| ID | 原文 | 译文 |
| 5764 | 但是其在进行端元提取时,采用的计算公式仍需进行矩阵求逆,随着端元的逐个求解,矩阵维数增多导致计算量增加。 | But the yuan extraction in the end, the calculation formula of would still need a matrix inversion, as the yuan to solve them one by one, matrix dimension increased lead to increase the amount of calculation. |
| 5765 | 由于端元提取时获得的端元Gram矩阵满足对称特性,引入埃尔米特矩阵(Hermitian matrix)分块求逆引理,简化矩阵求逆处理,优化快速Gram行列式端元提取方法。 | Due to the side when yuan extraction and matrix "gramm meet symmetrical features, the introduction of Hermitian matrix (Hermitian matrix) partitioned inversion lemma, simplified matrix inversion processing, optimization of rapid extraction method" gramm determinant components. |
| 5766 | 采用美国Cuprite矿区的机载可见光/红外成像光谱仪(airborne visible/infrared imaging spectrometer,AVIRIS)机载高光谱图像进行实验验证,并对该方法在不同初始化条件下的求解结果进行分析。 | Adopt the Cuprite mining area of the airborne visible light/infrared imaging spectrometer (fizzy visible/infrared imaging spectrometer, AVIRIS) airborne hyperspectral image experiment, and the method in different initial conditions were analyzed. |
| 5767 | 结果表明快速Gram行列式端元提取方法会受到初始条件的影响,在端元、像元数量增加时所提方法可提升计算效率。 | And the solution results showed that the rapid extraction method "gramm determinant components are influenced by the initial conditions, in the yuan, like yuan increase in the number of the proposed method can improve calculation efficiency. |
| 5768 | 针对高光谱图像(hyperspectral images,HSI)中缺损像元及条带影响图像后续处理及应用的问题,应用稀疏表示理论,将HSI修复问题建模为不完整观测下的信号稀疏重建问题,提出自适应稀疏编码实现的HSI修复算法。 | For hyperspectral images (hyperspectral images, HSI) of defect like yuan and stripe image processing and application, application of the theory of sparse representation and the HSI is modeled as incomplete repair problems of signal sparse reconstruction problem, HSI repair to realize adaptive sparse coding algorithm. |
| 5769 | 首先,对加性噪声假设下的HSI观测模型进行研究。 | First of all, under the assumption of additive noise HSI observation models are studied. |
| 5770 | 然后,通过引入基于随机近似的在线学习优化方法,提出新的从高光谱数据中直接构造字典的算法,从而获取光谱字典。 | Then, with the introduction of online learning optimization method based on stochastic approximation, put forward the new constructed directly from hyperspectral data dictionary algorithm, so as to obtain the spectral dictionary. |
| 5771 | 之后,应用变量分解和增广拉格朗日稀疏回归方法对图像进行稀疏编码求解。 | Later, using variable decomposition and augmented Lagrangian sparse regression method for sparse coding to solve. |
| 5772 | 最后通过稀疏重构求得修复后的HSI。 | Finally by sparse reconstruction after repair of the HSI. |
| 5773 | 实验结果表明,相对于现有算法,在不同噪声条件下,所提算法均能够更有效地修复缺损的HSI,且与其他字典学习类修复算法相比计算耗时更短。 | The experimental results show that, compared with the existing algorithms, under the condition of different noise, the proposed algorithm can more effectively repair defects of HSI, and compared with other dictionary them repair algorithm calculation time shorter. |