Assessment of distortion in watermarked geospatial vector data using different wavelets

Abstract

With the rapid growth of the Internet, the copyright protection problem occurs frequently, and unauthorized copying and distributing of geospatial data threaten the investments of data producers. Digital watermarking is a possible solution to solve this issue. However, watermarking causes modifications in the original data resulting in distortion and affects accuracy, which is very important to geospatial vector data. This article provides distortion assessment of watermarked geospatial data using wavelet-based invisible watermarking. Eight wavelets at different wavelet decomposition levels are used for accuracy evaluation with the help of error measures such as maximum error and mean square error. Normalized correlation is used as a similarity index between original and extracted watermark. It is observed that the increase in the strength of embedding increases visual degradation. Haar wavelet outperforms the other wavelets, and the third wavelet decomposition level is proved to be optimal level for watermarking.