Commercial Hyperspectral Analysis Tools
N-FINDR Automated Endmember Finder
| << Hyperspectral Visualization Environment |
Back to Contents | Crisp Hyperspectral Sharpening >> |
 |
 |
 |
The TRA N-FINDR program provides rapid detection of both scene elements (endmembers) and spectral anomalies. Endmembers can often be identified with the purest examples of particular materials in a hyperspectral image. The N-FINDR algorithm is an automated technique for finding these endmembers in an image. Once these endmember have been found, the often hundreds of bands of a hyperspectral data set can be projected into approximately 10 to 15 images for analyst review. Since these images are maps of occurrences of each material, they are a great aid to scene classification. Subsequent optimizations of the algorithm have allowed the use of this algorithm at speeds exceeding those required by real-time applications.
N-FINDR example using AHIApplication to Agricultural Remote Sensing