论文部分内容阅读
为实现对受到小麦条锈病菌侵染而尚未表现明显症状的小麦叶片进行早期检测,利用近红外光谱技术结合定性偏最小二乘法建立了小麦条锈病潜育期叶片定性识别模型。获取健康叶片30片、条锈病潜育期叶片330片(每天取30片,共11天)和发病叶片30片,扫描获得其近红外光谱曲线。采用内部交叉验证法建模,研究了不同谱区、建模比(建模集∶检验集)、光谱预处理方法和主成分数对建模识别效果的影响。在5 400~6 600和7 600~8 900cm-1组合谱区内,建模比为4∶1、预处理方法为“散射校正”和主成分数为14时,所建模型识别效果较理想,建模集的识别准确率、错误率和混淆率分别为95.51%,1.28%和3.21%;检验集的识别准确率、错误率和混淆率分别为100.00%,0.00%和0.00%。结果表明,利用近红外光谱技术可在接种1天后(即提前11天)识别出健康小麦叶片和受到条锈病菌侵染的小麦叶片,并且可以识别不同潜育期天数的叶片。因此,利用近红外光谱技术对条锈病菌潜伏侵染检测是可行的,为该病早期诊断提供了一种新途径。
In order to realize the early detection of wheat leaves which have not shown the obvious symptom by the infection of wheat stripe rust, a qualitative identification model of wheat leaf rust during the latent period of wheat stripe rust was established by using near infrared spectroscopy combined with partial least square method. Thirty leaves of healthy leaves, 330 pieces of leaves of stripe rust (30 pieces per day for 11 days) and 30 leaves of diseased leaves were obtained, and their NIR spectra were obtained by scanning. The effects of different spectral regions, modeling ratios (model set: test set), spectral preprocessing methods and principal components on model recognition were studied by using internal cross validation method. In the spectral regions of 5 400-6 600 and 7 600-8 900 cm -1, the modeling ratio was 4:1, and the pretreatment method was “Scatter Correction” and the number of principal components was 14, the model recognition effect Ideally, the recognition accuracy, error rate and confusion rate of the model set are 95.51%, 1.28% and 3.21% respectively. The recognition accuracy, error rate and confusion rate of the test set are 100.00%, 0.00% and 0.00% respectively. The results showed that healthy wheat leaves and wheat leaves infected with stripe rust can be identified one day after inoculation (ie, 11 days in advance), and leaves of different incubation periods can be identified. Therefore, the use of near infrared spectroscopy to detect latent infection of stripe rust fungi is feasible, providing a new way for the early diagnosis of the disease.