Distortion optimization for wide-angle computational cameras

An imaging system design procedure for miniature wide-angle computational cameras with subsequent software correction of distortion is described. Such miniature wide-angle computational cameras have a broad range of applications, including eye-tracked extraocular cameras for retinal prostheses, and also wearable visual aids for the blind and those with low vision. As significant (typically barrel) distortion is commonplace in wide-field-of-view imaging systems, digital post-processing is often employed to generate rectilinear output images. Relaxation of the constraint on distortion during the optical system design process is shown to allow for improved optimization of other image-degrading aberrations. Analysis of the effects of distortion on the software-corrected final image during optical design is accomplished by using comprehensive image quality metrics such as the correlation coefficient and the spatial frequency response. Selection of a surprisingly large exact initial distortion value as a constraint allows for the design of a miniature wide-angle imaging system that yields significantly enhanced final image quality.