空间碎片激光测量图像跟踪技术研究

发布时间：2018-05-20 16:25

本文选题：SLR + 图像能量累积　；参考：《中国地震局地震研究所》2013年硕士论文

【摘要】：空间碎片又称太空垃圾，因其对太空环境造成极大的危害，而受到各国重视，目前对空间碎片采取的主要措施就是观测。SLR(Satellite Laser Ranging)技术是现代空间大地测量最先进技术之一。在SLR观测中，带激光反射棱镜的观测对象称为合作目标，像空间碎片等不带有反射镜的观测对象称为非合作目标。本论文要研究空间碎片SLR图像跟踪技术，由于SLR观测技术是被动跟踪，故图像跟踪技术的核心是目标识别，研究的主要内容是对SLR图像增强器上采集下的非合作目标图像进行目标识别。本研究主要分为5个部分：图像采集，图像去噪，图像增强，目标识别和目标分析。第一部分：图像采集是基于MATLAB中ImageAcquisition Toolbox工具实现的，由于MATLAB自带图像采集模块，只需设置采集参数。第二部分：图像去噪部分是研究的重点，研究中针对卫星运行多样性，选取目标易识别和目标不易识别两种情况下的观测图像作为研究对象，并分别采用形态学去噪中的开运算、闭运算，空间域去噪法中的中值滤波、均值滤波、自适应性滤波，频域去噪法中的离散余弦去噪对样本进行单帧图像的去噪实验，效果均不明显，后采用有限序列图像累加的方式进行图像去噪，效果明显。第三部分：图像增强采用laplace算子、sobel算子、prewitt算子和log算子进行实验，均达不到实验目的，后采用负相变换结合线性运算，实现目标周围灰度差增大。第四部分：目标识别采用阈值分割法及边缘检测法中的基于sobel算子和robinson算子的边缘分割对图像进行目标识别，由于噪声干扰过多，对于小目标图像识别有困难，经过有效去噪后，进行简单的阈值分割，实现目标识别，，后利用目标形态进行目标筛选，精确识别目标。第五部分：目标分析是标识和计算目标位置，最后可返回目标坐标。最终通过研究实现了对SLR观测图像的目标识别，为今后SLR跟踪观测技术的发展提供参考。
[Abstract]:Space debris is also known as space garbage. Because of its great harm to the space environment, it has been paid much attention by all countries. The main measure to take space debris is the observation of.SLR (Satellite Laser Ranging) technology is one of the most advanced technologies in modern space geodetic. In the SLR observation, the object of observation with a laser reflective prism is called cooperation. The object, such as space debris and non reflecting mirrors, is called non cooperative target. In this paper, we should study space debris SLR image tracking technology. Because SLR observation technology is passive tracking, the core of image tracking technology is target recognition. The main content of the research is the non cooperative target image acquisition on the SLR image intensifier. Line target recognition.
This study is mainly divided into 5 parts: image acquisition, image denoising, image enhancement, target recognition and target analysis.
The first part: image acquisition is based on ImageAcquisition Toolbox tool in MATLAB. Since MATLAB has its own image acquisition module, it only needs to set acquisition parameters.
The second part: the image denoising part is the focus of the study. In the study, the diversity of the satellite operation is selected, and the observation images under two conditions are selected and the target is easy to recognize and the target is not easily recognized as the research object, and the open operation in the morphological denoising, the closed operation, the mean filtering, the mean filtering and the self-adaptive in the space domain denoising method are adopted respectively. Filtering, the discrete cosine de-noising in the frequency domain denoising method is used to denoise the single frame of the sample, and the effect is not obvious. Then the image denoising is carried out with the finite sequence image accumulation.
The third part: the image enhancement uses the Laplace operator, the Sobel operator, the Prewitt operator and the log operator to carry out the experiment, which can not reach the experimental purpose, and then the negative phase transformation is combined with the linear operation, and the gray scale difference around the target is increased.
The fourth part: the target recognition uses the threshold segmentation method and the edge detection method based on the Sobel operator and the Robinson operator to recognize the image. Because of the excessive noise interference, it is difficult for the small target image recognition. After effective denoising, a simple threshold segmentation is carried out to realize the target recognition, and then the target shape is used. The state carries out the target screening and accurately identifies the target.
The fifth part: target analysis identifies and calculates the location of the target, and finally returns the target coordinates.
Finally, the target recognition of SLR observation images is realized, which provides reference for the development of SLR tracking technology in the future.
【学位授予单位】：中国地震局地震研究所
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P225.2;X738

【参考文献】