Laplacian sparse coding exhibits good performance for image classification, because of its outstanding ability to preserve the locality and similarity information of the codes. However, there still exists two drawback...
详细信息
Automatic image registration method is important because data amount acquired and processed is increasingly heavy. For SAR and optical sensor images, it is important to align the same object shown in different images ...
详细信息
ISBN:
(纸本)9781510818934
Automatic image registration method is important because data amount acquired and processed is increasingly heavy. For SAR and optical sensor images, it is important to align the same object shown in different images from multi sensors. Considering extracting and matching features in SAR and optical images, a template registration strategy is proposed in this paper. A particular similarity measurement which is robust in matching remote sensing images of general sceneries is hardly to construct. Being the invertible and essential tendency nowadays in the field of earth observation, high pixel-resolution images can offer much more information of the ground. Thus subpixel image registration methods have been researched, among which can successfully achieve 1/50 pixel level accuracy in matching homologous sensor image. This paper firstly investigated the performance of 2D Gaussian fitting for low correlated synthetic images and real satellite images acquired by Radat SAT-2, COSMO and TerraSAR-X satellites. Three algorithms for subpixel locating have been utilized with consideration of the low correlated coefficient peak and corrupted shape of neighbourhood around it. The results show 2D Gaussian fitting is the most efficient on noisy coefficient maps. Then a subpixel image registration method for SAR and optical images is proposed. Normalized gradient correlation (NGC) algorithm is used to solve the problem of different intensity property at first, and by combining 2D Gaussian fitting to relocate matched points at a subpixel level accuracy. Experiment results have shown that the proposed method has enough capabilities in handling differences between images from multi-sensors. The proposed approach has been found to be simple and efficient in implementation with a low computation load. And the most important performance is the results presented here have achieved matching accuracy within 1 pixel.
Asteroid landing is a new research field in deep space exploration, which has important scientific and economic values in learning the origin of the solar system and exploring the asteroid. It is microgravity and litt...
详细信息
ISBN:
(纸本)9781467391054
Asteroid landing is a new research field in deep space exploration, which has important scientific and economic values in learning the origin of the solar system and exploring the asteroid. It is microgravity and little media properties are obtained on the asteroid surface, thus good landing performance is demanded to realize safe landing of the landing mechanism. Landing mechanism in the paper has three legs, and the damping device is located between the landing gear and the equipment base. Simulation method is introduced to analyze the landing performance of the landing mechanism. Landing performance of the landing mechanism on 30° landing slope with the largest landing velocity is simulated in three typical landing modes, and the landing performance is reflected by stability time of the landing gear and overloading acceleration of the equipment base. Landing simulation has important significance in learning landing performance and guiding the landing of the landing mechanism.
Spatial markers can be used to not only complete coordinate transformation between two or more different Cartesian coordinate systems, but also to provide the error evaluation of the above transformation. On the one h...
详细信息
Spatial markers can be used to not only complete coordinate transformation between two or more different Cartesian coordinate systems, but also to provide the error evaluation of the above transformation. On the one hand, a group of transforming markers is responsible for accurate parameters i.e. rotation matrix and translation vector;on the other hand, a group of checking markers can be used to quantify the accuracy of the results between coordinate transformation by Euclidean distance between markers. In this article, an invalidation of the error evaluation for coordinate transformation by relative Euclidean distance has been concluded and proven that the indexes are always constant because of an invariant group of checking markers, independent of the other group checking markers.
暂无评论