| タイトル | A superior edge preserving filter with a systematic analysis |
| 本文（外部サイト） | http://hdl.handle.net/2060/19910018548 |
| 著者（英） | Holladay, Kenneth W.; Rickman, Doug |
| 著者所属（英） | NASA Stennis Space Center |
| 発行日 | 1991-08-01 |
| 言語 | eng |
| 内容記述 | A new, adaptive, edge preserving filter for use in image processing is presented. It had superior performance when compared to other filters. Termed the contiguous K-average, it aggregates pixels by examining all pixels contiguous to an existing cluster and adding the pixel closest to the mean of the existing cluster. The process is iterated until K pixels were accumulated. Rather than simply compare the visual results of processing with this operator to other filters, some approaches were developed which allow quantitative evaluation of how well and filter performs. Particular attention is given to the standard deviation of noise within a feature and the stability of imagery under iterative processing. Demonstrations illustrate the performance of several filters to discriminate against noise and retain edges, the effect of filtering as a preprocessing step, and the utility of the contiguous K-average filter when used with remote sensing data. |
| NASA分類 | COMPUTER PROGRAMMING AND SOFTWARE |
| レポートNO | 91N27862
NASA-TM-4311
NAS 1.15:4311 |
| 権利 | No Copyright |
| URI | https://repository.exst.jaxa.jp/dspace/handle/a-is/131576 |
