Real-time Detection And Classification Of Ship By Computer Vision

Since the 70's, Navigational development is rapidly growing; and so the number of ships become more and more increasing and traffic density in sea area and the hazardous goods loading also increase. Marine accidents frequently occur and serious threat to ships navigation security and sea ecological environment. Therefore, Safer Navigation becomes very important issue to be considered. In the long-term research for safer navigation, people realize more and more profoundly that the importance of information exchanging and ship classification. Ship detection and classification is very important to prevent the ships from causing marine accidents. Ship Classification is a system for safeguarding life, property and the environment at sea. For Marine Watch System, in addition to radar and the automatic identification system (AIS), the Electronic Chart Display and Information System (ECDIS) has been applied wisely in recent years. For the sea surveillance, if we only want to detect the ships that are approaching the shore or other ships, radar system is the best choice. But radar only knows some objects are approaching and it is hard to classify what kind of object it belongs to. Moreover, it cannot detect object in its blind area..(AIS) is new equipment to identification and avoiding ship collision. Compared with radar, it is a revolution in tracking technique. Automatic identification system (AIS) is required for the convention ships of 300 GT and over when the SOLAS Chapter V as revised comes into force on July 1, 2002. Since the ships that weigh below 300 GT have no duty to install, that is, not all vessels are installed with AIS and ships without AIS installed cannot be detected and identified. Besides, AIS may be switched off under the dangerous condition. This might be the case in sea area where pirates and armed robber are known to operate. Therefore, we should consider other possible ways helping to get vessel information and identification of vessels that are not equipped with AIS and cannot be detected by radar at sea. Under these circumstances, other marine watch support systems should be constructed and used extensively.Computer Vision is a dynamic and well-developed research field. Many researches are currently in progress on the application of Computer Vision to Traffic Surveillance System. Our study approaches the Safer Navigation via Computer Vision and Image Processing techniques, that is, monitoring the other ships and tries to get information via these techniques. We can apply the computer vision for marine watch system two ways; shore-based (shore-to-ship) system and ship-to-ship system. In our work, we only consider for the shore-based system and the image acquisition camera is stationary and stands alone. Firstly, we acquire the ship image via CCD camera and then we consider the classification method to recognize and classify the ship. In our work, we use image processing techniques and neural network to classify the ship. By using Visual C++ software development tool, we make image-processing operations easy to write in a compact and clear manner. Before extraction the special features of ship image, we make some pre-processing works on image in order to get better result.. Then, we extract ship's feature information from image such as total number of edge pixels, the overall length, height and perimeter, object region area, ship color information,etc.and these are used as the input of neural network which classify the ships.In our work, we use the feed-forward neural network trained using BP learning algorithm to classify the ships. We show in a clear way how our 'Feed-Forward Back Propagation Neural Network' works for ship classification. We can classify the ships into three categories: small, medium and large and can't classify the approximate DWT of ship. We also provided the experimental results using actual data of ship which demonstrate the effectiveness of the method. We also train this classification network with the actual dimensions of ships. We use the actual data of the ship dimensions such as overall length and height of the ships to the input of network. We considered the minimum, maximum and average dimensions for each type of vessel under various DWT. The data we use are obtained from 'Navigational Channel Side Slope & Design Ship Size'. We classify the ships into three categories, that is, ships under 5000 DWT, ships between 10000 DWT to 50000 DWT and ships over 70000 DWT. The network is well trained and can perform successfully the classification of above three categories. Moreover, the network can predict the class of ship, that is, the approximate DWT of ship, by giving its dimensions to input of network. Our proposed method is effective and can give the correct classification rate 88% when we use the parameters in image and higher rate of 96% when we use the actual data of ship.Moreover, we present a method to recognize the type of ship. There are altogether 96 ship patterns are used as the target ships to be recognized. In our work, a feed forward neural network trained using the back-propagation learning algorithm is also used to classify the ship patterns. There are altogether 96 ship patterns are used; 32 ship patterns are used as training pattern sets and 96 ships are used as the input target to recognize. Target input ship patterns contain the ideal image of ship patterns in training sets and their noisy image patterns. We test the network to be able to handle noise because in practice, the network cannot receive a perfect ship image pattern as input. To obtain the network not sensitive to noise, we trained with the ideal copies and the noisy copies of images in ship patterns. The noisy image is made by adding 10% and 20% of noise to ideal image. This force the neurons to learn how to properly identify the noisy image; while requiring that it can still response well to ideal image. Our proposed network is robust, not sensitive to noise, easy to implement and suitable for real-time use. The experiment proved that the network performs well on ship patterns and classification can be done correctly by using the proposed method. As the experimental result, the average recognition (classification) rate for ideal ships was obtained 97 % and 85.7% for noisy images. Moreover, we train the network with the ship patterns collected under various environmental conditions; sunny, rainy, foggy and night. We also provided the graphical user interface (GUI) which allows the user to simulate the classification result. The simulation results show that the performance of the proposed method is easy to use and much more intuitive.In conclusion, along with the development in science and technology; VTS will become Big information network with high technologies and the important marine management system in near future. But ship identification and classification still important issue for VTS. Therefore, we tried to approach the application of computer vision to marine watch and management system. The objective of this study is to achieve the integration of the traditional navigation equipment (system) with a computer vision system Our proposed method can detect and classify the ships in port area and helps to improve the efficiency of VTS; surveillance of the sea traffic area, improve the safety and to reduce the risk of vessel traffic accidents in port areas and approaches, monitor accidental situation and improve efficiency and operational activities of vessel traffic flow in sea area.