论文部分内容阅读
为了研究快速识别轻微损伤壶瓶枣与完好壶瓶枣的有效方法,本文以轻微损伤壶瓶枣和完好壶瓶枣为研究对象,动态采集轻微损伤壶瓶枣和完好壶瓶枣的近红外光谱数据。采用S-G平滑与多元散射校正(MSC)相结合的方法预处理光谱数据,分别以预处理后的全光谱(FS)数据和采用主成分分析(PCA)法提取主成分、采用连续投影算法(SPA)提取特征波长作为输入变量,建立偏最小二乘判别分析(PLS-DA)和最小二乘支持向量机(LS-SVM)模型,比较4种损伤壶瓶枣及完好壶瓶枣的判别准确性。结果表明:采用PCA提取主成分有较明显的优势,对4种损伤壶瓶枣的判别准确性均能满足实际要求,且采用PCA-LS-SVM模型对4种轻微损伤壶瓶枣和完好壶瓶枣的正确判别率最佳,分别达到100%、86%、100%、100%和100%,总的正确判别率为97.2%。该研究为轻微损伤壶瓶枣的动态判别提供了新的理论基础。
In order to study the rapid identification of minor damage to the bottle and jujube pot pomelo effective methods, this paper, a minor damage to the bottle and jujube pachytene jujube as a research object, the dynamic acquisition of a small bottle and jujube pachinko nenarrhea NIR spectra data. The spectral data were preprocessed by the combination of SG smoothing and multivariate scatter correction (MSC). The preprocessed whole spectrum (FS) data and principal components analysis (PCA) were used to extract the principal components. The continuous projection algorithm (SPA (PLS-DA) and least square support vector machine (LS-SVM) models were established to compare the discriminant accuracy of the four kinds of jugulosa pots with jujube damage . The results showed that PCA was more effective in extracting principal components, and the accuracy of discriminating four kinds of bottle-necked jujube could meet the practical requirements. PCA-LS-SVM The correct identification rate of bottle jujube was the best, reaching 100%, 86%, 100%, 100% and 100% respectively. The total correct discrimination rate was 97.2%. This study provides a new theoretical basis for the dynamic discriminant of slight damage to pot bottle jujube.