Contour Image Match and Its Application in Underwater Navigation and Positioning

Abstract

This paper presents a new method to fulfill the contour match of seabed topography through image match algorithms. Firstly, the generation of topographic contour image is studied. Then, the theory of image matching is introduced. Based on above, a complete matching process is given out. Finally, a positioning experiment is implemented by means of contour image match. Positioning accuracy is better than 5 meters. Different from the navigation above water surface, the underwater navigation requires the navigation has high accuracy, stability and continuity, at the same time, emphasizes the concealment. Besides, many approaches used in the navigation above water surface, such as radio, satellite and light, cannot be used in underwater navigation due to the median propagation in sea water. Although sound wave has better propagation performance than others, larger-size sound array will exposure the underwater vehicle. The existing methods of underwater navigation include inertial navigation system (INS), gravity and geomagnetic matching navigation and so on. INS is a traditional independent navigation system, and is often used as an assistant navigation system. However INS often suffers the influence of inertial drift and results in accumulation error in orientation. The longer the time lasts, the bigger the accumulation error will become. The character decides that INS cannot be used as a main unit in long-time navigation, only as an assistant unit. Underwater topographic match navigation is a kind of independent navigation methods. By means of real-time sounding result and background seabed topography, underwater vessel position can be determined by the match between real-time sounding series and background topography. Nowadays, multibeam bathymetric system (MBS) is widely used in hydrographic measurement. MBS observation presents seabed topography with complete- coverage swath (1) . The data quantity of MBS swath is far more than that of single-beam sounding system, which provides a possibility to fulfill the high-accuracy underwater navigation based on seabed topographic match. These characters make that the navigation based on seabed topographic match can be hardly implemented by means of the gravity match algorithm. For the algorithms used for matching navigation, two traditional algorithms, terrain contour matching (TERCOM) and iterated closest contour point (ICCP), are often adopted. However both of them often produce low-quality matching results even error matching due to their mechanism defect, especially in the area where the characteristic changes of marching background are similar (2) . In order to fulfill the rapid and accurate match between the observed seabed topography and the background seabed topography, we refer the method of image match, and give out the method of contour image match for underwater navigation (3) . The new method is studied and depicted as following. II. GENERATION OF TOPOGRAPHIC CONTOURS IMAGE MBS can acquire a large number of scatter points in a short time. Using these scatter points, underwater contour can be drawn. There are many traditional methods for drawing contour of seabed topography based on building regular grid such as inverse weighted distance average method and Kriging and so on. Considering the contour precision, interpolation speed and regularity of the points, we adopt the method of inverse weighted distance average method, which is an interpolation method by the weighted average (4) . The smaller the distance between sampling point and estimated point (grid node) is, the greater the impact of sampling point to the estimated points will be. The interpolation method can be expressed as following.