washington, 1998. -P. 94-105.
[17]	Nagesh, H. MAFIA: Efficient and Scalable Subspace Clustering for Very Large Data Sets / H. Nagesh, S. Goil, A. Choudhary // Technical Report Number CPDC-TR-9906-019, Center for Parallel and Distributed Computing, Northwestern University, 1999. 20 p.
[18]	Rongjun Qin, Tao Liu A Review of Landcover Classification with Very-High Resolution Remotely Sensed Optical Images—Analysis Unit, Model Scalability and Transferability, Remote Sens. 2022, 14, 646. https://doi.org/10.3390/rs14030646

Cite This Article

Plain Text BibTeX RIS

APA Style

Anatolii Kuzmin, Leonid Grekov, Nataliia Kuzmina, Oleksii Petrov. (2022). About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms. American Journal of Remote Sensing, 10(2), 30-38. https://doi.org/10.11648/j.ajrs.20221002.11

Copy | Download

ACS Style

Anatolii Kuzmin; Leonid Grekov; Nataliia Kuzmina; Oleksii Petrov. About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms. Am. J. Remote Sens. 2022, 10(2), 30-38. doi: 10.11648/j.ajrs.20221002.11

Copy | Download

AMA Style

Anatolii Kuzmin, Leonid Grekov, Nataliia Kuzmina, Oleksii Petrov. About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms. Am J Remote Sens. 2022;10(2):30-38. doi: 10.11648/j.ajrs.20221002.11

Copy | Download

@article{10.11648/j.ajrs.20221002.11,
  author = {Anatolii Kuzmin and Leonid Grekov and Nataliia Kuzmina and Oleksii Petrov},
  title = {About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms},
  journal = {American Journal of Remote Sensing},
  volume = {10},
  number = {2},
  pages = {30-38},
  doi = {10.11648/j.ajrs.20221002.11},
  url = {https://doi.org/10.11648/j.ajrs.20221002.11},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajrs.20221002.11},
  abstract = {Applied problems of studying the earth's surface using satellite images of remote sensing of the Earth are considered for the study of forest, agricultural, water and other natural resources, where clustering and classification algorithms are instrumental research methods. It is noted that the most well-known procedures for classifying and segmenting multispectral space images in GIS systems, such as ArcGIS, ERDAS, ENVI, are built-in. The need to expand the toolkit for a more efficient solution of applied problems of this class is noted. New universal clustering and classification algorithms based on a unified approach are proposed. Both methods belong to grid-type algorithms, and at the first stage of their work they group points of a set of n - dimensional vectors into grid cells, each cell saves only the numbers of points belonging to it and is characterized by a unique code. The vector grid spacing is a parameter of the method and is set by the user using a single integer value. At the next stage, the clustering algorithm combines the cells and the points belonging to them into clusters using the cell neighborhood principle. In this case, the algorithm does not attach the next cell to the cluster in the case when its density is less than the specified value. The classification algorithm refers the points of the cell of the main set to the class to which the cell with the same code of the training set belongs. The algorithms can be used to process large data sets of large spatial dimensions, including satellite images. Clustering and classification algorithms do not require a preliminary specification of the number of clusters and information about the nature of the distribution of points in the input set.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms
AU  - Anatolii Kuzmin
AU  - Leonid Grekov
AU  - Nataliia Kuzmina
AU  - Oleksii Petrov
Y1  - 2022/09/05
PY  - 2022
N1  - https://doi.org/10.11648/j.ajrs.20221002.11
DO  - 10.11648/j.ajrs.20221002.11
T2  - American Journal of Remote Sensing
JF  - American Journal of Remote Sensing
JO  - American Journal of Remote Sensing
SP  - 30
EP  - 38
PB  - Science Publishing Group
SN  - 2328-580X
UR  - https://doi.org/10.11648/j.ajrs.20221002.11
AB  - Applied problems of studying the earth's surface using satellite images of remote sensing of the Earth are considered for the study of forest, agricultural, water and other natural resources, where clustering and classification algorithms are instrumental research methods. It is noted that the most well-known procedures for classifying and segmenting multispectral space images in GIS systems, such as ArcGIS, ERDAS, ENVI, are built-in. The need to expand the toolkit for a more efficient solution of applied problems of this class is noted. New universal clustering and classification algorithms based on a unified approach are proposed. Both methods belong to grid-type algorithms, and at the first stage of their work they group points of a set of n - dimensional vectors into grid cells, each cell saves only the numbers of points belonging to it and is characterized by a unique code. The vector grid spacing is a parameter of the method and is set by the user using a single integer value. At the next stage, the clustering algorithm combines the cells and the points belonging to them into clusters using the cell neighborhood principle. In this case, the algorithm does not attach the next cell to the cluster in the case when its density is less than the specified value. The classification algorithm refers the points of the cell of the main set to the class to which the cell with the same code of the training set belongs. The algorithms can be used to process large data sets of large spatial dimensions, including satellite images. Clustering and classification algorithms do not require a preliminary specification of the number of clusters and information about the nature of the distribution of points in the input set.
VL  - 10
IS  - 2
ER  -

Copy | Download

Author Information

Anatolii Kuzmin

Faculty of Computer Science and Cybernetics, Taras Shevchenko National University of Kyiv, Kyiv, Ukraine
Leonid Grekov

Scientific Production Enterprise “Agroresurssystems”, Kyiv, Ukraine
Nataliia Kuzmina

Faculty of Mathematics, Informatics and Physics, National Pedagogical Dragomanov University, Kyiv, Ukraine
Oleksii Petrov

Scientific Production Enterprise “Agroresurssystems”, Kyiv, Ukraine

Download PDF

Sections

Plain Text BibTeX RIS

APA Style

Anatolii Kuzmin, Leonid Grekov, Nataliia Kuzmina, Oleksii Petrov. (2022). About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms. American Journal of Remote Sensing, 10(2), 30-38. https://doi.org/10.11648/j.ajrs.20221002.11

Copy | Download

ACS Style

Anatolii Kuzmin; Leonid Grekov; Nataliia Kuzmina; Oleksii Petrov. About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms. Am. J. Remote Sens. 2022, 10(2), 30-38. doi: 10.11648/j.ajrs.20221002.11

Copy | Download

AMA Style

Anatolii Kuzmin, Leonid Grekov, Nataliia Kuzmina, Oleksii Petrov. About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms. Am J Remote Sens. 2022;10(2):30-38. doi: 10.11648/j.ajrs.20221002.11

Copy | Download

@article{10.11648/j.ajrs.20221002.11,
  author = {Anatolii Kuzmin and Leonid Grekov and Nataliia Kuzmina and Oleksii Petrov},
  title = {About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms},
  journal = {American Journal of Remote Sensing},
  volume = {10},
  number = {2},
  pages = {30-38},
  doi = {10.11648/j.ajrs.20221002.11},
  url = {https://doi.org/10.11648/j.ajrs.20221002.11},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajrs.20221002.11},
  abstract = {Applied problems of studying the earth's surface using satellite images of remote sensing of the Earth are considered for the study of forest, agricultural, water and other natural resources, where clustering and classification algorithms are instrumental research methods. It is noted that the most well-known procedures for classifying and segmenting multispectral space images in GIS systems, such as ArcGIS, ERDAS, ENVI, are built-in. The need to expand the toolkit for a more efficient solution of applied problems of this class is noted. New universal clustering and classification algorithms based on a unified approach are proposed. Both methods belong to grid-type algorithms, and at the first stage of their work they group points of a set of n - dimensional vectors into grid cells, each cell saves only the numbers of points belonging to it and is characterized by a unique code. The vector grid spacing is a parameter of the method and is set by the user using a single integer value. At the next stage, the clustering algorithm combines the cells and the points belonging to them into clusters using the cell neighborhood principle. In this case, the algorithm does not attach the next cell to the cluster in the case when its density is less than the specified value. The classification algorithm refers the points of the cell of the main set to the class to which the cell with the same code of the training set belongs. The algorithms can be used to process large data sets of large spatial dimensions, including satellite images. Clustering and classification algorithms do not require a preliminary specification of the number of clusters and information about the nature of the distribution of points in the input set.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - About One Approach to the Construction of Clustering and Classification Grid-Type Algorithms
AU  - Anatolii Kuzmin
AU  - Leonid Grekov
AU  - Nataliia Kuzmina
AU  - Oleksii Petrov
Y1  - 2022/09/05
PY  - 2022
N1  - https://doi.org/10.11648/j.ajrs.20221002.11
DO  - 10.11648/j.ajrs.20221002.11
T2  - American Journal of Remote Sensing
JF  - American Journal of Remote Sensing
JO  - American Journal of Remote Sensing
SP  - 30
EP  - 38
PB  - Science Publishing Group
SN  - 2328-580X
UR  - https://doi.org/10.11648/j.ajrs.20221002.11
AB  - Applied problems of studying the earth's surface using satellite images of remote sensing of the Earth are considered for the study of forest, agricultural, water and other natural resources, where clustering and classification algorithms are instrumental research methods. It is noted that the most well-known procedures for classifying and segmenting multispectral space images in GIS systems, such as ArcGIS, ERDAS, ENVI, are built-in. The need to expand the toolkit for a more efficient solution of applied problems of this class is noted. New universal clustering and classification algorithms based on a unified approach are proposed. Both methods belong to grid-type algorithms, and at the first stage of their work they group points of a set of n - dimensional vectors into grid cells, each cell saves only the numbers of points belonging to it and is characterized by a unique code. The vector grid spacing is a parameter of the method and is set by the user using a single integer value. At the next stage, the clustering algorithm combines the cells and the points belonging to them into clusters using the cell neighborhood principle. In this case, the algorithm does not attach the next cell to the cluster in the case when its density is less than the specified value. The classification algorithm refers the points of the cell of the main set to the class to which the cell with the same code of the training set belongs. The algorithms can be used to process large data sets of large spatial dimensions, including satellite images. Clustering and classification algorithms do not require a preliminary specification of the number of clusters and information about the nature of the distribution of points in the input set.
VL  - 10
IS  - 2
ER  -

Copy | Download