ID	原文	译文
50577	避免了局部运动对全局运动估计的影响，	avoid the local sports effect on the global motion estimation,
50578	然后在定制的L1范数约束下构建相机的最优化运动路径，	and then under the L1 norm constraint of custom build camera motion path of optimization,
50579	最后利用新的路径进行运动矫正获取稳定的视频图像序列，实验结果验证了算法的有效性。	finally using a new path for motion correction for the stability of the video image sequence, the experimental results verify the effectiveness of the algorithm.
50580	针对基于传统第一原理理论的诊断方法中，由于忽略了自主诊断时历史诊断结果对当前诊断的影响而导致无谓消耗大量计算量的问题，提出了新的诊断方法。	For diagnosis method based on the theory of the traditional principle of the first, due to ignoring the independent diagnosis history to the influence of diagnosis to the current diagnosis of meaningless consume a large amount of computation problem, puts forward a new diagnostic method.
50581	通过在传统的方法中引入状态记忆机制，扩展了诊断与冲突的含义。	The article introduces the status in the traditional way memory mechanism, extends the meaning of diagnosis and conflict.
50582	根据诊断解的结构分析了基于第一原理诊断方法中诊断解的分布，	According to the principle of diagnosis of solution structure is analyzed based on the first diagnosis method in the diagnosis of solution of distribution,
50583	并利用分析结果证明了新的诊断方法在效率上优于传统的诊断方法。	and use the analysis results show that the new diagnostic method is superior to the traditional diagnosis methods in efficiency.
50584	通过引入"时间戳"和"状态锁"，使得诊断方法能够自主地处理遥测数据在不同输入方式下的诊断问题。	By introducing "timestamp" and "state of the lock", the diagnostic method can independently deal with remote sensing data under different input methods of diagnosing the problem.
50585	为避免由于网络负载抖动而造成的频繁网络选择，本文为无线异构网络提出了一种预测网络未来负载的自适应负载均衡算法。	In order to avoid because of the network load jitter caused by the frequent network selection, this paper for heterogeneous wireless network presents a prediction network load adaptive load balancing algorithm in the future.
50586	通过马尔可夫链预测负载状态空间的概率，将预测到的概率通过负载趋势函数映射为趋势值，利用趋势值进行网络选择和自适应触发门限的调整。	By markov chain to predict the probability of load state space, will predict the trend of probability through the load function mapping for trend value, using the trend values for network selection and adjustment of adaptive trigger threshold.