The effect of rainfall on feature points extraction and image stitching

The aim of this paper is to investigate the effects of rainfall on image stitching by measuring the performance of the Harris corner detector, Scale-Invariant Feature Transform (SIFT) detector, and the Speeded Up Robust Feature (SURF) detector. A set of images were captured by using two cameras during light rain and during heavy rain. The number of detected features points was used as the performance measure. The results indicate that all the detectors are capable to extract a substantial amount of feature points. However, the rainfall does reduce the number of detected feature points by 48 and 67% when the Harris corner detector and the SURF detector were adopted respectively. At the end, the images were stitched together based on the feature points extracted by using the Harris corner detector. The results indicate that the effect of rainfall is negligible as long as there is a substantial amount of distinctive feature points.