期刊文献+

基于高光谱成像技术的富士苹果轻微机械损伤检测研究 被引量:8

Detection of Slight Mechanical Damage of Fuji Apple Fruits Based on Hyperspectral Imaging Technology
下载PDF
导出
摘要 为降低苹果损伤造成的商业损失,延长苹果的储存期,利用高光谱成像技术实现了基于特征波段的苹果表面轻微机械损伤的快速、无损检测。以120个富士苹果为研究对象,首先利用波段范围在400~1000nm的高光谱成像光谱仪获取完好和轻微损伤0、2、4 h的富士苹果的高光谱图像,并提取感兴趣区域的平均光谱数据,然后通过两次连续投影法进行分析,去除光谱波段间的冗余信息,找到共线性最小的波段组合(821 nm和940 nm);其次,对特征波段图像进行主成分分析,选择完好与损伤区域差异明显的第二主成分(PC2)作为检测损伤的有效图像;最后,对有效图像进行固定阈值分割和形态学处理,得到苹果表面机械损伤的检测结果。利用该方法对验证组40个正常和轻微损伤不同时间段的苹果进行测试,总体正确率达到94.4%。 To reduce the commercial loss caused by apple fruit damage and prolong the storage period of apple fruits,the hyperspectral imaging technology was used to realize rapid and non-destructive detection of slight mechanical damage on the surface of apple fruits based on feature bands.In the experiment,120 Fuji apple fruits were taken as the research object.Firstly,the hyperspectral images of intact and damaged samples after 0,2 and 4 hours were obtained by the hyperspectral imaging system across the wavelength range of 400~1000 nm.The reflectance of all pixels in the region of interest(ROI)was extracted by ENVI 5.2 software and analyzed by successive projections algorithm(SPA)to remove the redundant information between spectral bands and find the collinear minimum band combination(821 and 940 nm).Then,the PCA was conducted based on two images corresponding to the feature bands,and the second component(PC2)with obvious differences between intact and damaged regions was selected as the effective image for damage detection.Finally,the methods of threshold segmentation and morphological processing were used for the PC2 image to obtain the slightly damaged area on the surface of apple fruits.Using the developed algorithm to detect 40 intact and damaged samples,the average accuracy was 94.4%.
作者 沈宇 房胜 郑纪业 王风云 张琛 李哲 Shen Yu;Fang Sheng;Zheng Jiye;Wang Fengyun;Zhang Chen;Li Zhe(Science and Technology Information Institute,Shandong Academy of Agricultural Sciences,Jinan 250100,China;College of Computer Science and Engineering,Shandong University of Science and Technology,Qingdao 266000,China)
出处 《山东农业科学》 2020年第2期144-150,共7页 Shandong Agricultural Sciences
基金 山东省农业科学院农业科技创新工程项目(CXGC2017B04) 中国农业科学院与山东省农业科学院科技创新工程协同创新任务(CAAS-XTCX2018023)
关键词 高光谱成像技术 苹果 轻微机械损伤 连续投影法 特征波段 图像处理 Hyperspectral imaging technology Apple Slight mechanical damage Successive projections algorithm Feature band Image processing
  • 相关文献

参考文献11

二级参考文献117

共引文献254

同被引文献91

引证文献8

二级引证文献28

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部