基于非刚性连接位置姿态数据的机载扫描图像的姿态校正技术研究

发布时间：2018-04-14 02:10

本文选题：机载扫描图像 + 位置姿态数据　；参考：《山东大学》2017年硕士论文

【摘要】：随着高光谱遥感技术的成熟与发展,该成像技术已经广泛应用于气象和环境监测、农作物长势和土壤侵蚀分析、矿物资源探测、军事侦察与服务等各个方面。机载遥感又称航空遥感,是指利用飞机、飞艇等航空工具作为运载手段的遥感技术。机载遥感图像的成像质量是阻碍该技术发展的重要问题,导致这一问题的因素主要有以下四个方面:(1)载体飞行速度和高度;(2)成像仪的拍摄角度;(3)载体姿态的变化;(4)环境因素,包括光线变化和环境干扰等。这些干扰因素可能会导致图像具有几何变化大、重叠区域小、相似特征多等特点。本系统采用低精度稳像平台结合高精度位姿测量的方法以提高图像质量。其中,低精度稳像平台保证采集的图像数据变形不会太大,高精度位姿测量则用于事后纠正图像数据的变形。通常,机载扫描图像的姿态校正大都以位置姿态系统与成像系统之间为刚性连接为基础。而本系统成像功能安装在小型稳像平台上,位置姿态系统安装在运载工具上,两者属于非刚性连接。该稳像平台为主动姿态调节系统,稳定范围有限,超限后会主动复位,相对运动情况更加复杂。本系统上述特点使得在图像姿态校正过程中存在更多的干扰特征,对图像姿态校正带来极大的困难。从以上方面可以看出,图像姿态校正技术的应用前景十分广阔,并具有新的挑战性和丰富的研究价值。本文针对带稳像平台的机载高光谱成像仪,研究了该成像仪的几何定标、位姿数据预处理和几何校正等问题,并实现了图像的姿态校正软件系统。主要工作如下:(1)简要介绍图像采集系统的重要部分,包括成像系统、稳像平台系统、位姿数据测量系统。(2)针对多视场高光谱成像仪的结构特点和高精度的几何定标需求,利用全色光源、反光镜、反光镜旋转平台、平行光管、计算机等其他实验仪器创建了一种新的几何定标方法。几何定标的结果表明,成像仪的像元误差达到亚像素精度,符合应用要求。(3)姿态校正是利用位姿数据对图像数据进行姿态的几何校正,正确的姿态校正需要位姿数据与图像数据同步,系统有可能存在同步的偏差,本文提出多种方法消除该偏差。经过筛选,可获得处理效果最好的高光谱图像数据,并保存参数,用于以后相关的数据处理。(4)讨论了基于配准处理后的位姿数据的几何校正算法,包括外方位元素的计算、空间坐标转换、重采样等。其中重采样采用了新的局部距离倒数算法,并使用多线程工具OpenMP加快图像处理速度。经过校正后的图像效果良好。(5)本文详述了机载扫描图像姿态校正软件系统的框架设计和详细设计,并实现了该软件系统。
[Abstract]:With the maturity and development of hyperspectral remote sensing technology, the imaging technology has been widely used in meteorological and environmental monitoring, analysis of crop growth and soil erosion, mineral resources exploration, military reconnaissance and various services. Also called the airborne remote sensing aerial remote sensing, refers to the use of aircraft, airship aviation remote sensing technology tools as a means of transport. The imaging quality of airborne remote sensing image is an important problem hindering the development of this technology, the factors leading to this problem mainly in the following four aspects: (1) the carrier flight speed and altitude; (2) imager camera angle; (3) the change of attitude; (4) environmental factors, including changes in light and environmental disturbance. These factors may cause image with geometric changes, overlapping area, characteristics of similar characteristics. The system adopts multi low precision stabilization platform with high precision method of pose measurement to improve The high quality of the image. The low accuracy of image stabilization platform to ensure the image data acquisition and the deformation is not too large, high precision position and attitude measurement is used after correcting the distortion of image data. Usually, the image scanning airborne attitude correction between mostly position system and imaging system for rigid connection based imaging function of the system. Installed in the small image stabilization platform, position and attitude system installed in the vehicle, both belong to the non rigid connection. The stabilization platform for active attitude control system, the stability of limited scope, will take the initiative to limit reduction, more complex. The relative motion characteristics of the system so that there is more interference characteristics in image attitude correction in the image, attitude correction very difficult. As can be seen from the above aspects, application of image correction technology attitude is very broad, and has a new challenge and HSBC Study on the value of the rich. The stabilization of Airborne Hyperspectral Imaging Platform, studied the geometric imaging instrument calibration, pose data preprocessing and geometric correction, and to achieve the image of the attitude correction software system. The main work is as follows: (1) an important part is the brief introduction of the image acquisition system, including imaging system, image stabilization platform system, pose data measurement system. (2) according to the structure characteristics of multi view hyperspectral imager and high accuracy of the geometric calibration, the panchromatic light source, mirror, mirror rotating platform, collimator, computer and other instruments to create a a new geometric calibration method. Geometric calibration results show that the error pixel imager sub-pixel precision, meet the application requirements. (3) the attitude correction is the geometric attitude for image data based on pose data correction, the correct attitude should be corrected To pose data and image data synchronization, the system may have the synchronization error, this paper presents a variety of methods to eliminate the deviation. After screening, high spectral image data obtained the best treatment effect, and save the parameters for later related data. (4) discussed the geometric correction algorithm of pose data processing after registration based on the calculation, including the elements of exterior orientation, space coordinate conversion, resampling. The resampling using local inverse distance algorithm, and uses multi threading tool OpenMP accelerate the image processing speed. After the image effect after correction. (5) this paper describes the framework design and detailed design of the correction system airborne scanning image attitude, and implementation of the software system.

【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP751

【参考文献】