ID	原文	译文
6504	仿真结果表明,所提方法相比Voronoi图的路径参数更加优化,相比二叉树算法更加安全,同时生成的航路规划网能够为多枚导弹协同攻击提供全局性航路参考。	Simulation results show that the proposed method is compared with the more optimal path parameters of Voronoi diagram, is more secure than binary tree algorithm, and the generated route planning network can provide global route for missile attack coordinates.
6505	针对装备危险耦合传递随机Petri网(stochastic Petri nets,SPN)模型存在指数分布限制、求解过程复杂的问题,提出广义任意分布SPN(general arbitrary distribution SPN,GADSPN)模型。	For equipment risk transfer stochastic Petri net (stochastic Petri nets, SPN limitations exist exponential distribution model, the complex problem solving process, put forward generalized arbitrary distribution SPN (general arbitrary distribution SPN, GADSPN) model.
6506	通过引入熵概率密度函数、矩母函数(moment generating function,MGF)将变迁分布推广至一般任意分布;	By introducing the entropy probability density function, moment generating function (moment generating function, MGF) will change to general arbitrary distribution;
6507	采用耦合度公式、极大熵模型分别求解GADSPN变迁使能概率及其危险度概率密度函数。	By coupling formula, the maximum entropy model respectively to solve GADSPN change can make probability and risk probability density function.
6508	基于GADSPN等效MGF性质改进传统SPN模型解析分析能力,提出了装备系统等效危险度、危险状态敏感度、危险路径恶化度等分析参数,并结合SPN模型的状态机、可达图、T-不变量,构建GADSPN分析步骤;	Based on equivalent MGF GADSPN properties improve traditional SPN model analytic ability, puts forward the equipment system equivalent risk, risk sensitivity analysis of parameters, such as, dangerous path deterioration degree, and combining the SPN model of state machine, to figure, T - not variable, build GADSPN analysis steps;
6509	最后,通过飞机大表速低空俯冲实例和基于蒙特卡罗与ExSpect平台仿真验证表明,GADSPN模型能充分反映装备危险活动耦合传递随机交互过程,具有较好的求解精度与适用性。	Finally, through the plane dived large table speed low examples and simulation based on monte carlo and ExSpect platform shows that GADSPN model can fully reflect equipment dangerous activity coupling random interaction process, has good precision and applicability.
6510	提出了基于可变形部件模型(deformable part model,DPM)的高分二号(GaoFen-2,GF2)遥感影像船只检测方法,并与区域卷积网络(regional convolutional neural network,R-CNN)进行比较。	Is proposed based on deformable component model (deformable part model, DPM) high score 2 ships (GaoFen - 2, GF2) remote sensing image detection method, and with the regional convolution network (regional convolutional neural network, R - CNN) are compared.
6511	先将遥感影像分段以获得船只的粗略感兴趣区域(regions of interest,ROI),然后在ROI内计算方向梯度直方图(histogram of oriented gradients,HOG)和卷积特征,再分别由DPM和R-CNN采用HOG和卷积特征。	Remote sensing image segmentation for ships to a rough first interested area (regions of interest, ROI), and then within the ROI calculation gradient direction histogram (the histogram of oriented gradients, HOG) and characteristics of convolution and then respectively by using HOG and DPM and R - CNN convolution characteristics.
6512	为测试R-CNN的最佳性能,将具有5个卷积层(ZF网)和具有13个卷积层(VGG网)的网络应用于船只检测。	For test R - CNN's best performance, will have five convolution layer (ZF) and has 13 convolution layer (VGG network) network is applied to ship detection.
6513	使用8张GF2遥感影像的3 523艘船只的实验结果表明,DPM和R-CNN都能以高召回率和正确率检测水中的船只,但对于聚集船只而言,DPM的效果优于R-CNN。	Using eight GF2 remote sensing images of 3 523 ships of the experimental results show that the DPM and R - CNN can high recall ratio and accuracy in detecting the ships in the water, but for ships gathered, the effect of DPM is better than that of R - CNN.