基于近红外光谱的大枣表面损伤检测方法  

Detectionmethod of jujube surface damage based on near infrared spectroscopy

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10期210-216,共7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：为了提升大枣表面损伤检测精度,提出基于近红外光谱的大枣表面损伤检测方法。构建近红外光谱采集系统,采集大枣样本的近红外光谱数据,利用SpectraSuite软件实现光谱曲线的可视化呈现;对光谱数据进行黑白校正处理,提取大枣样本的感兴趣区域,剔除包含大量噪声的波长;采用基于竞争性自适应重加权算法提取大枣样本的光谱特征波长;构建基于SVM的损伤检测模型,实现大枣表面损伤的检测。实验结果表明:该方法可实现大枣表面损伤区域近红外光谱反射率的变化分析,选取的光谱特征波段较为精准,自然条件损伤检测误差率仅为3.32%,人为损伤检测存在1个误识别和1个漏识别样本,提高了检测精度。In order to improve the detection accuracy of jujube surface damage,a detection method of jujube surface damage based on near infrared spectroscopy was proposed.A near-infrared spectrum acquisition system was built to collect the near-infrared spectrum data of jujube samples,and the spectral curve was visualized by spectrasuite software;The black-and-white correction processing is performed on the spectral data to extract the region of interest of the jujube sample and eliminate the wavelength containing a large amount of noise;The competitive adaptive reweighting algorithm is used to extract the spectral characteristic wavelength of jujube samples;A damage detection model based on SVM is constructed to detect the surface damage of jujube.The experimental results show that this method can realize the change analysis of the near-infrared spectral reflectance of the damaged area on the surface of jujube.The selected spectral characteristic band is relatively accurate.The error rate of damage detection under natural conditions is only 3.32%.There is one false identification and one missing identification sample in the artificial damage detection,which improves the detection accuracy.

关 键 词：近红外光谱 表面损伤检测 无损识别 光谱仪 感兴趣区域 特征波段 

分 类 号：TN209[电子电信—物理电子学]