TDOA likelihood surfaces for a simulated 2D, two animal dataset corresponding to (a) two TDOAs heard on hydrophones shown with triangles; (b) two TDOAs heard on a different pair of hydrophones shown with triangles; (c) entry-wise multiplication of the sub-surface from (a) and (b); (d) the use of all hydrophone pairs. The two peaks in (d) correspond to the position of the two simulated animals. Black/red/white indicate high/medium/low likelihood values, respectively.
TOA likelihood surfaces at a single time for the simulated dataset corresponding to (a) and (b) multiple TOAs detected on the hydrophone indicated with a triangle; (c) entry-wise multiplication of the sub-surfaces from (a) and (b); (d) the use of all hydrophones. Only one simulated animal is found, corresponding to a call that was emitted by the animal at the search time. Several lower-amplitude spurious local peaks are also evident. The second animal appears when the surface is recalculated at a time corresponding to a click emitted by the second animal.
TDOA vs click time scatterplot for two phones from the multiple-animal dataset (top) and TDOAs extracted using the TDOA tracking method described in Sec. IV C (bottom).
TDOA vs click time scatterplot for the single sperm whale dataset demonstrating persistent lines corresponding to direct-direct arrival associations, reflected-reflected associations, and mis-associated clicks (direct with surface reflected and vice versa). Also evident toward the end of the recording is a sudden offset in TDOA, corresponding to a clock offset on one of the hydrophones.
TDOA (black dots) and TOA (red dots) tracking results. All tracks progress from right to left.
ICIs obtained from TOA position estimates for the leftmost whale in Fig. 5 . ICIs are slowly varying around 1 s, as expected for sperm whale usual clicks.
Article metrics loading...
Full text loading...