Synthetic Observer Approach to Multispectral Sensor Resolution Assessment

Resolution is often provided as one of the key parameters addressing the quality characteristic of a sensor. One traditional approach when determining the resolution of a sensor/display system is to use a resolution target pattern to detect the smallest element that can be ?resolved? using the system. This requires a human in the loop to make the assessment. This study investigated the use of a custom designed software approach to generate an effective resolution value for a sensor. Landolt Cs were selected as the resolution target, which were imaged at multiple distances with different sensors. The images were analyzed using custom software to determine the orientation of the C at each distance, which resulted in a probability of correct orientation detection curve as a function of distance. This curve was used to generate a resolution for the sensor without involving human vision. Resolution results for four different spectral band sensors were obtained as well as effective resolution of fused images from select pairs of sensors. These results and the possible use of this synthetic observer resolution approach are presented and discussed, as well as possible future research relating this resolution to human visual performance with fused image sources. Performed in cooperation with Ball Aerospace & Technology Corp., Fairborn, OH.