Piracy Boat Intelligent Identification Method Based On Shipborne Infrared Video Analysis

The infrared imaging technology can compensate for Radar and AIS's weakness in small target detection at sea, and provide new method for the seaman to prevent piracy boat invasion. To protect the lives and property at sea, the infrared cameras are used to monitor the targets around the own ship in all weather, the infrared video would be analyzed to detect, track and identify these targets' behavior, and then to judge whether there is any piracy boat or not. If there is a piracy boat, the system would give an alarm to the deck officer.In the aspect of shipborne multi-camera placement, by analyzing the imaging model, coverage, and surveillance capability of shipborne camera were found, the placement method of single shipborne camera was proposed. And then the shipbore multi-camera optimization model was constructed, which was solved by this paper with genetic algorithm, and the result is a shipborne multi-camera placement scheme.In the shipborne infrared image enhancement, by analyzing its spatial domain characteristic and frequency domain characteristic, the cross-like shape feature of shipborne infrared images' Fourier spectrum was found, and based on this feature the Gauss band pass filter with cross-like shape was proposed, which can repress the sea clutter and enhance the detail of shipbore infrared image.In the infrared target detection and tracking at sea, the method based on FAST corner and region growing was proposed. Firstly, the area containing target or sea clutter is decided by finding the FAST corners of the image. Secondly, the FAST corners are classified into target corners and sea clutter corners by the Support Vector Machine classifier according to their strength and congestion degree. Thirdly, the target corners are regarded as seeds to grow by the region growing method, and the foreground pixels are taken morphological operation in order to get the complete structure of targets. At last, the FAST corners on the target, the position of the target are used to track it.In the domain of intelligent identification for piracy boat, the representation method of target's behavior at sea was proposed. By setting unequal spacing warning circle around the own ship, the target's behavior was represented with the bearing variations when it crossed every two adjacent warning circles. Beside this, the behavior identification methods for piracy boat based on motion and vision were forwarded, and the intelligent identification system for piracy boat was developed.In a word, the problem about camera placement, shipborne infrared image enhancement, infrared target detection and tracking at sea, and piracy boat intelligent identification in the intelligent identification system for piracy boat based shipborne infrared video analysis were studied in this paper.