Enteromorpha detection method and enteromorpha detection device

Abstract

The invention provides an enteromorpha detection method and an enteromorpha detection device. The enteromorpha detection method comprises the steps as follows: using a deep learning method to design a model for selected sample data to obtain a default convolutional neural network model; iteratively training the obtained default convolutional neural network model for multiple times, and determining the default convolutional neural network model as an actual application convolutional neural network model when the precision of the default convolutional neural network model reaches a preset threshold; and detecting green tide information of a collected to-be-detected area according to the determined convolutional neural network model, and outputting a detection result indicating whether the green tide information is enteromorpha. According to the invention, a large number of data is calculated automatically by using a convolutional neural network to obtain hyper-high-dimensional features, the staff participation is reduced, and the calculated sample features are highly reliable relative to the corresponding thresholds. The method and the device are of high robustness. The problem concerning the influence of the environment, satellites and other factors is solved, the workload is reduced greatly, and the identification effect is improved.

Description

technical field [0001] The invention relates to the technical field of marine remote sensing recognition, in particular to a method and device for detecting Enteromorpha based on a convolutional neural network. Background technique [0002] Green tide is a harmful ecological phenomenon in which certain large green algae (such as Enteromorpha) in seawater proliferate explosively or accumulate highly under certain environmental conditions, causing water discoloration. Among them, green tide can lead to marine disasters. When ocean currents roll a large amount of green tide algae to the coast, the green tide algae decay and produce harmful gases, destroying the coastal landscape, and may also cause damage to the intertidal ecosystem. From 2008 to 2012 in China, green tide disasters occurred in summer for five consecutive years in the Yellow Sea of ​​China. Therefore, in order to better determine the occurrence process of green tide, and achieve the purpose of preventing, contr...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a method and device for detection of Enteromorpha, which uses convolutional neural network to automatically calculate a large amount of data to obtain ultra-high-dimensional features, reduces the participation of personnel, and has high robustness; and solves the problem of The influence of multiple factors such as the environment and the satellite itself greatly reduces the manual workload and improves the recognition effect

Images