摘要
鸭梨黑心病是鸭梨的主要生理病害,直接影响鸭梨的出口创汇;准确快速鉴别鸭梨黑心病对于梨的出口业具有重要现实意义。探讨可见/近红外漫透射光谱在线检测鸭梨黑心病的可行性,选取80个正常鸭梨和70个黑心鸭梨作为建模集,建模集被分为校正集和预测集以求获得最优模型。未参与建模的30个正常梨和20个黑心梨作为预测集,用于评价模型的预测能力。鸭梨的可见/近红外漫透射光谱,在5个/秒的速度下采集。建模集能量谱经标准正交变换(SNV)和多元散射校正(MSC)处理后,分别建立健康梨和黑心梨的偏最小二乘判别模型(DPLS)、峰面积判别模型(DPA)、主成分判别模型(DPCA)。用建模集模型判别预测集鸭梨,经比较,DPLS模型的判别准确率最高,黑心梨正确识别率达到100%。实验结果表明:可见/近红外漫透射光谱结合DPLS判别方法,可以实现黑心鸭梨的在线检测。相比传统的人工破损方法,在线检测可为梨出口贸易提供技术支撑和科学参考依据。
Pears which suffer from the main physiological disease as blackheart directly affected the export of pears ,so fast and accurate identification of blackheart pears for pear export is of practical significance .Vis/NIR spectroscopy combining new online technique was proposed to identify exterior and interior properties of objects in the present study .The purpose of this paper is to have a research on the feasibility of using visible/near-infrared diffuse transmission spectrum on-line detector to detect blackheart pears .80 normal pears and 70 pears in disease are taked as modeling set .All samples were divided into the calibration set and prediction set for developing optimal results and assessing their performance .The visible/near-infrared diffuse transmission spec-trum was collected under the speed of 5 pears per minute .Energy spectrum processed by the standard orthogonal transformation （SNV） and multiple scatter correction （MSC） respectively established the model of blackheart pears by partial least-squares dis-criminant model（DPLS） ,peak area discriminant model （DPA） and PCA discriminant model（DPCA） .30 normal pears and 20 blackheart pears were taken as prediction set was used to evaluate the ability to predict .By comparison ,the model of DPLS has the highest accuracy .All the blackheart pears were recognised correctly .The experiment shows that visible/near-infrared diffuse transmission spectrum with DPLS discriminant method realized the on-line detection of blackheart pears which provided technical support and reference for export trade of＇Yali＇ pear products .In comparison to destruction and time-consuming chemical meth-ods ,the results in the feasibility study may provide technical support and scientific reference in testing blackheart pears .
出处
《光谱学与光谱分析》
SCIE
EI
CAS
CSCD
北大核心
2017年第12期3714-3718,共5页
Spectroscopy and Spectral Analysis
基金
国家(863)高技术研究发展计划项目(2012AA101906)
赣鄱英才555工程领军人才培养计划(2011-64)
江西省光电检测工程技术研究中心项目(赣科发财字[2012]155号)
江西省研究生创新专项资金项目(YC2013-S166)
江西省杰出青年人才计划(20171BCB23060)资助
关键词
可见/近红外光谱
漫透射
鸭梨
黑心病
在线
Visible/near-infrared diffuse transmission spectrum
Pear
Blackheart disease
On-line detctor
