异构 WSN 中基于参考值的可恢复隐私保护数据融合

文章介绍了无线传感网络的安全数据融合技术,针对当前数据融合算法安全性不高、开销过大以及融合精度低等问题,提出了基于参考值的可恢复隐私保护数据融合算法。该算法利用公钥同态加密机制为数据提供端到端的机密性和完整性认证。此外,该算法动态地为网络中的每个节点设定下次传输的参考值,并通过传输采样值和参考值之间的差值来减少网络中的数据传输量。仿真结果表明,该算法可以有效地减少数据传输量,提高整个网络的能量和带宽效率。     

基于分布式计算的电网运行可靠性评估算法研究

针对电网运行可靠性评估计算量大,但又要求计算时间短的问题,提出了一种基于分布式计算的电网运行可靠性评估算法。该算法采用改进的状态选择和分析方案,解决了计算终端间随机数序列的关联性问题,有效地降低了通信量,提高了计算速度。基于该算法编制的评估软件已通过RTS-24系统和某省级电网的实际运行测试,测试结果表明,该算法不仅能够有效评估电网运行可靠性,而且计算速度随计算终端数增加呈一定倍数提高。     

Aggregation pheromone characterization and comparison inDrosophila ananassae andDrosophila bipectinata

(Z)-11-Octadecenyl acetate (Z11–18Ac) and (Z)-11-eicosenyl acetate (Z11–20Ac) were identified as the aggregation pheromones ofDrosophila ananassae, andZ11–20Ac was identified as the aggregation pheromone ofDrosophila bipectinata. Z11–18Ac andZ11–20Ac were not attractive alone; however, in combination with fermenting food odors, the acetates attracted flies of both sexes in a wind-tunnel olfactometer. The pheromones were present in the ejaculatory bulb of sexually mature male flies and transferred to the female during mating. MaleD. bipectinata released little if anyZ11–20Ac to the food; however, recently mated females releasedZ11–20Ac to the surrounding surfaces in just a few hours after mating.D. ananassae males, on the other hand, appeared to release moreZ11–18Ac andZ11–20Ac to the surroundings than mated females. AlthoughD. bipectinata males had noZ11–18Ac, flies were as attracted toZ11–18Ac as to an equal quantity ofZ11–20Ac.D. ananassae were attracted toZ11–18Ac but not toZ11–16Ac orZ11–20Ac. However,Z11–20Ac in combination withZ11–18Ac was significantly more attractive thanZ11–18Ac alone.     

GPU近实时线性双目立体代价聚合

目的 近年来双目视觉领域的研究重点逐步转而关注其“实时化”策略的研究,而立体代价聚合是双目视觉中最为复杂且最为耗时的步骤,为此,提出一种基于GPU通用计算(GPGPU)技术的近实时双目立体代价聚合算法。方法 选用一种匹配精度接近于全局匹配算法的局部算法——线性立体匹配算法(linear stereo matching)作为代价聚合策略;结合线性代价聚合的原理,对其主要步骤(代价计算、均值滤波及系数求解等)的计算流程进行有针对性地并行优化。结果 对于相同的实验样本,用本文方法在NVIDA GTX780 实验平台上能在更短的时间计算出代价矩阵,与原有的CPU实现方法相比,代价聚合的效率平均有了数十倍的提升。结论 实时双目立体代价聚合方法,为在个人通用PC平台上实时获取高质量双目视觉深度信息提供了一个高效可靠的途径。     

Field Evaluation of Plant Odor and Pheromonal Combinations for Attracting Plum Curculios

The attractiveness of different synthetic host odors and a synthetic aggregation pheromone (grandisoic acid [GA]) to overwintered adult plum curculios (PCs), Conotrachelus nenuphar (Herbst) (Coleoptera: Curculionidae), was examined using two types of traps (sticky panels and black pyramids) placed in border areas surrounding an unsprayed section of an apple orchard in Massachusetts. In 2001, we evaluated the response of PCs to three synthetic fruit volatiles (benzaldehyde [BEN], ethyl isovalerate [EIV], and limonene [LIM]) assessed alone and in combination with GA, as well as the response to GA alone and a no-odor (control) treatment. BEN was the only host volatile that synergized the response of PCs to GA for both trap types. For both trap types, GA was as attractive to PCs as a single component as when in combination with either EIV or LIM. In 2002, four release rates of BEN (0, 2.5, 10, and 40 mg/day) and two release rates of GA (1 and 2 mg/day) were evaluated for attractiveness to PCs using panel and pyramid traps. For panel traps, an increase in amount of GA released (from 1 to 2 mg/day) was associated with a 35% increase in captures. However, PC captures by pyramid traps were similar regardless of the amount of GA released. For panel traps, 10 and 40 mg/day of BEN were the most attractive release rates regardless of the amount of GA released. For pyramid traps baited with GA. PC captures were enhanced by the presence of BEN, regardless of release rate. In 2003, GA at 1 mg/day + BEN at 80 mg/day of release did not enhance PC captures by panel traps relative to lower release rates of BEN. Pyramid traps releasing GA at 1 mg/day performed best when baited with BEN at 10 mg/day of release; a release rate of 80 mg/day of BEN decreased the attractiveness of the binary combination of BEN + GA. Combined results suggest that BEN at 10 mg/day + GA at 1 mg/day of release constitutes an attractive lure that may improve the effectiveness of monitoring traps for PCs.     

碱溶透析和酸析处理对碱木质素物化性质和Pb2+去除性能的影响

采用XPS、静态接触角、颗粒表面电荷、扫描电镜、官能团分析等方法来研究碱木质素经过碱溶透析和碱溶酸析处理后的物化性质的差异，探讨碱木质素分子聚集程度对Pb²⁺去除性能的影响。结果表明，碱溶透析木质素的比表面积是碱溶酸析木质素的4.77倍，碱溶透析样品较酸析样品的表面氧含量高48%、水/木质素薄膜的静态接触角低13°，反映出碱溶透析处理后的木质素比碱溶酸析处理后的木质素分子聚集程度低、分子链疏松、羧基和羟基等含氧官能团裸露程度增加。碱溶透析处理后的木质素与Pb²⁺的静电作用增强，对Pb²⁺的去除能力显著提高。Pb²⁺浓度为100 mg·L^-1时，碱溶透析木质素对Pb²⁺的去除量达136 mg·g^-1，去除率为81.8%，且去除性能受体系pH的影响小。与化学改性和表面功能化相比，利用碱溶透析处理以提高木质素对Pb²⁺的去除性能具有操作简单、无毒、能耗低的优点。     

面向部件分割的PointNet注意力加权特征聚合网络

为解决PointNet最大池化损失次要特征导致部件分割精度降低的问题,提出一种面向部件分割的PointNet注意力加权特征聚合网络,能够充分利用点云的不同特征进行部件分割。首先利用多层感知机提取点云的空间几何特征,将特征乘以共享权重矩阵,以获取每个点的每一个特征的注意力分数;接着把归一化的分数作为权重乘以对应的全局特征并求和,得到聚合的全局特征;最后使用多层感知机将聚合的特征映射到部件分割结果。实验结果表明,相比于传统PointNet方法,该方法提升了部件分割的总平均交并比,同时在网络鲁棒性和计算复杂度方面具有显著优势。该方法有效优化了PointNet。     

基于自适应聚合与深度优化的三维重建算法

针对现有基于多视图的三维重建方法未充分考虑像素点在其余视图的可见性,从而导致重建完整度不足,且在弱纹理和遮挡区域重建困难等问题,提出了一种应用于高分辨率的三维重建网络。首先提出了一种引入可见性感知的自适应成本聚合方法用于成本量的聚合,通过网络获取视图中像素点的可见性,可以提高遮挡区域重建完整性;采用基于方差预测每像素视差范围,构建空间变化的深度假设面用于分阶段重建,在最后一阶段提出了基于卷积空间传播网络的深度图优化模块,以获得优化的深度图;最后采用改进深度图融合算法,结合所有视图的像素点与3D点的重投影误差进行一致性检查,得到密集点云。在DTU 数据集上与其他方法的定量定性比较结果表明,提出的方法可以重建出细节上表现更好的场景。     

基于模糊强化学习和果蝇优化的WSN数据聚合算法

在无线传感器网络中,传感器的能量时有限的,如果传感器的能量耗尽,那么无线传感网络的鲁棒性和寿命就会大大降低.因此,提出了基于模糊强化学习和果蝇优化的数据聚合机制,以最大限度地延长网络寿命,并进行高效数据聚合.首先,网格聚类用于簇的形成和簇头的选择,接着评估各个网格簇所有可能的数据聚合节点,然后采用模糊强化学习选取最佳数据聚合节点,最后利用果蝇优化算法动态定位整个无线传感网络的数据汇聚节点.仿真结果表明,提出的数据聚合方案在能耗和网络鲁棒性方面优于对比方案.