4179小行星雷达模型的修正

发布时间：2018-07-09 21:58

本文选题：嫦娥二号 + 4179小行星　；参考：《哈尔滨工业大学》2015年硕士论文

【摘要】：在嫦娥二号飞越4179图塔蒂斯小行星之前,人们对于它的了解和观测主要是基于地基雷达的观测。2012年12月13日嫦娥二号飞越图塔蒂斯小行星,这是人类第一次近距离通过光学设备直接观测近地小行星,并获取了清晰的图塔蒂斯小行星的清晰的可见光图像。在1992年和1996年4179小行星近距离飞越地球时,天文学家们积累了大量地基雷达观测数据,从而推导出小行星的详细形状模型,本文中将此模型称作图塔蒂斯小行星的雷达模型。可见光图像上图塔蒂斯小行星的表面细节信息十分丰富,这些表面信息可以用来增强图塔蒂斯小行星的雷达模型的表面结构。在此应用背景下,本文提出了一种使用嫦娥二号得到的图塔蒂斯小行星的可见光图像来修正其三维雷达模型的方法。文中首先仿真生成了不同光照角度与旋转相位的雷达图像,接着使用不变矩对生成的图像和小行星可见光图像进行归一化,让小行星雷达图像与可见光图像具有相同的尺度与形心。然后通过链码、边界距离和不变矩的方法,对小行星雷达图像和可见光图像进行了姿态定位,得到了两组图像的最佳拟合姿态。在此基础上利用泰森多边形建立了雷达模型上顶点和可见光图像的像素之间的对应关系。其次,因为可见光图像是二维图像,而雷达模型是由三维空间上的点组成的。所以需要计算可见光图像的第三维深度信息,本文选取了从明暗恢复形状(Shape From Shading)方法来计算可见光图像的深度信息。最后,本文采用了数据融合的方法将可见光图像深度信息代表表面细节的部分加载到了雷达模型上。频率域中高频信息代表表面结构的细节信息,而低频信息代表表面结构的轮廓信息。可见光图像上表面结构的细节信息十分丰富,本文利用这种思想在频率域中将雷达模型的深度的频率域信息和通过高通滤波器的可见光图像深度信息,最终得出了有更丰富的表面信息的雷达模型。
[Abstract]:Before Chang 'e-2 flew over the 4179 Tutatisse asteroid, people's understanding and observation of it was mainly based on ground-based radar observations. On December 13, 2012, Chang 'e-2 flew over the Tutatis asteroid, This is the first time that the near Earth asteroid has been observed directly by optical equipment and a clear visible light image of the Tutatisse asteroid has been obtained. In 1992 and 1996, when 4179 asteroid flew over the earth at close range, astronomers accumulated a large amount of ground-based radar observation data, thus deducing the detailed shape model of asteroid. This model is called the radar model of Tutatisse asteroid in this paper. The surface details of the Tatisse asteroid above the visible image are abundant and can be used to enhance the surface structure of the radar model of the Tuttis asteroid. In this context, this paper presents a method to correct the 3D radar model of the Tutatisse asteroid by using the visible light image of Chang 'e-2. In this paper, the radar images with different illumination angles and rotation phases are first simulated, and then the generated images and asteroid visible light images are normalized using invariant moments. Let the asteroid radar image and the visible image have the same scale and centroid. Then, the method of chain code, boundary distance and invariant moment is used to locate the attitude of asteroid radar image and visible light image, and the best fitting attitude of two groups of images is obtained. Based on this, the correspondence between vertices and pixels of visible image in radar model is established by using Tyson polygon. Secondly, because the visible image is a two-dimensional image, the radar model is composed of three-dimensional points. Therefore, it is necessary to calculate the third depth information of visible light image. In this paper, the shape from shading method is selected to calculate the depth information of visible light image. Finally, the method of data fusion is used to load the depth information of visible image to the radar model. The high frequency information in the frequency domain represents the detail information of the surface structure, while the low frequency information represents the contour information of the surface structure. The details of the surface structure on the visible light image are very rich. In this paper, the depth information of the radar model in frequency domain and the depth information of the visible light image through the high-pass filter are analyzed in the frequency domain. Finally, the radar model with more surface information is obtained.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P185.7

【参考文献】