Multiple Index Information Extraction of Marine Oil Spills
ZOU Yarong,ZOU Bin,LIANG Chao,CUI Songxue,ZENG Tao
DOI: https://doi.org/10.3724/sp.j.1047.2012.00265
2012-01-01
Abstract:The oil spills bring great damage to the marine ecological environment even taking decades to repair.Using remote sensing technologies for marine oil spill detection has become a major direction.In this paper,through analyzing the marine oil spill remote sensing detection methods,a new one based on the multivariate index of oil spill information extraction with the SAR data is proposed.First,segmenting the images,and then establishing the shape parameters,the texture feature indexes,and the physical indexes of the spots,the indexes weights were given based on the hierarchical analysis method.For each category index,on the image processing with the on-site validation information,select shape parameters,e.g.the perimeter-to-area ratio,complexity,to establish oil spill shape interpretive level.Choose the texture characteristics parameters from the gray level co-occurrence matrix,e.g.relevance,entropy and change,to establish the oil spill texture feature judgment level.Choose the physical characteristics,e.g.the standard deviation,RMS and contrast,to establish the oil spill physical parameters judgment level.Finally,by calculating the remote sensing information extraction index of the segmentations image dark spots,we can evaluate the credibility of the oil spill remote sensing information extraction,even get a significant basis for the oil spill identification.From the paper,oil spill in the SAR image performance mechanism,the image characteristics and the shape aspects can comprehensively reflect oil spill remote sensing detection characteristics,and the three combination confidence in the remote sensing monitoring has certain practical value.