Research Of Ship Recognition And Motion Tracking

With the development of shipping system,the situation of water surface is becoming more and more complicated.Water environment protection,traffic management,water rescue,crime fighting and other water conditions have caused great difficulties to managers.Therefore,effective recognition and efficient management of surface vessels can improve the efficiency of personnel and improve the management of water surface.In the study of ship recognition and tracking,this paper discusses the complex weather conditions on the surface of the water,and analyzes the problems encountered in the real-time tracking and recognition of the ships.The target image recognition and tracking are realized effectively.The research content of this paper is divided into the following parts:The complex water conditions were discussed,and the weather conditions on the water surface were divided into three categories: conventional mode,night mode and dense fog mode.Gauss convolution is used to denoise the conventional pattern while preserving most of the details of the image.For the night mode,the gray value is stretched,and then Laplasse is used for convolution operation to enhance the image.Retinex algorithm is used to fog the dense fog model,and the multi-scale Retinex algorithm is used to deal with it.The comparison of the two methods is carried out.The preprocessing of the ship image under the weather is effectively realized,and the quality of the image is improved.In order to identify the ship,the LBP cascade classifier is used to train the ship image.Then the ship image identified by the LBP cascade classifier is matched with the standard template library,and the matching coefficient is observed to determine the type of the ship.The two algorithms are combined for efficient data training and fast data recognition.Aiming at ship image tracking,the Mixtore of Gaussian is used to separate the foreground and the foreground,and extract the target area.Then,the morphological operation of the region is processed to remove the small targets in the image.The target area is transformed into HSV color space,and the H component is extracted,and the histogram of the image is obtained to obtain the color probability distribution.Finally,the MeanShift algorithm is compared with the CamShift algorithm,and the tracking effect is compared by analysis and experiment.It realizes the automatic separation of the target from the background and effectively tracks the ship.