论文部分内容阅读
拉曼光谱判别具有特征维数高等特点,若不进行特征提取可能会出现“维数灾”问题,为了提高拉曼光谱判别的准确性和效率,提出一种核主成分分析与神经网络的拉曼光谱判别模型。首先采集拉曼光谱判别数据,并对数据进行预处理,然后采用主成分分析提取拉曼光谱判别的重要特征,简化神经网络的结构,最后采用神经网络对重要特征进行训练,得到拉曼光谱判别结果。仿真测试结果表明,在相同实验条件下,相对于其他拉曼光谱判别模型,本文模型得到了更高的拉曼光谱判别正确率,减少了计算量,加快了拉曼光谱判别的速度。
In order to improve the accuracy and efficiency of Raman spectrum discrimination, a principal component analysis (PCA) and neural network (ANN) are proposed to solve the Raman spectrum discrimination problem. Raman spectrum discriminant model. Firstly, the discriminant data of Raman spectra were acquired, and the data were preprocessed. Then, the principal components analysis was used to extract the important features of Raman spectrum discrimination to simplify the structure of the neural network. Finally, the neural network was used to train the important features to get the Raman spectrum discrimination result. Simulation results show that under the same experimental conditions, compared with other Raman spectroscopy discriminant models, the proposed model achieves a higher Raman spectrum discrimination accuracy, reduces the computational complexity and speeds up the discrimination of Raman spectra.