A Relaxed Matrix Inversion Method for Retrieving Water Constituent Concentrations in Case II Waters: The Case of Lake Kasumigaura, Japan
A Relaxed Matrix Inversion Method for Retrieving Water Constituent Concentrations in Case II Waters: The Case of Lake Kasumigaura, Japan
复制标题
反演案例 II 水域水成分浓度的松弛矩阵反演方法：日本霞浦湖案例
DOI：
10.1109/tgrs.2011.2126048
复制
发表时间：
2011-05
影响因子：
8.2
通讯作者：
Fukushima, Takehiko
中科院分区：
文献类型：
--
作者：
Chen, Jin;Yang, Wei;Matsushita, Bunkei;Fukushima, Takehiko
矩阵求逆法（MIM）是估算II类沃茨中水组分浓度的一种有效算法。为了应用这种方法，水中每种成分的适当和准确的特定固有光学特性（SIOP）是必不可少的。然而，许多湖泊水质的常规观测实际上并不提供SIOP，从而限制了MIM的应用。本文提出了一种基于线性矩阵求逆理论的MIM方法，以放宽对SIOP测量的要求。为此，所谓的ESIOP（估计的SIOP），首先推导出一个不寻常的应用MIM的基础上，足够的校准样品，然后在整个研究区域的水成分浓度检索的标准应用MIM的基础上派生的ESIOP。对于每一个校准样品，需要测量反射光谱和相应的水成分浓度，这可以从定期的卫星数据和例行的实地调查中获得。利用Hydrolight和三个MERIS图像的仿真数据对该方法的性能进行了评估。结果表明，该方法对噪声污染的模拟数据集的水质组分浓度得到了令人满意的估计。对于我们研究区域（日本霞浦湖）的MERIS数据，每个水组分浓度的平均偏差（平均归一化偏差或MNB）和相对随机不确定性（归一化均方根误差或NRMS）范围为-11.2%至3.4%和4.8%至29.7%。这些结果表明，在本研究中提出的算法是理论上合理的，实际上是适用的。
The matrix inversion method (MIM) is an effective algorithm for estimating water constituent concentrations in case II waters. To apply this method, appropriate and accurate specific inherent optical properties (SIOPs) for each constituent in water are essential. However, many routine observations of lake water quality do not in fact provide SIOPs, thus limiting the application of the MIM. In this paper, an alternative MIM method based on linear matrix inversion theory was proposed to relax the requirement of SIOPs measurement. For this, so-called ESIOPs (Estimated SIOPs) were first derived by an unusual application of MIM based on adequate calibration samples; then the water constituent concentrations for the whole study area were retrieved by the standard application of MIM based on the derived ESIOPs. For each calibration sample, measurement of the reflectance spectrum and corresponding water constituent concentrations, which can be obtained from periodical satellite data and routine field surveys, is required. The performance of the proposed method was evaluated using the simulation data from Hydrolight and three MEdium Resolution Imaging Spectrometer Instrument (MERIS) images. The results showed that this method yielded satisfactory estimations of the water constituent concentrations for the noise-contaminated simulation data sets. For MERIS data in our study area (Lake Kasumigaura, Japan), the average bias (mean normalized bias or MNB) and relative random uncertainty (normalized root mean square error, or NRMS) were in the range of -11.2% to 3.4% and 4.8% to 29.7% for each water constituent concentration. These findings imply that the algorithm proposed in this study is theoretically reasonable and practically applicable.
登录
查看更多内容
影响因子：
3.2
作者：
B. Matsushita;T. Fukushima
通讯作者：
B. Matsushita;T. Fukushima
影响因子：
1.9
作者：
HALE, GM;QUERRY, MR
通讯作者：
QUERRY, MR
DOI：
--
发表时间：
2000
期刊：
--
影响因子：
--
作者：
G. Fargion;J. Mueller
通讯作者：
G. Fargion;J. Mueller
影响因子：
13.5
作者：
Gitelson, Anatoly A.;Dall'Olmo, Giorgio;Holz, John
通讯作者：
Holz, John
影响因子：
11.4
作者：
Gons, HJ
通讯作者：
Gons, HJ