Abstract:We put forward a method to easily generate a single or a sequence of fully developed speckle patterns with pre-defined correlation distribution by utilizing the principle of coherent imaging. The few-to-one mapping between the input correlation matrix and the correlation distribution between simulated speckle patterns is realized and there is a simple square relationship between the values of these two correlation coefficient sets. This method is demonstrated both theoretically and experimentally. The square relationship enables easy conversion from any desired correlation distribution. Since the input correlation distribution can be defined by a digital matrix or a grayscale image acquired experimentally, this method provides a convenient way to simulate real speckle-related experiments and to evaluate data processing techniques.

To detect the weak sources of interest area which are obscured by strong sources, a new algorithm of near-field localization of underwater noise source based on matrix filter is proposed. This algorithm use matrix filter as a pretreatment to filter array data of vector sensor, optimized design method is used to design the matrix filter, and then formulate the problem into a second-order cone programming model to calculate the matrix equation, which can strictly control the attenuation of stop band, and it can avoid increment of noise power by restrain the norm of matrix filter. With this filter matrix, a new covariance matrix is produced for the following beam-forming processing with vector sensor array. Compared with the previous work, this algorithm can accurately locate the weak sources of interest that are obscured by strong sources; in addition, higher location accuracy and spatial resolution, better spatial anti-aliasing performance, as well as stronger capability to distinguish the starboard and larboard benefit from using vector sensor array. The simulation and real data show that this algorithm can precisely locate the weak source of interest among some strong sources.