Object ensonification with active ultrasonic wave

This introduces a curated ultrasonic echo database designed for sonar cross section (SCS)\u2013aware object recognition and representation learning. The corpus comprises 880-echo recordings acquired at 250 kHz sampling using a wideband emitter and a binaural microphone pair at fixed range (50 cm). Targets include three 3-D objects (e.g., cube, cylinder, and sphere-cone) each measured over 1~3 views aspect angles to capture view-dependent SCS signatures (plat surface, edge, and round etc). For each target\u2013view, 110 - echoes were recorded and annotated in file name as \object_channel_view in degree_distance in meter_idx.wav\.