PRE AND POST PROCESSING APPROACHES FOR EFFICIENT FEATURE MATCHING PROCESS

Generally, extracting keypoint descriptors andfeature matching are the steps that
are used to detect moving objects. Unfortunately, the existing feature matching
algorithms may not work well under noisy environments such as changing
illumination, view distortion, geometry transformation and less overlap features.
Varying the environment has to be considered as an option because surveillance
systems monitor scenes from day to night in an indoor or outdoor environment. The
inability to provide accurate matched point would affect the overall performance of a
surveillance system. In this paper, the accuracy of SURF feature descriptors used in
feature matching between two input images of extreme illumination levels are
evaluated. Based on the evaluationresults, a novel pre-processing method to equalize
both images intensity with respect to each other while maintaining the image content
is proposed.