Mission-driven evaluation of imaging system quality

Area of Science:

• Optics and Vision Science
• Remote Sensing Technology
• Image Analysis

Background:

• Traditional image-quality criteria optimize for general image fidelity at a fixed sampling rate.
• These criteria are inadequate for applications prioritizing the detection of specific object characteristics.
• A need exists for quality metrics tailored to the specific observational goals.

Purpose of the Study:

• To develop a novel, object-oriented imaging quality criterion for observation system design.
• To shift focus from image-centric to object-centric quality assessment.
• To establish a calibrated numerical scale for rating observation system performance.

Main Methods:

• Developed a quality criterion based on the intrinsic properties of the objects being observed.
• Proposed an object-oriented approach to imaging system design and evaluation.
• Introduced a method for calibrating a numerical scale to quantify observation system service quality.

Main Results:

• The proposed object-oriented criterion is better suited for applications requiring detection of geometric and radiometric properties.
• The new criterion offers an alternative to traditional image-oriented metrics.
• A scalable framework for objectively rating observation system quality has been established.

Conclusions:

• Object-oriented imaging quality criteria enhance the design and evaluation of observation systems for specific tasks.
• The developed criterion and numerical scale provide a more relevant measure of performance for detection-critical applications.
• This work facilitates objective comparison and calibration of diverse observation systems.