油桃外部缺陷的高光谱成像检测

Application of Hyperspectral Imaging for Detection of Defective Features in Nectarine Fruit

采用高光谱（420~1 000 nm）成像技术对＂中油9号＂油桃的4种外部缺陷（裂纹果、锈病果、异形果和暗伤果）进行检测判别。对400个样本（4种外部缺陷样本和完好样本）运用偏最小二乘回归（PLSR）从全波段中分别提取了10条特征波长,分别为497、534、657、677、696、709、745、823、868、943 nm。缺陷样本的高光谱图像经过主成分分析后,对876 nm下的单波段图像通过掩膜、Sobel算子处理,并对主成分图像经过区域生长算法实现缺陷样本的缺陷区域分割。对光谱数据进行主成分分析得到前10个主成分值,并对图像数据采用灰度共生矩阵（GLCM）提取得到6项图像纹理指标（均值、对比度、相关性、能量、同质性、熵值）。将主成分值和纹理值融合建立极限学习机（ELM）模型对油桃外部缺陷进行检测判别。结果表明,该模型对缺陷样本的判别正确率为91.67%,完好样本的正确率为100%。

Hyperspectral imaging,an emerging analytical technology for quality and safety inspections of agricultural and sideline products,combines the advantages of digital image or computer vision with spectroscopy technology in the whole system. Hyperspectral imaging can simultaneously acquire both spatial and spectral information,which deliver chemical,structural and functional information from the samples. In this work,hyperspectral imaging technology was applied to determine a classifier that can be used for nondestructive defection of the defective features in ＂No. 9 of Zhongyou＂nectarine fruit. There were 400 samples from a nectarine planting garden in the Wanan Village in Yuncheng City of Shanxi Province,China,including： crack（ 50）,peel spots（ 50）,malformation（ 50）,hidden damage（ 50） and normal（ 200） samples. Hyperspectral imaging technology covered the range of 420 ~ 1 000 nm was employed to detect defects（ crack,peel spots,malformation and hidden damage） of nectarine fruit. 400 RGB images were acquired through a total of 400 samples,which included four types of defective features and sound samples. After acquiring hyperspectral images of nectarine fruits,the spectral data were extracted from region of interest（ ROI）. Using Kennard-Stone algorithm,all kinds of samples were randomly divided into training set（ 280） and testing set（ 120）. First of all,based on the calculation ofpartial least squares regression（ PLSR）,10 wavelengths at 497 nm,534 nm,657 nm,677 nm,696 nm,709 nm,745 nm,823 nm,868 nm and 943 nm were selected as the optimal sensitive wavelengths（ SWs）,respectively. Subsequently,the image of the 876 nm wavelength was named as the feature image,then principal component analysis（ PCA）,mask process,＂Sobel＂edge detector and ＂region grow＂algorithm were carried out among defective and normal nectarines to extract the defective region.Moreover,ten principal components（ PCs） were selected based on PCA,and seven textural feature variables（ mean,contrast,correlation,energy,homogeneity and entropy） were extracted by using gray level co-occurrence matrix（ GLCM）, respectively. Finally, the ability of hyperspectral imaging technique was tested by using the extreme learning machine（ ELM） models. The ELM classification model was built based on the combination of PCs and textural features. The results show the correct discrimination accuracy of defective samples was 91. 67%,and the correct discrimination accuracy of normal samples was 100%. The research revealed that the hyperspectral imaging technique is a promising tool for detecting defective features in nectarine,which could provide a theoretical reference and basis for designing classification system of fruits in further work.