江苏沿海滩涂利用模式遥感影像解译和时空变化分析

为了弄清江苏沿海滩涂利用模式分布情况,定量分析其时空变化趋势,本文基于Google Earth Engine云计算平台和Landsat卫星遥感影像数据,结合随机森林分类方法,通过机器学习建立了滩涂利用模式自动识别分类方法,样本训练后识别准确率达93％以上.利用该方法提取了2002年、2014年和2019年江苏沿海的土地...     

基于航摄影像的滇池捞鱼河湿地植物类型提取

为了获取滇池湿地的植物类型分布信息,以捞鱼河湿地为示范区,采用分辨率为0.15 m的航摄影像为影像数据源,利用人机交互和面向对象相结合的影像解译方法,实现捞鱼河湿地植物类型分布信息的提取.在植物类型边界提取过程中,发现过高的影像分辨率(≤0.3m)不利于成片具有空间异质性植物类型边界的正常提取,容易出现同一类型过度分割情况.通过邻域降分辨率重采样技术,将影像分辨率降至0.6m时,能有效解决过度分割问题.当分辨率降至更低时,虽然能解决过度分割问题,但会导致地物边界出现不同程度的锯齿而与地物边界不符的后果.研究所应用的基于Feature Extaction模块的植物类型边界提取方法,能有效减少人机交互方法中人工勾绘边界的工作量,在保证分类精度的同时减少人力投入,对航摄影像在大范围湿地植物类型的提取中具有重要借鉴意义.     

基于遗传小波神经网络的海底声学底质识别分类

分割海底声纳探测图像,提取单元特征向量进行主成份分析,选取均值、标准差、对比度、相关系数、能量及同质性作为训练特征向量,构建小波神经网络。利用遗传算法优化小波神经网络的初始权值及小波参数,对砂、礁石、泥3种底质类型分别进行训练,并得到3种底质的测试精度都在90%以上,优于单独利用小波神经网络进行训练时的测试精度,克服了小波神经网络训练时易陷入局部极小的固有缺陷,表明基于遗传算法的小波神经网络可有效用于海底底质声纳图像的识别和分类。     

海岛礁光学和雷达影像基于改进非子采样小波变换的自动配准

Homologous feature point extraction is a key problem in the optical and synthetic aperture radar （SAR） image registration for islands. A new feature point extraction method using a threshold shrink operator and non-subsampled wavelet transform （TSO-NSWT） for optical and SAR image registration was proposed. Moreover, the matching for this proposed feature was different from the traditional feature matching strategies and was performed using a similarity measure computed from neighborhood circles in low-frequency bands. Then, a number of reliably matched couples with even distributions were obtained, which assured the accuracy of the registration. Application of the proposed algorithm to SPOT-5 （multi-spectral） and YG-1 （SAR） images showed that a large number of accurately matched couples could be identified. Additionally, both of the root mean square error （RMSE） patterns of the registration parameters computed based on the TSO-NSWT algorithm and traditional NSWT algorithm were analyzed and compared, which further demonstrated the effectiveness of the proposed algorithm. The algorithm can supply the crucial step for island image registration and island recognition.     

Granular experiments of thrust wedges: Insights relevant to methane hydrate exploration at the Nankai accretionary prism

The accumulation mechanism of methane hydrates has been a central issue in previous hydrate research regarding the Nankai accretionary prism, southwest of Japan. Expulsion of formation fluids is significant during the prism accretion process, and the migration of these methane-bearing fluids exerts a strong control on the accumulation of hydrates. Two types of fluid pathways, inter-granular porosity and faults, need to be evaluated to understand hydrate accumulation. Fluid migration along faults can be partly modeled by examining faulting activity. Our study modeled the accretion process by using two granular methods that approximated the geologic body as an assemblage of particles: (1) analog experiments using granular materials, and (2) a numerical simulation based on the distinct element method. The analog experiments closely reproduced the prism geometry observed in seismic profiles across the Nankai accretionary prism. Digital image correlation analysis indicated that the frontal thrust is generally active but older structures are also frequently reactivated. The numerical simulations produced prism geometries similar to those of the analog experiments. The velocity distributions of the particles showed evidence of episodic faulting and reactivation, but the internal stress field exhibited little change in the deeper part of the prism during deformation. The frequent and substantial changes in fault activity displayed by the models indicate episodic fluid flow along fault surfaces. Active frontal thrusting suggests that formation fluids generally migrate from deep within the prism to the deformation front, but may move along reactivated older faults. Inter-granular permeability also fluctuates, as it is controlled by temporal and spatial variations in the internal stress field. However, fluid flow is likely to be relatively stable in the deeper segment of the prism.     

数字影像分析技术在海滩环境监测的应用

运用视频监测技术对青岛石老人海滩进行环境变化的监测,将基于高斯混合模型的聚类方法应用于数字影像中进行浒苔目标分割,并计算出浒苔的面积。分析结果表明,该项技术能够对海滩浒苔等海滩环境进行实时的监测,从而避免浒苔在海滩的堆积,对浴场海滩的环境及旅游业的发展都具有重要意义。     

基于PIV技术不同形式防波堤周围涡旋场特性研究(英文)

深水防波堤周围流场存在复杂的紊动变化,防波堤较容易发生失效破坏。本文利用PIV技术,基于模型实验,对规则波作用下防波堤周围的速度流场及其涡流特性进行了实验研究,获得了规则波流场的瞬时速度分布。通过对不同型式防波堤在不同的水力条件下的波浪流场实验数据的分析,定量地研究了防波堤堤身水域的涡旋流场的水力特性。     

Determination of the crop row orientations from Formosat-2 multi-temporal and panchromatic images

This paper presents a technique developed for the retrieval of the orientation of crop rows, over anthropic lands dedicated to agriculture in order to further improve estimate of crop production and soil erosion management. Five crop types are considered: wheat, barley, rapeseed, sunflower, corn and hemp. The study is part of the multi-sensor crop-monitoring experiment, conducted in 2010 throughout the agricultural season (MCM’10) over an area located in southwestern France, near Toulouse. The proposed methodology is based on the use of satellite images acquired by Formosat-2, at high spatial resolution in panchromatic and multispectral modes (with spatial resolution of 2 and 8 m, respectively). Orientations are derived and evaluated for each image and for each plot, using directional spatial filters (45° and 135°) and mathematical morphology algorithms. “Single-date” and “multi-temporal” approaches are considered. The single-date analyses confirm the good performances of the proposed method, but emphasize the limitation of the approach for estimating the crop row orientation over the whole landscape with only one date. The multi-date analyses allow (1) determining the most suitable agricultural period for the detection of the row orientations, and (2) extending the estimation to the entire footprint of the study area. For the winter crops (wheat, barley and rapeseed), best results are obtained with images acquired just after harvest, when surfaces are covered by stubbles or during the period of deep tillage (0.27 > R² > 0.99 and 7.15° > RMSE > 43.02°). For the summer crops (sunflower, corn and hemp), results are strongly crop and date dependents (0 > R² > 0.96, 10.22° > RMSE > 80°), with a well-marked impact of flowering, irrigation equipment and/or maximum crop development. Last, the extent of the method to the whole studied zone allows mapping 90% of the crop row orientations (more than 45,000 ha) with an error inferior to 40°, associated to a confidence index ranging from 1 to 5 for each agricultural plot.     

Multi-class geospatial object detection and geographic image classification based on collection of part detectors

The rapid development of remote sensing technology has facilitated us the acquisition of remote sensing images with higher and higher spatial resolution, but how to automatically understand the image contents is still a big challenge. In this paper, we develop a practical and rotation-invariant framework for multi-class geospatial object detection and geographic image classification based on collection of part detectors (COPD). The COPD is composed of a set of representative and discriminative part detectors, where each part detector is a linear support vector machine (SVM) classifier used for the detection of objects or recurring spatial patterns within a certain range of orientation. Specifically, when performing multi-class geospatial object detection, we learn a set of seed-based part detectors where each part detector corresponds to a particular viewpoint of an object class, so the collection of them provides a solution for rotation-invariant detection of multi-class objects. When performing geographic image classification, we utilize a large number of pre-trained part detectors to discovery distinctive visual parts from images and use them as attributes to represent the images. Comprehensive evaluations on two remote sensing image databases and comparisons with some state-of-the-art approaches demonstrate the effectiveness and superiority of the developed framework.     

天宫一号高光谱数据辐射校正的CUDA并行优化

高光谱图像经过辐射校正后,消除了探测元的响应差异,能更好地满足专题信息提取的数据要求.利用探测元的列均值、列标准差等统计信息对天宫一号高光谱短波红外数据进行辐射校正检验,并基于GPU CUDA计算模型对均值归一化、矩匹配、相邻列均衡等3种相对辐射校正算法进行了并行计算优化.通过辐射校正计算流程拆分,CPU控制流程逻辑,GPU执行数据级并行计算,并建立CUDA的计算单元与数据单元的映射关系,获得5—7倍的计算加速比,这些辐射校正算法依据图像自身统计信息,且易于进行并行计算优化,满足实时校正的处理时效要求,为未来高光谱数据在轨实时辐射校正提供了新思路.