An efficient approach for satellite image building detection based on Tetra wavelet

摘要

Building detection is an important task for disaster management and urban area planning. Many papers are already published to provide the solution of building detection but they are failed with insufficient accuracy. This paper provides a advanced and efficient solution for building detection process using the novel Tetra wavelet method. Tetra wavelet process deals with tetra images which are derived from the original grayscale image. Also it handles the pixel values in between 0 to 3 which is also in terms of tetra. The Tetra wavelet process approaches wavelet procedure which is constructed by XOR operation. The final building detection is controlled by the eight component connectivity method. This approach is suitable for all types of satellite image building detection problems. The experimental result shows that the proposed method improves the segmentation accuracy at a significant level.