【摘要】：潮溝系統(tǒng)作為潮灘上最活躍的地貌單元,是潮灘與水下岸坡區(qū)物質(zhì)與能量交換的通道,在潮流、波浪、風(fēng)暴潮、人類活動(dòng)等因素的綜合作用下,潮溝系統(tǒng)敏感、多變,對(duì)鄰近海堤、港口等沿海工程構(gòu)成極大威脅。江蘇中部沿海是我國(guó)潮溝系統(tǒng)最為發(fā)育的區(qū)域,在區(qū)域沉積動(dòng)力環(huán)境快速變化、人類活動(dòng)影響加劇等背景下,亟需掌握潮溝系統(tǒng)地貌演化過(guò)程與規(guī)律,為沿海經(jīng)濟(jì)社會(huì)平穩(wěn)、持續(xù)發(fā)展提供決策支持。針對(duì)潮溝系統(tǒng)因海岸帶復(fù)雜地理環(huán)境與自組織網(wǎng)絡(luò)結(jié)構(gòu)而導(dǎo)致的提取難、表征難、分析難等關(guān)鍵問(wèn)題,本文完成了如下研究：(1)結(jié)合目標(biāo)潮溝系統(tǒng)在高程屬性中作為連續(xù)負(fù)地形的幾何屬性,研究提出了針對(duì)機(jī)載LiDAR DEM數(shù)據(jù)的顧及剖面形態(tài)特征的全自動(dòng)潮溝提取算法(APMECF)。算法流程包含：基于多窗口鄰域分析的局部潮灘高程趨勢(shì)面均衡、基于多方向多尺度高斯匹配濾波的潮溝增強(qiáng)、基于兩輪自適應(yīng)閾值分割潮溝提取、多輪潮溝提取結(jié)果融合策略。該算法攻克了在海岸帶高程低對(duì)比度環(huán)境下提取目標(biāo)潮溝系統(tǒng)的難題,通過(guò)兩套不同空間分辨率LiDARDEM數(shù)據(jù)的測(cè)試,展示了APMECF算法在處理潮溝系統(tǒng)細(xì)小尺度、保持潮溝系統(tǒng)拓?fù)渚W(wǎng)絡(luò)完整性等方面具有較好的優(yōu)勢(shì)。通過(guò)針對(duì)類潮溝系統(tǒng)的拓展延伸性實(shí)驗(yàn)(如陸地溝壑提取實(shí)驗(yàn)、干旱區(qū)河流提取實(shí)驗(yàn)等),驗(yàn)證了APMECF算法在信息載體背景相對(duì)單一環(huán)境下對(duì)于類潮溝系統(tǒng)目標(biāo)地物提取應(yīng)用的可移植性。(2)國(guó)產(chǎn)資源三號(hào)高分辨率遙感影像數(shù)據(jù)固有的高頻訪問(wèn)周期、光譜信息豐富、觀測(cè)成本低廉等特點(diǎn),使其相較于機(jī)載LiDAR數(shù)據(jù)而言,在海岸帶長(zhǎng)期遙感動(dòng)態(tài)監(jiān)測(cè)應(yīng)用中更具優(yōu)勢(shì)。但是,多光譜影像中由光譜維度的間接信息載體記錄的海岸帶復(fù)雜地理相帶對(duì)潮溝系統(tǒng)提取的背景干擾是遙感動(dòng)態(tài)監(jiān)測(cè)應(yīng)用的重大挑戰(zhàn)。為了解決上述難題,研究從目標(biāo)潮溝系統(tǒng)的幾何特征信息入手,提出了顧及局部微分幾何結(jié)構(gòu)的多尺度潮溝系統(tǒng)提取算法(MLDSI)。算法流程包含：復(fù)雜異質(zhì)潮灘背景均一化、多尺度潮溝系統(tǒng)增強(qiáng)、譜問(wèn)二元抑制規(guī)則、融合區(qū)域特征空間粘滯性的水平集演化模型。該算法能夠在復(fù)雜異質(zhì)性環(huán)境中針對(duì)形態(tài)多變的潮溝系統(tǒng)構(gòu)建具備一致性度量的描述子LFI,從而實(shí)現(xiàn)潮溝系統(tǒng)的有效提取應(yīng)用。通過(guò)針對(duì)近岸潮灘(90km2)與離岸沙洲(520 km2)等多個(gè)研究區(qū)的綜合實(shí)驗(yàn),驗(yàn)證了MLDSI算法與傳統(tǒng)經(jīng)典分類／分割提取算法在處理復(fù)雜異質(zhì)背景下線狀地物提取的優(yōu)越性。(3)研究在完成上述潮溝系統(tǒng)提取的基礎(chǔ)上構(gòu)建了能夠定量統(tǒng)計(jì)與分析潮溝系統(tǒng)靜態(tài)幾何形態(tài)參數(shù)和動(dòng)態(tài)演化過(guò)程水動(dòng)力特征參數(shù)的方法,即定量表征潮溝系統(tǒng)各向異性及待渠化密度兩種維度的特征信息。總體而言,潮溝系統(tǒng)各向異性具有由海洋至內(nèi)陸不斷遞增的趨勢(shì)。通過(guò)比對(duì)相鄰潮盆區(qū)域,以及同一潮盆內(nèi)部的子潮溝區(qū)域,其待渠化平均徑流長(zhǎng)度具有高度動(dòng)態(tài)變異的特點(diǎn),同時(shí)表現(xiàn)出明顯的無(wú)標(biāo)度分布特征。各個(gè)潮盆區(qū)域具有唯一標(biāo)識(shí)的水動(dòng)力特征參數(shù)待渠化徑流長(zhǎng)度,其能夠用來(lái)作為潮溝系統(tǒng)網(wǎng)絡(luò)分叉率和蜿蜒度的定量測(cè)度,對(duì)潮溝系統(tǒng)靜態(tài)特征表述提供了新的分析維度與技術(shù)支撐。
[Abstract]:The tidal channel system, as the most active physiognomy unit on the tidal flat, is a channel of material and energy exchange between the tidal flat and the underwater shore slope area. Under the comprehensive action of the factors such as tidal current, wave, storm surge and human activity, the tidal channel system is sensitive and changeable, Coastal engineering, such as the port, poses a great threat. The coastal area of the central part of Jiangsu is the most developing area of the tidal channel system in our country. In the background of the rapid change of regional sedimentary dynamic environment and the intensification of the influence of human activity, it is urgent to master the evolution and law of the geomorphic evolution and the law of the tidal channel system, and provide the decision support for the stable and sustainable development of the coastal economy. In order to solve the key problems such as difficulty, difficulty and difficulty in the extraction from the complex geographical environment of the coastal zone and the self-organized network structure, this paper has completed the following research: (1) as the geometric property of the continuous negative topography in the elevation property of the target tidal channel system, A full-automatic channel extraction algorithm (APMECF), which takes into account the profile characteristics of airborne LiDAR DEM data, is presented in this paper. The algorithm flow comprises the following steps of: based on a multi-window neighborhood analysis local tide beach elevation trend surface equalization, a multi-directional multi-scale Gaussian matched filtering tidal channel enhancement, and a multi-wheel tidal channel extraction result fusion strategy based on two-wheel adaptive threshold segmentation. The method solves the problem that the target tidal channel system is extracted under the low contrast environment of the elevation of the coastal zone, and the small scale of the APMECF algorithm is shown in the processing of the tidal channel system by the two sets of test of the LiDARDEM data with different spatial resolutions, And has good advantages in maintaining the integrity of the topological network of the tidal channel system and the like. In this paper, the portability of the APMECF algorithm for the extraction and application of the target features of the class-tidal channel system under the relatively single environment is verified by extending the extension experiment (such as the land groove extraction experiment, the river extraction experiment in the arid area, etc.) for the class-tidal channel system. (2) The high-frequency access period, the rich spectrum information and the low observation cost of the high-resolution remote sensing image data of the domestic resource 3 make the high-resolution remote sensing image data more advantageous in the long-term remote sensing application of the coastal zone compared with the on-board LiDAR data. However, the background interference of the complex geographical phase of the coastal zone, which is recorded by the indirect information carrier of the spectral dimension in the multi-spectral image, is a major challenge for the remote sensing dynamic monitoring application. In order to solve the above-mentioned problems, a multi-scale tidal channel system extraction algorithm (MLDSI), which takes into account the local differential geometry, is proposed based on the geometric feature information of the target tidal channel system. The algorithm process includes the following steps: the complex heterogeneous tidal flat background is uniform, the multi-scale tidal channel system is enhanced, the spectrum is asked by the binary suppression rule, and the horizontal set evolution model of the characteristic space viscosity of the fusion area is obtained. The algorithm can construct a description sub-LFI with a consistent measurement in a complex heterogeneous environment for the variable-form tidal channel system, thereby realizing the effective extraction and application of the tidal channel system. In this paper, the superiority of the MLDSI algorithm and the traditional classical classification/ segmentation extraction algorithm in the processing of the linear feature extraction under the complex heterogeneous background is verified by the comprehensive experiment of the near-shore tidal flat (90 km2) and the offshore sandbank (520km2). and (3) the method for quantitatively counting and analyzing the static geometric shape parameters and the dynamic evolution process water power characteristic parameters of the tidal channel system is constructed on the basis of completing the extraction of the tidal channel system, Namely, the characteristic information of the two dimensions of the anisotropy of the tidal channel system and the density to be channelized is quantitatively characterized. In general, that anisotropy of the tidal channel system has an increasing trend from the sea to the inland. The average runoff length to be channelized has the characteristics of high dynamic variation through comparison to the adjacent tidal basin area and the sub-tidal channel area inside the same tide basin, and meanwhile, the obvious scale-free distribution characteristic is displayed. Each tidal basin area has a uniquely identified water power characteristic parameter to be channelized runoff length, which can be used as a quantitative measure of the network bifurcation rate and the winding degree of the tidal channel system, and provides a new analysis dimension and technical support for the static characteristic expression of the tidal channel system.
【學(xué)位授予單位】：南京大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：P715.7;P737

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 孫超;劉永學(xué);李滿春;趙賽帥;周e，

本文編號(hào)：2501776