A seafloor objects detection method based on lacunarity texture feature

For rapid detecting suspected target area in sonar images, reducing the worklord of artificial or computer aided target recognition, a new method of detecting the seafloor objects using the lacunarity texture feature in sonar images is presented, which is helpful to reduce the false alarm rate of object identification. The objects on seafloor usually show significant texture features relative to the seafloor background and shadow in a sonar image, which leads to the variation of the lacunarity value. Therefore, the objects in sonar images deserving to further investigate can be marked by estimating the lacunarity values in different regions of sonar images. The effect of the computational window size on the lacunarity value is analyzed. Finally, the conclusions are verified by using real side-scan sonar images. The results show that the proposed method has has better detection performance. © 2015, China Ordnance Society. All right reserved.