ID 原文 译文
44436 为了进一步提高多光谱图像与全色图像的融合质量，提出了一种结合向导滤波和非下采样剪切波变换(NSST)的融合算法。 In order to improve fusion quality of the multi-spectral (MS) and the panchromatic (PAN) images by combining the spectral information with spatial details efficiently, a fusion method based on guided filtering and non-subsampled shearlet transform (NSST) was proposed.
44437 该算法首先对多光谱和全色图像进行 NSST 分解； Firstly, the MS and PAN images were decomposed by NSST respectively.
44438 采用基于形态学滤波亮度调节的融合策略对低频分量进行融合，采用基于改进脉冲耦合神经网络(PCNN)和向导滤波的融合策略指导高频分量的融合； Secondly, the fusion rule of low-frequency coefficients which based on morphological filter-based intensity modulation technology was proposed. For the fusion rule of the high-frequency coefficients, an improved pulse-coupled neural net-work (PCNN) and guided filter was proposed.
44439 最后将高、低频分量系数的融合结果进行 NSST 逆变换，得到最终的融合图像。 Thirdly, the fusion coefficients were reconstructed with inverse NSST to obtain the fusion image.
44440 实验结果表明，提出的算法在主观视觉效果和客观定量评价指标上均优于其他 5 种经典及主流算法，能够在提高融合图像的空间连续性和空间分辨率的同时，很好地保持光谱信息。 Experiments demonstrate that proposed method can improve the spatial continuity and resolution effectively while maintaining the spectral information well. In general, the proposed method is superior to five traditional and state-of-the-art fusion methods from the overall effect of the subjective visual aspects and the objective evaluation indicators.
44441 Web 服务器作为提供政务服务和承载数据的重要平台，安全上面临重大挑战，现阶段面向数据的攻击手段如拖库、篡改、截取、非法访问等层出不穷。 Web service, as the important platform for providing government affairs and hosting the data, the security of which faces significant challenges. Common data-oriented attacks, such as Drag, Tamper, Interception and Illegal Accesses is emerging at this stage.
44442 提出了基于攻击转移的拟态安全网关技术与实现方法，在不直接接触政务数据基础上将访问流量进行策略性转移，为政务数据提供了一种安全可靠的有效保障途径。 To address these security problems, mimic security gateway technology and implement method based on attack diversion model were proposed and studied. On the basis of not accessing the government information directly, strategically transfer the traffic flow was proposed, which provides security insurance.
44443 安全性和性能测试结果表明，拟态攻击转移模型能够完整地模拟政务服务，并在较小开销的前提下有效防御测试中多项攻击类型。 Tests on performance and security show the results that the Mimic Attack Diversion model can completely simulate the government services and defense most of the attacks in the tests in the premise of low overhead.
44444 针对无线网络节点的个体识别问题，从网络节点在信号层的细微特征入手，提出了一种基于深度学习的网络节点个体识别方法。 In view of the identification problem of wireless network nodes, a network node identification method based on deep learning was proposed, based on the subtle characteristics of the network nodes in the signal layer.
44445 首先，采用增强局部学习方法对原始信号采样数据进行特征提取以及降维，在减小数据计算复杂度的同时，保持数据样本的分类特征； Firstly, the enhanced local learning method was used to feature extraction and dimensionality reduction for the original signal data,so as to reduce the complexity of the data and maintain the classification characteristics of the data.