Print

发布时间: 2017-06-16
摘要点击次数:
全文下载次数:
DOI: 10.11834/jig.160637
2017 | Volume 22 | Number 6




    综述    




  <<上一篇 




  下一篇>> 





遥感图像中尺度海洋锋及涡旋提取方法研究进展
expand article info 黎安舟1,2, 周为峰1, 范秀梅1
1. 中国水产科学研究院渔业资源遥感信息技术重点实验室, 上海 200090 中国;
2. 上海海洋大学海洋科学学院, 上海 201306

摘要

目的 中尺度海洋锋及涡旋均是重要的中尺度海洋环境特征。中尺度海洋锋及涡旋的提取及其时空分布、变化的研究对海洋生态系统的研究、渔业资源评估、渔情预报及军事等都有重要意义。遥感技术能在同一时间获取大面积海洋要素观测数据,遥感数据具有优良的连续性、同步性,因此遥感数据被广泛应用于中尺度海洋锋及涡旋提取的研究中。 方法 对基于遥感数据进行中尺度海洋锋提取的梯度法、Canny算法、小波分析法和基于引力模型的方法,以及涡旋的提取的Okubo-Weiss法(OW法)、Winding-Angle法(WA法)、基于海面高度的无阈值等值线法和Hybird Detection(HD法)进行总结和分析,并提出对中尺度海洋锋面及涡旋提取方法的见解及新思路。 结果 利用2014年2月南海北部海表温度(SST)数据,分别采用梯度法中的Gradient法、Sobel算法以及Canny算法对南海北部温度锋进行提取并得到该区域温度锋分布图。结果明在多种锋面提取方法中,Canny算法具有较高的效率且其提取结果的连续性和精度更好。中尺度涡的提取方法中,WA法的提取结果具有更好的准确性。早期的中尺度涡提取方法忽略了多中心结构涡旋存在的情况,而后来的HD法能较好地识别多中心结构涡旋。 结论 阈值选取是中尺度海洋锋及涡旋提取的难点和提取结果好坏的关键。然而海洋要素图像弱边缘的特点使得传统边缘检测方法不一定适用于中尺度锋提取。文章通过对不同锋面及涡旋提取方法的总结与分析,为海洋锋面及涡旋提取的研究提供了参考依据。

关键词

中尺度锋; 中尺度涡旋; 遥感; 数字图像处理; 弱边缘提取

Research progress of methods for the extraction of mesoscale ocean fronts and eddies based on remote sensing data
expand article info Li Anzhou1,2, Zhou Weifeng1, Fan Xiumei1
1. Key Laboratory of Fisheries Resources Remote Sensing and Information Technology, Chinese Academy of Fishery Sciences, Shanghai 200090, China;
2. College of Marine Sciences, Shanghai Ocean University, Shanghai 201306, China
Supported by: National Natural Science Foundation of China(31602206);Shanghai Municipal Natural Science Foundationl(16ER1444700)

Abstract

Objective Mesoscale ocean fronts and eddies are important mesoscale marine environment characteristics. The ocean front is the interface of water masses with different properties. In the area where an ocean front exists, corresponding hydrological factors (e.g., temperature, chlorophyll concentration, and salinity) present a high horizontal gradient. Seawater convergence and vertical motion are enhanced in fields where fronts appear; this enhancement leads to the enrichment of nutrients and provides a rich diet to plankton, fish, and so on. Therefore, the sea area where fronts appear can serve as a good fishing ground (e.g., Zhoushan and Minnan fishing grounds in China). Mesoscale eddy plays an important role in ocean circulation and is an important undertaker of energy transport and ocean material transfer in oceans. Furthermore, eddies can influence the distribution of hydrological factors, such as temperature and salinity, and is thus one of the important factors of marine hydrological variation. Eddies associated with local lifting flow, such as the upwelling associated with cold eddies, carry nutrients to euphotic zones from the bottom of the ocean, greatly improve the primary productivity of the ocean, and influence the distribution of the fishing ground. Thus, they affect the development of the marine economy. Remote sensing data demonstrate excellent continuity and synchronization and can effectively reflect the spatial distribution characteristics of marine hydrological elements and the sea surface height. Therefore, remote sensing data, such as sea surface temperature, sea surface height, and sea level anomaly, have been widely used in the extraction of mesoscale ocean fronts and eddies. Research on the extraction of mesoscale ocean fronts and eddies based on remote sensing data is thus significant to marine ecosystem research, fishery stock assessment, and fishing condition forecasting. We aim to provide a reference and ideas for the extraction of mesoscale ocean fronts and eddies by summarizing and analyzing the methods of extracting mesoscale ocean fronts and eddies. Method This study summarizes and analyzes the methods of front extraction, such as gradient method, canny algorithm, wavelet analysis method, and algorithms based on the law of universal gravity, as well as the methods of eddy extraction, such as OW, WA, SSH-based and HD methods. Insights and new ideas are then provided. To show the difference among various front extraction methods intuitively, fronts are extracted from same area by using gradient, Sobel, and canny algorithms. The extraction results are presented in a figure. When gradient and Sobel algorithms are used to extract fronts, the thresholds used to distinguish the background and front pixels are obtained by the iterative method. Afterward, the Zhang-Suen method is applied to implement binary image thinning. The center line of the front is then obtained. When the canny algorithm is utilized to extract fronts, 0.25 is selected as the low threshold and 0.9 as the high threshold. Result With sea surface temperature (SST) data on the northern South China Sea (SCS) in February 2014, gradient, Sobel, and canny algorithms are used to extract the temperature front in northern SCS. A figure is subsequently drawn for the temperature front distribution of this area.Results show that among the various front extraction methods, the gradient method is simple but influenced greatly by noise. The canny algorithm presents a great advantage in front positioning accuracy, continuity, and computational efficiency. Wavelet analysis can be used in multi-scale analysis, but the computation is complex. The algorithm based on the law of universal gravity considers the influence of the value and position of the center and neighborhood pixels, so it has better anti-noise ability and accuracy than the gradient method. Among the various eddy extraction methods, the OW method can identify the eddy core region well, but the extraction result is greatly influenced by the selection of the W value threshold. The WA method presents excellent accuracy but requires extensive calculation to obtain the streamline. The SSH-based method is simple but can only extract the eddy boundary and not the core area of the eddy. Early eddy extraction methods ignore the condition that multi-core eddy structures may appear in oceans, and these methods are unable to identify multi-core eddy structures. The HD method combines the advantages of the OW and SSH-based methods. It can extract the boundary and core area of an eddy simultaneously and can identify multi-core eddy structures. Conclusion According to the summary and analysis of various front and eddy extraction methods, threshold selection for front and eddy extraction is difficult but is important to the quality of the extraction result. Many researchers have examined the threshold selection method. In addition, edge detection methods, such as gradient method and the canny algorithm, that are used widely in front extraction are designed for extracting sharp edges (e.g., solid edge). However, sea water is fluid and characterized by a weak edge. In other words, the edge of water masses with different properties is not obvious and more difficult to identify than a solid edge. Therefore, traditional edge detection methods are unsuitable for mesoscale front extraction due to the weak edge characteristic of marine environment element images. Given that the ocean front is the interface of different water masses, the region growing algorithm, which is used widely in image segmentation, can be utilized to segment the marine environmental element image of the study area into several independent parts that represent different water masses. Then, the boundaries we wish to extract (i.e., the front) are searched.

Key words

mesoscale front; mesoscale vortex; remote sensing; digital image processing; weak edge detection

0 引言

中尺度海洋锋及涡旋均是重要的中尺度海洋环境特征。海洋锋面所在的海域海水辅聚,垂直运动加强,因此往往有各种营养物质富集[1],从而为浮游生物、鱼类等提供丰富的饵料,使得锋面处通常能形成良好的渔场[2-6],例如中国的舟山渔场和闽南渔场存在较强的海洋锋[7]。中尺度涡对海洋环流起着重要作用,是海洋中能量输送和物质传递的重要承担者,同时能影响温度、盐度等海洋水文要素的分布,是决定海洋水文变化的重要因子之一[8-9]。此外,中尺度涡的发生常伴随着局地升降流,如冷涡伴随的上升流能将海洋下层的营养盐携带至真光层,大大提高海洋的初级生产力,影响大洋渔场的分布,从而影响海洋经济的发展[10]。因此,中尺度海洋锋及涡旋的提取及其时空分布、变化的研究对海洋生态系统的研究、渔业资源评估、渔情预报及军事等都有重要意义。中尺度海洋环境特征时间尺度在数天至数月之间,空间尺度则在数十到数百公里之间[11-12]。遥感技术能在同一时间获取大面积海洋要素观测数据,遥感数据具有优良的连续性、同步性,因此遥感数据被广泛应用于中尺度海洋锋及涡旋提取的研究中。本文对基于遥感数据进行中尺度海洋锋面提取的梯度法、Canny算法、小波分析法和基于引力模型的方法,以及涡旋的提取的OW(qkubo-weiss)法、WA(winding-angle)法、基于海面高度的无阈值等值线法和HD(hlybird detection)法进行总结和分析。并提出对中尺度海洋锋面及涡旋提取方法的见解及新思路。

1 中尺度海洋锋提取

1.1 中尺度海洋锋提取研究进展

海洋锋面是不同的水团或水系间的分界面,在锋面处,对应的水文要素(如温度、叶绿素浓度、盐度等)呈现出较高的水平梯度[13]。随着卫星遥感技术的发展,卫星遥感数据被广泛的应用于中尺度海洋锋提取的研究中。在海洋水文要素遥感图像中,海洋锋面可视为图像的边缘信息,因此海洋锋的提取可视为对海洋水文要素遥感图像的边缘检测。目前,关于中尺度海洋锋提取的研究主要集中在温度锋的提取,也有国外学者开始对叶绿素锋的提取进行研究,而国内关于叶绿素锋提取的研究较少。海洋中尺度锋提取的数据源主要为通过卫星遥感反演得到海洋水文要素的数据,如SeaWiFS水色数据[5, 14-16]、MODIS-SST[17-19]和AVHRR-SST[20-22]遥感影像数据。目前基于遥感数据提取中尺度海洋锋的方法主要有统计直方图法[23-24]、梯度法[25-26]、Canny算法[27-28]、小波分析[29-30]、基于引力模型的方法[31-32]等。

1.2 梯度法

梯度法提取海洋中尺度锋的原理是锋面处水文要素呈现较高的水平梯度,因此通过计算图像像元梯度并选择高梯度像元以实现锋面提取。其基本步骤分为梯度计算、阈值选取及图像二值化3步。目前,计算水文要素度梯度的方法主要Gradient[33]法、Sobel梯度算法[15]等。

海洋水文要素遥感影像中,像元$T\left( {i,j} \right)$梯度为

$G = \sqrt {D_x^2 + D_y^2} $ (1)

Gradient法中

${D_x} = \frac{{T\left( {i,j + 1} \right) - T\left( {i,j - 1} \right)}}{{2\Delta X}}{\rm{ }}$ (2)

${D_y} = \frac{{T\left( {i + 1,j} \right) - T\left( {i - 1,j} \right)}}{{2\Delta Y}}$ (3)

式中,${\Delta X}$${\Delta Y}$分别是$X$$Y$方向上的像元大小,及遥感图像的空间分辨率。Gradient法计算简单,但是没有考虑到其他方向上的相邻像元的影响,对噪声较敏感因此用Gradient法计算图像梯度前需要对图像进行滤波处理以减少噪声的影响。

Sobel梯度算法则利用模板$\left[ {\begin{array}{*{20}{c}} { - 1}&0&1\\ { - 2}&0&2\\ { - 1}&0&1 \end{array}} \right]$$\left[ {\begin{array}{*{20}{c}} 1&2&1\\ 0&0&0\\ { - 1}&{ - 2}&{ - 1} \end{array}} \right]$对图像进行卷积运算求得像元${D_x}$${D_y}$值。Sobel梯度算法充分考虑到了不同位置邻近像元对梯度影响及影响程度的差异,从而抑制了噪声对锋面提取的影响。Sobel梯度算法可有效增强栅格图像的边缘可视性,在图像梯度计算中应用广泛[15]

海洋水文要素遥感影像中,锋面像元对应图像中的高梯度像元。因此得到梯度图像后,需要选取合适的阈值并对图像进行二值化处理,即将梯度值大于或等于阈值的像元作为锋面像元,其余为背景像元。

图 1为利用Gradient法和Sobel算法对2014年2月南海北部温度锋的提取结果。锋面提取前对SST图像进行了中值滤波处理以抑制噪声。阈值通过迭代法自动选取,分别为:0.047 5 ℃/km,0.043 ℃/km。此外利用zhang-suen法对二值化后图像进行了细化处理以获取锋面中心线。对比图 1(a)(b)可看出,Sobel算法锋面提取结果中出现的破碎锋较少,说明Sobel算法有更好地抑制噪声效果。

图 1 不同方法锋面提取结果
Fig. 1 Front extration results of different methods
((a) Gradient algorithm; (b)Sobel algorithm)

1.3 Canny算法

Canny算法是近几年在锋面提取中应用日趋广泛的边缘检测方法。与梯度法一样,Canny算法是基于锋面处水文要素呈现高水平梯度的原理来实现锋面提取的,不同的是Canny算法通过非极大值抑制和双阈值检测提高了锋面的定位精度和连续性,此外通过高斯滤波一定程度上去除了噪声的影响。Canny算法的基本步骤包括:对图像进行高斯滤波;梯度计算;非极大值抑制和双阈值检测。Canny算法中,通过高斯滤波去除噪声;通过非极大值抑制去除伪边缘;通过高、低阈值实现锋面的检测和连接,并进一步去除伪边缘[34-35],在有效抑制噪声的同时保证了锋面的连续性。此外,用Canny算法提取海洋锋面可获得海洋锋面的中心线,进而可获取海洋锋面的位置信息。图 2为利用Canny算法对2014年2月南海北部温度锋的提取结果,其中高低阈值分别为0.25、0.9。从图 2可以看出,Canny算法的锋面提取结果具有良好的连续性。

图 2 Canny算法锋面提取结果
Fig. 2 Front extration results of Canny algorithm

1.4 小波分析法

水文要素图像经过小波分析,可得到一系列尺度的近似图像和细节图像,细节图像对应原图像的高频部分,包括边缘和噪声,通过设定合理的阈值可从边缘信息中进一步提取锋面信息[29]。锋面信息在各尺度空间的分布上具有连续性,而噪声没有,即海洋水文要素图像经一系列小波变换后,锋面信息得到保留,噪声信息则逐渐消失,因此可以根据锋面信息和噪声在尺度空间分布上连续性的差异来剔除噪声对锋面提取的影响。通过细节图像的小波逆变换,即可得到研究区域的锋面信息,步骤如图 3所示。

图 3 小波分析法提取海洋锋流程图
Fig. 3 Flow chart of extracting front by wavelet analysis method

1.5 基于引力模型的方法

Sun等人[31, 36]提出了基于引力模型的边缘检测方法,在此方法的基础上可进一步提取海洋锋面信息。基于引力模型的锋面提取方法引用力学的概念,将像元看做天体,像元值则对应像元“质量”,根据牛顿万有引力定律

${F_x} = \frac{{{m_{i,j}}{m_{k,l}}k}}{{{{\left( {\sqrt {{k^2} + {l^2}} } \right)}^3}}}$ (4)

${F_y} = \frac{{{m_{i,j}}{m_{k,l}}l}}{{{{\left( {\sqrt {{k^2} + {l^2}} } \right)}^3}}}$ (5)

$F = \sqrt {F_x^2 + F_y^2} $ (6)

计算邻域像元对中心像元的“引力”矢量和,随后通过设定合适的阈值提取“合力”较大的像元作为锋面像元。

1.6 锋面提取方法对比分析

海洋锋面提取的方法中,梯度法、Canny算法是根据图像梯度的大小对锋面进行提取;小波分析是根据细节图像中像元的连续性差异对锋面进行提取;基于引力模型的锋面提取方法是根据像元所受“引力”的大小进行锋面提取。梯度法提取像元原理简单,在早期的锋面提取中应用最广。但梯度法对噪声较为敏感,提取结果连续性欠佳,且由于梯度算子大小固定,梯度法只能对单一尺度锋面信息进行提取。Canny算法锋面定位精度高,提取结果连续性比梯度法好,且效率高,如图 2中锋面提取结果连续性明显优于图 1,尤其在23°N 27°N,127.5°E 130°E处锋面提取结果连续性的优势最为明显。此外Canny算法提取锋面可以获取锋面中心线,无需对图像做细化处理,不仅能提高效率,更为进一步研究锋面位置提供方便。Canny算法的这些优势使得其在近几年被广泛应用,并有研究者对传统Canny算法进行改进,如王娜等人[37]利用3×3邻域求梯度代替传统算法中2×2求梯度以进一步抑制噪声,并利用8邻域双线性插值的非极大值抑制代替传统算法的非极大值抑制以提高边缘检测精度和准确性;王植等人[38]利用自适应动态阈值代替传统算法中的高、低阈值以防止假边缘出现。但Canny算法也只能对单一尺度锋面信息进行提取。小波分析法可对不同尺度锋面信息进行提取,且能有效区别锋面与噪声信息,但原理较复杂,计算量大,若要获取锋面中心线则需要对图像进行细化处理。基于引力模型的锋面提取方法同时考虑了邻域像元位置及像元值大小、中心像元的像元值大小的影响,抗噪能力强,且能保证锋面定位精度,但也只能对单一尺度锋面信息进行提取,此外也需对图像进行细化才能提取锋面中心线。前述提到的锋面提取方法对阈值选取的依赖性均较大,而目前在海洋中尺度锋提取领域中阈值的选取仍以经验为主,受主观因素影响较大,且阈值选取自动化程度不高。此外,除Canny算法可直接提取锋面中心线外,其他方法只能对锋面形态进行识别,若要提取锋面中心线,则需要对图像进行细化处理(如Zhang-Suen法[39],外部压力法[40]等)。

2 中尺度涡提取

2.1 中尺度涡提取研究进展

中尺度涡是海洋中普遍存在的一种中尺度海洋现象,是指尺度显著小于背景流场的长期封闭式气旋或反气旋环流[10]。中尺度涡出现处海表面高度有明显变化,因此中尺度涡在海表面高度或海表面高度异常遥感图像上特征明显。目前,应用于中尺度涡提取的遥感数据主要有:海表面高度SSH(sea surface height) [41-43]、海平面高度异常SLA (sea level anomaly)等[44-47]。近年来应用广泛的基于遥感数据的中尺度涡提取方法主要有:OW法、WA法、基于海面高度的无阈值等值线法、HD法等。

2.2 OW法

OW法是长期被使用一种海洋中尺度涡提取方法,OW法计算量相对较小,并能有效识别涡旋核心区,因此在海洋中尺度涡提取中被广泛应用。OW法引入一个$W$值来判断某一时刻流场的状态(变形和旋转)[48],并通过$W$值确定涡旋的核心区域,OW法的基本步骤为[49-50]

1) 计算$W$值,即

$W = S_n^2 + S_s^2 - {\omega ^2}$ (7)

式中,${S_n} = \frac{{\partial u}}{{\partial x}} - \frac{{\partial v}}{{\partial y}}$,为剪切形变率,${S_s} = \frac{{\partial v}}{{\partial x}} + \frac{{\partial u}}{{\partial y}}$,为拉伸形变率。$\omega = \frac{{\partial v}}{{\partial x}} - \frac{{\partial u}}{{\partial y}}$,为相对涡度[51]$S_n^2 + S_s^2$为平方变形率,用以描述流体的变形;${\omega ^2}$是涡度拟能,用以描述流体的旋转[52]

2) 确定涡旋核心区域。OW法中,$W$>0时,时,流场以变形为主,$W$<0时,流场以旋转为主,即以涡旋形式存在。因此,可以通过选取合适的阈值对研究区域的涡旋进行提取。以往的研究中常取$W < 0.2{\sigma _w}$的像元作为涡旋核心区,${\sigma _w}$$W$值的标准偏差。

2.3 WA法

WA法是通过SLA计算研究区域流场流线,并通过流线几何形状进行涡旋的识别的。WA法最初由Sadarjoen和Post提出[53],后由Chaignuea等人[54]进行改进,在WA法提取中尺度涡的研究中,以使用Chaignuea等改进后的WA法为主。改进后的WA法利用3×3的移动窗口搜索SLA图像的局部极值,并将搜索到的SLA极值作为涡旋中心,同时将包含局部极值的最外圈流线作为涡旋边界(图 4)。其中图像流线的计算为

图 4 WA法提取中尺度涡流程图
Fig. 4 Flow chart of extracting eddy by WA method

$U' = - \frac{g}{f}\frac{{\partial \left( {{\rm{SLA}}} \right)}}{{\partial y}}$ (8)

$V' = \frac{g}{f}\frac{{\partial \left( {{\rm{SLA}}} \right)}}{{\partial x}}$ (9)

式中,$U′$$V′$分别为地转速度异常分量,$f$为科氏参数,g为重力加速度。

2.4 基于海面高度的无阈值等值线法

WA法中流线的获取需要较大的计算量,而涡旋的流线往往与闭合的海面高度异常等值线平行或重合,因此Chelton等人[43]提出了基于海面高度的无阈值等值线法。基于海面高度的无阈值等值线法将局部SSH极值作为涡旋中心,并将包含涡旋中心的最外圈SSH等值线作为涡旋边界,从而达到涡旋识别的目的[55]。其基本步骤如图 5所示。

图 5 基于海面高度的无阈值等值线法流程图
Fig. 5 Flow chart of extracting eddy by SSH-based method

2.5 HD法

HD法是OW法与基于海面高度的无阈值等值线法的综合。HD法通过OW法的方式确定涡旋核心区,通过基于海面高度的无阈值等值线法的方式确定涡旋边界。HD法提取涡旋的基本步骤是[56]

1) 计算SLA图像的$W$值并将$W < 0.2{\sigma _w}$的像元作为涡旋核心区。

2) 用大小为3×3的活动窗口检测SLA图像的局部极值。

3) 将位于涡旋核心区的极值点设为涡旋中心。

4) 去除浅层涡旋中心。

5) 以0.5 cm为间隔提取SLA等值线。

6) 去除非闭合等值线及直径大于500 m的闭合等值线。

7) 将包含涡旋中心的涡旋核心区的最内层闭合等值线设为涡旋的有效边界。

8) 如果没有完全包含该涡旋核心区的闭合等值线,则选取与核心区域相交的最外层闭合等值线设为涡旋的有效边界。

9) 如果没有与该涡旋核心区相交的闭合等值线,则选取核心区域边界作为涡旋的有效边界。

10) 多中心结构涡旋的检测和边界重建。

HD法综合了OW法与基于海面高度的闭合等值线法的优点,能更准确的对涡旋进行提取并对涡旋中心位置和边界进行检测,并能有效检测到多中心结构涡旋。

2.6 中尺度涡提取方法对比分析

OW法从流体的物理状态出发,引用$W$值对流体瞬时间的状态进行判断,进而对涡旋进行提取。OW法能较好地识别涡旋核心区,然而其提取结果受$W$值阈值选取影响较大,且提取结果往往比涡旋实际半径小,同时可能有检测过剩的现象[54]。WA法通过计算流场流线对涡旋进行识别,能很好地识别涡旋边界,并有很好的准确性,但WA法中流线的获取需要较大的计算量,从而导致该方法涡旋检测效率较低,且当涡旋所在区域流线几何特征不明显时,难以检测到涡旋。基于海面高度的无阈值等值线法计算简单,且无需阈值的选取,从而减少了阈值选取对涡旋提取结果的影响,其缺点是只能对涡旋边界进行识别而不能识别涡旋核心区。HD法是OW法与基于海面高度的无阈值等值线法的结合,因此融合了两种方法的优势,能对涡旋核心区及涡旋边界进行识别,保证了涡旋检测的准确性,并且能识别多中心结构涡旋。

3 结论

在海洋中尺度锋及涡旋提取中需根据研究区域情况和实际需要选择合理的方法进行锋面及涡旋的提取,或对提取方法进行改进。此外,阈值的选取是锋面和涡旋提取的难点,目前阈值的选取易受主观因素的影响,所选阈值过高会导致海洋环境特征信息存在缺损,阈值过低则会降低海洋环境特征信息定位精度。目前仍没有一种理想的、客观的、自适应程度高的阈值选取方法。因此,海洋环境特征信息提取中合理阈值的选取仍需进行大量的工作和研究。目前已有研究者对海洋环境特征提取中阈值选取方法展开研究,如Bo等人[28]根据图像累计直方图特征对锋面提取阈值进行选取并利用双阈值对锋面进行检测。此外,目前锋面提取中应用最广泛的梯度法和Canny算法在算法设计之初均是用于对强边缘(如固体边缘)的提取[15, 34],研究的对象边缘均很明显。而海洋属于流体,海洋中各水文要素在卫星影像中呈现出弱边缘的特征(如图 6),即边缘不明显,因此梯度法和Canny算法并不一定适用于流体边缘的提取[35]。由于海洋锋为两个不同性质水团的分界面,因此可以考虑利用图像分割中区域增长的方法将研究区域的海洋要素图像分割成若干个特征相似的区域,即不同性质的水团,然后提取各水团之间的边界,该边界即为锋面。在对研究区域的海洋要素图像进行分割时,需要选择合适的相似性准则,并通过该相似性准则将特征相似的区域合并。经过分割后的图像边缘特征得到强化,从而解决了原始图像弱边缘的问题。因此先对海洋要素图像进行分割,再对分割后图像进行边缘检测,能取得更好的锋面提取效果。

图 6 2014年2月南海北部SST分布图
Fig. 6 SST distribution of north SCS in February, 2014

参考文献

  • [1] Bost C A, Cotté C, Bailleul F, et al. The importance of oceanographic fronts to marine birds and mammals of the southern oceans[J]. Journal of Marine Systems, 2009, 78(3): 363–376. [DOI:10.1016/j.jmarsys.2008.11.022]
  • [2] Chen X J. Fisheries Resources and Fishing Grounds[M]. Beijing: China Ocean Press, 2004. [ 陈新军. 渔业资源与渔场学[M]. 北京: 海洋出版社, 2004.]
  • [3] Liu Z. Thespatio-temporal variability of oceanic fronts offshore China seas and analysis of marine observations[D]. Beijing:The Institute of Oceanology, Chinese Academy of Sciences, 2012. [刘泽. 中国近海锋面时空特征研究及现场观测分析[D]. 北京: 中国科学院研究生院, 2012.]
  • [4] Pan Y Q, Cao X Z, Xu J P. A preliminary investigation of the cause and characteristics of the upwelling front zone off Zhejiang[J]. Transactions of Oceanology and Limnology, 1982(3): 1–7. [潘玉球, 曹欣中, 许建平. 浙江沿岸上升流锋区特征及其成因的初步探讨[J]. 海洋湖沼通报, 1982(3): 1–7. ] [DOI:10.13984/j.cnki.cn37-1141.1982.03.001]
  • [5] Ning X R, Liu Z L, Cai Y M. A review on primary production studies for China seas in the past 20 years[J]. Donghai Marine Science, 2000, 18(3): 13–20. [宁修仁, 刘子琳, 蔡昱明. 我国海洋初级生产力研究二十年[J]. 东海海洋, 2000, 18(3): 13–20. ] [DOI:10.3969/j.issn.1001-909X.2000.03.003]
  • [6] Wang WY, Shao Q Q, Xue Y C, et al. On the relationship between the resources of ommastrephes bartrami and marine environment in the northwest pacific ocean based on GIS[J]. Geo-Information Science, 2003, 5(1): 39–44. [王文宇, 邵全琴, 薛允传, 等. 西北太平洋柔鱼资源与海洋环境的GIS空间分析[J]. 地球信息科学, 2003, 5(1): 39–44. ] [DOI:10.3969/j.issn.1560-8999.2003.01.008]
  • [7] Ren S H, Wang H, Liu N. Review of ocean front in Chinese marginal seas and frontal forecasting[J]. Advances in Earth Science, 2015, 30(5): 552–563. [任诗鹤, 王辉, 刘娜. 中国近海海洋锋和锋面预报研究进展[J]. 地球科学进展, 2015, 30(5): 552–563. ] [DOI:10.11867/j.issn.1001-8166.2015.05.0552]
  • [8] Cui F J. Mesoscale eddies in the south China sea:identification and statistical characteristics analysis[D]. Qingdao:Ocean University of China, 2015. [崔凤娟. 南海中尺度涡的识别及统计特征分析[D]. 青岛: 中国海洋大学, 2015.]
  • [9] Chen M. Numerical simulation of the circulation and the mesoscale eddies in the east China seas inwinter[D]. Beijing, China:The Institute of Oceanology, Chinese Academy of Sciences, 2002. [陈敏. 冬季东中国海环流及中尺度涡旋数值模拟[D]. 北京: 中国科学院研究生院, 2002.]
  • [10] Yan D C.Automatic identification algorithms for mesoscale eddies based on satellite altimeter[D]. Beijing:National Marine Environmental Forecasting Center, 2015. [燕丹晨. 基于卫星高度计的中尺度涡自动识别算法研究[D]. 北京: 国家海洋环境预报中心, 2015.]
  • [11] Wunsch C. Low-frequency variability of the sea[M]//Warren B A, Wunsch C. Evolution of Physical Oceanography:Scientific Surveys in Honor of Henry Stommel. Cambridge, MA, USA:MIT Press, 1981:342-374.
  • [12] Li L. A review on mesoscale oceanographical phenomena in the South China Sea[J]. Journal of Oceanography in Taiwan Strait, 2002, 21(2): 265–274. [李立. 南海中尺度海洋现象研究概述[J]. 台湾海峡, 2002, 21(2): 265–274. ] [DOI:10.3969/j.issn.1000-8160.2002.02.021]
  • [13] Belkin I M, Cornillon P C, Sherman K. Fronts in large marine ecosystems[J]. Progress in Oceanography, 2009, 81(1-4): 223–236. [DOI:10.1016/j.pocean.2009.04.015]
  • [14] Qu L Q, Guan L, He M X. The global availabilities ofSeaWiFS, MODIS and merged chlorophyll-a data[J]. Periodical of Ocean University of China, 2006, 36(2): 321–326. [曲利芹, 管磊, 贺明霞. SeaWiFS和MODIS叶绿素浓度数据及其融合数据的全球可利用率[J]. 中国海洋大学学报, 2006, 36(2): 321–326. ] [DOI:10.3969/j.issn.1672-5174.2006.02.031]
  • [15] Belkin I M, O'Reilly J E. An algorithm for oceanic front detection in chlorophyll and SST satelliteimagery[J]. Journal of Marine Systems, 2009, 78(3): 319–326. [DOI:10.1016/j.jmarsys.2008.11.018]
  • [16] Wan X F, Pan A J, Guo X G. Seasonal and interannual variability of the chlorophyll concentration in the surface ocean of the Taiwan Strait[J]. Journal of Oceanography in Taiwan Strait, 2007, 26(1): 36–45. [万小芳, 潘爱军, 郭小钢. 台湾海峡表层叶绿素含量的季节、年际变化特征研究[J]. 台湾海峡, 2007, 26(1): 36–45. ] [DOI:10.3969/j.issn.1000-8160.2007.01.005]
  • [17] Jia Y J. Remote sensing and numerical simulation of the Kuroshio front instability process in east china sea[D]. Beijing:The Institute of Oceanology, Chinese Academy of Sciences, 2010. [贾永君. 东海黑潮锋面不稳定过程遥感与数值模拟研究[D]. 北京: 中国科学院研究生院, 2010.]
  • [18] Fan X M, Wu Y M, Cui X S, et al. Relation betweenchlorophy Ⅱ and SST fronts of the North Pacific and the distribution of squid fishing grounds[J]. Fishery Information & Strategy, 2016, 31(1): 44–53. [范秀梅, 伍玉梅, 崔雪森, 等. 北太平洋叶绿素和海表温度锋面与鱿鱼渔场的关系研究[J]. 渔业信息与战略, 2016, 31(1): 44–53. ] [DOI:10.13233/j.cnki.fishis.2016.01.008]
  • [19] Nieto K, Demarcq H, McClatchie S. Mesoscale frontal structures in the Canary Upwelling System:new front and filament detection algorithms applied to spatial and temporal patterns[J]. Remote Sensing of Environment, 2012, 123: 339–346. [DOI:10.1016/j.rse.2012.03.028]
  • [20] Shaw A G P, Vennell R. Measurements of an oceanic front using a front-following algorithm for AVHRR SST imagery[J]. Remote Sensing of Environment, 2001, 75(1): 47–62. [DOI:10.1016/S0034-4257(00)00155-3]
  • [21] Liu C Y, Wang F. Distributions and intra-seasonal evolutions of the sea surface thermal fronts in the Yellow Sea warm current origin area[J]. Marine Sciences, 2009, 33(7): 87–93. [刘传玉, 王凡. 黄海暖流源区海表面温度锋面的结构及季节内演变[J]. 海洋科学, 2009, 33(7): 87–93. ]
  • [22] Ta N. Numerical simulation of the change in the yellow sea warmcurrent[D]. Qingdao:The First Institute of Oceanography, SOA, 2013. [塔娜. 冬季黄海暖流变异的数值研究[D]. 青岛: 国家海洋局第一海洋研究所, 2013.]
  • [23] Cayula J F, Cornillon P. Edge detection algorithm for SST images[J]. Journal of Atmospheric and Oceanic Technology, 1992, 9(1): 67–80. [DOI:10.1175/1520-0426(1992)009<0067:EDAFSI>2.0.CO;2]
  • [24] Belkin I M, Cornillon P. SST fronts of the Pacific coastal and marginal seas[J]. Pacific Oceanography, 2003, 1(2): 90–113.
  • [25] Breaker L C, Mavor T P, Broenkow W W. Mapping and monitoring large-scale ocean fronts off the California coast using imagery from the GOES-10 geostationary satellite[J]. California Sea Grant College Program, 2005.
  • [26] Kostianoy A G, Ginzburg A I, Frankignoulle M, et al. Fronts in the Southern Indian Ocean as inferred from satellite sea surface temperature data[J]. Journal of Marine Systems, 2004, 45(1-2): 55–73. [DOI:10.1016/j.jmarsys.2003.09.004]
  • [27] Liu C Y. Distribution and variations of the thermal fronts in the eastern Chinaseas[D]. Beijing:The Institute of Oceanology, Chinese Academy of Sciences, 2009. [刘传玉. 中国东部近海温度锋面的分布特征和变化规律[D]. 北京: 中国科学院研究生院, 2009.]
  • [28] Ping B, Su F Z, Meng Y S, et al. A model of sea surface temperature front detection based on a threshold interval[J]. Acta Oceanologica Sinica, 2014, 33(7): 65–71. [DOI:10.1007/s13131-014-0502-x]
  • [29] Xue C J, Su F Z, Zhou J Q. Extraction of ocean fronts based on wavelet analysis[J]. Marine Science Bulletin, 2007, 26(2): 20–27. [薛存金, 苏奋振, 周军其. 基于小波分析的海洋锋形态特征提取[J]. 海洋通报, 2007, 26(2): 20–27. ] [DOI:10.3969/j.issn.1001-6392.2007.02.003]
  • [30] Chen J, Chen B, Xu S Q. Detection ofocean features in SAR images with 2D continuous wavelet power spectrum[J]. Acta Electronica Sinica, 2010, 38(9): 2128–2133. [陈捷, 陈标, 许素芹. 基于二维连续小波变换的SAR图像海洋现象特征检测[J]. 电子学报, 2010, 38(9): 2128–2133. ]
  • [31] Ping B, Su F Z, Du Y Y, et al. Application of the model of universal gravity to oceanic front detection near the kuroshio front[J]. Journal of Geo-Information Science, 2013, 15(2): 187–192. [平博, 苏奋振, 杜云艳, 等. 基于引力模型的海洋锋信息提取[J]. 地球信息科学学报, 2013, 15(2): 187–192. ] [DOI:10.3724/SP.J.1047.2013.00187]
  • [32] Ping B, Su F Z, Du YY. Bohai sea front detection using BJ-1 small satellite data[J]. Journal of Remote Sensing, 2014, 18(3): 686–695. [平博, 苏奋振, 杜云艳, 等. 北京一号数据检测渤海海洋锋[J]. 遥感学报, 2014, 18(3): 686–695. ] [DOI:10.11834/jrs.20143221]
  • [33] Cui X S, Zhou W F, Wang D, et al. Fronts extraction and vectorization from raster image of sea surface temperature[J]. Journal of Image and Graphics, 2012, 17(11): 1384–1390. [崔雪森, 周为峰, 王栋, 等. 海面温度栅格图的锋面提取与矢量化[J]. 中国图象图形学报, 2012, 17(11): 1384–1390. ] [DOI:10.11834/jig.20121107]
  • [34] Canny J. A computational approach to edgedetection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1986, 8(6): 679–698. [DOI:10.1109/TPAMI.1986.4767851]
  • [35] Zhang W, Cao Y, Luo Y. An ocean front detection method based on the Canny operator and mathematicalmorphology[J]. Marine Science Bulletin, 2014, 33(2): 199–203. [张伟, 曹洋, 罗玉. 一种基于Canny和数学形态学的海洋锋检测方法[J]. 海洋通报, 2014, 33(2): 199–203. ] [DOI:10.11840/j.issn.1001-6392.2014.02.013]
  • [36] Sun G Y, Liu Q H, Liu Q, et al. A novel approach for edge detection based on the theory of universal gravity[J]. Pattern Recognition, 2007, 40(10): 2766–2775. [DOI:10.1016/j.patcog.2007.01.006]
  • [37] Wang N, Li X. An improved edge detection algorithm based on the Canny operator[J]. Journal of Shenzhen University:Science and Engineering, 2005, 22(2): 149–153. [王娜, 李霞. 一种新的改进Canny边缘检测算法[J]. 深圳大学学报:理工版, 2005, 22(2): 149–153. ] [DOI:10.3969/j.issn.1000-2618.2005.02.010]
  • [38] Wang Z, He S X. An adaptive edge-detection method based on Cannyalgorithm[J]. Journal of Image and Graphics, 2004, 9(8): 957–962. [王植, 贺赛先. 一种基于Canny理论的自适应边缘检测方法[J]. 中国图象图形学报, 2004, 9(8): 957–962. ] [DOI:10.11834/jig.200408183]
  • [39] Zhang T Y, Suen C Y. A fast parallel algorithm for thinning digital patterns[J]. Communications of the ACM, 1984, 27(3): 236–239. [DOI:10.1145/357994.358023]
  • [40] Cui X S, Wu Y M, Dai Y, et al. Binary image skeleton extraction by external pressure method (EPM)[J]. Computer Engineering and Applications, 2013, 49(13): 138–141, 227. [崔雪森, 伍玉梅, 戴阳, 等. 外部压力法(EPM)的二值图像骨架提取[J]. 计算机工程与应用, 2013, 49(13): 138–141, 227. ] [DOI:10.3778/j.issn.1002-8331.1110-0668]
  • [41] Wang X. Research on automatic identification method andspatio-temporal process case construction of ocean mesoscale eddies[D]. Jinan:Shandong University of Science and Technology, 2012. [王欣. 海洋中尺度涡自动识别与时空过程案例构建方法研究[D]. 济南: 山东科技大学, 2012.]
  • [42] Du YY, Wang L J, Fan X, et al. GIS-based analysis on temporal-spatial features of typical processes of meso-scale eddies in the South China Sea[J]. Marine Sciences, 2014, 38(1): 1–9. [杜云艳, 王丽敬, 樊星, 等. 基于GIS的南海中尺度涡旋典型过程的特征分析[J]. 海洋科学, 2014, 38(1): 1–9. ] [DOI:10.11759/hykx20140101]
  • [43] Chelton D B, Schlax M G, Samelson R M. Global observations of nonlinear mesoscale eddies[J]. Progress in Oceanography, 2011, 91(2): 167–216. [DOI:10.1016/j.pocean.2011.01.002]
  • [44] Bi J W, Dong Q, Xue C J, et al. Extraction algorithm applied to northern Pacific mesoscale eddies based on altimetric remotely sensed data[J]. Journal of Remote Sensing, 2015, 19(6): 935–946. [毕经武, 董庆, 薛存金, 等. 基于高度计遥感数据的北太平洋中尺度涡提取[J]. 遥感学报, 2015, 19(6): 935–946. ] [DOI:10.11834/jrs.20154282]
  • [45] Yan D C, Ji T Y, Li Y, et al. Preliminary analysis of the anti-cyclonic eddy off southeast Vietnam by WA method during summer and autumn in 2013[J]. Marine Forecast, 2015, 32(5): 53–60. [燕丹晨, 仉天宇, 李云, 等. 基于WA方法的2013年夏秋越南东南外海暖涡初步分析[J]. 海洋预报, 2015, 32(5): 53–60. ] [DOI:10.11737/j.issn.1003-0239.2015.05.007]
  • [46] Sun W Z. Years of temporal and spatial variation of vortexin the sea area northeast of Taiwan base on the sea satellite altimeter data[D]. Shanghai:Shanghai Normal University, 2012. [孙闻政. 基于海面卫星测高数据的台湾东北海域涡旋多年时空变化[D]. 上海: 上海师范大学, 2012.]
  • [47] Bi J W. Kriging algorithms applied to the identification of mesoscale eddies in the Pacific Ocean on multivariate remotely sensed data[D]. Jinan:Shandong Normal University, 2015. [毕经武. 基于多源遥感数据的太平洋中尺度涡克里金提取方法研究[D]. 济南: 山东师范大学, 2015.]
  • [48] Isern-Fontanet J, Font J, García-Ladona E, et al. Spatial structure of anticyclonic eddies in the Algerian basin (Mediterranean Sea) analyzed using the Okubo-Weiss parameter[J]. Deep Sea Research Part Ⅱ:Topical Studies in Oceanography, 2004, 51(25-26): 3009–3028. [DOI:10.1016/j.dsr2.2004.09.013]
  • [49] Okubo A. Horizontal dispersion of floatable particles in the vicinity of velocity singularities such asconvergences[J]. Deep Sea Research and Oceanographic Abstracts, 1970, 17(3): 445–454. [DOI:10.1016/0011-7471(70)90059-8]
  • [50] Weiss J. The dynamics ofenstrophy transfer in two-dimensional hydrodynamics[J]. Physica D:Nonlinear Phenomena, 1991, 48(2-3): 273–294. [DOI:10.1016/0167-2789(91)90088-Q]
  • [51] Yang G. A study on the mesoscale eddies in the northwestern PacificOcean[D]. Beijing:The Institute of Oceanology, Chinese Academy of Sciences, 2013. [杨光. 西北太平洋中尺度涡旋研究[D]. 北京: 中国科学院研究生院, 2013.]
  • [52] Cui F J, Kuang X D, Wang Y. The analysis on interannual variation characteristics of eddy activities and its dynamic mechanism in the south China sea[J]. Oceanologia et Limnologia Sinica, 2015, 46(3): 508–516. [崔凤娟, 匡晓迪, 王玉. 南海中尺度涡年际变化特征及动力机制分析[J]. 海洋与湖沼, 2015, 46(3): 508–516. ] [DOI:10.11693/hyhz20140900242]
  • [53] Sadarjoen I A, Post F H. Detection, quantification, and tracking of vortices using streamline geometry[J]. Computers & Graphics, 2000, 24(3): 333–341. [DOI:10.1016/S0097-8493(00)00029-7]
  • [54] Chaigneau A, Gizolme A, Grados C. Mesoscale eddies off Peru in altimeter records:identification algorithms and eddy spatio-temporal patterns[J]. Progress in Oceanography, 2008, 79(2-4): 106–119. [DOI:10.1016/j.pocean.2008.10.013]
  • [55] Wang X, Du YY, Zhou C H, et al. An improved, SSH-based method to automatically identify mesoscale eddies in the ocean[J]. Journal of Tropical Oceanography, 2013, 32(2): 15–23. [DOI:10.11978/j.issn.1009-5470.2013.02.002]
  • [56] Yi J, Du Y, He Z, et al. Enhancing the accuracy of automatic eddy detection and the capability of recognizing the multi-core structures from maps of sea levelanomaly[J]. Ocean Science, 2014, 10(1): 39–48. [DOI:10.5194/os-10-39-2014]