猕猴桃便携式无损检测仪影响因素及其精度校准分析OACSTPCD
Analysis of Influencing Factors and Accuracy Calibration of Portable Nondestructive Testing Instrument of Kiwifruit
为探究产地、成熟度对便携式可溶性固形物无损检测精度的影响,以同一品种3个产地的猕猴桃为试材,分别采用便携式无损检测仪和常规有损检测仪对其可溶性固形物含量进行测定和比较,并运用人工神经网络对其进行分析.结果表明,在使用便携式无损检测仪对猕猴桃可溶性固形物检测时,3个产地的无损检测R2分别为0.26、0.52、0.61,成熟度会对其检测精度造成一定影响,但在不同产地间没有稳定、统一的规律;基于人工神经网络学习、建模产生的随机预测值可将R2提高5%～49%.因此,可利用预测值来作为中间校准程序,以辅助提高便携式无损检测精度.
In order to explore the influence of cultivation bases and maturity on the precision of portable nondestructive testing for the soluble solid of kiwifruit,the samples of the same kiwifruit variety from 3 cultivation bases were used as the materials,and the soluble solid of kiwifruit was determined by portable nondestructive testing instrument and conventional lossy testing respectively.The result was analyzed by artificial neural network.The results showed that when the portable nondestructive testing instrument was used to detect the soluble solid of kiwifruit,the nondestructive testing R2 of 3 cultivation bases were 0.26,0.52 and 0.61 respectively.Maturity would have a certain impact on its detection accuracy,but there was no stable uniform law among different cultivation bases.The random predicted value based on artificial neural network learning and modeling could increase R2 by 5%～49%.Therefore,the predicted value can be used as an intermediate calibration procedure to help improve the accuracy of portable nondestructive testing.
张琛;郗笃隽;裴嘉博;刘辉;汪末根
杭州市农业科学研究院 园艺研究所,浙江 杭州 310024淳安县农业农村发展服务中心,浙江 淳安 311700
园艺学与植物营养学
猕猴桃可溶性固形物便携式无损检测精度人工神经网络
KiwifruitSoluble solidPortable nondestructive testingAccuracyArtificial neural network
《江西农业学报》 2024 (004)
53-58 / 6
浙江省科技计划项目(2021C02066-8-3);浙江省农业重大技术协同推广计划(2022XTTGGP03);杭州市农科院科技创新与示范推广基金项目(2022HNCT-14);杭州市农业科技协作与创新攻关项目(市院合作项目).
评论