In the reliability evaluation of many high-reliability and high-value products， product reliability often cannot be accurately evaluated due to the lack of objective test data. Aiming at this problem， a reliability evaluation method for high-reliability products based on improved evidence fusion was proposed in order to make full use of reliability information from different sources. Firstly， combining the characteristics of reliability engineering， the modified weight of each evidence was determined by the consistency of the evidence at credal level， pignistic level and the uncertainty of the evidence itself. Secondly， the optimal comprehensive weight was obtained by linear combination of each weight vector based on game theory. Finally， the Dempster’s combination rule was used to fuse the modified evidence， and the probability distribution of the product reliability index was obtained through the Pignistic probability transformation formula to complete the product reliability evaluation. The reliability evaluation results of one electronic device show that compared with the results of Jiang’s combination method and Yang’s combination method， which also consider multi-dimensional weight modification， the credibility of the conflict interval given by the proposed method is reduced by 69.6% and 54.6% respectively， and the credibility of the overall frame of discrimination given by the proposed method is reduced by 5.6% and 3.7% respectively. Therefore， in the application of reliability engineering， the performance of the proposed method in solving evidence conflict and reducing the uncertainty of fusion results is better than that of the comparison methods， and this method can fuse multi-source reliability information effectively and improve the credibility of the results of product reliability evaluation.

Due to the increasingly high performance requirements on the Advanced Encryption Standard (AES) algorithm which was widely used, software-based cryptographic algorithms have been increasingly difficult to meet the demands of high-throughput ciper cracking. As a result, more and more encryption algorithms have been accelerated by using Field-Programmable Gate Array (FPGA) platform. Focused on the issue that the development of AES algorithm based on FPGA has high complexity and long development cycle, with High-Level Synthesis (HLS) design methodologies, AES hardware acceleration algorithm was designed by using high-level programming language. Firstly, loop unrolling, etc were used to improve operation parallelism. Secondly, to make full use of on-chip memory and circuit resources, the resource balance optimization technology was used. Finally, the full pipeline structure was added to improve the clock frequency and throughput of the overall design. The detailed analysis and comparison of the benchmark design and different optimized designs with structural expansion, resource balance and pipeline were decribed. The experimental results show that the clock frequency of AES algorithm is up to 127.06 MHz and the throughput eventually achieves 16.26 Gb/s on Xilinx xc7z020clg484 platform, compared with the benchmark AES design, performance increases by three orders of magnitude.

Through the analysis on the characteristics of the moving objects in video sequence, a method to acquire background was proposed. Moving point accumulation was applied to extract the background, and then background subtraction was employed to detect moving objects. Because of visual comparability, the result included shadow. Shadow filter function was used to remove the shadow to get the completely moving objects. Experimental results show that the algorithm is efficient, adaptive, and establishes a foundation for advanced image processing.

The problems and defections of the existing methods of SVM multi-class classification were analyzed.A multi-class classification based on binary tree was put forward.A modified self-organization map(SOM),KSOM(kernel-based SOM),was introduced to convert the multi-class problem into binary tress,in which the binary decisions were made by SVMs.The results show that it has higher multiclass classification accuracy than the multi-class SVM approaches with "one-versus-one" and "one-versus-the rest".