航拍图像车辆检测中的圆形滤波器HOG特征快速计算

航拍图像中车辆一般近似为矩形结构,因此通过统计检测窗口中的梯度方向直方图并根据梯度主方向估计车辆朝向,将检测窗口旋转到相应方向进行分类器判别。车辆检测采用级联boosting分类器和梯度方向直方图特征,针对旋转窗口中梯度方向直方图特征的计算,设计一种基于圆形滤波器的梯度方向直方图特征。与传统基于积分直方图的梯度方向直方图特征提取方法相比,显著提高了旋转窗口中梯度方向直方图特征的计算效率,在计算每个像素的梯度时采用查找表代替梯度向量的求模和角度计算也减小了计算量。使用真实图像进行的实验表明,该车辆检测算法快速高效。     

利用支持向量机的矢量跟踪通道故障检测算法

为了克服卫星导航矢量跟踪接收机中故障通道对其他正常通道的影响,提高矢量跟踪接收机的鲁棒性,提出一种利用支持向量机的通道故障检测算法,利用矢量跟踪接收机的导航滤波器的新息序列作为支持向量机的输入。支持向量机的输出为通道的状态,根据支持向量机的输出通道的状态来判断该通道是否纳入导航滤波器,用来跟踪滤波器状态值,这样能够有效地避免故障通道对导航结果的污染。仿真实验结果表明:该方法能准确地检测出有故障的通道,提高矢量跟踪环路的鲁棒性。     

用于机动目标跟踪的自适应交互式多模型算法

针对当前统计模型对弱机动或非机动目标跟踪效果不理想等问题,提出了一种修正当前统计模型与匀速模型的自适应交互式多模型算法,可在线修正当前统计模型的加速度极限值,调整过程噪声方差,提高了当前统计模型的自适应性。同时,通过在常规匀速模型中引入机动检测机制,抑制了常规匀速模型对机动目标跟踪的滤波发散,通过引入强跟踪算法,增强了模型对目标突发机动的自适应跟踪能力。仿真结果表明,该算法充分发挥了当前统计模型和交互式多模型算法的优势,对强机动和弱机动目标都具有很好的效果。     

基于CUDA GPU的多摄像机场面航空器识别加速算法

基于摄像机的航空器识别是机场场面监视的重要工具。针对多摄像机场面航空器识别算法存在的计算效率低等缺点,提出基于GPU CUDA的加速算法。利用CUDA线程并行处理能力与GPU计算能力,对算法进行了重新设计与优化。通过实地对多路场面视频监视数据进行了多次实验,验证了在NVIDIA Geforce 8800GTS显卡上可实现10倍以上的加速性能,提高了航空器目标识别效率,可以满足机场场面监视中对航空器识别与跟踪的实时性要求。     

The resource-conflict debate revisited: Untangling the case of farmer–herdsman clashes in the North Central region of Nigeria

The resource debate is easily discerned as part of the ongoing history of farmer–herdsman conflict in the North Central region of Nigeria. Scarcity theorists are adept at linking scarcity with the onset of livelihood conflict while on the other hand resource abundance pundits insist it is profusion and not scarcity that impels conflict. This article traverses these wrangles and proceeds to downplay the resource polemic altogether. It is proposed that the resource debate, despite its profoundness, presents a narrow reading of farmer–herdsman clashes in Nigeria’s North Central region. It is suggested that a number of other factors, including elite land grabbing, ethno-religious identity construction, weak state capabilities, the citizenship question, corrupt traditional institutions, the lack of an effective land tenure system and a widespread culture of impunity, make for better readings of the conflict. Owing to the negative impacts of the conflict on state and society, it is recommended that the state in Nigeria should commit itself to addressing the citizenship question, strengthening the capabilities of security institutions, extirpating the culture of impunity and revaluing its policy on land redistribution.     

基于ICA的油库阀门内漏声发射定量检测

基于声发射原理的阀门内漏检测作为一种动态无损检测方法得到了广泛应用,但是定量检测一直是研究难点。现有的内漏率定量检测模型计算方法因参数不易测定,导致计算困难,精度不高。采用独立分量分析(ICA)方法提取声发射信号特征向量,并与参考样本集的特征向量相比较,相似程度最高的即可作为待测阀门的内漏率。内漏模拟实验分别将常规参数、频谱成分和ICA提取特征作为特征向量进行比较,结果表明ICA提取的特征向量维数为15时,均方根误差可以达到0.01 L/min,能够满足工程检测需要。     

Resource rents,institutions, and violent civil conflicts

Natural resources have been blamed for inducing slow growth and sparking civil conflicts and violence. This paper first develops a model to account for the hazard of armed civil conflicts as a manifestation of the natural resource curse, which is mediated by the quality of both economic and political institutions. We then use recently published data on institutional quality and natural resource rents to measure the potential impact of the resource curse on violent civil conflicts using a panel of data for over 100 countries in the period 1970–2010. Our model explicitly accounts for the role of good economic and political institutions in deterring the recourse to violence as well as the extent to which they might weaken the resource rents effect.     

Resource Dependence and Armed Violence: Impact on Sustainability in Developing Countries

The dependence on oil, gas, and mineral exports arguably has a negative impact on economic growth in resource-rich, developing countries. This article looks at the impact of resource dependence on adjusted net savings (ANS) as an indicator of weak sustainability. Our results, based on a panel of 104 developing countries during the recent commodity price boom, confirm a negative relationship between resource extraction and sustainable development as measured by ANS. We further look at the specific role of armed conflict and armed violence as captured by the homicide rate. Armed conflict, which is positively associated with resource dependence, negatively affects ANS per capita according to both our OLS and instrumental variables (IV) estimates. Similarly, armed violence has a detrimental effect on sustainable development. Our IV estimate suggests that a one-point increase in the homicide rate decreases ANS per capita by $60. Since education expenditures are a critical ANS component, we further examine the impact of resource dependence and violence on human capital. Consistent with previous findings, resource-dependent countries underinvest in education but armed conflict and violence do not affect the instantaneous share of education expenditures, hinting at a detrimental effect working through physical and social capital rather than education.     

基于AHP-PCR5的多传感器目标识别

目标识别系统中所获得的信息常常是高度冲突和不确定的。基于DSmT理论的多传感器目标识别,可以解决证据高度冲突情况下的信息融合问题。然而由于DSmT理论融合结果分类精细而不利于判决,需要将某些分类结果进行重新分配。层次分析法(AHP)包含不确定知识矩阵,生成基本信度分配函数。基于此提出AHP-PCR5方法进行证据高度冲突情况下多传感器综合目标识别,不仅提高识别精度,降低识别过程的不确定因素,同时引入折扣系数灵活处理识别过程中具有不同可信度的多传感器融合问题。