查询词典 gradient

Second, the theory of scale space was used to design a new operator for target detection and tracking. Then, the property of scale invariant was analyzed and proved, and a tracking scheme based on the gradient LOG operator was proposed. Experimental results show that the gradient LOG operator is appropriate to describe small target in complex background, and the detection precision is in the level of sub-pixel. Furthermore, it is efficient and robust to noise.

然后根据尺度空间的基本理论构造了一种针对该成像模型的目标检测算子，分析并证明了该算子具有尺度不变性的优点，设计并实现了基于梯度LOG算子的小目标跟踪方案，实验结果表明：梯度LOG算子能够较好地跟踪复杂背景中的光团目标，定位精度可达亚像素级，且具有计算量小，抭嗓声能力强等优点。

Second, the theory of scale space was used to design a new operator for target detection and tracking. Then, the property of scale invariant was analyzed and proved, and a tracking scheme based on the gradient LOG operator was proposed. Experimental results show that the gradient LOG operator is appropriate to describe small target in complex background, and the detection precision is in the level of sub-pixel.

然后根据尺度空间的基本理论构造了一种针对该成像模型的目标检测算子，分析并证明了该算子具有尺度不变性的优点，设计并实现了基于梯度LOG 算子的小目标跟踪方案，实验结果表明：梯度LOG 算子能够较好地跟踪复杂背景中的光团目标，定位精度可达亚像素级，且具有计算量小，抗噪声能力强等优点。

For the image with random field feature, the description language is probability distribution model, and the description length equals the negative base-two logarithm of the probability of the image data; For the region homogeneity properties of active contour, the description language is chosen as a sign gradient map, and the description length is the summation of sign gradient inside the active contour; For the active contour regularity, the description language is smoothness, and the corresponding description length is the length of active contour curve.

随机场图像的描述语言采用概率分布模型，描述长度是围道内外区域中图像数据概率的以2为底对数的负数；活动围道区域一致性的描述语言为有符号的梯度图，描述长度是围道内区域有符号梯度的总和；围道形态的描述语言采用平滑性，描述长度正比于围道的长度。

Bars are made of 100% Green gradient and the gradient 50 percent of the magenta.

色条是由100％的青渐变和50％的品红渐变所组成的。

The choice of step-length strongly affects the convergence rate of the Gradient Method, the classical Gradient Method—the method of steepest descent converges rather slowly in most cases, the poor behavior of the method is due to the optimal choice of step-length.

步长的选取对梯度法的收敛速度影响非常大，经典的梯度法-最速下降法在大多数情况下收敛得相当慢的原因在于最优步长的选取。

The results of cloud image analysis show that the TBB is lower than 230K (-44℃) when the convection grows most prosperously and the area with the brightness temperature difference of lower than -4℃ corresponds to the place where is most affected by the strong thunderstorm. The center of low value with T≤-7℃ corresponds to the strong echo of radar. There is a good correspondence between the strong convective weather and maximum gradient of TBB, which distributes along the moving cloud cluster. The moving path of strong storm is similar to that of the maximum positive gradient of T.

FY-2C卫星云图定量分析结果表明这次飑线过程对流最旺盛期TBB低于230K(-44℃)；长波红外分裂窗双谱组合T≤-4的低值区与强风暴天气影响区域相对应，T≤-7低值中心与强雷达回波区相对应；强对流天气区域与分布在沿对流云团前进方向的TBB梯度最大区域有很好的对应关系，强风暴天气发生区的移动路径与T正梯度最大区域的移动路径相似。

Across the altitudinal gradient, a total of 12635 individuals belonging to 6 phyla, 15 classes and 33 orders were collected. With increasing altitude, the density of soil fauna decreased, but the group number was less affected. There was a higher co-occurrence of soil faunal communities along the altitudinal gradient, and Nematoda, Acarina, and Collembola were the dominant groups. Below the altitude 370 m, the Shannon-Wiener index, Pielou index, and Margalef abundance index increased with increasing altitude; while above the altitude 370 m, the Shannon-Wiener index and Pielou index decreased but the abundance index increased after an initial slight decrease.

结果表明：土壤动物密度随海拔升高而下降，类群数受海拔影响较小，各海拔群落的共有度较高，优势类群基本均为线虫、蜱螨目和弹尾目；Shannon-Wiener指数、Pielou均匀度指数及Margalef丰富度指数仅在海拔370 m以下随海拔升高而增加，而在海拔370 m以上，Shannon-Wiener指数与Pielou均匀度指数下降，丰富度指数在略有下降后再次上升。

In order to solve problem, this paper proposes the natural gradient algorithms with nonholomic constraints. We derive it by incorporating a nonholonomic constraint into the natural gradient learning algorithm.

章提出了不完整约束的自然梯度算法，该算法是由不完整约束条件与自然梯度算法的结合而推导出来的。

In the last part of this dissertation, we discuss how to model an image is discussed and how to segment an image by EDEK.Zero-Cross principle is a classical concept in signal processing. We generalize the concept to be zero normal curvature after the geometry frame built. Furthermore, we present the explanation to choose the dual parameters in SPMVIF. And we research why the Laplacian gradient superior to edge gradient.

梯度方向二阶导数过零原理是信号处理中的经典结论，在运用微分几何的知识建立SPMVIF的几何框架的同时，将这个结论推广到零-法曲率的形式；另外我们对SPMVIF的双参数选取规则进行了讨论并给出它们的几何意义，并且对SPMVIF中法向矢量选取时拉普拉斯梯度优于边缘图像梯度的结论给予统计信息熵的解释。

This spatial information is normalized gradient vector information instead of ordinary gradient information,and it can enhance the noise resistibility greatly during coarse registration.

文中利用的空间信息不同于一般的图像梯度信息，它是归一化的GVF矢量场信息，把它用于粗配准，抗噪性能良好。

Chrysanthemum of 10 thousand birthday is lax to edaphic requirement, with the arenaceous qualitative loam with fecund, good drainage had better.

万寿菊对土壤要求不严，以肥沃。排水良好的砂质壤土为好。

He unstepped the mast and furled the sail and tied it.

他拔下桅杆，把帆卷起，系住。