Research On Key Technologies Of Signal Processing For Phased Array Three-dimentional Imaging Sonar System

Phased array three-dimensional (3-D) imaging sonar system is a new type of underwater imaging system. It ensonifies the whole viewing scene with one single acoustic bing, and gathers the backscattered signals using a planar transducer array. More than ten thousands of beams are simutaneously generated with the help of the phased array technology, and then real-time3-D images for the underwater scenario are obtained. As a result of the huge number of both the transducers and the beam signals that require to be computed in real time, the underwater signal processor of phased array3-D imaging sonar system is confronted with the problems of quite complicated hardware and too large computational burden. In addition, the imaging quality of the system are degraded greatly by the inevitable transducer array perturbations and the difficulty to identify valid data among the output results of the system beamformer exactly. Therefore, it is of the theoretical significance and practical value to carry on an in-depth study in these key technologies of signal processing for phased array3-D imaging sonar system.A method to design sparse planar arrays considering both the far-field and the near-field conditions is proposed to reduce the huge transducer number and the associated hardware complexities. Based on the study of the near-field beam pattern, a new energy function is brought forward for near-field array optimization according to practical engineering applications. It is also simplified appropriately to reduce the computation. Then the near-field sparse array is obtained by minimizing the simplified energy function. A second-step optimization approach could be implemented if an even higher sparse ratio for far-field were required.A novel calibration method for gain and phase errors in large uniform rectangle arrays is proposed for underwater3-D sonar imaging systems. It requires only one calibrator source at an unknown position in the far field. An efficient and speedy three-step-iteration algorithm is performed first to provide a robust direction of arrival estimator in the presence of gain and phase errors. It then follows with the gain and phase errors estimation using a spatial matched filter. Finally, a maximum a posteriori exercise is executed to further adjust the estimated phase parameters and the source direction. Comparisons illustrate the advantages of the proposed method in computational complexity, scope of application and calibration accuracies. An optimized Chirp Zeta Transform(CZT) beamforming algorithm is developed, which is based on the effecient pruned split-radix FFT algorithm. The orignial routine of2-D CZT beamforming algorithm is firstly optimized and implemented with a series of one dimentional (1-D) discrete convolutions. Then those1-D convolutions are caculated through an efficient pruned split-radix FFT algorithm based on the partial input and output property of the convolutions. The optimized CZT algorithm, which calculates the beam signals accurately, shows advantages in computation with respect to other typical beamforming algorithms, and also keeps computationally efficient for sparse arrays.An adaptive threshold method is proposed for phased array3-D imaging systems. Firstly, beam-strength based confidence image was modified by checking the neighborhood symmetry of each beam. Next, the principle of maximum entropy was introduced to help searching the optimal threshold adaptively for the new confidence image. Then the objects and backgrounds can be distinguished effectively. For underwater3-D imaging systems, the proposed method presents better performances than other representive algorithms, but just has a computation burden similar to traditional spatial filters.