Search

Search Results

• Unsupervised classification of high resolution satellite imagery by self-organizing neural network

  37-44

  Árpád Barsi

  Katalin Gáspár

  Zsuzsanna Szepessy

  Views:

  85

  The current paper discusses the importance of the modern high resolution satellite imagery. The acquired high amount of data must be processed by an efficient way, where the used Kohonen-type self-organizing map has been proven as a suitable tool. The paper gives an introduction to this interesting method. The tests have shown that the multispectral image information can be taken after a resampling step as neural network inputs, and then the derived network weights are able to evaluate the whole image with acceptable thematic accuracy.

• Urban vegetation classification with high-resolution PlanetScope and SkySat multispectral imagery

  66-75

  Loránd Szabó

  Dávid Abriha

  Kwanele Phinzi

  Szilárd Szabó

  Views:

  813

  In this study two high-resolution satellite imagery, the PlanetScope, and SkySat were compared based on their classification capabilities of urban vegetation. During the research, we applied Random Forest and Support Vector Machine classification methods at a study area, center of Rome, Italy. We performed the classifications based on the spectral bands, then we involved the NDVI index, too. We evaluated the classification performance of the classifiers using different sets of input data with ROC curves and AUC values. Additional statistical analyses were applied to reveal the correlation structure of the satellite bands and the NDVI and General Linear Modeling to evaluate the AUC of different models. Although different classification methods did not result in significantly differing outcomes (AUC values between 0.96 and 0.99), SVM’s performance was better. The contribution of NDVI resulted in significantly higher AUC values. SkySat’s bands provided slightly better input data related to PlanetScope but the difference was minimal (~3%); accordingly, both satellites ensured excellent classification results.

• Mapping aquatic vegetation of the Rakamaz-Tiszanagyfalui Nagy-Morotva using hyperspectral imagery

  1-10

  Péter Burai

  Gabriella Lövei

  Csaba Lénárt

  Ildikó Nagy

  Enyedi Enyedi

  Views:

  251

  Rapid development in remote sensing technologies provides more and more reliable methods for environmental assessment. For most wetlands, it is difficult to walk-in without disturbing the endangered species living there; therefore, application of opportunities provided by remote sensing has a great importance in population-mapping. One effective tool of vegetation pattern estimation is hyperspectral remote sensing, which can be used for association and species level mapping as well, due to high ground resolution. The Rakamaz-Tiszanagyfalui Nagy-morotva is an oxbow lake, located in the north-eastern part of Hungary. For this study, a wetland area of 1.17 km2 containing the original water bad and shoreline was selected. For the image analysis, images taken by an AISA DUAL system hyperspectral sensor were used. At the same time, 7 main vegetation classes were separated, which are typical for the sample plot designated on the test site. Classification was performed by the master areas signed by the most common associations of the Rakamaz-Tiszanagyfalui Nagy-morotva with determined spectrums. During the image analysis, SAM classification method was used, where radian values were optimized by the results of classification performed at the control area.