The Research On Navigational Vision Enhancement System

Sea and Fog are common weather phenomenon on sea, which decrease the visibility at sea and cause serious degradation on the ship monitoring equipment obtained video image streaming. So it will decrease the reliability of vision navigational information acquired by seamen and has a serious impact on navigation safety in fog. In order to improve ship navigation safety in fog, the thesis researches on navigational vision enhancement system.The following main works have been finished in this thesis:(1) Present an improved haze removal algorithm of sea image. Through establishing fog degraded physical model and analyzing the reasons of the degraded image in fog and the ambiguity of single image haze removal, the algorithm betters the estimated atmospheric light of He proposed dark channel prior haze removal method. Dark channel prior knowledge is preliminarily used to estimate sea image transmission map, then use guider image filtering to refine the preliminary transmission map. The characteristics of large areas of sky light of sea image are used to improve the estimation of global atmospheric light. Finally, the haze removal sea image is recovered based on the above physical mode.(2) Present an improved algorithm of real-time video stream haze removal. Inequality scaling theory is preliminarily used to estimate transmission and the global atmosphere of one frame of marine surveillance video in this algorithm.Finally, recovering the haze-removal surveillance video is based on the above physical mode. According to the marine surveillance video with large area of sky and sea, we estimate the average filtering radius and adjust intensity parameters of the algorithm based on the factors of navigation target details and image contrast. So it achieves real-time marine video stream haze removal at the resolution of 640 multiply 480.(3) Present the architecture of the navigational vision enhancement system and prelimin-arily implement the navigational information fusion into virtual navigational environment and solve series of problems of navigational information real-time fusion into virtual scene.The Thesis implements the real-time video stream haze removal modular of navigational vision enhancement system, which includes the image haze removal modular and video haze removal modular. It also realizes the real-time rendering of virtual navigational environment and navigational information real-time fusion. It will lay the foundation of further research and development of navigational vision enhancement system.