论文部分内容阅读
为了快速无损地获取微藻生物膜生长信息,本文利用高光谱成像技术在线原位监测了不同时期微藻生物膜的特征光谱,提取了25个光谱特征参数;利用相关系数对光谱特征参数与生物膜量进行了分析,提取了相关系数绝对值较大的四个特征参数:(SDr-SDy)/(SDr-SDy)、SDr、OSAVI、NDVI;同时对比分析了单一特征的曲线拟合模型、BP神经网络模型及多特征参数曲线拟合模型对生物膜生长预测所需时间及精度。结果表明:单一特征拟合模型预测精度的SDr参数为5.12%,预测时间为0.018s;BP神经网络模型预测的精度为2.68%,预测时间为0.873s;多特征参数曲线拟合模型的预测精度提高到3.01%,预测时间缩短至0.024s。实验结果及理论分析表明多光谱特征参数拟合模型对生物膜量的预测较好,可为微藻生物膜高效培养提供参考。
In order to quickly and non-destructively obtain the microalgae biofilm growth information, this paper used in-situ monitoring of the characteristic spectra of microalgae biofilm in different periods by using hyperspectral imaging technology and extracted 25 spectral characteristic parameters. By using the correlation coefficient, (SDr-SDy) / SDr-SDy, SDr, OSAVI and NDVI. The curve fitting model with single feature was compared and analyzed. BP neural network model and multi-parameter curve fitting model for biofilm growth prediction time and accuracy. The results show that the prediction accuracy of single feature fitting model is 5.12% and the prediction time is 0.018s. The prediction accuracy of BP neural network model is 2.68% and the prediction time is 0.873s. The prediction accuracy of multi-parameter curve fitting model To 3.01%, and the forecast time is shortened to 0.024s. Experimental results and theoretical analysis show that the multi-spectral characteristic parameter fitting model predicts the amount of biofilm well and provides a reference for the efficient cultivation of microalgae biofilms.