Research On Automatic Detection And Tracking Measurement System For Ship Target

In modern sea battle, the problem of effective moving target detection and tracking measurement on the moving platform has become an important research area that has been taken care of national navy of many countries. The research of relative methods is an urgent requirement of developing martial technology on the sea. The research achievements are of very important meanings of enhancing the ability of detecting and recognizing targets on the sea and improving campaign efficiency. It has many advantages to observe, detect, track and measure ship target automatically using the pattern of digital imaging such as low cost, good real time ability and cover-up ability.This dissertation has done some research at the respects of display processing of sea-sky background image, detection of sea-sky line, detection and segmentation of target area, detection of corner and tracking measurement based on digital imaging technology and proposes some new ideas and methods. The dissertation includes some sections as follow:1. The research of display processing methods of sea-sky background image. The features of sea-sky background image and the application of common image enhancement methods on it are researched. A system of digital image processing adapt to various imaging environment is designed and realized. Meanwhile, the method of stability of target in images is researched and realized.2. The research of sea-sky line detection. The application of methods of Hough transform, linear fit and wavelet transform in sea-sky line detection is researched. A new sea-sky line detection method based on area variance and RANSAC algorithm is proposed.3. Three methods of target area detection are proposed on the foundation of sea-sky line having been detected, which are line histogram, area variance and edge detection. Some common methods of ship target segmentation such as region growing and maximum variance between two classes are researched on the foundation of target area detection having been detected. The segmentation methods based on genetic algorithm are researched emphatically and analyzed with an example.4. Several common corner detection methods are researched on the foundation of ship target has been segmented. The Harris method is selected based on the contrast of the experimental results for its good performance.5. Tracking methods of corners and measuring methods based on the position of corners are researched. Common tracking methods are researched and a new track method which syncretizes kalman filter, pyramid algorithm and combines fixed-template with variable-template is proposed. Meanwhile, ship database is established, and the function of camera calibration is relized, and Measuring methods of motion parameters of ship target are given.