共查询到20条相似文献，搜索用时 15 毫秒
1.
利用卫星多光谱数据反演浅海水深是水深测量的一种重要手段。已有水深反演方法是在研究区建立统一的数学参数的反演模型,未考虑由于海底底质和水质变化导致的空间非平稳性问题。本文使用地理加权回归模型(Geographically Weighted Regression, GWR)对回归参数在空间上进行估计,针对GWR模型的带宽对反演精度的影响,使用了交叉验证(Cross Validation,CV)的方法来确定最佳带宽,并以南海永兴岛和甘泉岛海域为实验区,基于WordiVew-2多光谱数据对使用GWR模型的可行性和精度进行了验证。实验结果表明:永兴岛研究区GWR模型精度较线性回归模型提高了36.05%,在0~5,5~10,10~15和15~20 m区间,精度分别提高了49.46%,39.97%,12.36%和49.68%;甘泉岛研究区GWR模型精度较线性回归模型提高了8.08%,在0~5,5~10,10~15和15~20 m区间,精度分别提高了12.05%,16.23%,4.49%和12.23%,表明GWR模型具有更好的水深反演效果。 相似文献
2.
卫星遥感反演水深（Satellite Derived Bathymetry, SDB）是获取浅海水深信息的有效手段。然而,其有效范围只限于光学浅水区域,在深水区域呈现“伪浅海”的失真现象。因此,如何准确识别SDB数据的有效范围对其广泛应用至关重要。本文基于高空间分辨率多光谱卫星影像,在深入分析深/浅水辐射亮度统计分布特征差异的基础上,提出一种数据驱动的水深反演有效性评价方法。该方法以卫星影像辐亮度信息的局域标准差作为特征,基于K-S检验方法对光学深水区域统计特征进行模型优选,并使用假设检验方法对深水无效区域对应的SDB进行识别。甘泉岛水域实验结果表明,该方法通过统计分布划分光学浅水与深水区域边界,可以有效识别光学深水区域产生的无效水深反演数据。在剔除无效区域数据后,光学浅水有效区域内水深反演平均绝对误差（MAE）为1.01,均方根误差（RMSE）为1.52。实验结果表明,本文提出的方法可准确识别SDB结果的有效区域,进而为浅海地形解译提供方法支撑。 相似文献
3.
利用多光谱卫星遥感影像反演浅海水深是水深测量的一种重要方式。提出一种基于主成分分析的地理加权回归模型(PCA-GWR),采用WorldView-2多光谱卫星遥感影像数据,对经过数学变换后的波段反射率数据先进行主成分分析,将得到第一主成分量进行地理加权回归分析,并与双波段比值模型、多波段线性模型和地理加权回归模型(GWR)的水深反演结果进行比较。结果显示,各个反演模型反演水深值与实测水深值的相关系数r均大于0.75,其中PCA-GWR模型水深反演结果最好,r为0.96、RMSE为1.56 m、MAE为1.06 m。研究表明,PCA-GWR模型可有效去除数据变换后的冗余信息,降低数据空间非平稳性,具有较高的反演精度与可靠性,适用于浅海水深反演。 相似文献
4.
《海洋技术学报》2023,(4)
水深是重要的海洋要素,水深遥感反演是获取浅水水深的重要手段。当前水深遥感反
演应用以国外卫星数据为主,国产卫星数据的研究和应用较少。本文针对国产高分六号卫星
（GF-6） 数据,以三亚南山港为研究区域,分别建立单波段回归模型、双波段比值模型、多波段
回归模型,进行多光谱影像的水深反演能力研究,并与国外主流哨兵2 号卫星（Sentinel-2） 数
据进行实验比较。实验结果表明：GF-6 遥感影像具有较好的浅水水深反演能力和一定的反演精度,各波段水深探测能力依次为：绿波段跃蓝波段跃红波段跃近红外波段,反演方法效果依次为：多波段模型跃双波段模型跃单波段模型。相较于Sentinel-2 数据,GF-6 数据水深反演精度与其一致,这表明GF-6 影像具备替代国外遥感数据进行水深反演的能力和大规模应用的潜力。本文针对GF-6 影像水深反演能力的研究方法和分析,结果将为国产高分系列卫星数据的水深反演研究和应用提供有益的参考。 相似文献
5.
利用WorldView-2四波段卫星数据和电子海图数据,基于改进的耀斑改正算法和双波段比值算法,反演获得了3处典型海域的浅水水深。通过不同海域、不同耀斑条件下水深反演实验,探讨了RED,NIR波段在典型四波段水深反演中的作用和影响,发现在双波段比值法水深反演中,引入RED+NIR波段进行耀斑改正处理,可以增加珊瑚、海藻等绿色物质覆盖海底的反射率,有效地提高该类海域的水深反演精度。基于耀斑改正的多光谱水深反演方法,适用于中轻度耀斑条件下,水质较清澈的浅海水深反演,可在国内外典型四波段卫星数据水深反演中推广应用。 相似文献
6.
Michelle L. McIntyre David F. Naar Kendall L. Carder Brian T. Donahue David J. Mallinson 《Marine Geophysical Researches》2006,27(2):129-136
We present a large-scale quantitative test of a hyperspectral remote-sensing reflectance algorithm. We show that coastal bathymetry
can be adequately derived through model inversions using data from the Airborne Visible-Infrared Imaging Spectrometer instrument.
Data are analyzed from a shore-perpendicular transect 5 km offshore Sarasota, Florida at water depths ranging from 10 m to
15.5 m. Derived bottom depths are compared to a high-resolution multibeam bathymetry survey. Model-derived depths are biased
4.9% shallower than the mean of the multibeam depths with an RMS error of 7.83%. These results suggest that the model performs
well for retrieving bottom depths from hyperspectral data in subtropical coastal areas in water depths ranging from 10 m to
15.5 m. 相似文献
7.
多鳞鱚（Sillago sihama）是山东近海重要的渔业种类之一。本研究根据2016年秋季（10月）在山东近海开展渔业资源底拖网调查取得的数据,分析该海域多鳞鱚的空间分布特征,并运用广义可加模型（GAM）和地理加权回归（GWR）模型探究影响其分布的因素及其与环境因子的非线性和空间非平稳性关系。GAM拟合结果显示,影响秋季多鳞鱚分布的环境因子主要有水深、底层水温和底层盐度,水深的偏差解释率最大,为23.50%。GWR模型拟合结果显示,多鳞鱚分布与水深和底层水温之间存在空间非平稳性关系。水深与多鳞鱚相对资源量呈负相关关系,底层水温与多鳞鱚相对资源量呈正相关关系。赤池信息准则和决定系数（R2）指标对比结果显示,GWR模型的表现优于GAM,在渔业生态数据分析中表现出较好的发展潜力。本研究为今后开展渔业生物空间分布提供了一种新的方法。 相似文献
8.
根据遥感水深反演原理,利用海南岛龙湾港的WorldView-2多光谱卫星数据和海图水深资料,通过对水深进行0~2,2~5,5~10,10~15和15~20 m的分区处理、潮汐改正和海图水深数据与相应图像波段反射率值的相关性分析及回归分析,建立了浅海水深线性回归反演模型,开展了浅海水深的实际计算与精度分析。结果表明:对不同水深范围分别建立线性回归模型反演的水深精度要高于未分区建立的模型;分区模型中,多波段模型在0~5 m的反演精度最高,而双波段比值模型在5~20 m的反演精度最高,但是反演水深在最浅处的精度还有待提高。本文方法提取的水深与海图水深数据变化趋势基本相似,可以满足海洋科学研究对大范围浅水水下地形探测的要求。 相似文献
9.
This paper investigates the use of data assimilation in coastal area morphodynamic modelling using Morecambe Bay as a study site. A simple model of the bay has been enhanced with a data assimilation scheme to better predict large-scale changes in bathymetry observed in the bay over a 3-year period. The 2DH decoupled morphodynamic model developed for the work is described, as is the optimal interpolation scheme used to assimilate waterline observations into the model run. Each waterline was acquired from a SAR satellite image and is essentially a contour of the bathymetry at some level within the inter-tidal zone of the bay. For model parameters calibrated against validation observations, model performance is good, even without data assimilation. However the use of data assimilation successfully compensates for a particular failing of the model, and helps to keep the model bathymetry on track. It also improves the ability of the model to predict future bathymetry. Although the benefits of data assimilation are demonstrated using waterline observations, any observations of morphology could potentially be used. These results suggest that data assimilation should be considered for use in future coastal area morphodynamic models. 相似文献
10.
在多光谱遥感浅海水深反演过程中,考虑到水体和底质影响,水深值和海水表面辐射亮度之间的线性关系不成立。本文以甘泉岛南部0～25m范围的沙质区域为研究区域,利用GeoEye-1多光谱遥感影像和多波束实测水深数据构建XGBoost非线性水深反演模型,研究了XGBoost算法用于水深反演的性能。以决定系数(R~2),均方误差(MSE)和平均绝对误差(MAE)作为评价指标,并与3种传统线性回归模型进行了对比分析。结果表明, XGBoost非线性水深反演模型的R~2、MSE和MAE分别为0.991、0.33m和0.44m,拟合程度最好,精度优于线性回归模型。为进一步探究各模型在不同水深的反演精度,将水深范围分成3段(0～8 m, 8～15 m, 15～25 m)分别进行精度验证和误差分析。结果表明, XGBoost模型在各分段的精度均优于线性回归模型, MSE依次为0.56 m, 0.14 m和0.43 m。可见,在单一底质区域下XGBoost模型的水深反演精度更高,且反演效果更稳定。 相似文献
11.
传统的湖泊、海岸带测深主要是基于船载多波束系统或者机载激光雷达测深系统,但这些方式测量成本较高。因此提出了一种仅利用卫星观测数据,实现高分辨率动态水域地形图的获取方法,该方法基于ICESat-2单光子激光点云和Landsat图像数据的全球地表水数据集(GSWD),对所获取的高精度激光沿轨轮廓线与多年期湖泊水域边界等高线进行融合匹配。以美国最大的水库米德湖为实验区域,生成高程范围约为34 m的地形图结果,覆盖面积超过307 km~2,水平分辨率为30 m;在与机载激光雷达数据等现场实测结果的对比中,所绘制地形图均方根误差约为2 m。研究方法有望为水位波动较大或水质相对较好的内陆水体(例如湖泊)和沿海地区(例如潮间带)提供一种新的水陆交界区域地形图获取方法。 相似文献
12.
针对浅海测深的数据特点和应用需求,以我国南海甘泉岛为例,研究了利用ICESat-2(Ice,Cloud,and Land Elevation Satellite-2)激光卫星数据和光学遥感影像开展主被动融合水深测量的方法。首先通过信号点提取、水面/水底识别、水下点折射改正等步骤处理ICESat-2数据,获得水深值,随后以激光点作为控制,计算光学水深反演模型参数,最后由点及面地获取大范围高精度水深。实验表明,甘泉岛区域主被动融合测深中误差优于1.30m,基于激光卫星数据的主被动融合测深方法能够为浅水水深测量提供新手段。 相似文献
13.
文中首先介绍了DOT2008A模型建立的原理和方法,利用我国48个长期验潮站19 a连续观测数据验证DOT2008A模型的精度,其均方差为0.118 m,残差的标准偏差为0.086 m;联合多源测高卫星建立的海面地形模型数据验证DOT2008A的精度,均方差为0.106 m,残差的标准偏差为0.099 5 m,且其残差分布图基本符合正态分布。从实验结果来看,DOT2008A模型在我国周边海域精度较高,其建立方法较为合理,转化应用价值较高;同时与多源测高卫星建立的海面地形比较结果来看,验证了自主建立的海面地形模型的正确性,对建立我国高精度海面地形模型具有重要的借鉴意义。 相似文献
14.
Regular surveys of coastal zone seabed deliver important information about geomorphologic processes such as silting of waterways. The recent introduction of the Sentinel series of sensors has allowed for the use of satellite sensing for shallow bathymetry morphology monitoring. In this context, this article presents a dedicated Geographic Information System for Baltic Sea shallow water depth monitoring on the basis of Sentinel-2 imagery. The system employs Geovisual Analytics for differential analysis of bathymetry changes as well as monitoring the visibility of known wrecks in the coastal waters of Southern Baltic Sea. Results are verified with regard to known changes in shallow water bathymetry between 28 June 2015 and 3 March 2017. 相似文献
15.
Adjustments to satellite constrained navigation are required to match SeaBeam bathymetric data at track crossings due to errors in dead reckoning and inaccuracies in satellite fixes. By shifting one of the SeaBeam swaths involved in a track crossing relative to the other and calculating the sum of the squares of the differences in bathymetry within the area of overlapping coverage, we map a two-dimensional error surface whose minimum corresponds to the best estimate of the correction to navigation required at the crossing point. Estimates of the covariance of this correction are derived from the error surface. We employ the curve fitting technique of Tarantola and Valette (1982) to invert for a smooth correction function to a starting model of the position of the ship as a function of time. This technique incorporates formal errors assigned to dead reckoning, satellite fixes, and the shifts required to match bathymetric swaths at crossing points in a simultaneous inversion for the correction function for all tracks within the study area.In a test of the method in a study area on the southern Mid-Atlantic Ridge, a data set involving two cruises, 30 days of SeaBeam data, and 753 track crossings, we found that crossing SeaBeam swaths can potentially resolve the relative position of the ship on the two tracks to within 30 to 70 m. The inversion procedure yielded a much better constrained navigation function and much improved match of bathymetry. The final model of the navigation fit crossing shifts about as well as satellite data (with respect to their assigned data errors) with the RMS value of the crossing shifts decreasing from 1200 m in the original satellite-constrained navigation to 200 m in the final solution. However, the potential resolution of position using SeaBeam swaths was not fully achieved in the solution because there are systematic bathymetric artifacts in SeaBeam data, multiple local minima in the error surfaces in highly lineated topography, inadequate dead reckoning data, occasional bad satellite fixes, and limitations on the short period corrections allowed in the model. 相似文献
16.
《Marine Geodesy》2012,35(1):23-43
AbstractRemote sensing is becoming common in the estimation of bathymetry for navigational charting through a process known as Satellite Derived Bathymetry (SDB). Most SDB techniques currently used by hydrographic offices employ an empirical approach, requiring the use of in-situ data to calibrate a relationship between spectral information and coincident depths. This article reports on a multi-site test of an alternative SDB method which uses photogrammetry to extract depths from stereo WorldView-2 imagery. In areas with heterogeneous seafloors, the empirical approach faces difficulties in establishing the relationship between colour and depth, while the photogrammetric approach uses the contrasting seafloor features for triangulation. Additionally, the photogrammetric method may be applied in areas lacking previous survey data. Five study areas in Nunavut, Canada were selected to test the robustness of the method in different environments and under different imaging conditions. Study areas were (with resulting RMSE/Bias given in metres) Coral Harbour (0.84/?0.47), Cambridge Bay (1.16/?0.15), Queen Maud Gulf (0.97/0.06), Arviat (0.99/?0.009), and Frobisher Bay, where extraction largely failed due to environmental conditions. Accuracies demonstrated here are similar to those seen using the empirical approach, suggesting that these two methods may be used in conjunction, each applied to regions where they are better suited. 相似文献
17.
卫星水深反演是水深测量的一种重要手段,其中Stumpf比值算法和Lyzenga多项式算法应用广泛并诞生了大量改进算法,但这些算法没有顾及不同光谱的测深极限与适用范围,为此本文提出一种基于光谱分层的水深反演方法。首先,根据红、绿、蓝光谱对水体的穿透能力差异,提出一种基于影像本身的无参数光谱分层策略,提取红光层、绿光层、蓝光层;然后,根据不同光谱层的波段测深性能,分光谱层构建水深反演优化模型,获取浅海水深反演结果。以我国南沙海域长线礁和美属维尔京群岛巴克岛为实验区,本文方法对经典Stumpf比值算法和Lyzenga多项式算法进行改进后,水深均方根误差、平均绝对误差、平均相对误差分别降低了0.41～0.89 m、0.35～0.65 m、4%～19%,尤其在红光层,即水深较浅区域,平均相对误差降低了58%～149%,精度提升明显。因此,改进算法在提高卫星水深反演效果方面具有可行性和有效性。 相似文献
18.
19.
Synthetic Aperture Radar(SAR)has become one of the important tools for shallow water bathymetry surveys.This has significant economic efficiency compared with the traditional bathymetry surveys.Numerical models have been developed to simulate shallow water bathymetry SAR images.Inversion of these models makes it possible to assess the water depths from SAR images.In this paper,these numerical models of SAR technique are reviewed,and examples are illustrated including in the coastal areas of China.Some issues about SAR technique available and the research orientation in future are also discussed. 相似文献