ically like a PSF, but needn’t be the same function. For narrow kernels, the pixel-mixel

“voting” is almost 1-to-1, but for diffuse kernels, each mixel value is the weighted combination of

information from many pixels, leading to a “blurred” composite. In fact, if a small kernel is used

that accurately models the actual PSF, and the noise content of the imagery is relatively small, this

becomes a quick method for producing a super-resolved image-ically like a PSF, but needn' t be the same function. For narrow kernels, the pixel-mixel " voting"　is almost 1-to-1, but for diffuse kernels, each mixel value is the weighted combination of information from many pixels , leading to a " blurred"　composite. In fact, if a small kernel is used that accurately models the actual PSF, and the noise content of the imagery is relatively small, this becomes a quick method for producing a super-resolved image