ID 原文 译文
12114 针对无损检测红外热波图像对比度低、边缘模糊、含大量噪声的问题,提出了基于Contourlet变换和混沌变异双粒子群优化(adaptive chaotic variation particle swarm optimization,ACPSO)的自适应增强方法。 For the infrared thermal wave nondestructive detection, image contrast is low, edge blur, containing a lot of noise, is proposed based on Contourlet transform and chaos mutation particle swarm optimization (adaptive chaotic variation particle swarm optimization, ACPSO) methods of adaptive enhancement.
12115 红外热波图像经Contourlet变换分解成低通和带通方向子带。 ‭Infrared thermal wave image by Contourlet transform into subband low-pass and band-pass direction.
12116 低通子带系数依据一种适应于人类视觉系统的灰度级变换调整,待定参数由ACPSO确定,为了得到最佳增强效果,适应度函数由一种对比度测量函数确定; ‭Low-pass subband coefficients based on an adaptation of gray level transformation to adjust the human visual system, undetermined parameters are determined by ACPSO, in order to get the best effect, the fitness function determined by a contrast measurement function;
12117 带通方向子带系数的调整则采用非线性增益函数实现,从而抑制噪声并增强细节。 Bandpass direction of subband coefficients, by a nonlinear gain function is to restrain noise and enhance the details.
12118 大量红外热波图像增强实验结果表明,与现有的4种增强方法相比,能大大提高缺陷和背景之间的对比度,增强缺陷的边缘细节。 ‭A large number of experimental results show that the infrared thermal wave image enhancement compared with the existing four enhancement methods, can greatly improve the contrast between the defects and background, enhance the edges of the defect details.
12119 进一步采用倒数熵多阈值分割方法时,能更有效地提取缺陷,为后续准确进行缺陷识别和尺寸测量奠定了基础。 ‭Further using inverse entropy threshold segmentation method, more effectively extract defect, for subsequent accurate defect recognition and measurement laid a foundation.
12120 提出一种稀疏描述与结构特征相结合的极化合成孔径雷达(polarimetric synthetic aperture radar,PolSAR)图像斑点抑制算法。 Put forward a kind of sparse description combined with structure characteristics of polarization synthetic aperture radar (polarimetric synthetic aperture radar, PolSAR) image speckle suppression algorithm.
12121 首先利用图像的极化信息对原图像按结构特征分类,形成分类标记图; ‭ First use of polarization information of the image of the original image according to the structure characteristics of classification, form classification mark figure; ‭
12122 然后采用正交匹配追踪(orthogonal matching pursuit,OMP)算法对图像进行稀疏分解,利用K奇异值分解(K-singular value decomposition,K-SVD)算法对过完备字典进行训练更新,得到图像相应的训练字典和稀疏系数,重构图像; ‭ Then the orthogonal matching pursuit (orthogonal matching pursuit, OMP) algorithm for image sparse decomposition, using singular value decomposition (SVD) K (K - singular value decomposition, K - SVD) training algorithm for complete dictionary update, get the image of the corresponding training dictionary and sparse coefficient, reconstructing images;
12123 最后在重构图像中按分类图增强相应的点线目标。 ‭Finally according to the classification figure in the reconstructed image enhancement corresponding target point line.