The Scale-invariant Feature Transform (SIFT) by David Lowe is broadly used for image keypoint detection and matching. SIFT is invariant to rotation, scale, illumination, and small changes of viewpoint, but not to radial distortion. Thus it provides poor results when used in images acquired by cameras with a fish-eye, a mini-lens, or a endoscope.
sRD-SIFT is an enhancement of SIFT that provides invariance to non-linear distortion in addition to scale and rotation. It provides much better results when dealing with images like the ones enumerated above. The improvements are both in terms of number of matches and sub-pixel accuracy in localizing keypoints.