Offshore oil spill detection method based on multi-scale conditional adversarial network

Abstract

The invention discloses an offshore oil spill detection method based on a multi-scale conditional adversarial network, belongs to the field of oil spill detection, and solves the problem of low detection precision under a small sample training condition. The method comprises the following steps: (1) constructing a small sample training set; (2) constructing a multi-scale conditional adversarial network; (3) taking a sample pair and downsampling the sample pair to different scales, and taking the sample pair as input components of the adversarial networks of all levels respectively; (4) carrying out adversarial training independently step by step according to a coarse-to-fine scale, and introducing an edge constraint term into a generator loss function to enhance an edge detection effect; (5) setting the output of the current scale generator as the input component of the next scale generator; (6) repeating the steps (3) to (5), and circularly traversing the training set to preset training times; and (7) inputting the oil spill image of the test set into the multi-scale generator model, and outputting an oil spill detection result. In conclusion, the oil spill detection performance under small sample training is ensured through effective fusion of image multi-scale features and enhancement of edge constraint terms on edge detection.

Description

technical field [0001] The invention relates to the field of offshore oil spill detection, in particular to an oil spill detection method based on a multi-scale conditional confrontation network. Background technique [0002] Marine oil spill detection technology is a unique frontier technology, occupying an important position in the field of remote sensing monitoring of marine disasters. Synthetic Aperture Radar (SAR), as the main sensor for marine oil spill monitoring, has the characteristics of all-day, all-weather, strong penetrating power, and wide coverage, and can effectively obtain the location information of the oil spill area. Oil spill detection technology based on SAR images plays a vital role in oil spill range assessment, drift and spread prediction, and oil spill disposal decision-making. In recent years, deep learning algorithms have performed well in oil spill detection in SAR images, and can realize an end-to-end automatic detection mechanism. However, th...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to provide a marine oil spill detection method based on multi-scale conditional confrontation network to solve the problem of low detection accuracy under small sample training conditions

Images