二阶微分方程Neumann边值问题解的存在性

研究了Banach空间中二阶Neumann边值问题解的存在性,利用非紧性测度的性质和凝聚映射的Sadovskii不动点定理,获得了若干解的存在性定理。     

基于SOM与BP神经网络的葡萄酒质量评价

葡萄酒有着非常高的营养价值,而决定其营养价值的主要因素是葡萄酒的质量指标。葡萄酒的主要质量指标大体可分为感官指标和理化指标。根据酿酒葡萄和葡萄酒的理化指标,对给定的葡萄酒样品进行质量评价。采用自组织特征映射网络(SOM)对给定的酿酒葡萄样品进行分类,并对酿酒葡萄和葡萄酒的理化指标以及葡萄酒的质量进行聚类,同时采用BP神经网络对结果进行了验证。研究表明,红葡萄与红葡萄酒的理化指标之间的相关性比白葡萄与白葡萄酒的理化指标之间的相关性要大得多;酿酒葡萄和葡萄酒的某些理化指标(白葡萄酒的总糖、红葡萄酒的蛋白质和pH值)、葡萄酒的果香、发酵香物质对葡萄酒质量有显著影响。     

Conditional QTL Mapping of Sedimentation Volume on Seven Quality Traits in Common Wheat

To evaluate the possible genetic interrelationships between flour components and the sedimentation volume（SD）,a doubled haploid（DH） population comprising 168 lines were used to identify the conditional quantitative trait loci（QTLs） for SD in three environments.Ten additive QTLs and 15 pairs of epistatic QTLs were detected for SD through unconditional and conditional QTL mapping.Three major additive QTLs were detected for SD conditioned on the seven quality traits.Two additive QTLs were found to be independent of these traits.Three additive QTLs were suppressed by three of the seven traits because of non-detection in unconditional mapping.Three pairs of epistatic QTLs were completely affected by the seven traits because of detection in unconditional mapping but no-detection in conditional mapping.Twelve pairs of epistatic QTLs were detected in conditional mapping.Our results indicated that conditional mapping could contribute to a better understanding of the interdependence of different and closely correlated traits at the QTL molecular level,especially some minor QTLs were found.The conditional mapping approach provides new insights that will make it possible to avoid the disadvantages of different traits by breeding through molecular design.     

基于特征融合与局部判别映射的苹果叶部病害识别方法

针对利用植物病害叶片图像特征识别病害类别的复杂性,提出一种基于特征融合与局部判别映射的植物叶部病害识别方法。首先,在中心对称局部二值模式(CS-LBP)的基础上,设计了一种自适应中心对称局部二值模式(ACS-LBP),由此分割病害叶片的病斑图像;然后提取并融合病斑图像的纹理、形状和颜色特征;再利用局部判别映射算法对融合特征进行维数约简;最后利用支持向量机进行病害类别分类。在3种常见苹果病害叶片图像数据库上进行病害识别验证试验,结果表明,该方法能够有效识别苹果叶部病害,平均识别率高达96%以上。     

利用径向基函数神经网络通过高光谱数据估算植被生物物理参数

The radial basis function (RBF) emerged as a variant of artificial neural network. Generalized regression neural network (GRNN) is one type of RBF, and its principal advantages are that it can quickly learn and rapidly converge to the optimal regression surface with large number of data sets. Hyperspectral reffectance (350 to 2500 nm) data were recorded at two different rice sites in two experiment fields with two cultivars, three nitrogen treatments and one plant density (45 plants m^-2). Stepwise multivariable regression model (SMR) and RBF were used to compare their predictability for the leaf area index (LAI) and green leaf chlorophyll density (GLCD) of rice based on reffectance (R) and its three different transformations, the first derivative reffectance (D1), the second derivative reffectance (D2) and the log-transformed reflectance (LOG). GRNN based on D1 was the best model for the prediction of rice LAI and GLCD. The relationships between different transformations of reffectance and rice parameters could be further improved when RBF was employed. Owing to its strong capacity for nonlinear mapping and good robustness, GRNN could maximize the sensitivity to chlorophyll content using D1. It is concluded that RBF may provide a useful exploratory and predictive tool for the estimation of rice biophysical parameters.     

基于图像分割映射的农业机器人视觉去雾方法

视觉导航农业机器人在雾天作业容易受前端含雾图像的影响,严重时无法有效工作。提出了一种基于图像分割映射的农业机器人视觉去雾方法。对前端采集图像进行近景与远景区域分割,并通过亮度信息的分段映射获取大气散射函数的预测估计值;采用导向滤波对大气散射函数的估计值进行优化,进一步增强图像的边缘信息,改善大面积天空背景引起的去雾残留问题。基于实际的农业智能导航平台对实测的含雾前端图像进行了去雾分析,并同传统的去雾方法进行了综合比较,显示所提方法具有较高的去雾精度和实时性。两段视频的图像去雾综合指标分别改善了28.9%和29.1%,时间消耗分别减少了34.4%和53.9%。     

含脉冲的时标神经网络的平衡解的存在唯一性（英文）

讨论了一般性的具有脉冲和时滞的时标神经网络解的存在条件.利用非光滑分析和李雅普诺夫泛函,在不假定激励函数有界和可微的条件下,得到了这类神经网络存在唯一的平衡点的充分条件.作为特例,得到了具有脉冲和时滞的时标Hopfield神经网络、细胞神经网络、双向联想记忆神经网络等系统的平衡解的存在唯一性条件.     

基于网格映射法实现修边线的精确计算

在有限元仿真计算模拟翻边过程的基础上,提出了网格映射法来实现修边线的精确计算.将设计翻边轮廓线离散成点并投影到仿真后的网格模型上,确定投影点所在单元编号及相对于单元的位置,利用仿真过程建立的单元和节点映射关系计算出这些投影点在翻边坯料网格模型上的新坐标,然后构造出修边线,通过两至三次迭代计算后就可以实现修边线的精确计算.     

深拖式地震数据采集记录系统软件设计与实现

为应对深拖系统采集记录软件连续采集模式的新要求,在现有的海洋地震勘探系统软件的基础上,采用多线程技术和内存映射技术,实现了连续采集模式下的实时监控和海量地震数据的快速存储,提高了系统工作效率。软件经过近海海试,运行稳定,连续采集和数据存储功能完好,对其他类似的海洋地震数据采集系统软件设计具有一定参考价值。     

基于深度图像的大菱鲆体重估测模型研究

为了对大菱鲆体重进行无损估测,提出了一种基于深度图像的大菱鲆体重估测模型。该方法首先对大菱鲆深度图像进行图像预处理,提取出深度信息与生长数据进行映射建模,拟合出目标特征,并结合网格搜索优化支持向量回归(GS-SVR)算法,实现大菱鲆体重估测。估测结果与实际测量结果进行对比,两者的决定系数(R2)达到0.990 1,均方根误差(RMSE)为0.029 7。该模型具有简单灵活、估测精度高和易于实现等特点,同时具有很好的应用前景。