ID 原文 译文
8354 该算法从雾天图像模型出发,首先利用估计天空区域更加准确的明暗像素先验获取介质传输图粗估计,然后在随机游走模型的框架下,将粗估计的介质传输图作为先验约束传统随机游走能量模型,进一步优化介质传输图,得到最终的无雾图像。 Starting from the image contrast model, the algorithm is first used to estimate the sky area pixel prior to obtain more accurate light and shade transmission medium coarse estimation, and then under the framework of the random walk model, the coarse estimation of transmission medium figure as the prior energy constraints, the traditional random walk model, further optimize the medium transmission diagram, get the final no fog image.
8355 实验结果表明,该算法有效地在随机游走模型下结合了明暗像素先验对雾天图像估计的优缺点,证实了所提方法的可行性和有效性。 The experimental results show that the proposed algorithm effectively under the random walk model is a combination of the advantages and disadvantages of pixel prior to estimate the image contrast of light and shade, confirmed the feasibility and effectiveness of the proposed method.
8356 为高效处理3D视线跟踪技术中的非线性优化问题,使系统满足实时准确及稳定性需求,以差分进化(differential evolution,DE)为核心,结合混合蛙跳算法(shuffled frog-leaping algorithm,SFLA)及Nelder-Mead单纯形法算法思想,提出了一种新型混合算法,即DE-SFL-NM混合算法。 For efficient processing nonlinear optimization problem in 3 d eye tracking technology, make the system meet the demand of real-time accuracy and stability, with differential evolution (differential evolution, DE) as the core, combined with hybrid leapfrog algorithm (shuffled frog leaping - algorithm, SFLA) and Nelder Mead simplex method algorithm thought, a new hybrid algorithm is proposed, namely DE - SFL - NM hybrid algorithm.
8357 利用无穷乘积的性质对DE-SFL-NM进行了收敛性分析,并得出依概率收敛结论。 By using the properties of infinite product of DE - SFL - NM convergence is analyzed, and the convergence in probability conclusions.
8358 使用包含单、多模态的10个基准测试函数的数值实验结果进行比较,验证了该算法在收敛速度、求解精度及鲁棒性能方面的有效性和进步性。 Using include single and multiple mode of 10 benchmark test function compares the results of numerical experiment, shows that the algorithm in convergence speed, accuracy and validity of the robust performance and progressive.
8359 同时,应用DE-SFL-NM快速且精准地求解了3D视线跟踪系统中的角膜曲率中心。 At the same time, the application of DE - SFL - NM quickly and accurately solve the corneal curvature center of 3 d vision tracking system.
8360 方位信号重构是多通道合成孔径雷达(sythetic aperture radar,SAR)成像过程中关键的一个步骤。 The azimuth signal reconstruction is the multichannel synthetic aperture radar (sythetic aperture radar, SAR) imaging is a key step in the process.
8361 文中在现有算法的基础上进行改进和扩展,提出了一种基于多普勒谱结构估计的多通道SAR盲重构方法,能够有效解决模糊分量个数和位置变化,导向矢量未知时信号重构的问题。 In this paper, on the basis of the existing algorithm was improved and extended, this paper proposes a estimated based on the doppler spectrum structure of multi-channel SAR blind reconstruction method, can effectively solve the fuzzy weight number and position change, oriented vector unknown signal reconfiguration problem.
8362 该方法借助Capon谱估计的思想获得混叠多普勒谱的结构图,然后根据结构图构造导向矢量,实现方位信号无模糊重构。 The method with the help of a Capon spectrum estimates the thoughts of aliasing doppler spectrum chart, then according to the structure diagram oriented vector structure, realize the azimuth signal fuzzy refactoring.
8363 所提出的算法能够尽可能保持多普勒谱的完整性,并抑制多普勒谱边缘的噪声。 The proposed algorithm can as far as possible to maintain the integrity of the doppler spectrum, and suppress noise on the edge of the doppler spectrum.