Genetic algorithm-based deep reinforcement learning model for estimating chlorophyll-a concentration from remote sensing data

被引：0

作者：

Lan, Xinmei ^{[1
,2
]}

Ye, Yang ^{[3
]}

An, Jianfeng ^{[4
]}

Fang, Xinxin ^{[1
,2
]}

Liu, Yuanda ^{[1
,2
]}

Xue, Guokun ^{[3
]}

机构：

[1] School of Geomatics, Information Service, Liaoning Technical University, Liaoning, Fuxin, China

[2] School of Geospatial Information Service, Liaoning Technical University, Liaoning, Fuxin, China

[3] Dalian Huangbohai Marine Surveying Data Information Co., Ltd, Liaoning, Dalian, China

[4] College of Geomatics, Xi'an University of Science and Technology, Shaanxi, Xi'an, China

来源：

Journal of Biotech Research | 2023年 / 15卷

基金：

中国国家自然科学基金;

关键词：

Computer programming - Deep learning - Genetic algorithms - Learning algorithms - Plankton - Reinforcement learning;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

ChlorophyII concentration relies on the plankton present on the ocean surface that is affected by climatic changes. Remote sensing applications deploy sensing devices on the ocean surface for detecting the characteristics of plankton behaviors. Such a process generates time-dependent data for analyzing the chlorophyII concentrations. This article introduced a Hybrid ChlorophyII Concentration Estimation Scheme (HC2ES) using remote sensing application data. The hybrid process was employed by using a genetic algorithm and deep reinforcement learning for data segregation and concentration analysis. The genetic process assimilated the low and high chlorophyII (sensed) data for identifying the climatic impact on vegetation. The no-concentration data was segregated by using the mutation operation of the genetic process. The mutated data was recurrently analyzed by using current and previous concentration levels (data) in the reinforcement layers, which provided further classifications on high and low chlorophyII concentrations for providing insights into remote sensing expansions. The joint hybrid process filtered the fewer concentration data preventing its analysis from identifying chlorophyII and its associated inputs. The proposed scheme was validated by using analysis rate, classification, concentration detection, and computation time. © (), (). All Rights Reserved.

引用

页码：204 / 218