The invention provides an
oil film detection method and
system based on feature enhancement and
hybrid optimization, and relates to the technical field of target detection. The method comprises the following steps: extracting multi-dimensional features capable of representing
oil film physical scattering characteristics from original
radar image data; clustering is carried out based on the multi-dimensional features, and an
oil film interest area
mask is obtained; acquiring an optimal segmentation threshold by using a
hybrid optimization
algorithm, and performing optimal segmentation on the oil film interest area
mask by using the optimal segmentation threshold to acquire an oil film binary
mask so as to acquire an oil film detection result; wherein the
hybrid optimization
algorithm is a
beaver behavior optimization
algorithm and a
simulated annealing algorithm. The method has the advantages of being high in detection robustness, low in
false alarm rate, high in segmentation precision, high in self-adaption, high in
processing efficiency, capable of meeting real-time performance, high in unknown scene adaptability, capable of being applied to ship-borne
radar real-time marine
oil spill monitoring and important in practical value.