%0 Journal Article %T Automatic Sea-Surface Obstacle Detection andTracking in Forward-Looking SonarImage Sequences %+ Lab-STICC_ENSTAB_CID_PRASYS %A Karoui, Imen %A Quidu, Isabelle %A Legris, Michel %< avec comité de lecture %@ 0196-2892 %J IEEE Transactions on Geoscience and Remote Sensing %I Institute of Electrical and Electronics Engineers %S Geoscience and Remote Sensing, IEEE Transactions on %V 53 %N 8 %P 4461-4469 %8 2015-06-03 %D 2015 %R 10.1109/TGRS.2015.2405672 %K Multitarget tracking %K obstacle avoidance %K seasurface obstacle detection %Z Engineering Sciences [physics]/Signal and Image processingJournal articles %X Automatic sea-surface object detection and trackingfor safe autonomous underwater vehicle and submarine surfacingis a critical issue in relation to the accidents reported in the lastdecades. Here, we propose an efficient tool to detect and tracksea-surface obstacles by processing forward-looking sonar images.The proposed method can detect either still or moving objects withand without wake. For each image sequence, a sequential procedureis proposed to detect various obstacle signatures. Then, targetpositions and velocities are estimated in Cartesian coordinatesusing the debiased converted measurement Kalman filter and thejoint probabilistic data association filter. Detection and trackingstages exchange information in order to reduce the number of falsealarms. Promising results are obtained using real data collected atsea with various objects and scenarios. %G English %L hal-01203532 %U https://hal.science/hal-01203532 %~ UNIV-BREST %~ INSTITUT-TELECOM %~ ENSTA-BRETAGNE %~ CNRS %~ UNIV-UBS %~ ENSTA-BRETAGNE-STIC %~ ENIB %~ LAB-STICC %~ INSTITUTS-TELECOM