基于遥感TM影像纹理特征的长江河口流态信息分析

发布时间：2018-09-05 10:33

【摘要】：长江口是我国最大的潮汐河口,受径流、潮流、波浪、科氏力、复杂地形等环境因子的综合影响,动力条件多变,泥沙输运复杂。卫星遥感技术具有大尺度观测、快速同步更新等优势,可为宏观定期监测该地区水动力和泥沙输运提供有效途径,前人已经做了大量有效工作,已取得众多成果。但有关基于光学遥感影像的流态信息研究多停留在定性层面,尚未深入到定量和半定量的研究。本文收集了1995～2013年间多时相遥感TM影像数据和多年典型潮位站的潮位资料,通过遥感影像的空间纹理特征分析,实现长江河口表层涨落潮流流向和流速两种主要流态信息从定性描述到半定量化的初步研究。同时,利用课题组建立的长江河口和杭州湾北岸MIKE21_FM数值模型模拟获得的流场,验证由遥感TM影像纹理特征分析得到的流态信息解译结果,发现流向的均方差误差在落潮和涨潮情况下,分别为6.2。和18.5。,流速分级的吻合度在落潮和涨潮情况下分别为84.6%和61.5%,落潮情况下的流态解译结果较好。因此,基于遥感TM影像线状纹理特征的河口流态信息定量分析有一定的参考价值,并有一定的创新性。主要结论如下： 1、Gabor滤波方法可增强遥感TM影像纹理信息,Gabor滤波后图像线状纹理信息量越大的区域越接近主流方向,而纹理信息量越少的区域越偏离实际水流方向,由此推断,遥感TM影像线状纹理方向可解译为水流方向。与传统的纹理特征统计方法相比,Gabor滤波方法在增强流向解译标志、提高流向判释准确度方面有一定的优势,可用于河口地区表层流向信息的定量提取。 2、鉴于遥感TM影像线状纹理熵值越大的区域纹理越密集,而熵值较小的地方纹理较稀疏,其与Garbor滤波方法获得的线状纹理信息量越大的区域越接近水流主流区域较为一致。由此推断,纹理熵值可能与流速存在一定的正相关关系,可利用同一景TM影像的纹理熵分级从宏观上表征河口地区流态的强弱。 3、遥感影像定量解译流态信息的方法在悬沙浓度适中的河口区域精度较高,且Gabor滤波图像提取流向的效果较纹理熵表征流速分级的效果更好。在时序的适用上,枯季的TM影像纹理特征较洪季更为显著,更有利于流态解译,枯季大潮时的TM影像较枯季小潮更有利于流态信息的遥感解译。 4、将Gabor滤波方法应用于长江河口及其邻近水域遥感TM影像的流态信息解译,探讨1995年以来枯季大小潮期间长江河口南北槽、南汇嘴和杭州湾北岸东海大桥所在水域的流态特征和流态变化,表明深水航道治理工程、圈围工程和东海大桥等大型工程对其所在水域的流态产生一定的影响。不足与展望：纹理熵值是表征潮汐河口流速的重要参数,但还需考虑更多的流态解译因素来提高精度；限于遥感影像数据的精度和大范围同步实测数据的缺乏,不能实时捕捉长江口流态变化的影响因素,有待进一步后续研究和深入探讨。
[Abstract]:The Yangtze Estuary is the largest tidal estuary in China. It is influenced by runoff, tidal current, wave, Coriolis force and complex topography. The dynamic conditions are changeable and the sediment transport is complex. A great deal of effective work has been done and many achievements have been made. However, most of the researches on the flow information based on optical remote sensing images remain at the qualitative level, and have not yet gone deep into the quantitative and semi-quantitative research.
In this paper, TM image data of multi-temporal remote sensing and tide level data of typical tide stations for many years from 1995 to 2013 are collected. Through the analysis of spatial texture characteristics of remote sensing images, the qualitative description and Semi-quantitative Study on the flow direction and velocity of surface fluctuation tidal current in the Yangtze Estuary are carried out. The flow field obtained by MIKE21_FM numerical model in the Yangtze Estuary and the Northern Bank of Hangzhou Bay is validated. The results show that the mean square error of the flow direction is 6.2.0 and 18.5. Therefore, quantitative analysis of estuarine flow regime information based on linear texture features of remote sensing TM images has certain reference value and innovation.
1. Gabor filtering can enhance the texture information of remote sensing TM images. The larger the linear texture information is, the closer the region is to the mainstream direction, and the less the texture information is, the more the region deviates from the actual flow direction. Therefore, the linear texture direction of remote sensing TM images can be interpreted as the flow direction. Compared with the Gabor filtering method, the Gabor filtering method has some advantages in enhancing the interpretation marks of flow direction and improving the accuracy of flow direction interpretation, and can be used for quantitative extraction of surface flow information in estuaries.
2. In view of the fact that the larger the entropy value of the linear texture, the denser the texture is and the sparser the local texture is, the closer the area with the larger information of the linear texture obtained by Garbor filtering is to the mainstream area of the flow. Therefore, it is inferred that the texture entropy value may have a positive correlation with the flow velocity, which is beneficial. The texture entropy classification of the same scene TM image is used to characterize the flow pattern in the estuary area macroscopically.
3. The method of quantitative interpretation of flow regime information from remote sensing images is more accurate in estuaries with moderate suspended sediment concentration, and the effect of extracting flow direction from Gabor filter image is better than that of texture entropy to characterize flow velocity classification. TM image is more favorable for remote sensing interpretation of flow information than dry season neap tide.
4. The Gabor filtering method is applied to interpret the TM images of the Yangtze Estuary and its adjacent waters, and the flow characteristics and flow regime changes in the waters of the North-South Channel of the Yangtze Estuary, Nanhui Zui and Donghai Bridge on the north coast of Hangzhou Bay during the dry season tide period since 1995 are discussed. The results show that the deepwater channel harnessing project, encirclement project and Donghai Bridge are located in the Yangtze Estua And other large projects have a certain effect on the flow pattern of the water area.
Deficiencies and prospects: texture entropy is an important parameter to characterize tidal estuary flow velocity, but more factors of flow regime interpretation need to be considered to improve the accuracy; limited to the accuracy of remote sensing image data and the lack of large-scale synchronous measured data, it can not capture the real-time influencing factors of flow regime changes in the Yangtze Estuary, which needs further study and in-depth study. Discuss.
【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TV148;P343.5

【参考文献】