Detection of Ships in Marine Environments by Square Integration Mode and Multilayer Perceptrons

作者：Vicen Bueno Raul^*; Carrasco Alvarez Ruben; Jarabo Amores Maria Pilar; Nieto Borge Jose Carlos; Alexandre Cortizo Enrique

来源：IEEE Transactions on Instrumentation and Measurement, 2011, 60(3): 712-724.

DOI：10.1109/TIM.2010.2078330

• Artificial intelligence
• artificial neural networks (ANNs)
• clutter
• multilayer perceptrons (MLPs)
• nonlinear signal processing
• radar
• remote sensing
• sea waves

A novel method for detecting ships in marine environments is presented in this paper. For this purpose, the information contained in the marine images obtained by a measuring and monitoring marine system is used. The ship detection is done by multilayer perceptrons (MLPs). In the first approach, the MLP processes the information extracted from the images using horizontal or vertical integration modes. However, if a suitable combination of these integration modes is done, better detection performances are achieved. Therefore, the use of an improved integration mode is proposed, which is based on a square shape. These modes are also used in a commonly used detector, the cell averaging constant false alarm rate (CA-CFAR) detector, which is taken as reference in our experiments. The comparison of the performances of both detectors shows how the MLP-based detector outperforms the CA-CFAR detector in all the cases under study. This comparison is based on objective (probabilities of false alarm and detection) and subjective estimations of their performances. The MLP-based detector also presents another advantage, particularly when the square integration mode is considered: high-performance robustness against changes in the marine environmental conditions.

• 出版日期2011-3