Hierarchical Processor And Matched Filters For Range Image Processing

The use of range imagery and new-processing techniques for multi-class object invariant recognition are presented here. Techniques investigated involve different hierarchical processors with emphasis given to processors using correlation filters and histogram processing. The use of range imagery is shown to significantly improve classification in many cases especially with severe distortions in the object orientation. Histogram operations are appropriate for providing aspect angle estimates with aircraft objects. Initial experimental results are shown for each technique and with several data bases.