ID 原文 译文
47956 所提方法在收敛速度和对准精度上具有一定优势,并从计算量的角度讨论了算法的实用性和可行性, Proposed method has certain advantages in the convergence speed and alignment accuracy, and from the Angle of the calculation discussed the practicability and feasibility of the algorithm,
47957 在一定程度上兼顾了对准精度和快速性,对惯导设备的快速反应能力具有重要的现实意义。 to a certain extent, the alignment accuracy and rapidity, the quick reaction capability of inertial device has important practical significance.
47958 通过实测数据对所提算法的有效性进行了验证,同时使用不同数据段对准和不同初始姿态下的对准进一步展示了算法的一般性。 Through the measured data on the effectiveness of the algorithm are verified, at the same time using different data under different initial alignment and the alignment of further shows the general algorithm.
47959 为更好地实现图像跟踪,寻找更具鲁棒性和计算简便的特征描述子,提出了一种基于核局部不变映射的尺度不变特征转换(scale-invariant feature transform,SIFT)特征描述算法。 To better realize the image tracking, looking for more robustness and simple calculation of character description, this paper proposes a local invariant mapping based on nuclear scale invariant feature transform (scale - invariant feature transform, SIFT) feature description algorithm.
47960 该算法在继承SIFT算法良好性质的基础上, The algorithm inherited the SIFT algorithm, on the basis of good properties,
47961 依据不同空间尺度下能量特征差异性,对尺度内的子图像层数进行细化,以提高稳定特征点的数量。 according to different spatial scales energy difference, the scale of the layer number of sub image thinning, to increase the number of stable feature points.
47962 此外,借助核方法的映射特性,解决了局部不变映射法丢失非线性高维特征的问题, Mapping features, in addition, with the aid of kernel methods to solve the local constant loss problem of nonlinear high-dimensional feature mapping method,
47963 形成一种基于核局部不变映射的非线性降维法,进而对特征描述子进行特征重划。 forming a kind of nonlinear dimension reduction method based on kernel partial constant mapping, and characteristics of feature descriptor to redraw.
47964 实验结果表明,在图像尺度缩放、旋转、模糊、亮度变化等多种场景下, Experimental results show that the image size scaling, rotating, fuzzy, brightness change wait for a variety of scenarios,
47965 相较现有的主成分分析-SIFT算法,该描述子不但取得更多的稳定特征点,而且计算速度也得到大幅提升。 compared with the existing principal component analysis - the SIFT algorithm, the descriptor is not only more stable feature points, and the calculation speed has also been improved.