高分辨率数字航测相机现状和思考

本文根据国内外数字航测相机发展的实际情况，对于新型数字航测相机视场角、影像分辨率、像移补偿、摄影间隔和影像存贮器容量、相机系统的动态定位与定向以及真彩色与多光谱摄影等技术问题进行了分析并给出了相应的结论。     

应用数字近景摄影测量提取岩体裂隙迹长信息方法研究

针对传统以皮尺为主要工具的岩体裂隙迹长测量方法精度及效率较低、尤其是遇有危险场地不能遥测等缺点,研究应用数字近景摄影测量的方法获取岩体裂隙迹长信息,以普通非量测数码相机为主要工具,以JX-4A数字摄影测量工作站为依托进行信息处理,可以快速获取隧道施工掌子面裂隙迹线图。经实验比较分析,数字近景摄影测量方法获取岩体裂隙信息在精度和效率方面都优于传统方法。     

基于数理统计分析的数字摄影资料空三布点方案选择

采用DMC(DigitalMappingCamera)数字航摄相机的数字摄影资料取代传统的模拟航片进行空三加密,基于数理统计分析原理,提出用绘制误差折(曲)线图和误差频率曲线图来综合评价DMC空三加密成果,并将该评价方法用于比较DMC不同布点方案的优劣。试验证明,该可视化评价方法表达直观,降低了平差结果分析的复杂度,有助于空三加密成果的定性、定量评价。     

大面阵CCD影像多通道不一致性消除算法

大面阵CCD多通道影像转移过程中,由于通道硬件的偏置和增益的不一致,将造成各通道影像的亮度和反差的微小差异,以致重建后的大面阵CCD影像出现条带效应。提出基于随机序列分析的模拟运动影像拼接算法,并将这种方法与基于影像直方图均衡化和影像直方图规定化的影像拼接算法进行比较。实验证明,基于随机序列分析的模拟运动影像拼接算法不仅能保持拼接后的影像整体一致性,而且能较好的消除拼接缝。该算法已经应用于我国第一台大面阵彩色CCD数字航测相机的影像预处理。     

采用热像仪测量玉米冠层半球方向亮温的四种方法比较

为了提高玉米冠层半球方向亮温（DBT）测量的准确性，对一套由工业用起重机平台和热红外成像仪组成的地面观测系统的四种测量方法进行了对比和评价。这些方法基于不同的测量原理：（1）通过移动起重机和平台，从不同的角度去测量一个样区的DBT；（2）固定起重机和平台的位置，通过旋转热像仪，从不同的区域收集不同角度的DBT；（3）对方法2的改进，转动热像仪朝向的同时，吊车沿作物垄行方向运动，得到大区域内平均的DBT；（4）当热像仪沿着垂直于垄行方向的起重机臂运动时，用高成像频率对目标热辐射进行采样。在对这些方法进行比较时，着重考虑了玉米冠层亮温分布的时间和空间变化所产生的影响，对本观测系统的优点和局限性进行了初步探讨。最后，本文从实际应用的角度，提出了一个基于这个系统且优化了的DBT测量方法。     

Geometric test field calibration of digital photogrammetric sensors

Test field system calibration will be a fundamental part of the future photogrammetric production line. Accurate calibration and performance evaluations are necessary for fully assessing the stability and accuracy of digital sensing techniques. In this paper, a method of comprehensive geometric calibration in a test field has been developed and empirically tested using eight image blocks collected with three UltraCamD digital large format photogrammetric cameras. Permanent photogrammetric test fields form the basis of the method. Important components of the method are determination of system parameters, evaluation of systematic errors, and assessment of geometric accuracy. The results showed that UltraCamD images contained systematic deformations that could not be modeled with single lens additional parameter models. Good point determination accuracy was obtained despite the systematic errors; the typical accuracy was 2–3 μm in image space in the horizontal coordinates and 0.05–0.09‰ of the object distance in height. One of the cameras had significantly poorer performance. In the worst cases, the horizontal accuracy was 5 μm in image space and the height accuracy was 0.18‰ of the object distance. The analog cameras gave better results than the UltraCamD, but the development of appropriate mathematical models for UltraCamD as well as improvements in digital sensors may change the situation in the near future.     

Practical Influences of Geometric and Radiometric Image Quality Provided By Different Digital Camera Systems

Imaging systems founded on current digital camera technology are finding widespread use in high precision measurement applications. A single digital CCD camera, or an array of such cameras, equipped with ring lighting equipment is commonly used to acquire imagery of high contrast retroreflective targets placed on the object at discrete locations to signalize points of interest. The precise and accurate measurement of each imaged target location is a fundamental requirement if suitable measurement tolerances are to be obtained. Whilst such systems are undoubtedly capable of producing excellent results, the practical effects of target image quality on the photogrammetric measurement process is in need of careful consideration. This paper revisits some fundamentals of the optical imaging of retrotargets and investigates some abilities of a range of digital camera systems to provide images of retrotargets that are appropriate to the measurement process. Some experimental results are presented including the imaging of planar arrays of differing sized retrotargets at differing angles and exposures and a series of network analyses in which the level of target image intensity has been varied systematically.     

CCD阵列相机的几何标定

提出了一种ＣＣＤ阵列相机几何标定的多级处理方法。根据几何成像原理，该方法把相机内外部参数分开求解；而在内部参数标定过程中，又把各内部参数的标定分阶段进行。这种标定方法是建立在严格的几何约束关系之上的，它克服了传统的整体参数标定方法中各未知参数的相关性影响，从而保证了标定结果的数据一致性和精度。实验验证了本文方法的优越性。     

利用地图与单影像进行建筑物三维重建的新途径

首先利用直接线性变换(DLT）、单影像空间后方交会确定影像方位元素的初值．然后利用影像上全部建筑物的直线（特别是建筑物的铅垂线)来确定影像方位元素的精确值．以达到对多栋建筑物同时建模的目的。