基于GPU的SIFT特征点快速匹配

影像匹配是计算机视觉及数字摄影测量的核心问题,基于特征的SIFT匹配算法由于其较强的匹配能力和良好的健壮性成为研究的热点.但SIFT算子的多量性及提取特征点维度较高的特点,直接影响了匹配速度.因此,本文利用GPU强大的并行处理能力,就SIFT特征点匹配在GPU上的设计与实现进行了详细的介绍,并通过对比实验说明利用GPU对SIFT特征点进行匹配所带来的优势.     

遥感影像几何精校正中控制点选取的研究

在采用多项式模型进行遥感影像的几何精校正中,控制点的选取对校正精度有重要影响。本文通过实例计算,分析比较控制点的数量、分布对校正结果的影响,分析结果表明:适当增加控制点的数量会提高校正精度,但控制点数量并非越多越好;选择均匀分布的控制点有利于提高几何校正的精度。     

基于RPC模型的ZY-3正射纠正研究

采用基于RPC模型进行ZY-3影像正射纠正处理时,当使用全局布点控制方案时控制点增加到12个以上后,对整景影像的纠正精度影响不明显,且由于选取多个高精度的控制点非常困难,而有时控制点的增加并不一定能带来精度的提高。本文通过采用几种不同的控制点布设方案对同一景ZY-3影像进行正射纠正,并对几种方案的纠正成果进行了精度比较和分析,从而进一步证实了选择合适的控制点布设方案和控制点数量大大提高ZY-3影像的纠正质量和纠正效率。     

利用改进的SUSAN算法提取航空影像中孤立的特征点

本文针对地形起伏较大、无明显建筑物的航空影像,分析了SUSAN算法角点检测理论,提出一种提取孤立特征点的方法.该方法先对图像进行梯度幅值运算,然后对梯度幅值进行Otsu法阈值分割,设计模板并对孤立特征点进行套合,最后利用SUSAN算法计算原始影像的角点初始响应,经过非极大值抑制提取孤立特征点.经实骑证明,与传统的Har...     

浅谈延清高速公路1∶2000数字高程模型制作

随着计算机技术的不断发展,数字高程模型应用越来越广泛,在测绘领域发挥着越来越重要的作用,具有广阔的应用前景与潜力。本文阐述了数字高程模型的生产方法、数字高程模型生产注意事项及DEM原始数据采集误差来源。     

国家级强对流潜势预报业务进展与检验评估

为了推动我国现阶段天气预报业务向专业化和精细化方向发展,国家气象中心自2009年组建了强天气预报中心并开展了国家级强对流落区潜势预报业务。开发了基于加密自动气象站WS报、全国闪电定位监测网、FY系列卫星以及雷达组网等多种实况观测资料的强对流实时监测产品,制定了基于MICAPS 3.0业务平台的《中尺度天气分析规范》,研发了基于全球模式T639以及区域中尺度模式GRAPES-RUC、WRF-EPS等模式输出量的强对流动力热力参数的诊断分析产品及潜势预报方法,建立了国家级强对流天气实时预报业务并发布雷暴、雷雨大风和冰雹、短时强降水等分类落区预报指导产品。对2010年4—9月国家级预报产品进行的客观检验结果表明：6小时间隔雷暴TS评分为18%,短时强降雨为2.6%,冰雹和雷雨大风为2.1%;12小时间隔雷暴TS评分为18.4%,短时强降雨为4.1%,冰雹和雷雨大风为1.3%。     

CINRAD/SB雷达接收机技术特点及故障诊断方法

介绍了CINRAD/SB雷达接收系统技术特点。根据接收机的主通道、测试通道、故障定位通道的信号流程,从监测信息、报警信息、关键点参数测量入手,总结了CINRAD/SB雷达接收系统故障诊断方法和技巧。详细介绍了利用RDA计算机的报警信息直接查找故障、通道测试法和关键点波形测试法3种诊断故障的方法,并列举了接收机主通道前端、后端以及测试通道3个不同类型的典型故障个例的分析和处理步骤。提出了接收系统维修方面的一些建议,为新一代天气雷达技术支持和保障提供借鉴。     

海滩上的新一代天气雷达系统防雷要点浅析

通过湛江新一代天气雷达系统的防雷设计与实际防护效果,总结地处海滩的雷达站防雷设计的要点．提出相关防雷设计思路。     

开钻新工艺介绍及其优越性分析

介绍自升式钻井平台勘探二号在渤海某井作业中采用的一、二开合并钻进的开钻新工艺,并对新工艺与传统工艺对比分析,找出新工艺存在的优越性,以及使用新工艺的必要条件和技术要点。     

Sequence stratigraphy based on high-resolution seismic profiles in the late Pleistocene and Holocene deposits of the Venice area

The sequence-stratigraphic investigation by Very High-Resolution (VHR) seismic profiles allowed recognition of the detailed architecture of the late Pleistocene and Holocene succession of the Venice area. In this way deposits previously known by the analyses of scattered cores, mainly taken along the lagoon margin and the littoral strips, have been correlated at regional scale including the near offshore sector and the result has pointed out the lateral variability of the stratal architecture. Late Pleistocene deposits consist of an aggrading floodplain and fluvial channel fills accumulated during decreasing eustatic sea level, and they are coeval with offlapping forced regressive marine wedges in the Central Adriatic basin. The Holocene sequence is composed of three main seismic units separated by major stratal surfaces. Unit 1 (up to 9 m thick) is formed by channelized deposits separated by areas showing sub-horizontal and hummocky reflectors, and is bounded at the base by a surface that records prolonged conditions of subaerial exposure and at the top by a flatter surface resulting from erosion by marine processes. Deposits of Unit 1 are interpreted as estuarine and distributary channel fills, and back-barrier strata. Unit 2 is well distinguishable from Unit 1 only in the offshore area and at the barrier island bounding the Venice Lagoon, and is composed of a prograding marine wedge (up to 10 m thick) that interacts laterally with ebb tidal deltas. Unit 3 consists of a tidal channel complex and inlet deposits, which testify the evolution of the lagoon area. Tidal channels are entrenched in the lagoon mud flat (coeval with Units 1–2) and cut the Pleistocene–Holocene boundary in several places.Following current sequence-stratigraphic concepts, the Holocene sequence is composed of a paralic transgressive systems tract (TST) (Unit 1) overlying a sequence boundary (the Pleistocene–Holocene boundary) and overlain by a marine highstand systems tract (HST) (Unit 2) in seaward locations and by highstand lagoonal deposits landwards. TST and HST are separated by a downlap surface that is amalgamated with a wave ravinement surface in several places. Unit 3 is coeval with the upper part of Unit 2, and its development has been favoured by human interventions, which led to a transgression limited to the lagoon area.Local factors during the deposition, i.e. subsidence, sediment supply, physiography, and current/wave regimes, led to a significant lateral variability in the architecture of the Holocene sequence, as evidenced by the extreme thickness variation of the TST along both depositional strike and dip. The HST, instead, shows less pronounced strike variations in the stratal architecture. Also, present data clearly evidence that the human impact has a great relevance in influencing the late Holocene sedimentation.