Affiliation:
1. Department of Industrial Engineering and Operations Research Viriginia Polytechnic Institute and State University Blacksburg, Virginia 24061
Abstract
This paper investigated the efficacy of digital image enhancement/restoration processes as measured by human performance. The human performance measures consisted of an information extraction task and a subjective scaling task. Various enhancement/restoration processes which addressed image degradations of poor contrast, blur, and noise were selected and evaluated. The digital images consisted of military and civilian scenes, each of which were degraded by different levels of blur and noise. Ten Air Force photointerpreters participated in the investigation. The photointerpreters viewed the images on CRTs and performed both tasks. The results indicated that image degradations deterred the quality of digital imagery for interpretation. However, the addition of enhancement/restoration processing improved information extraction performance by 12%. On the other hand, subjective scaling results revealed that processing did not improve nor worsen perceived interpretability of the images.