Cross-view scene matching refers to the discovery of images of the same geographical target from different platforms （such as drones and satellites）. However， different image platforms lead to low accuracy of UAV （Unmanned Aerial Vehicle） positioning and navigation tasks， and the existing methods usually focus only on a single dimension of the image and ignore the multi-dimensional features of the image. To solve the above problems， GAMF （Global Attention and Multi-granularity feature Fusion） deep neural network was proposed to improve feature representation and feature distinguishability. Firstly， the images from the UAV perspective and the satellite perspective were combined， and the three branches were extended under the unified network architecture， the spatial location， channel and local features of the images from three dimensions were extracted. Then， by establishing the SGAM （Spatial Global relationship Attention Module） and CGAM （Channel Global Attention Module）， the spatial global relationship mechanism and channel attention mechanism were introduced to capture global information， so as to better carry out attention learning. Secondly， in order to fuse local perception features， a local division strategy was introduced to better improve the model’s ability to extract fine-grained features. Finally， the features of the three dimensions were combined as the final features to train the model. The test results on the public dataset University-1652 show that the AP （Average Precision） of the GAMF model on UAV visual positioning tasks reaches 87.41%， and the Recall （R@1） in UAV visual navigation tasks reaches 90.30%， which verifies that the GAMF model can effectively aggregate the multi-dimensional features of the image and improve the accuracy of UAV positioning and navigation tasks.

The meteorological ground minute data has characteristics including various elements, large amounts of information and high frequency generation, therefore the traditional relational database system has some problems such as server overload and low read and write performance in data storage and management. With the research of storage model of distributed databases HBase, the database model of the meteorological ground minute data was proposed to achieve distributed storage of massive meteorological data and meta-information management, in which the row key was designed by the method of time plus station number. When processing the complex meteorological query case, the response time of unique index in HBase is too long. To address this defect and meet the requirements of retrieval time efficiency, with considering the query case, API interface offered by search engine solr was used to establish secondary index for related field. The experimental results show that this system has high efficiency of storage and index, the maximum storage efficiency can be up to 34000 records/s. When generic query cases return, the time consuming can be down to millisecond level. This method can satisfy the performance requirements of large-scale meteorological data in business applications.

For the numerical simulation of one-dimensional Burgers' equation based on Lattice Boltzmann method, there had been 2-bit and 4-bit models. In this paper, an equilibrium distribution function was constructed by choosing the proper kind of discrete velocity model. And then, using Lattice Bhatnagar-Gross-Krook (LBGK) model, Chapman-Enskog expansion and multiscale technique, a 3-bit Lattice Boltzmann Method (LBM) called D1Q3 model was proposed for the one-dimensional Burgers equation. Some numerical experiments were carried out and the numerical results were in good agreement with analytical solutions, therefore the effectiveness of the new method was verified.

In the cognitive radio networks, the main functions of MAC protocols include channel sensing, channel selection and access, the detection time and transmission time has the important influence on network performance. In dynamic wireless network environment, how to reasonably distribute detection time and transmission time is a challenge. A two radios multi-channel distributed MAC(TM-MAC) protocol are proposed, which is not necessary to detect channels before node transmit data. Node can detect the spectrum resource while other nodes transmit data. Then, the node can transmit data use the idle spectrum resource. A model is built to analyze the throughput of TM-MAC under the saturation condition. The analysis results show that TM-MAC improves the throughput in cognitive networks.

The resolution and frame rate of the video captured by the current camera arrays is not high enough to create the high quality 3DTV programs. We designed and implemented a high resolution camera array. It can be used to capture the video whose frame rate was 27fps and resolution was 1280×1024, or which frame rate was 93fps and resolution was 640×480. The two capture schemes can be chosen by the user applications. Some issues about the architecture, camera unit, synchronization control and real-time storage of the camera array were presented.

A dual-core DMA communication method between DSP and FPGA based on reverse wave analysis was proposed. In high definition and real-time dualcore embedded systems for catadioptric omnidirectional imaging, the dual-core DMA communication method was used to provide high speed panoramic data communication between DSP and FPGA. Reverse engineering methods were applied to perform reverse wave analysis on traditional DMA operations, and dual-core DMA communication was accomplished according to the results of wave analysis. Experiments indicate that the proposed dual-core DMA communication method provides a maximum data transmission speed between DSP and FPGA at 4.704Gbps (i.e. 588MBps).