Acoustic image improvement by confidence levels

This paper proposes a method for improving the quality of 2D and 3D underwater acoustic images, operating directly on the information extracted from each beam signal computed by a beamforming process. Some problems may arise when an acoustic image is built directly:on the basis of the envelope of the beam signals (as it is generally done) and this affects the quality of the results. The image quality can improve by using a general criterion operating as an interface between the beam signals and the image pixels. For each beam signal, this measure, called confidence level, assesses the reliability of the actual echoes in function of the similarity with expected echoes. Results obtained, in both real underwater 2D and simulated 3D images, are reported