| SPATIAL-SPECTRAL MORPHOLOGICAL FEATURE EXTRACTION FOR HYPERSPECTRAL IMAGES CLASSIFICATION |
| Paper ID : 1172-SMPR |
| Authors: |
|
Mohammad Dowlatshah *1, hassan ghassemian2, maryam imani3 1Image Processing and Data Analysis Laboratory Tarbiat Modares University Tehran, Iran 2Image Processing and Data Analysis Laboratory Tarbiat Modares University Tehran, Iran 3Maryam Imani Image Processing and Data Analysis Laboratory Tarbiat Modares University Tehran, Iran maryam.imani@modares.ac.ir |
| Abstract: |
| Remote sensing image classification is one of the methods used to group pixels to show details of the Land cover. One of the methods for spatial feature extraction is the use of morphological filters. The basic idea of morphological filters is the comparison of the structures within the image with a reference form called the structural element. Four types of important morphological filters are included (dilation, erosion, opening, and closing) in this method. Opening morphological filter is used to extract spatial features; this filter is created with two successive sequences dilation and erosion filters. This filter in binary images removes the light areas that are smaller than the structural element, and in the gray surface images, the areas that are smaller than the structural element and are brighter than the neighboring regions are removed using this filter. In the proposed method, the principal component analysis is used to reduce the dimensions of the data. Differential morphology filters are another important morphological filters, which are also used in this proposed method, and an SVM classifier is used to classify the data. |
| Keywords: |
| Feature Extraction, Principal Component Analysis (PCA), Hyperspectral, Morphology Profiles |
| Status : Conditional Accept (Poster) |
