Visualization Of 3D Underwater Acoustic Data

Underwater exploration is becoming more and more important for the growing application requirements involving marine resources exploration, as well as geological and archaeological research. Unfortunately, the complexity of the marine environment and the special properties of seawater as propagating medium have placed limitations on the means of the underwater exploration. As an advanced underwater imaging sonar, synthetic aperture sonar (SAS) has the ability to acquire a high azimuth resolution from a physically small array, thus has a broad prospect. However, compared to optical image, SAS data has lower contrast and weaker ability to distinguish targets due to the interferences. Therefore it's significant to effectively enhance the data, detect the targets and show the processed result.Based on the analysis of classic enhancement and segmentation algorithms, in combination with SAS data characteristics, a set of complete and effective processing scheme involving data enhancement,target segmentation and 3D visualization is proposed by effectively combing a variety of algorithms.First of all, the idea of separating water and seafloor firstly and then processing water and seafloor respectively is proposed. Separation of water and seafloor is achieved by the combination of log transformation,Gaussian smoothing,thresholding and the largest connected component labeling.As for the water, an algorithm based on the inter-scale correlation of wavelet coefficients and fuzzy theory is proposed to enhance the targets. Compared with single-level fuzzy enhancement and adaptive histogram equalization, this algorithm has the best performance and the results are the foundation of segmentation. Then, based on the research of targets characteristics, an automatic segmentation algorithm based on 3-D adaptive region growing is proposed, experimental results show that the algorithm has good performance. Finally, because there are a lot of interferences around the targets in water, thresholding is adopted to remove interferences to achieve better visualization result.As for the seafloor, the segmentation algorithm based on 3-D adaptive region growing which is used to segment the water is adopted to segment the targets in seafloor. Because seafloor has low contrast, Gamma transformation is used to enhance the seafloor to achieve better visualization result.Finally, the processed water and seafloor data is combined, and then the visualization of SAS volume data is achieved by the ray casting algorithm of direct volume rendering.