Abstract

Traditionally, imaging of ensonified objects is done by beamforming using a fully populated receiver array. When a sparse array is used, grating lobes appear and this is not longer possible. The grating lobe problem can be solved by using a broadband signal: here the grating lobes for different frequencies are in different locations and thus average out. When only one reflective point exists time delay beamforming can correctly identify the location. A new problem arises when multiple reflections occur, creating ambiguities in locating the correct reflective points. These ambiguities are caused by many reflections arriving within the same time window at each sensor of the array. Thus, the number of possible combinations of each arrival in the different channels increases exponentially with the number of arrivals. Most of these combinations do not have a location in physical space, but the remaining ones (expect for the original points) are spurious and place energy in 3d-space that does not align with the original source point Â- making the reconstruction of an object impossible. Incoherent synthetic aperture in combination with new algorithms allows distinguishing between real and false reflective points. Reflection points of an object change only little during the movement of the source creating the synthetic aperture while spurious points Â"flickerÂ" on and off and change location often. These new algorithms allow thus to create an image of the object and to filter out the false aliases.