基于BP神经网络算法的内蒙古岱海水体透明度遥感估算

刁瑞翔; 青松; 越亚嫘; 王芳; 刘楠; 郝艳玲; 包玉海

引用本文:	刁瑞翔,青松,越亚嫘,等.基于BP神经网络算法的内蒙古岱海水体透明度遥感估算[J].灌溉排水学报,0,():-.
	DiaoRuixiang,QingSong,YueYalei,et al.基于BP神经网络算法的内蒙古岱海水体透明度遥感估算[J].灌溉排水学报,0,():-.

【打印本页】【下载PDF全文】【查看/发表评论】【EndNote】【RefMan】【BibTex】

过刊浏览高级检索

本文已被：浏览 757次下载 0次
分享到：微信更多字体:加大+\|默认\|缩小-
基于BP神经网络算法的内蒙古岱海水体透明度遥感估算
刁瑞翔¹, 青松¹, 越亚嫘¹, 王芳¹, 刘楠¹, 郝艳玲², 包玉海¹
1.内蒙古师范大学地理科学学院;2.内蒙古大学生态与环境学院

摘要:

【目的】湖泊在地球上有着重要的作用，水体透明度（塞氏盘深度）是衡量湖泊水质的重要指标，可以直观地反映出湖泊的清澈与浑浊程度，为湖泊生态系统提供了水质的有效信息。本文利用BP神经网络算法对内蒙古岱海水体的透明度进行遥感估算。【方法】基于内蒙古岱海实测透明度和光谱数据（地面遥感反射率和卫星遥感得到的反射率），建立BP神经网络水体透明度反演模型，并将此模型应用于Sentinel-2 MSI和Landsat-8 OLI卫星数据，遥感反演岱海水体透明度。【结果】1）本文建立的BP神经网络模型中，最优模型OLI_insitu_220模型的测试集决定系数R2=0.66，均方根误差RMSE=0.23 m，平均绝对百分比误差MAPE=21.56%。2）与传统计算方法相比，BP神经网络算法更适合岱海水体透明度的估算（R2>0.81，RMSE<0.18 m，MAPE<14.97%），反演透明度值与实测值有较高的一致性。【结论】实测与卫星匹配的独立验证进一步显示该算法的有效性，能够客观地反映湖泊水体透明度状况，证明了BP神经网络算法运用在内陆湖泊反演水体透明度的可行性。

关键词: 遥感；透明度；BP神经网络；岱海

DOI：

分类号:TP79

基金项目:国家自然科学基金项目(No.41961057)、内蒙古自治区高等学校青年科技英才支持计划项目(NJYT-17-B04)和内蒙古自然科学基金项目(2019MS04013)联合资助。

Remote Sensing Estimation of Transparency of Daihai Lake in the Inner Mongolia Based on Back Propagation Neural Network Algorithm

DiaoRuixiang¹, QingSong¹, YueYalei¹, Wang Fang¹, Liu Nan¹, Hao Yanling², Bao Yuhai¹

1.College of Geography Science, Inner Mongolia Normal University;2.College of Ecology and Environment,Inner Mongolia University

Abstract:

【Objective】Lakes play an important role in the earth. The water transparency (Depth of The Secchi disk) is an important index to measure the water quality of lakes. It can directly reflect the clarity and turbidity of lakes, and provide effective information about the water for lake ecosystem.In this study, the BP neural network algorithm is used to estimated the transparency of Daihai Lake in Inner Mongolia by remote sensing.【Method】Based on the measured transparency and spectral data(reflectance obtained from ground remote sensing and satellite remote sensing) of Daihai Lake, Inner Mongolia, a BP neural network water transparency inversion model was established, and the model was applied to Sentinel-2 MSI and Landsat-8 OLI satellite data to invert the water transparency of Daihai Lake. 【Result】1) In the BP neural network model established in this paper, the test set determination coefficient of the optimal model OLI_insitu_220 model R2=0.66, the root mean square error RMSE=0.23 m, and the average absolute percentage error MAPE=21.56%. 2) Compared with the traditional calculation method, the BP neural network algorithm is more suitable for the estimation of the transparency of the Daihai Lake (R2>0.81, RMSE<0.18 m, MAPE<14.97%), and the inversion transparency value has a high consistency with the measured value. .【Conclusion】The independent verification of satellite and measured data further shows the effectiveness of the algorithm, which can objectively reflect the water transparency of the lake, and proves that the BP neural network algorithm is a feasible method for inversion of water transparency in inland lakes.

Key words: remote sensing; transparency; BP neural network; Daihai Lake