Wavelet enhanced robot navigation using image sequences

Abstract

Here we propose a framework that exploits the continuity of image sequences that is enhanced by wavelet-based image indexing and searching. We employ an algorithm that characterises the colour variations over the spatial extent of the image in a manner that provides semantically meaningful image comparisons. The indexing algorithm applies a Daubechies wavelet transform for each of the three opponent colour components. The wavelet coefficients in the lowest frequency bands, and their variances are stored as feature vectors. These vectors obtained are much smaller in size than the original image and give faster and more accurate results than traditional algorithms, thereby improving the robot's response.