Sea surface oil spilling detection method based on a multi-scale feature deep convolutional neural network

Abstract

The invention provides a sea surface oil spilling detection method based on a multi-scale feature deep convolutional neural network. The sea surface oil spilling detection method comprises the following steps: establishing a deep convolutional neural network structure for sea surface oil spilling detection; constructing a deep convolutional neural network model and selecting a training sample to train the deep convolutional neural network model; carrying out sea surface oil spilling detection by using the trained deep convolutional neural network model through the image; wherein the image is afusion image with multi-scale characteristics, which is obtained by carrying out wavelet reconstruction on an original hyperspectral sea surface image based on Daubechies wavelet transformation. According to the sea surface oil spilling detection method based on the multi-scale feature deep convolutional neural network, the problem that single-scale features are greatly interfered by sea surfaceflare is considered, through combination of multi-scale spatial features, flare, noise and other high-frequency components can be restrained, and the accuracy of oil spilling detection is improved.

Description

technical field [0001] The invention relates to the field of remote sensing applications, in particular to a hyperspectral remote sensing detection method for sea oil spills based on a multi-scale feature deep convolutional neural network. Background technique [0002] Oil spills are caused by the leakage of crude oil or oil products during offshore oil exploration, development, and transportation. It has been listed as one of the 32 scientific problems to be solved before 2030 announced by the American Academy of Sciences. In recent years, sea oil spills have occurred frequently around the world, the marine environment has been seriously polluted, and public health is facing a huge threat. Accurate monitoring of sea oil spills has become the premise and basis for rapid response and countermeasures. [0003] Hyperspectral has the advantage of combining maps and spectra. The detection and identification capabilities of oil spills can be improved through the subdivision of spe...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a sea oil spill hyperspectral remote sensing detection method based on multi-scale feature DCNN with high detection accuracy for the single-scale feature of the existing oil spill detection method that is greatly disturbed by flares and low detection accuracy.

Images