Application of Wavelet Transform, Fractal Theory, and Statistics to Automatic Target Recognition Using L

ORINCON proposes to investigate the application of the wavelet transform, fractal analysis, and statistical measures of texture to automatic target identification (ATI) for laser radar (ladar). We intend to use multiscale edge analysis (MEA) based on the wavelet transform to decompose a ladar intensity image into an "edge" component and a complementary "texture" component. The boundary information captured in the MEA edge component will be used to detect and classify objects in ladar intensity images. The textural information captured in the MEA texture image will be analyzed using fractal and statistical methods to further enhance object detection and classification. Neural networks will be used to evaluate the effectiveness of combining edge, fractal, and statistical features for ladar ATI. A database of ladar intensity images together with extracted features will be compiled. Multifeature data fusion strategies for combining edge, fractal, statistical, and range images will be investigated.