Proceedings; XXI International Congress for Photogrammetry and Remote Sensing: Proceedings; XXI International Congress for Photogrammetry and Remote Sensing

Bibliographic data

Multivolume work

Persistent identifier:: 1838419977

Title:: Handbuch der speciellen Pathologie und Therapie

Year of publication:: 1854

Place of publication:: Erlangen; Hannover

Identifier (digital):: 1838419977

Language:: German

Additional Notes:: Bände 1-7 erschienen von 1854-1876

Publisher of the original:: Verlag von Ferdinand Enke

Contributor:: Bamberger, Heinrich; Biermer, Anton; Falck, Karl Philipp; Friedreich, Nikolaus; Hasse, Karl Ewald; Hebra, Ferdinand; Lebert, Hermann; Pitha, Franz; Simon, Friedrich Alexander; Stiebel, Salomon Friedrich; Veit, Gustav; Vogel, Julius; Westphal, Wilhelm; Wintrich, Anton

Editor:: Virchow, Rudolf

Publisher of the digital copy:: Technische Informationsbibliothek (TIB)

Document type:: Multivolume work

Volume

Persistent identifier:: 1838465901

Author:: Wintrich, Anton

Title:: Krankheiten der Respirationsorgane

Scope:: 384 Seiten

Year of publication:: 1854

Place of publication:: Erlangen

Identifier (digital):: 1838465901

Signature of the source:: a 3177(5,1,1/2)

Language:: German

Publisher of the original:: Verlag von Ferdinand Enke

Usage licence:: Public Domain Mark 1.0

Other person:: Virchow, Rudolf

Publisher of the digital copy:: Technische Informationsbibliothek (TIB)

Place of publication of the digital copy:: Hannover

Year of publication of the original:: 2023

Document type:: Volume

Collection:: Medicine

Title page

Document type:: Multivolume work

Structure type:: Title page

EFFECTIVENESS OF BOOSTING ALGORITHMS IN FOREST FIRE CLASSIFICATION C. Özkan 3 ’*, F. Sunar b , S. Berberoglu c , C. Dönmez c a Erciyes University, Engineering Faculty, Geodesy and Photogrammetry Engineering Dept., 38039 Kayseri, Turkey - cozkan@erciyes.edu.tr b Istanbul Technical University, Civil Engineering Faculty, Remote Sensing Division, 34469 Maslak Istanbul, Turkey - fsunar@ins.itu.edu.tr c Cukurova University, Department of Landscape Architecture, Faculty of Agriculture, 01330 Adana, Turkey - (suha, cdonmez)@cu.edu.tr KEY WORDS: Forest fire, SPOT 4, Adaboost, Logitboost, Multilayer Perseptron, Regression Tree. ABSTRACT: In this paper, it is aimed to investigate the capabilities of boosting classification approach for forest fire detection using SPOT-4 imagery. The study area, Bodrum in the province of Mugla, is located at the south-western Mediterranean coast of Turkey where recent largest forest fires occurred in July 2007. Boosting method is one of the recent advanced classifiers proposed in the machine learning community, such as neural networks classifiers based on multilayer perceptron (MLP), radial basis function and learning vector quantization. The Adaboost (AB) and Logitboost (LB) algorithms which are the most common boosting methods were used for binary and multiclass classifications. The effectiveness of boosting algorithms was shown through comparison with Bayesian maximum likelihood (ML) classifier, neural network classifier based on multilayer perceptron (MLP) and regression tree (RT) classifiers. The pre and post SPOT images were corrected atmospherically and geometrically. Binary classification comprised burnt and non-bumt classes. In addition to the pixel based classification, textural measures including, gray level co-occurrence matrix such as entropy, homogeneity, second angular moment, etc. were also incorporated. Instead of the traditional boosting weak (base) classifiers such as decision tree builder or perceptron learning rule, neural network classifier based on multilayer perceptron were adapted as a weak classifier. The accuracy of the MLP was greater than that of ML, AB, LB and RT both using spectral data alone and textural data. The use of texture measures alone was found to increase classification accuracy of binary and multi-class classifications. The accuracy of the land cover classifications based on either binary or multi-class was maximised using a MLP approach. This was slightly greater than the accuracy achieved using AB and LB classifications. However, it was shown that AB and LB classifications hold great potential as an alternative to conventional techniques. 1. INTRODUCTION Forest fires, whether natural or initiated by man, have an important influence on the environment, human health (often leading to fatalities) and property. Fires destroy thousands of hectares of the worldwide forests each year, causing soil erosion and desertification processes, long-term site degradation and alteration of hydrological regimes, producing a significant amount of aerosols and carbon gases that may strong influence to the global climate changes. In terms of annual damage, Turkey is one of the most susceptible countries in the Mediterranean region. Turkish Ministry of Forestry reports that the most frequently burned regions in Turkey are Antalya, Balikesir, Çanakkale, Denizli, Isparta, Izmir and Mugla (im et al., 2006). The information needed for fire management includes fire risk map, fire detection and monitoring (Mazzeo et ah, 2007; Cuomo etc. 2001), damage assessment and planning post-fire recovery (Badarinath et ah, 2007; Isaev et ah, 2002; Phulpin et ah, 2002; Yu et al, 2004; Fraser et ah, 2003; Sunar and Òzkan 2001). Effective management requires the use of new techniques using remote sensing and Geographical Information Systems (GIS). In fire damage assessment, satellite imagery in conjunction with ground observations is the main source of information from which pre-fire vegetation condition can be determined. Pre-fire condition is important both in order to assess potential fire risk and to make rational judgements on the precautionary measures required to avoid potential soil erosion which topographic and meteorological conditions may promote on newly burnt land (Chuvieco and Martin, 1994; im et ah, 2006). Many remote sensing analyses comprise many digital image processing algorithms from simple spectral or spatial transformations such as image subtracting to complex procedures such as classification. The land cover changes such as fire burnt areas and the degree of severity because of fire can be effectively determined by the change detection approaches, image spectral transformations such as NDVI (Normalized Difference Vegetation Image) or classification methods such as ISODATA and maximum likelihood. The success of the change detection procedures mostly depends on the radiometric and geometric consistencies of the multi-temporal imageries. NDVI in which the vegetation is the primary reflectance object is a commonly used technique. Although change detection and NDVI are effective to extract the biomass and to discriminate the healthy and unhealthy vegetation, the classification phase is vital for the quantitative and qualitative precise analysis of these land covers. Among many supervised and unsupervised classification methods, the approaches based on the artificial intelligence such as artificial neural networks and ensemble algorithms such as boosting are mostly investigated. Boosting method is one of the recent advanced classifiers proposed in the machine learning community, such as neural networks classifiers based on multilayer perceptron, radial basis function Corresponding author.

Access restriction

Copyright

fullscreen: Proceedings; XXI International Congress for Photogrammetry and Remote Sensing (Part B7-2)

Multivolume work

Volume

Title page

Table of contents

Downloads

Full record

This page

Image fragment

Formats and links

Formats

Links

Cite

Full record

This page

Citation recommendation

Search results