Histogram Based Reflection Detection

Reflection appears as a layer that partly covers theoriginal image. This phenomenon may cause failure in extractinginformation from images of reflecting objects. This work presentsan automated technique for determining of reflection area as wellas its severity to define reliability of extracted information. Thisis done by analyzing histogram and objects found in the image.We present a strong detector based on combining the results ofthese two procedures. We presented new reflection detectionalgorithm and a new method to find a good threshold value forconverting any image into a binary image. The average reflectiondetection accuracy of the proposed algorithm is more than 95%for different types of reflection.

Keywords:

reflection detection , histogram , boundary detection