Visualizing 3 - D Texture : a Three Dimensional Structural

The evaluation of texture in digital remote sensing, is traditionally accomplished by applying statistical, rather than structural models. This approach is partially due to the difficulty in determining structural placement rules for natural scenes, and to early psychophysical texture theories that demonstrated humans were sensitive to second-order statistics. A literature review indicates that these statistical theories are now considered inaccurate discriminators of visual texture, yet they are still incorporated within current remote sensing texture practices. As a result, four recognizable areas of discrepancy exist between human texture perception, and digital remote sensing texture methodology. These discrepancies include: 1. Defining appropriate primitives. 2. Quantifying the spatial arrangement of primitives. 3. Determining appropriate scale / spatial resolution. 4. The inclusion of depth information. To overcome these discrepancies, a structural texture methodology referred to as the triangulated primitive neighborhood method (TPN) has been developed based on current psychophysical texture theory. Though similar to the grey level cooccurrence matrix method (GLCM), this technique reduces many of the user specified requirements and inadequacies of the GLCM by incorporating object specific height and area data, location specific primitives, and a variable sized and shaped moving window to produce digital texture features approximating human perception. In this study, a simple unsupervised classification of Neighborhood Height and Neighborhood Area data sets provide a visually correlative perception of the withinand between-stand texture of two different aged forest stands.