| 基于光谱与图像信息的杏鲍菇多糖含量检测 |
| |
| 引用本文： | 宋镇,姬长英,张波.基于光谱与图像信息的杏鲍菇多糖含量检测[J].华南农业大学学报,2019,40(3):104-110. |
| |
| 作者姓名： | 宋镇 姬长英 张波 |
| |
| 作者单位： | 南京农业大学工学院,江苏南京,210031;南京农业大学工学院,江苏南京,210031;南京农业大学工学院,江苏南京,210031 |
| |
| 基金项目： | 江苏省重点研发计划（SBE2015310266）；江苏省自然科学基金（BK20140729） |
| |
| 摘 要： | 【目的】利用高光谱成像技术实现杏鲍菇Pleurotus eryngii多糖含量的快速无损检测。【方法】利用高光谱图像采集系统获取350～1 021 nm波长范围内的杏鲍菇高光谱图像,同时利用苯酚–硫酸法测定对应样本的多糖含量。通过波段运算和阈值分割构建掩膜图像,使样本与背景相分离。采用主成分分析(PCA)处理原始高光谱图像,获得代表原始图像99%信息的2个主成分图像(PC1、PC2),然后利用连续投影算法(SPA)选出554.4、772.8、811.4、819.1、855.6、986.3和1 019.5 nm 7个特征波长及对应的光谱特征,分别提取7个特征波长图像和2个主成分图像的纹理与颜色特征,最后利用偏最小二乘回归(PLSR)建立杏鲍菇样本基于不同图像特征与多糖含量之间的关系模型。【结果】从校正集决定系数(Rc2)来看,基于特征光谱+特征波长图像特征+主成分图像特征的模型效果最好,Rc2=0.954,RMSEc=0.341;从预测集决定系数Rp2来看,基于特征光谱+特征波长图像特征的模型效果最好,Rp2=0.868,RMSEP=0.539。【结论】该研究结果可为杏鲍菇多糖含量的快速、无损检测提供一定的参考。
|
| 关 键 词： | 高光谱图像 杏鲍菇 多糖 主成分分析 特征波长 纹理特征 |
| 收稿时间： | 2018/7/26 0:00:00 |
Detection of polysaccharide content in Pleurotus eryngii based on spectral and image information |
| |
| SONG Zhen,JI Changying and ZHANG Bo.Detection of polysaccharide content in Pleurotus eryngii based on spectral and image information[J].Journal of South China Agricultural University,2019,40(3):104-110. |
| |
| Authors: | SONG Zhen JI Changying and ZHANG Bo |
| |
| Institution: | College of Engineering, Nanjing Agricultural University, Nanjing 210031, China,College of Engineering, Nanjing Agricultural University, Nanjing 210031, China and College of Engineering, Nanjing Agricultural University, Nanjing 210031, China |
| |
| Abstract: | Objective To quickly and non-destructively detect polysaccharide content in Pleurotus eryngii using hyperspectral imaging technology.Method Hyperspectral images of P. eryngii in the visible and near infrared (390-1050 nm) regions were acquired using the hyperspectral imaging system. Polysaccharide contents in corresponding P. eryngii samples were measured by phenol sulfuric acid method. The binary mask image was constructed by the method of band operation and threshold segmentation to separate the sample area from the background area. Principal component analysis (PCA) was used to process the original hyperspectral images, and two principal component images (PC1, PC2) representing 99% information of the original image were obtained. Seven characteristic wavelengths of 554.4, 772.8, 811.4, 819.1, 855.6, 986.3, and 1 019.5 nm were selected using the successive projection algorithm (SPA). Texture and color data were extracted from two principal component images and seven characteristic wavelength images, and spectral data were also extracted from seven characteristic wavelength images. Using partial least squares regression (PLSR) models were established based on the correlations of different image features and polysaccharide contents in P. eryngii samples Result According to the determination coefficient of the calibration set (Rc2), the best model is the one based on characteristic spectra, characteristic wavelength images and principal component images with Rc2=0.954 and RMSEC=0.341. According to the determination coefficient of the prediction set (Rp2), the best model is the one based on characteristic spectra and characteristic wavelength images with Rp2=0.868 and RMSEP=0.539.Conclusion This study provides references for fast and non-destructive detection of polysaccharide content in Pleurotus eryngii. |
| |
| Keywords: | hyperspectral image Pleurotus eryngii polysaccharide principal component analysis characteristic wavelength texture feature |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《华南农业大学学报》浏览原始摘要信息 |
| 点击此处可从《华南农业大学学报》下载免费的PDF全文 |
|
