基于时刻独立脉冲耦合神经网络的高空间分辨率摇感影像分割

高空间分辨率遥感影像中地物目标内部光谱信息复杂性的增强,使得传统基于光谱特征值的数据处理方法效果不再显著,影像分割为解决这一问题提供了一种思路,成为当前高空间分辨率遥感影像处理的研究焦点.时刻独立脉冲耦合神经网络具有状态相近、空间相邻神经元相互耦合同步脉冲激发和区域之间神经元脉冲激发时刻独立两大特点,已被应用于非遥感影像分割中,并取得较好效果.本文结合高空间分辨率遥感影像特点,通过对网络参数进行实验和分析,提出一个基于时刻独立脉冲耦合神经网络的高空间分辨率遥感影像分割方法,并利用空间分辨率0.3m的航空影像进行了数据试验,将分割结果进行讨论并与现有时刻独立脉冲耦合神经网络方法和ISODATA方法分割结果进行对比分析.结果表明:时刻独立脉冲耦合神经网络在高空间分辨率遥感影像分割处理中具有很好的应用前景.     

遥感影像特征提取与选择及在影像分类中的应用

利用模式识别中特征提取和特征选择的相关理论对遥感影像的纹理特征进行遴选和变换处理,得到描述纹理的二次特征。实验证明这些新特征能够提高影像分类精度和分类效率。     

基于LBV变换的遥感影像多步骤分类法研究

面对与日俱增的遥感数据,如何快速准确地进行影像分类成为遥感领域亟待解决的问题。本文在前人的基础上提出了一种基于LBV变换的遥感影像多步骤分类方法。L:地物的总辐射水平,集中反映了裸地的信息;B:地物的可见-红外光辐射平衡,是地面水分状况和水体存在的一个良好指标;V:地物辐射随波段变化的方向和速度向量,能集中反映地面植被状况。多步骤分类法遵循由易到难的原则,能克服类别之间的互相影响,从而提高分类精度。文章结合LBV变换和多步骤分类法的优点对影像进行了分类,结果表明该方法简单易行,且能达到良好的分类精度。     

遥感图像线性影纹理解专家系统设计与实现

线性影纹信息是遥感图像中的一类重要信息,线性影纹信息自动提取是遥感图像智能解译的重要研究领域。笔者采用基于特征对象的专家系统技术来完成线性影纹信息的简单类别提取,设计并实现了线性影纹理解专家系统;阐明了系统结构,解译规则获取、表达,基于消息的不确定性推理机及解释机的设计;最后利用系统做了线性影纹类别提取实验,结果表明基于特征对象的专家系统用于线性影纹类别自动提取是切实可行的,并具有较高的提取精度。     

水平井含水率上升影响因素

利用塔里木油田塔中4油田(TZ4)底水油藏相关的地质、流体数据建立数值模型。在所建模型的基础上,应用数值模拟计算的累积产油、产水和产液量回归俞启泰水驱特征曲线,以反映水平井见水特征的参数b。以参数b为研究对象,采用正交试验的方法研究不同因素对b值的影响,筛选影响水平井见水特征的主要因素,认为原油粘度、油层厚度、非均质性及水平井水平段在油藏中的位置是影响水平井含水上升趋势的主要指标。最后建立主要因素与b值的关系式,结合俞启泰曲线b值图版,提出预测水平井见水规律的公式——图版法(F-b法)。应用F-b法对塔里木水平井含水率进行预测,并与其他相关方法及实际生产数据对比,认为F-b法可作为预测塔里木油田水平井含水率、估算可采储量的一种有效方法。     

Study on rain-runoff process in the peripheral mountainous area of the Sichuan Basin

Studies on rain-runoff process in the peripheral mountainous area of the Sichuan Basin, which is regarded as a key ecological shelter, will contribute to flood control and environmental protection for the Upper Yangtze River Basin. In two typical catchments--the Fujiang River Catchment and the Wujiang River Catchment, rainfall simulations have been conducted to study the rain-runoff processes of yellow soil and limestone soil in three types of land use--forestland, farmland and grassland. Results showed that （1） within the same rainfall process, overland flow occurs first on farmland, then on grassland, and finally on forestland; （2） soil surface coverage has a great impact on the occurrence and amount of overland flow. The runoff amount can increase 2-4 times after the coverage is removed; （3） the infiltration before the occurrence of overland flow will decrease because of higher gravel contents of soil, but it takes no effect on infiltration once overland flow becomes stable; （4） the runoff coefficient of the limestone soil forestland is greater than that of the yellow soil forest land, but less than that of the farmland; （5） three empirical infiltration models, including Horton＇ model, Kostiakov＇ model, and modified Kostiakov＇ model, were compared by using the observed results under rainfall simulation. The results showed that the Kostiakov＇ model performed better than both the Horton＇ model and modified Kostiakov model. According to the results of this research, the Kostiakov＇s model can be used to simulate rainfall infiltration when water erosion is modeled in the peripheral mountainous area of the Sichuan Basin.     

Approaches for comparative evaluation of raster GIS-based landslide susceptibility zonation maps

Evaluation of maps generated from different conceptual models or data processing approaches at spatial level has importance in many geoenvironmental applications. This paper addresses the spatial comparison of different landslide susceptibility zonation (LSZ) raster maps of the same area derived from various procedures.     

Simulation experiment on hydrocarbon generating potential of various source rocks on the northern margin of the Qaidam Basin,Qinghai, China

The northern margin of the Qaidam Basin is one of the main oil-gas exploration areas in western China, where source rocks are composed mostly of Middle and Lower Jurassic dark mud shale, carbargillite and coal. A large number of subsurface and outcrop samples differing in lithology with different types of organic matter were selected for resource evaluation, research and calculation. And among them, 13 samples were used for simulation experiment on hydrocarbon generating potential of various source rocks. At first, two kinds of heating modes were compared through simulation experiment, including single temperature-step heating and continual heating. Perhaps, the process of hydrocarbon generation and expulsion occurred naturally between a close system and an open system. In addition, the first heating mode was convenient, and all its reaction products were involved in the whole thermal evolution, and the final simulation experimental results were basically in consistency with the natural evolution trend Therefore, the first heating mode was adopted and the hydrocarbon yield of every sample was worked out. According to the type and lithology of organic matter and the hydrocarbon yield of samples for simulation experiment, hydrocarbon generation and expulsion mode with three kinds of lithology and five types of source rock has been established to provide the basis for hydrocarbon generation evaluation, research and resource calculation.     

Methodology and use of tensor invariants for satellite gravity gradiometry

Although its use is widespread in several other scientific disciplines, the theory of tensor invariants is only marginally adopted in gravity field modeling. We aim to close this gap by developing and applying the invariants approach for geopotential recovery. Gravitational tensor invariants are deduced from products of second-order derivatives of the gravitational potential. The benefit of the method presented arises from its independence of the gradiometer instrument’s orientation in space. Thus, we refrain from the classical methods for satellite gravity gradiometry analysis, i.e., in terms of individual gravity gradients, in favor of the alternative invariants approach. The invariants approach requires a tailored processing strategy. Firstly, the non-linear functionals with regard to the potential series expansion in spherical harmonics necessitates the linearization and iterative solution of the resulting least-squares problem. From the computational point of view, efficient linearization by means of perturbation theory has been adopted. It only requires the computation of reference gravity gradients. Secondly, the deduced pseudo-observations are composed of all the gravitational tensor elements, all of which require a comparable level of accuracy. Additionally, implementation of the invariants method for large data sets is a challenging task. We show the fundamentals of tensor invariants theory adapted to satellite gradiometry. With regard to the GOCE (Gravity field and steady-state Ocean Circulation Explorer) satellite gradiometry mission, we demonstrate that the iterative parameter estimation process converges within only two iterations. Additionally, for the GOCE configuration, we show the invariants approach to be insensitive to the synthesis of unobserved gravity gradients.     

A framework of region-based spatial relations for non-overlapping features and its application in object based image analysis

Object based image analysis (OBIA) is an approach increasingly used in classifying high spatial resolution remote sensing images. Object based image classifiers first segment an image into objects (or image segments), and then classify these objects based on their attributes and spatial relations. Numerous algorithms exist for the first step of the OBIA process, i.e. image segmentation. However, less research has been conducted on the object classification part of OBIA, in particular the spatial relations between objects that are commonly used to construct rules for classifying image objects and refining classification results. In this paper, we establish a context where objects are areal (not points or lines) and non-overlapping (we call this “single-valued” space), and propose a framework of binary spatial relations between segmented objects to aid in object classification. In this framework, scale-dependent “line-like objects” and “point-like objects” are identified from areal objects based on their shapes. Generally, disjoint and meet are the only two possible topological relations between two non-overlapping areal objects. However, a number of quasi- topological relations can be defined when the shapes of the objects involved are considered. Some of these relations are fuzzy and thus quantitatively defined. In addition, we define the concepts of line-like objects (e.g. roads) and point-like objects (e.g. wells), and develop the relations between two line-like objects or two point-like objects. For completeness, cardinal direction relations and distance relations are also introduced in the proposed context. Finally, we implement the framework to extract roads and moving vehicles from an aerial photo. The promising results suggest that our methods can be a valuable tool in defining rules for object based image analysis.