SAR图像不同油膜特征的最佳分割方法选择

SAR图像溢油分割是SAR溢油监测中一个重要环节。文中选取4种不同形状、尺寸和对比度的SAR油膜数据,分别采用双峰阈值分割法、最大熵分割法、区域生长法、分水岭算法、图割法、水平集方法等6种方法进行溢油信息提取,探讨适合于不同油膜特征的最佳提取方法。结合现有尺度分割标准,提出一种SAR图像溢油信息评价指数——有效分割指数(Effective Segmentation Index,ESI),对不同分割方法得到的溢油提取结果进行定量评价,得出了不同特征油膜所适合的最佳分割方法。     

基于人类感知的SAR图像海上溢油检测算法

在基于SAR图像的海上溢油检测中,识别的效率与准确率是关键。纹理特征是人类专家能够较好的判别SAR图像中油膜,类油膜以及海水的一个重要依据。本文算法一方面融合灰度共生矩阵与Tamura特征,直接对SAR原始图像进行特征提取,避免了对图像进行分割、降噪等预处理,提高了识别算法的可行性与识别效率。另一方面,应用深度信念网络(DBN)的分类方法,可以很好地解决溢油检测中小样本分类的问题,并且模仿人类感知系统高效准确的表示信息、获取本质特征。本文应用人类感知的思想对油膜、类油膜以及海水这3类样本进行分类识别。通过实验确定了DBN中利于分类的关键参数值。本算法对原始SAR图像中3类样本的识别准确率达到90.36%,具有较好的实用价值。     

Tsallis熵和改进CV模型的海面溢油SAR图像分割

为了解决海洋表面溢油监测中合成孔径雷达(SAR)图像分割精度不高的难题,提出一种基于Tsallis熵多阈值分割与改进CV(ChanVese)模型相结合的海面溢油图像分割方法。首先采用基于Tsallis熵的多阈值选取算法对海面溢油图像进行粗分割;然后分别将得到的溢油区域和溢油粗略轮廓作为CV模型的局部区域和初始轮廓,以降低CV模型的场景复杂度及其对初始条件的敏感性。CV模型仅考虑了图像各区域的均值信息而没有考虑图像的局部信息,尽管能够得到渐进型边界图像,但其分割结果存在误差。本文采用了加入移动因子的改进CV模型降低分割误差,提高收敛速度。实验结果表明,提出的海面溢油SAR图像分割方法具有分割边界定位准确、运行高效和无需设置初始条件等优点。     

相位一致性的SAR图像溢油分割算法研究

针对传统的SAR图像溢油分割方法对斑点噪声敏感,而且对疑似溢油区域区分能力较弱的问题,该文提出一种基于相位一致性的SAR图像溢油分割算法,在对海面溢油图像进行灰度预处理的基础上,计算图像相位一致性特征,再根据最大类间方差法对计算得到的相位一致性特征值进行阈值分割。实验表明:与目前常用的溢油分割方法相比,该方法既可以有效抑制海面斑点噪声,又对疑似溢油区域具有一定抗干扰能力,而且兼顾了分割的效率。     

基于机载多光谱遥感数据的溢油信息提取方法

为研究应用紫外-可见光-近红外-热红外遥感数据识别海上溢油的方法,利用机载多光谱溢油监测设备获取了舟山海域多光谱遥感数据,并利用野外地物光谱仪获取了水体和油膜的反射光谱。通过提取并分析机载多光谱遥感图像上多目标的影像特征和光谱响应特征,利用决策树分类法对海上油膜相对厚度的分布进行分类和制图。研究表明,基于机载多光谱遥感数据的溢油信息提取方法总体分类精度达93.7%,能够准确区分薄油膜和厚油膜,可有效提取海洋溢油污染信息,完全能够满足海洋溢油污染遥感监测需求。     

海洋溢油光谱分析与卫星信息提取

通过对海洋中原油、重柴油、轻柴油、润滑油和煤油溢油波谱特征测试分析,绘制各油种不同厚度时光谱特征曲线图,是提取溢油信息的基础.经过多年卫星影像处理实践,筛选出各种油膜不同厚度时的最佳卫星通道组合与增强处理模式,成为理想海洋溢油卫星影像制作的关键技术.并应用户需求在影像中解译出了油膜位置、面积、溢油量、漂移方向和速度.     

遥感监测墨西哥湾溢油目标识别算法

为了及时监测溢油事故相关信息,防止污染扩张,该文利用2010年4月29日、5月17日及5月24日3个时相的中分辨率MODIS数据,以2010年发生的墨西哥湾溢油事故为例进行目标提取与识别。提取算法考虑到油膜和背景海水易混分的现象,采用了辅加纹理特征量的光谱角匹配算法进行油膜提取;同时,对比了常用的最小距离和支持向量机分类方法;并基于混淆矩阵对结果进行精度验证。结果显示,改进的光谱角匹配算法可以更准确地提取海面目标,精度高达90%以上,能够较好地监测出大面积溢油,可以用于海面溢油灾害的动态监测。     

海洋溢油风险评价与区划系统的设计与实现

海洋溢油不仅对生态环境造成巨大的损害,而且对国民经济和人民生活产生很大影响。海洋溢油风险评价与区划系统以GIS技术和数据库技术为支撑,由溢油源数据信息、敏感区区划信息、溢油风险评价、溢油风险扩散模拟和溢油风险区划组成。实现对渤海海域溢油风险源进行溢油概率以及溢油后溢油量的计算,并对针对不同季节溢油发生后的扩散进行动态模拟,最终生成溢油风险区划图。为有效降低溢油风险、预防溢油事故发生后造成海洋水体污染提供决策支持。     

基于OpenCV的近景图像增强和分割算法研究

本文通过对近景图像的对比度和亮度进行分析实现了自动选择图像增强算法的方法,分别对金字塔均值漂移分割算法、分水岭分割算法和GrabCut分割算法进行了对比研究。经实验表明,图像对比度太小、亮度太高或太低对近景图像分割都会造成不利的影响,伽马变换增强算法适合调整亮度较高的图像,直方图均衡化增强算法适合调整亮度较低的图像。分水岭分割算法和GrabCut分割算法不适合近景图像多类别的自动分割,而金字塔均值漂移分割算法是基于核密度梯度估计的无参数快速统计迭代算法,可实现类别自动分割,分割精度较高,而且利用建立金字塔结构可进行多尺度分割。本文的研究成果对自动选择最佳的图像增强算法和分割算法以及参数设置具有重要的参考价值。     

基于HJ–CCD数据的海面溢油提取方法研究

快速、准确地获取溢油污染信息,对海洋的动态监测、保护和可持续利用具有重要意义。环境与灾害监测预报小卫星星座一号(HJ–1)是我国针对生态环境污染和灾害监测发射的新型卫星平台,但HJ–1 CCD多光谱数据的光谱波段较少,仅依赖光谱信息获取海面溢油范围的精度较低。因此,以墨西哥湾溢油事件为研究对象,在分析不同地物光谱特征的基础上,采用灰度共生矩阵,选择合适的纹理结构因子,提取HJ–1 CCD图像中影响溢油识别的地物纹理特征;建立光谱特征和纹理特征相结合的决策树模型,提取海面溢油信息,并与只考虑光谱信息的传统分类方法进行精度对比。结果表明,与最大似然分类法相比,决策树方法的油膜提取用户精度和制图精度分别提高了11.85%和4.28%。     

SAR影像中海洋浮油膜特征分割的Level Set方法

介绍了一种全新的区域影像分割技术--基于迎风格式偏微分方程（PDE）的Level Set方法进行海洋浮油膜特征提取.在该方法中,海洋浮油膜特征表示为扩散界面,影像灰度的梯度决定了界面扩散的方向和强度.界面边缘在影像灰度差异动力和曲率流的共同作用下不断向外扩散,能够有效地克服尖锐突起和裂缝等,在特征边缘趋于稳定和光滑.同时,该方法对高噪声具有一定抑制作用,适用于低对比度高噪声图像,特别是SAR图象中特征提取.以不同地区ERS-2 SAR图像中海洋浮油膜提取为例,进行方法验证,同时对传统的影像分割技术进行对比.     

Determining oil slick thickness using hyperspectral remote sensing in the Bohai Sea of China

Abstract

Determining oil slick thickness plays an important role in assessing oil spill volume and its environmental impacts on the ocean. In this study, we used a Hyperion image of an oil spill accident area and seawater and fresh crude oil samples collected in the Bohai Sea of China. A well-controlled laboratory experiment was designed to simulate spectral responses to different oil slick thicknesses. Spectral resampling and normalization methods were used to reduce the differences in spectral reflectances between the experimental background seawater sample and real background seawater. Fitting the analysis with laboratory experimental data results showed a linear relationship between normalized oil slick reflectance and normalized oil slick thickness [20th band (R ²=0.92938, n=49, p<0.01), 26th band (R ²=0.93806, n=49, p<0.01), 29th band (R ²=0.93288, n=49, p<0.01)]. By using these statistical models, we successfully determined the normalized oil slick thickness with the Hyperion image. Our results indicate that hyperspectral remote sensing technology is an effective method to monitor oil spills on water. The spectral ranges of visible green and red light were the optimal bands for estimating oil slick thickness in case 2 water. The high, stabilized spectral reflectance of background seawater will be helpful in oil slick thickness inversion.     

基于多特征的SAR影像溢油暗斑提取

由于星载合成孔径雷达(Synthetic Aperture Radar,SAR)溢油影像包含大量斑点噪声,仅依靠传统模糊聚类方法不能有效提取出其溢油区域。针对SAR图像存在的斑点噪声问题,本文提出了一种结合多特征与改进模糊C均值聚类(Fuzzy C-means,FCM)的溢油暗斑提取方法。该方法首先提取影像的多个特征,以便更加充分地反映影像信息;然后同时考虑像素与其邻域的强度和空间位置关系,以此来构造模糊加权因子,进而定义目标函数;最后通过迭代最小化目标函数,获得最佳溢油暗斑提取结果。文中对真实的SAR溢油影像进行了溢油暗斑提取实验,并分别与利用单一特征和加入邻域关系的模糊聚类方法得到的提取结果进行对比分析,实验结果证明了本文方法的有效性。     

GNSS反射信号海面溢油回波DDM仿真研究

海上溢油已成为影响海洋生态环境的重要污染物之一,我国近40年发生约3200起海上溢油事故。当今用于监测海上溢油的遥感主要是光学和雷达卫星,卫星遥感往往重访周期长,而海上溢油事件时常发生,给海洋带来严重的环境污染,需要快速、准确的监测其状况。GNSS R技术具有全天候监测海洋的特点,因此更适合用于海面溢油检测。为了验证GNSS R技术在检测海面溢油的可行性,利用2013年中国青岛海洋溢油事故的遥感图像的溢油结果,作为仿真实验检测目标,进行岸基的GNSS R海面溢油检测仿真研究。利用Z V散射模型和海水／溢油的均方坡度(MSS)模型结合,建立了能反映海面状况GNSS散射信号特征的时延 多普勒图(DDM)。仿真得到DDM中检测到海面溢油区域,验证了利用GNSS反射信号进行海面溢油检测的可行性。      

高分辨率影像基于影像对象的受损道路提取

基于影像对象对高分辨率影像中的受损道路进行提取,并重点介绍了易康软件中的影像分割,形成影像对象,建立道路规则库,进而提取出道路信息,并判断受损道路信息。以龙门乡受灾区为例进行试验,利用易康软件对影像进行分割并提取受损道路信息,表明基于影像对象的思想在易康对受损道路进行提取是可行的。     

Comparing supervised and unsupervised multiresolution segmentation approaches for extracting buildings from very high resolution imagery

Although multiresolution segmentation (MRS) is a powerful technique for dealing with very high resolution imagery, some of the image objects that it generates do not match the geometries of the target objects, which reduces the classification accuracy. MRS can, however, be guided to produce results that approach the desired object geometry using either supervised or unsupervised approaches. Although some studies have suggested that a supervised approach is preferable, there has been no comparative evaluation of these two approaches. Therefore, in this study, we have compared supervised and unsupervised approaches to MRS. One supervised and two unsupervised segmentation methods were tested on three areas using QuickBird and WorldView-2 satellite imagery. The results were assessed using both segmentation evaluation methods and an accuracy assessment of the resulting building classifications. Thus, differences in the geometries of the image objects and in the potential to achieve satisfactory thematic accuracies were evaluated. The two approaches yielded remarkably similar classification results, with overall accuracies ranging from 82% to 86%. The performance of one of the unsupervised methods was unexpectedly similar to that of the supervised method; they identified almost identical scale parameters as being optimal for segmenting buildings, resulting in very similar geometries for the resulting image objects. The second unsupervised method produced very different image objects from the supervised method, but their classification accuracies were still very similar. The latter result was unexpected because, contrary to previously published findings, it suggests a high degree of independence between the segmentation results and classification accuracy. The results of this study have two important implications. The first is that object-based image analysis can be automated without sacrificing classification accuracy, and the second is that the previously accepted idea that classification is dependent on segmentation is challenged by our unexpected results, casting doubt on the value of pursuing ‘optimal segmentation’. Our results rather suggest that as long as under-segmentation remains at acceptable levels, imperfections in segmentation can be ruled out, so that a high level of classification accuracy can still be achieved.     

星载红外数据应用于大型事故溢油

使用NOAA卫星数据(AVHRR)同时对“阿提哥”号和“塔斯曼海”号溢油事件做出分析和研究。对出事海域的卫星数据进行了拉伸和彩色合成等图像增强处理,突出了影像上的油污特征;并根据其理化特性,设计算法实现了计算机对油污的自动判别;最后,估算了事故所溢油量,为事故后的调查、取证工作提供了良好依据。     

Marine Radar Oil Spill Monitoring Technology Based on Dual-Threshold and C–V Level Set Methods

Marine radar oil spill monitoring technology has the advantages of wide-range target detection, a flexible working mode, and the ability to operate in severe weather. Based on oil spill data collected from the 7.16 accident in Dalian, China, two comprehensive oil spill monitoring methods using X-band horizontal polarized marine radar are presented. The improved Prewitt operator, Otsu algorithm, median filtering, and mean filtering were used to preprocess the original marine radar image. Subsequently, the recognition of offshore oil films were studied via the dual-threshold method and C–V level set method. Results show that the image data achieved the ideal noise reduction effect after pretreatment, and retained marine radar ocean wave information effectively. The C–V level set method required relatively lower levels of data preprocessing. However, it relied heavily on the energy-driven window by expert presupposition. It was therefore only suitable for obtaining continuous oil film information. Its calculation efficiency was related to the number of evolutions and the size of the image. The dual-threshold method required higher data preprocessing. Furthermore, the effective monitoring range of ocean waves must be determined in advance. This method could identify discrete oil film information at sea and had high computational efficiency. The dual-threshold method is better for the automatic identification of marine oil film information. The two kinds of monitoring methods described herein are useful for the identification and disposal of oil spills at sea.