FFT-PCA Image Fusion Based Flora and Vegetation Mapping Of Eshkevarat No Hunting Zone
محورهای موضوعی : فصلنامه علمی پژوهشی سنجش از دور راداری و نوری و سیستم اطلاعات جغرافیایی
Zeinab Hoseinnejad
1
(M.Sc. Environmental Assessment, Coolege of environment. Department of Environment karaj, Iran)
Hasan Hasani Moghaddam
2
(MSc of remote sensing, Kharazmi university, Tehran, Iran)
Zahra Parvar
3
(M.Sc Environmental Assessment, Department of Environment, Faculty of Natural Resources, Malayer University, Iran)
Kourosh Kavousi
4
(Phd , Department of Biology, Science and Research Branch, Islamic Azad University Tehran)
Hamid Gashtasb Meigooni
5
(Associated Professor, Department of Natural Environment and Biodiversity, College of Environment, Karaj, Iran)
کلید واژه: SVM, FFT-PCA, Flora and vegetation, Landsat ETM+, Eshkevarat,
چکیده مقاله :
fusion of remote sensing data is essential in order to obtain more information from different images. Mapping the vegetation of an area is very important due to its environmental importance. In this research, used Landsat ETM+ images and field surveying to identify vegetation states of the Eshkevarat No hunting zone. After applying necessary preprocessing like gap filling and atmospheric correction, the panchromatic and multi-spectral images were fused based on the FFT-PCA algorithm. In the next section, the fused image was classified based on the Support Vector Machine (SVM), algorithm into five classes. The results showed that the overall accuracy and kappa coefficient of classified images is 0.943% and 0.910 respectively. In order to field surveying of study area, 1-meter plots in 500-meter distance choose and 14 Flora and vegetation species were identified and mapped. The results showed that satellite images have good accuracy in this field but based on its spatial resolution limitations a large number of species present in the area have not been identified. In this research, it is suggested to use a combination of both satellite image sources and field surveys.
fusion of remote sensing data is essential in order to obtain more information from different images. Mapping the vegetation of an area is very important due to its environmental importance. In this research, used Landsat ETM+ images and field surveying to identify vegetation states of the Eshkevarat No hunting zone. After applying necessary preprocessing like gap filling and atmospheric correction, the panchromatic and multi-spectral images were fused based on the FFT-PCA algorithm. In the next section, the fused image was classified based on the Support Vector Machine (SVM), algorithm into five classes. The results showed that the overall accuracy and kappa coefficient of classified images is 0.943% and 0.910 respectively. In order to field surveying of study area, 1-meter plots in 500-meter distance choose and 14 Flora and vegetation species were identified and mapped. The results showed that satellite images have good accuracy in this field but based on its spatial resolution limitations a large number of species present in the area have not been identified. In this research, it is suggested to use a combination of both satellite image sources and field surveys.
