The challenges of severe Doppler effects in high-speed railway are considered. By building a cooperative antenna system, an algorithm of joint channel estimation and Doppler frequency offset (DFO) estimation is proposed based on Ricean channel model. First, a maximum likelihood estimation (MLE) algorithm for DFO is designed, showing that the Doppler estimation can be obtained by estimating moving velocity of the train and the path loss with the exploitation of pilots that are placed inside the frame. Then a joint detection algorithm for the receiver is proposed to exploit multi-antenna diversity gains. Last, the theoretical Crammer Rao bound (CRB) for joint channel estimation and DFO estimation is derived. The steady performance of the system is confirmed by numerical simulations. In particular, when the Ricean fading channel parameter equals 5 and the velocities of train are 100 m/s and 150 m/s, the estimation variances of DFO are very close to the theoretical results obtained by using CRB. Meanwhile, the corresponding signal to noise ratio loss is less than 1.5 dB when the bit error rate is 10-5 for 16QAM signals.
In this paper, we propose a fast registration scheme for remote-sensing images for use as a fundamental technique in large-scale online remote-sensing data processing tasks. First, we introduce priori-information images,and use machine learning techniques to identify robust remote-sensing image features from state-of-the-art ScaleInvariant Feature Transform(SIFT) features. Next, we apply a hierarchical coarse-to-fine feature matching and image registration scheme on the basis of additional priori information, including a robust feature location map and platform imaging parameters. Numerical simulation results show that the proposed scheme increases position repetitiveness by 34%, and can speed up the overall image registration procedure by a factor of 7:47 while maintaining the accuracy of the image registration performance.
Subspace appearance models are widely used in computer vision and image processing tasks to compactly represent the appearance variations of target objects. In order to ensure algorithm performance, they are typically stored in high-precision formats; this results in a large storage footprint, rendering redistribution costly and difficult. Since for most image and vision applications, pixel values are quantized to 8 bits by the acquisition apparatuses, we show that it is possible to construct a fixed-width, effectively lossless representation of the bases vectors, in the sense that reconstructions from the original bases and from the quantized bases never deviate by more than half of the quantization step-size. In addition to directly applying this result to losslessly compress individual models, we also propose an algorithm to compress appearance models by utilizing prior information on the modeled objects in the form of prior appearance subspaces. Experiments conducted on the compression of person-specific face appearance models demonstrate the effectiveness of the proposed algorithms.
