Research On Technologies Of Real-time Three- Dimentional Acoustic Imaging Based On Cross Array

As a key research at present, the underwater real-time 3D acoustic imaging has been widely used in the field of oceaneering, seabed topography, underwater construction research, survey of shipwrecks and marine applications. However, the large number of the transducers in the planar array leads to a huge hardware cost and computational load, which restricts the development of the underwater real-time 3D acoustic imaging system. As an optimized array configuration, the cross array can reduce the complexity of the system efficiently. However, the large time taken to scan the whole 3-D scene prohibits the cross array to be applied in real-time 3D acoustic imaging system. Thus, the technologies of real-time 3D acoustic imaging and sparse array based on cross array have been researched in this thesis, and a low-complexity real-time 3D acoustic imaging system has been designed based on the research results.1. To reduce the scanning time of the cross array, a multi-frequency transmitting beamforming algorithm is proposed. The algorithm subdivides the vertical direction into several sectors. In each sector, a series of narrowband fan-shaped beams with different frequencies are emitted sequentially, and each beam is steered to a specific vertical direction. Thus, the number of transmissions is decreased to the magnitude of subdivided sectors rather than steering directions. Moreover, a post-processing time-delay compensation is presented, which avoids the increment of performance cost and memory requirement. A comparison of the performance between the planar array and the cross array is given, and the result reveals the superiority of the proposed algorithm.2. To overcome the large computational load in the receiving beamforming procedure, a parallel subarray beamforming algorithm is proposed. The algorithm subdivides the receiving array into subarrays, and the first-stage beamforming is implemented in parallel on each subarray. Then, the divided subarray is treated as a basic element, and the second-stage beamforming in the beam-domain is performed. Due to the small size of the subarray and the two-stage parallel and pipeline framework, a great computational advantage can be achieved. Compared to the direct method, the proposed algorithm reduces the computational load of the system with an acceptable beam pattern performance.3. To further reduce the number of transducers in cross array without degradation in beam pattern performance, a sparse array with the fewest array elements is proposed. Based on the algorithms aforementioned, the energy function of the simulated annealing (SA) algorithm is redefined, and the optimized SA algorithm is used for thinning the cross array. Compared to the other planar sparse arrays, the sparse array based on cross array can obtain similar beam patterns with the fewest array elements.4. Based on the previous research, a low-complexity (in both hardware cost and computational load) real-time 3D acoustic imaging system with cross array has been designed. First, considering the manufacturing technique of the transducer, the requirement of system and other restriction, the parameters of the system have been decided. Second, the design of the hardware framework and algorithm logic is introduced in detail. Finally, the simulation and real experiment are carried out, which demonstrates the availability, reliability and the practical performance of the low-complexity real-time 3D acoustic imaging system.