Hybrid approach for accurate echo detection in the formation of acoustic images

This paper proposes a hybrid approach aimed at precise echo detection for improving short-range underwater images generated by beam-forming process. The final goal is the construction of the scene model as accurately as possible to be used by an autonomous vehicle to navigate in the environment. Our approach is essentially a voting method. It consists in the integration of two methods based on the estimation of the angle of arrival for each time instant and the estimation of the time of arrival for each fixed direction. Each sampled time instant of each beam signal contributes with an opportune quantity on a voting space. At the end of the process, maxima recovery in the voting space and post-processing methods are applied to recover the precise silhouette of the scene. Angular-scan sonar data were simulated in order to verify the validity of the idea. Results are reported in order to compare our method with commonly used echo estimation methods, showing advantages and drawbacks of our approach.