An improved Harris-FAST algorithm for underwater object corner detection

In order to ensure the images captured by UUV with low contrast and noise accurately matched, an improved corner detection method is proposed. First, The Harris algorithm combined with contourlet transform is applied to detect corners in multi-scale and multi-direction, which overcomes the impact of losing information of corner and position offset. Second, the FAST with adaptive threshold is used to eliminate false corners. Finally Non maxima suppression is applied to restrain corner clustering. Simulation studies are performed based on the experimentally captured gray images to demonstrate the effectiveness and feasibility o f the proposed algorithm.