Binary feature descriptors are a recent and promising trend in the computer vision field. Nonetheless, they are not yet enough studied when compared to the more established floating-point features. Thus, the need of testing this kind of feature descriptors arises. In particular, in the underwater domain very few works used binary feature descriptors. Therefore, this article tries to explore this recent trend and to test the latest algorithms of this kind. The context of application is Remotely Operated Vehicle (ROV) motion estimation. Experimental data is used to validate each approach and both a qualitative and quantitative analysis is shown. The results show that BRIEF is the best approach for this kind of application. © 2013 IEEE.
Binary visual features for ROV motion estimation
Veruggio Gianmarco;Caccia Massimo;Bruzzone Gabriele
2013
Abstract
Binary feature descriptors are a recent and promising trend in the computer vision field. Nonetheless, they are not yet enough studied when compared to the more established floating-point features. Thus, the need of testing this kind of feature descriptors arises. In particular, in the underwater domain very few works used binary feature descriptors. Therefore, this article tries to explore this recent trend and to test the latest algorithms of this kind. The context of application is Remotely Operated Vehicle (ROV) motion estimation. Experimental data is used to validate each approach and both a qualitative and quantitative analysis is shown. The results show that BRIEF is the best approach for this kind of application. © 2013 IEEE.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.