Radar image interpretability analysis
Abstract
The utility of radar images with respect to trained image interpreter ability to identify, classify and detect specific terrain features (linear, natural area, complex area features, and individual man-made features) was qualitatively determined. Further, radar images were evaluated with respect to their utility for determining vehicle movement potential and the level of activity within the test areas. Because there are no physical laws governing the relationship between image quality and measurable image characteristics, signal-to-noise ratio, dynamic range, image bandwidth, geometric fidelity and root-mean-square error, a second order linear statistical model was assumed. An experiment was designed and conducted to estimate the coefficients of the model given each image application. The result of this analysis was that statistically significant models were obtained to relate the measured image properties to the interpreters: ability to analyze linear features and to evaluate the utility of radar images for vehicle movement potential and activity level. Further, it was found that the relative importance of the measured image properties with respect to image utility changed with image application. This study has provided useful information as to how certain image characteristics relate to radar image utility as a function of several image applications.
- Publication:
-
Final Report
- Pub Date:
- January 1981
- Bibcode:
- 1981kans.rept.....F
- Keywords:
-
- Radar Imagery;
- Synthetic Aperture Radar;
- Target Recognition;
- Bandwidth;
- Range (Extremes);
- Signal To Noise Ratios;
- Communications and Radar