论文部分内容阅读
为了对冬枣损伤进行早期检测,采用近红外高光谱图像技术对损伤区域成像。针对高光谱图像波长多的特点,分别采用连续投影算法、相关特征选择算法、一致性(Consistency)算法选择冬枣损伤的特征波长,对提取的特征波长分别应用k-邻近、朴素贝叶斯(naive bayes,NB)、支持向量机(support vector machine,SVM)3种分类方法进行损伤区域识别。结果表明:所有方法选择的一致特征波长在1 353 nm和1 691 nm附近。Consistency算法选择的特征波长在SVM分类器下分类识别正确率达到95.16%,一致特征波长在NB分类器下分类识别正确率达到84.26%,验证了一致波长的有效性,为多光谱成像技术实现在线检测冬枣损伤提供参考依据。
In order to early detection of jujube injury, near-infrared hyperspectral imaging technique was used to image the damaged area. According to the characteristic of hyperspectral image wavelength, the continuous projection algorithm, the correlation feature selection algorithm and the Consistency algorithm were used to select the characteristic wavelengths of winter jujube damage respectively. The k-adjacent and naive Bayesian (naive) bayes, NB) and support vector machine (SVM). The results show that the consistent characteristic wavelengths of all the methods are around 1 353 nm and 1 691 nm. Consistency algorithm selects the characteristic wavelength classification accuracy of SVM classifier recognition rate of 95.16%, consistent characteristic wavelength classification accuracy of 84.26% under the NB classifier, verifies the consistency of the wavelength of the effective multi-spectral imaging technology to achieve online Detection of jujube injury provide a reference.