GLIDER Features

A.    GLIDER Views
GLIDER provides users three views to manage, visualize and analyze data.  In the Project Explorer View (Fig. 1), users can create projects to organize their data and analysis results.  If a particular data set is selected within this view, its metadata is displayed on the right panel.  This allows the user to browse useful metadata information such as the different fields contained within the data, its spatial bounds and the temporal range. Once the user has selected a file within the Project Explorer, a menu is presented, allowing the user to either view the data in the Image Analysis View or the Earth View. 
Using the Image Analysis View (Fig. 2), the user can visualize the entire satellite swath. This View provides powerful image analysis capabilities designed to enable interactive data exploration.  A user can open multiple Image Analysis Views to visualize different data files at the same time. 

From either the Project Explorer or the Image Analysis View, the user can open the Earth View to visualize the data projected on a 3D globe.  The Earth View, shown in Fig. 3, uses the navigation information to display the satellite imagery in the geographic coordinate system using the WGS84 datum.  The Earth View is dynamically synced with the Image Analysis View. Any changes made to the display in the Image Analysis View automatically cause the Earth View image to refresh.  Thus, a user can change spectral bands in the Image Analysis View and the image projected on the globe will automatically change. The results of image processing and pattern recognition operations can also be visualized in this view.  Moreover, the user can display images from multiple Image Analysis Views on the same globe. Additional ancillary data such as cities, political boundaries can be added to the display within the Earth View.

Figure 1.     Project Explorer View

Figure 2.     Image Analysis View

Figure 3.     Earth View

B.    Satellite Image Calibration
GLIDER has image calibration routines built into it’s data readers.  For example, while the MODIS data is read by the tool, it is calibrated into reflectance and brightness temperature based on the actual radiance values.  The release version 1.0 can read and calibrate MODIS satellite imagery. Other satellite data will sets be added in the future releases.

C.    Image Display and Analysis
GLIDER provides many different image display functionality in both Image View and the Earth View.  Within the Image View, the user can view the satellite data as a grey scale or create a three channel color composite using different spectral bands.  There are features such as linear contrast stretch, grey scale inversion and histogram equalization that can be applied to enhance the image.  Pan and scroll capabilities are also available within Image View.  The Earth View allows the user to display the satellite imagery on a 3D globe and be able to zoom, pan and rotate in 3D.  A user can display multiple images on the globe as layers and select or deselect different layers.  The Earth View also allows overlay of vector and model data in 3D.

There are several image analysis features that are currently available in GLIDER.  Histogram analysis feature allow users to create a histogram of the three selected spectral bands based on a user selected region or part of the image that is visible or the entire image.  User can also create scatter plots between two spectral bands.  The scatter plots are displayed using pixels from either user selected region or part of the image that is visible or the entire image.  GLIDER also provides the capability to analyze spectral profiles of individual pixels.  This feature is extremely useful in determining which spectral bands provide the maximum separation for different classes while performing supervised classification.  Spectral profiles of multiple pixel points can be generated on a single plot.  Linear transect feature allows the user to arbitrarily draw a line anywhere on the image. The spectral values of all the pixels coincident with this line channel being displayed are presented as a line plot.  Examples of the different image analysis features are shown in Fig. 4.

Figure 4: Image analysis features (Histogram, Scatter plot, Spectral Profile, Linear transect) available in GLIDER v1.0

D.    Image Processing
GLIDER also provides a set of image processing capabilities that are useful for extracting features from images as a precursor to mining. Since, filtering plays an important role in many image analysis applications, GLIDER contains spatial domain, median, mode and morphological filtering (erosion and dilation) capability. It also includes the pulse coupled neural network, which can be used for image smoothing and segmentation. GLIDER can be used to find boundaries, contiguous regions, and polygons in images.

E.    Unsupervised Classification

GLIDER provides users the capability to perform unsupervised classification or clustering using the spectral bands as features. There are several different clustering algorithms available within glider and these are Isodata, Maximin, Hiearchical, Kmeans and Kmediods.  All of these clustering algorithms take a set of patterns such as values at different spectral bands as input and group them into classes based on some similarity measure.  The use of K-Means clustering algorithm is depicted in Figures 5 and 6.  Figure 5 shows the dialog box where the user can specify the parameters required by the K-Means algorithm.  The result of the K-Means clustering is presented in Figure 6.

Figure 5: Dialog box for specifying different parameters for the K-Means algorithm

Figure 6: Result from the clustering algorithm displayed in GLIDER