Modeling the MTF and noise characteristics of an image chain for a synthetic image generation system
Frank Tantalo · RIT Scholar Works (Rochester Institute of Technology) · 1996
This is an approach for modeling sensor degradation effects using an image chain applied to a synthetic radiance image. The sensor effects are applied in the frequency domain by cascading modulation transfer functions (MTF) and phase transfer functions (PTF) from the different stages in the acquisition portion of the image chain. The sensor simulation is intended to not only degrade an image to make it look real, but to do so in a manner that conserves the image's radiometry. Some common transfer functions steps include; effects from the atmosphere, optical diffraction, detector size, and scanning motion. The chain is modeled in a modular format that allows for simplified use. AVS was chosen for the operating platform because of its "drag and click"user interface. The sen sor model includes the addition of noise from various stages and allows the user to include any noise type. The frequency representations of the images are calculated using the Fast Fourier Transform (FFT) and the optical transfer function (OTF) for the exit pupil function is calculated by an auto correlation of a digital representation of the exit pupil. Analysis of the simulation image quality is conducted by comparing the empirical MTFs between a truth image and a simulated image. Also, a visual comparison between the image features is made for further validation.