Study On Oil Spills Information Extraction And Inversion With Airborne Laser-induced Fluorescence On The Sea

It is extremely important to detect oil spills and identify oil source, and to obtain oil thickness timely and accurately on the occurrence of oil spills at sea, so that it can effectively reduce the hazards of the oil spill, and provide an trustworthy basic means to emergency response, incident assessment and identification. The detection of oil slicks on sea surface using airborne laser-induced fluorescence system which can compensate for response lag in satellite remote-sensing and slow search speed, low efficiency in the ship, is one of the most important and most effective oil spill monitoring techniques at present, and it also has the flexible and Real-time features. Through detecting the features of oil fluorescence spectrums, high spatial resolution and high-precision information of oil spills area can be obtained rapidly in real-time, including the type of oil spills, the oil thickness and so on.Based on the principle of laser-induced fluorescence detection, the extraction of oil spills fluorescent information on sea surface was studied in this paper. Via Ground experiments and sea trials respectively validated the airborne laser fluorescence detection system and proved that the entire system was stable and reliable in selection, integration and running chain. The airborne laser fluorescence detection system in the experiment mainly include the components of laser, collimating and beam expanding system, optical collection and spectral spectrophotometry system, photoelectric detection system, data acquisition and data processing system. Raman calibration is a suitable tool for calibrating fluorescence measurements onto a "global" scale that makes it possible to quantitatively compare measurements from different settings on one instrument or between instruments. The algorithm of Curvelet transform with SVM which took advantage of the oil fluorescence spectral shape characteristics and time characteristics was used in order to identify oil type. Except for a relatively high recognition rate, the algorithm has great advantages in small sample and processing speed. So it is especially appropriate for spectrum acquisition and inversion rapidly and simultaneously. In the last chapter of this paper two methods to measure oil film thickness using laser Raman were present, and through the research and experiment their limitations were pointed out becomingly. A method which combines the oil thickness instrument in laser Raman with multiband detector equipped on helicopter was present in order to eventually determine the quantities of oil spills.