5 3 Vol. 5 No. 3 2014 3 Journal of Food Safety and Quality Mar., 2014 潘磊庆, 屠 * 康 (, 210095) 摘要 : 目的 方法,, Matlab 7.0,, 结果 95.6% 结论 关键词 : ; ; ; ; Evaluation for the length and curvature of garden bean by computer vision PAN Lei-Qing, TU Kang * (College of Food Science and Technology, Nanjing Agricultural University, Nanjing 210095, China) ABSTRACT: Objective To evaluate the length and curvature of garden bean by computer vision. Methods In this paper, garden beans images were obtained by computer vision and processed by using Matlab 7.0. After images processing, the parameters with length and curvature of garden beans were extracted and graded based on computer vision information. Results The evaluation accuracy for garden beans quality could reach 95.6% by computer vision. Conclusion This method showed the feasibility to help the auto grading device development for garden beans. KEY WORDS: computer vision; garden bean; length; curvature; evaluation (Phaseolus vulgaris L.),,,, A C,,, [1],,,,, 3 t, 50 [2], 基金项目 : (31101282) (KYZ201120) Fund: Supported by Natural Science Foundation of China (NSFC, 31101282), Fundamental Research funds for the Central Universities (KYZ201120) and Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD) * 通讯作者 :,,, E-mail: kangtu@njau.edu.cn *Corresponding author: TU Kang, Professor, Doctoral Supervisor, College of Food Science and Technology, Nanjing Agricultural University, No. 1, Weigang Road, Xuanwu District, Nanjing 210095, China. E-mail: kangtu@njau.edu.cn
692 5,,,,,,, [3], [4], [5] [6-9] [10], () [11-12],, 1 材料与方法 1.1 试验材料, 150, 90 1.2 仪器设备 : R6, 480 640 ; : CPU P4 1.7 GHz, 512M, GeForce4 MX440; : 80 cm 80 cm 100 cm,,, ; : ; : Matlab 7.0 [13] 2 图像处理及参数提取 2.1 青刀豆人工分级, [11-12],, 1 1 Fig. 1 Grade of garden bean 2.2 图像采集和处理 2.2.1 图像采集,,,, JPEG, 2.2.2 图像处理 matlab imread, RGB, 2-a : R G B, ( 2b 2c 2d), B graythresh, ; graythresh, im2bw,,,, 1,, 0 3-a, : 3-a,, 0 1 bwmorph clean,,, imclose imopen, 3-b : bwmorph thin,, N=Inf,, 3-c 2.3 特征参数提取 2.3.1 长度的提取 bwlabel,, 3-c,, L 2.3.2 弯曲度的提取 L, 0.1L
3, : 693 0.25L 0.5L 0.75L 0.9L, A B C D E, l 1 l 2 l 3 l 4 ; A E l 4, F : F=l/(l 1 +l 2 +l 3 +l 4 ) (1) Fig. 2 2 RGB Images of garden bean and three-component histogram of RGB Fig. 3 3 Image processing of garden bean
694 5, (F) (3) 4 Fig. 4 Computation of garden bean curvature 3 结果与分析 3.1 长度的检测与分级,,, 445, 380 445, 380, (L) :,,, 2 2, 100%, 90.0%, 94.4%,, 表 2 青刀豆弯曲度分级准确率 Table 2 Grading accuracy of garden bean curvature / / /% 30 27 90.0 30 30 100 (2) 90,,,, 1, 96.7%, 93.3%, 94.4%, 90 85 94.4 3.3 长度和弯曲度综合品质分级 3.3.1 等级标准制定, 3.1 3.2, : Table 1 表 1 青刀豆长度分级准确率 Grading accuracy of garden bean length / / /% 30 29 96.7 90 85 94.4 3.2 弯曲度的检测与分级 4, L, F 3.3.2 综合品质分级验证, 90,,,, (4), 3 3, 96.7%, 93.3%, 95.6% (4)
3, : 695 表 3 青刀豆外观分级准确率 Table 3 Grading accuracy of garden bean external quality / / /% 30 29 96.7 30 29 96.7 90 86 95.6,,, ;,,,, 4 结论 1), R G B, B, 2), : L 445 F 0.95; L 445 0.85 F 0.95 380 L 445 F 0.95; L 380 F 0.85 3), 96.7%, 93.3%, 95.6% 参考文献 [1]. [M]. :, 2002. Zhou XQ. Eating beans processing and utilization [M]. Beijing: Chemical Industry Press, 2002. [2],. ISO22000: 2005 [J]., 2010, 17(1): 48-50. Yao HL, Zhou Y. Application of ISO22000: 2005 in processing of frozen kidney bean [J]. Cereal and Food Ind, 2010, 17(1): 48-50. [3],. [J]. ( ), 2004, 21(5): 497-506. Zhang RY, Liu SS. Researches and applications of computer vision technique in fruit and vegetables. commercialization after harvesting [J]. J Chongqing Technol Business Univ. (Nat Sci Ed), 2004, 21(5): 497-506. [4] Chong VK, Kondo N, Ninomiya K, et al. Feature extraction for eggplant fruit grading system using machine vision [J]. Appl Eng Agric, 2008, 24 (5): 675-684. [5],,. [J], 2009, 35(4): 529-531. Li EC, Li ZM, Liu WR. Image segmentation of cucumbers based on color and textural features [J]. Optical Technique, 2009, 35(4): 529-531. [6],,,. [J], 2013, 27(4): 100-103. Yin R, Gao Y, Meng GF, et al. A method of tomato shape classification with the application of circularity [J]. J Changshu Inst Technol(Natural Sciences), 2013, 27(4): 100-103. [7], [J], 2013, 26(2): 99-100. Liu YF, Tang XH. Research on method of tomato color grading based on machine vision [J]. Dev Innov Mach Electr Prod, 2013, 26(2): 99-100. [8],,. [J]., 2006, 37(2): 215-218. Wang SW, Zhang CL, Fang JL. Study on classification of tomato with bruise using computer vision [J]. J Northeast Agric Univ, 2006, 37(2): 215-218. [9],. [J]., 2011, 27(2): 355-359. Zhao HB, Zhou XH. Recognition of artificial ripening tomato and nature mature tomato based on computer vision [J]. Trans CSAE, 2011, 27(2): 355-359. [10] Kıvanç K, Hakki B, Hamit K, et al. A classification system for beans using computer vision system and artificial neural networks [J]. J Food Eng, 2007, 78: 897 904. [11] DB46/T 82-2007 [S].
696 5 DB46/T 82-2007 Kidney bean [S]. [12] NY/T 1062-2006 [S]. NY/T 1062-2006 Grades and specifications of Chinese kidney beans [S]. [13],. MATLAB 6.X - [M]. :, 2002. Xu D, Wu Z. System analysis and design based on MATLAB X-neural network [M]. Xi an: Xian University of Electronic Science and Technology Press, 2002. ( 责任编辑 : 赵静 ) 作者简介 潘磊庆, 博士, 副教授, 硕士生导师, 主要研究方向为农产品质量检测与控制 E-mail: pan_leiqing@njau.edu.cn 屠康, 博士, 教授, 博士生导师, 主要研究方向为农产品贮藏加工与果蔬采后生理 E-mail: kangtu@njau.edu.cn 粮油产品质量安全 专题征稿 GCIRC FAO/WHO 2014 6 2014 4 15 Email 投稿方式 : www.chinafoodj.com Email jfoodsq@126.com 食品安全质量检测学报 编辑部