检索规则说明:AND代表“并且”;OR代表“或者”;NOT代表“不包含”;(注意必须大写,运算符两边需空一格)
检 索 范 例 :范例一: (K=图书馆学 OR K=情报学) AND A=范并思 范例二:J=计算机应用与软件 AND (U=C++ OR U=Basic) NOT M=Visual
作 者:吴德刚[1] 赵利平[1] 陈乾辉[1] WU Degang;ZHAO Liping;CHEN Qianhui(Shangqiu Institute of Technology,Shangqiu Henan 476000,China)
机构地区:[1]商丘工学院,河南商丘476000
出 处:《激光杂志》2023年第7期212-217,共6页Laser Journal
基 金:河南省高等学校青年骨干教师培养计划项目(No.2018GGJS190);商丘工学院2022科研项目(No.2022KYXM02)。
摘 要:针对目前的农产品成熟度无损检测方法存在检测精度较差的问题,提出基于近红外光谱的大枣成熟度无损检测方法。利用近红外光谱仪采集大枣样品的近红外光谱,并计算采集到的近红外光谱的马氏距离,将超出阈值的马氏距离样品设置为异常样品,为此删除异常的大枣近红外光谱;利用连续投影算法对异常删除的大枣近红外光谱中的波段选择;将大枣近红外光谱波段选择结果作为卷积神经网络输入,通过卷积层和池化层的运算,输出大枣成熟度无损检测结果。实验结果表明:本方法的大枣成熟度无损检测精度最高可达99.12%,无损检测时间最高为11 s,无损检测误差最高为0.06,说明本方法无损检测精度好,具有较高的技术水平和应用价值。Aiming at the problem of poor detection accuracy of the current nondestructive testing methods for the maturity of agricultural products,this paper proposes a nondestructive testing method for the maturity of jujube based on near infrared spectroscopy.The near-infrared spectra of jujube samples were collected by near-infrared spectrometer,and the Mahalanobis distance of the collected near-infrared spectra was calculated.The samples with Mahalanobis distance exceeding the threshold value were set as abnormal samples,so the abnormal near-infrared spectra of jujube were deleted;Continuous projection algorithm is used to select the band in the near infrared spectrum of jujube with abnormal deletion;The results of near-infrared spectral band selection of jujube are used as the input of convolution neural network,and the nondestructive testing results of jujube maturity are output through the calculation of convolution layer and pool layer.The experimental results show that the highest accuracy of nondestructive testing of jujube maturity can reach 99.12%,the highest time of nondestructive testing is 11 s,and the highest error of nondestructive testing is 0.06,which shows that the method in this paper has good accuracy of nondestructive testing and has high technical level and application value.
关 键 词:近红外光谱 大枣成熟度 无损检测方法 连续投影算法 马氏距离 卷积神经网络
分 类 号:TN247[电子电信—物理电子学]
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在链接到云南高校图书馆文献保障联盟下载...
云南高校图书馆联盟文献共享服务平台 版权所有©
您的IP:216.73.216.222