Satellite Image Classification for Environmental Analysis using Image Processing
P. V. Goutham Reddy1, L. Rama Parvathy2
1P. V. Goutham Reddy, Department of Computer Science and Engineering, Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, Chennai, India.
2Dr. L. Rama Parvathy, Professor, Department of Computer Science and Engineering, Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, Chennai, India.
Manuscript received on December 15, 2019. | Revised Manuscript received on December 24, 2019. | Manuscript published on January 10, 2020. | PP: 1513-1517 | Volume-9 Issue-3, January 2020. | Retrieval Number: C8349019320/2020©BEIESP | DOI: 10.35940/ijitee.D1318.029420
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Conventional supervised classification of satellite pictures utilizes a solitary multi-band picture and incidental ground perceptions to build phantom marks of land spread classes. We contrasted this methodology with three choices that get marks from different pictures and timespans. signature speculation, in this unearthly marks, is gotten from various pictures inside one season, however maybe from various years. signature extension, in this phantom marks, is made with information from pictures obtained during various periods of that year; and mixes of development and speculation. Utilizing the information for India, we evaluated the nature of these various marks to characterize the pictures used to infer the mark, and for use in transient mark expansion, i.e., applying a mark acquired from the information of one or quite a long while to pictures from different years. While applying marks to the pictures they were gotten from, signature development improved exactness comparative with the customary strategy, and inconstancy in precision declined uniquely. Conversely, signature speculation didn’t improve grouping. While applying marks to pictures of different years (worldly expansion), the traditional technique, utilizing a mark got from a solitary picture, brought about extremely low characterization precision. Mark’s development additionally performed ineffectively yet multi-year signature speculation performed much better and this seems, by all accounts, to be a promising methodology in the transient augmentation of ghastly marks for satellite picture arrangements. This project summarizes the different audits on satellite picture characterization strategies and systems. The summary helps the analysts to choose suitable satellite picture characterization strategies or methods dependent on the requirements. Later on, the results acquired from the proposed technique will be an extraordinary measure for anticipating and examining the effect of floods. It will help salvage groups to address high caution regions first in this way, least or no loss of life will be accomplished. In the future, the technique can be adjusted to be utilized for coastline location, urbanization, deforestation, and seismic tremors.
Keywords: Supervised Classification, Signature Generalization, Signature Expansion, MATLAB, Localization, Segmentation, Feature Extraction
Scope of the Article: Classification