Three-dimensional Reconstruction Of Space Objects Based On Inverse Synthetic Aperture Radar

Inverse Synthetic Aperture Radar is widely used in our lives and has made some achievements, for example identify some objects such as aircraft, ships, missiles and other targets in recent years. Unlike early radar imaging, which could detect the object as point, two-dimensional imaging broke the radar early point limitation. However, this imaging form is essentially the projection of a three-dimensional structure of the target on a two-dimensional plane, which could not obtain comprehensive information of it. With the rapid development of science and technology and research work, people request the accuracy of radar imaging more higher. In order to extract more comprehensive information from the radar target echo, three-dimensional imaging has become a hot research trend. Three-dimensional reconstruction technique is one of the main methods to achieve the three-dimensional ISAR image. The subject has significance meaning for three-dimensional imaging theory based three-dimensional reconstruction technologies.The thesis studies based on single-antenna radar system, which completes the three-dimensional reconstruction of observation targets using two-dimensional ISAR images under different times. The method for three-dimensional reconstruction is called factorization method. The main completed work of it includes, two-dimensional ISAR imaging for observation objects, feature point extraction and matching, cross scaling of near-Earth orbit model, three-dimensional reconstruction of the target and so on. We have designed a series models for ISAR two-dimensional imaging turntable imaging such as rule graphics and airplane model. In order to obtain the two-dimensional coordinate information at different times, we need to extract feature points of the obtained two-dimensional images. Taking into account the changing targets, we must make sure that the feature point of the different moments of ISAR two-dimensional can be matched. The feature point matching algorithm must be seriously considered. This paper introduces two feature point extraction algorithms called the spatial extreme and Harris corner detection. After obtaining the feature points the scale-invariant transformation method that based on the multi-scale feature vector space is used for feature points matching. The algorithm need compute the eigenvectors of all these feature points of the different frames and complete the matching between different images based on the above descriptors. After the completion of the above work, we use factorization method to obtain a three-dimensional reconstruction based on the information of different times coordinates.Previous studies mainly used the grid method and the optical flow field matching tracing method to achieve the matching feature points of the ISAR two-dimensional images. This paper complete corner extraction and matching track based on the mentioned method above. Then use the factorization method for three-dimensional reconstruction. Finally, we verify the effectiveness of the feature point extraction, matching method and the three-dimensional reconstruction method using the reconstruction of the designed model and the measured data. Moreover, we also proved that we can obtain more accurate reconstruction effects based on more accurate matching points.