摘要
本文以猕猴桃为研究对象,对猕猴桃采摘过程中目标识别与定位的关键技术进行研究,基于机器视觉技术实现了果实的有效识别和空间的初步定位,从而为猕猴桃果实的采摘提供了必要信息。取得的主要的研究成果和结论如下:
(1)在基于机器视觉技术的猕猴桃果实识别中,为了确定有利于图像分割的颜色空间和颜色因子,对提取的40幅猕猴桃图像在RGB颜色空间中的颜色分量值进行了统计分析,通过对猕猴桃果实、树叶、树枝的感兴趣区域颜色特征的分析,研究表明R-B颜色因子对于猕猴桃果实分割最为有利。
(2)分别采用固定阈值法和自动阈值法对R-B分量图像进行了分割,分割效率分别大于86%和81%。其中固定阈值法具有较好的分割效率和实时性;自动阈值具有较好的分割效果和一定的自适应性,但耗时较多。针对分割处理后的二值图像存在残留物问题,利用图像填充结合形态学处理的方法进行了残留物去除,取得了较好的效果。
(3)根据果实采摘机器人对视觉系统的要求,确定了果实目标的边界、面积、圆形度、形心、外接矩形等作为图像识别特征参数。
(4)研究了基于猕猴桃形心特征的匹配方法,利用二维图像上猕猴桃果实的形心特征,实现了双目果实图像的立体匹配。
(5)在立体匹配的基础上,利用立体视觉原理实现了空间点三维信息的计算,初步确定了猕猴桃果实的空间位置。通过对猕猴桃特征点空间位置信息进行测量,实验表明:该定位方法在实际深度为800mm左右时,空间位置的计算误差为9.03mm。
With Kiwi fruit as the subject, the key technology of fruit recognition and location in developing the kiwifruit picking robot is studied; the machine vision technique is used to collect necessary information for effective recognition and location of kiwifruit. The main research results are as follows:
(1)In order to identify the best color spatial and color mode, based on robot vision technique in Kiwifruit recognition, the color components distilled in RGB color space of 40 pieces of Kiwifruit images are chosen for statistical analysis. With the analyses of color character in kiwifruit, leaves, branches and other important areas, the R–B color mode is the best to segment Kiwifruit;
(2)The fixed and automatic thresholds are adopted respectively to segment the R-B Component images and the segmentation efficiencies are higher than 86% and 81%. After comparison, the fixed threshold method has the better segmentation efficiency and real-time performance, automatic threshold has a better segment effect and adaptability, but it consumes more time. To solve the residual problem usually existing in segmented two value images, flood fill and morphology operation method are applied to eradicate the residual preferably;
(3)To meet the vision requirements of the fruit robot, outline, area, center of gravity and enclosing rectangle are confirmed as parameters for extracting the object's characters for image recognition;
(4)Develop the matching method of the Kiwifruit centroid; use the centroid characters of the fruit in 2D image to match the binocular images of Kiwifruit;
(5)Using the stereo vision principle after stereo matching, calculate the 3D information of spatial points and locate the position of the fruit. With precise dimensional measurements on the position data, the experiments show that calculation error of the special position is 9.03mm when the orientation depth is near 800 mm.
引文
[1]沈明霞,姬长英.农业机器人的开发背景及技术动向[J].农机化研究, 2000 (2): 31-35.
[2] Arima S, Kondo N. Cucumberha rvesting robot and plant training system[J]. Jounral of Robotics and Mechatronics,1999, 11(3): 208-212.
[3] Edan, Miles E. Design of an agricultural robot for harvesting melons[J]. Transactions of the ASAE, 1993, 36(2): 593-403.
[4] Monta, Kondo. Agricultural Robot in Grape Production System[J]. IEEE international Conference on Robotics and Automation, 1995, 2504-2509.
[5] Whittaker AD, Miles GE. Fruit location in a partially occluded image[J]. American Soci of Agricultural Engineers, 1987, 30(3): 591-596.
[6]耿端阳,张铁中,罗辉等.我国农业机械发展趋势分析[J].农业机械学报, 2004,35(4): 208-210.
[7]张瑞合,姬长英,沈明霞等.计算机视觉技术在番茄收获中的应用[J].农业机械学报, 2001, 32(5): 50-58.
[8] Kondo N, Ting. Robotics for bioproduction systems[M]. USA. ASAE publication, 1999.
[9] Hayashi S,Ganno K. Robotic harvesting system for egg plant[J]. JARQ, 2002, 36(3): 163-168.
[10] Guyer DE, Miles M. Machine Vision and Image Processing for Plant Identification[J]. Transactions of the ASAE, 1986, 29(6): 1500-1504.
[11] Gunasekaran S, Cooper, Berlage. Image Processing for Stress Cracks in Conr Kenrels[J]. Trans of the ASAE. 1988, 31(1): 257-263.
[12]方如明.计算机图像处理与米的品质检测[J].农业工程学报, 1992, 8(3): 104-112.
[13]王丰元,周一鸣.种子形状参数检测的计算机图像处理技术[J].农业机械学报, 1995, 26(2): 52-57.
[14] Delwiche M, Tang S. Prune Defect Detection by Line-Scan imaging[J]. Trans of the ASAE. 1990, 33(3): 950-954.
[15] Singh N, Delwiche. Machine Vision Methods for Defect Sorting Stonefruit[C]. Trans of the ASAE. 1994, 37(6): 1989-1997.
[16]何东健,杨青,薛少平.果实缺陷面积的计算机视觉测定研究[J].农业工程学报, 1997, 13(4): 156-160.
[17]刘禾,汪愚华.用计算机图像处理技术进行苹果坏损自动检测的研究[J].农业机械学报, 1998, 9(4): 81-86.
[18]应义斌,景寒松等.机器视觉技术在黄花梨尺寸和果面缺陷检测中的应用[J].农业工程学报, 1999, 15(1): 197-200.
[19]张建平,吴守一,方如明烟叶外观品质特征的定量检验[J].农业工程学报, 1996, 12(3): 158-162.
[20]Slaughter DC, Harrell RC. Discriminating fruit for robotic harvest using color in natural out door scenes[C]. Trans of the ASAE. 1989, 32(2): 757-763.
[21]张树槐,高乔照夫.农业机械学会. 1996, 58(1): 9-16.
[22] Molto E, Juste F. Vision system for the location of citrus fruit in a tree canopy[J]. Journal of Agricultural Engineering Research, 1992, 52: 101-110.
[23] Sites PW, Delwiche M. Computer vision to locate fruit on a tree[J]. Trans of the ASAE. 1988, 31(2): 257-263.
[24] Ness YA. Computer vision system for an orange harvesting robot[M]. Haifa, Israel, 1989.
[25] Whiltaker AD,Miles GE. Fruit locationin apartially occluded image[J]. Trans of the ASAE. 1987, 30(3): 591-596.
[26]孙明,高桥照夫.农业机械学会. 59(4): 53-60.
[27]德田胜,川村恒夫.农业机械学会. 59(4): 29-52.
[28] Jimenez AR, Ceres. Automatic fruit recognition: Asurvey and new results using range-atenuation images[C]. Patenr Recog. 1999, 32: 1719-1736.
[29]徐惠荣,叶尊忠,应义斌.基于彩色信息的树上柑橘识别研究[J].农业工程学报. 2005, 21(5): 98-102.
[30] Bulanon DM, Kataoka T. A Machine Vision System for the Apple Harvesting Robot. A gricultural Engineeiring Intenrational: the CIGR Jounral of Scientific Research and Development. Manuscirpt PM01006. Vol. II.
[31]袁国勇.黄瓜采摘机器人的研究-基于图像的黄瓜果实目标识别[D].中国科技论文在线. 2006.
[32] Woebbecke DM, Meyer GE. a Color In dices for Weed Identification under Various Soil, Residue, and Lighting Conditions[J]. Transacitons of the ASAE. 1995, 38(1): 259-269.
[33] Duke M, Bulanon. A color model for Recognition of apple by a robot charvesting system[J]. Jounral of JSAM. 2002, 64(5): 123-133.
[34] Yael Edan, Tamar Flash. Near-Minimum-Time Task Planning for Fruit-Picking Robots[J]. IEEE Transactions on Robotics and Automation. 1991, 7(1): 48-55.
[35] Pons J, Ceres R. A Tool for Manipulator Dynamics Improvement[J]. Journal of Intelligent And Systems, 1997: 277-287.
[36] Tateshi Fujiura, Kouji Ueda, Suk Hyun Chung. Vision System for Cucumber-harvesting Robot[M]. IFAC Bio-Robotics,Information Technology and Intelligent Control for Bio-Production Systems, Sakai, Osaka, Japan, 2000: 61-65.
[37] Keiok ada. Integration Of Real-Time Bin-ocular Steero Vision And Whole Body In formation For Dynamic Walking Navigation Of Humanoid Robot[J]. IEEEC on ferenceon Multisens or Fusion and Intergration for Intelligent Systems, 2003, 12: 156-161.
[38] Humberto Leaize, Jean Triboulet, Sylvie Lelandais. Christian Barat Match segmenting in stereoscopic vision[J]. IEEE Instrumentation & Measuerment Magazine, 2001, 3: 37-42.
[39]周东翔,蔡宣平,孙茂印.基于灰度段的立体匹配算法[J].软件学报, 12(7): 1101-1107.
[40]陈棣湘,罗飞路,潘孟春.体视觉测量中的图像匹配策略研究[J].光学技术, 2002, 5(28): 392-394.
[41]赵杰,彭京,蔡鹤皋.一种机器人立体视觉匹配方法的研究[J].机械与电子, 2004, 6: 33-37.
[42]郑小东,赵杰文,刘木华.基于双目立体视觉的番茄识别与定位技术[J].计算机工程, 30(22): 171-173
[43]仲琴.基于机器视觉的番茄收获机器人目标定位技术研究[D].硕士学位论, 2005.
[44]龚振邦.从国际大环境出发在我国机器人技术和产业发展战略上的若干思考[J].机器人, 1996, (6): 5-8.
[45]贾云德.机器视觉[M].北京:科学出版社, 2000.
[46] Rafael C, Richard E.数字图像处理[M].阮秋琦,阮宇智等译,北京:电子工业出版社, 2004.
[47]章毓晋.图像处理和分析[M].北京:清华大学出版社, 1999.
[48]陈兵旗,孙明.实用图像处理[M].北京:清华大学出版社, 2004.
[49] Arthur R, Weeks G. Color Segmentation in the HIS color space using the k-means algorithm[J]. SPIE, 1997, 30(26): 43-154,
[50]柏子游,张勇,虞烈.一种彩色图像的色彩分割方法[J].模式识别与人工智能, 1999 ,12(2): 241-244.
[51] Lim YW, Lee SU. On the color image segmentation algorithm based on the thresholding and the fuzzy-Means techniques[J]. Pattern Recognition, 1987, 23(9): 278-282.
[52]何金国,石青云.一种新的聚类方法[J].中国图像图形学报, 2000, (5): 401-405.
[53]黄建忠,李德华.一种室内彩色图像的分割方法[J].华中理工大学学报, 1997, (10): 97-100.
[54] Moghaddamzadeh A, Hourbakis N. A fuzzy growing approach for segmentation of color images[J]. Pattern Recognition, 1997, 30: 867-881.
[55] Tremeau A, Borel N. A region gorwing and merging algorithm to color segmentation[J]. Pattenr Recognition, 1997, 30(7): 1191-1203.
[56]席砺纯,闰宏伟.彩色图像的分割技术[J].微机发展, 2003, 13(4): 46-48.
[57]王爱民,赵忠旭.基于矢量Prewitt算子的多尺度图像边缘检测方法[J].中国图像图形学报, 1999, (12): 912-915.
[58]黄建忠,李德华.彩色边缘模板算子检测法[J].华中理工大学学报, 1998, 26(3): 26-28.
[59]许传样.二进小波变换及其在图像处理和计算机视觉中的应用[M].北京:北京大学, 1995.
[60]林通,石青云.一种基于边缘生长的灰度和彩色图像分割方法[J].中国图像图形学报, 2000, (11): 912-915.
[61] Moghaddamzadeh A, Bourbakis N. A fuzzy-like approach for smoothing and edged detection in color images[J]. Intenrational Jounralof Pattern Recogniiton and Artificial Interligence, 1998, 12(6): 801-806.
[62] Liltmann E, Ritter H. Adaptive color segmentation-a comparison of neural and statistical methods[J]. IEEE Trans on Neural Network, 1997, 8(1): 175-185.
[63] Ercal F. Neural network diagnosis of malignant melanoma from color images[J]. IEEE Trans on Biomed Eng, 1994, 41(9): 837-845.
[64] Sergey N, Krjukov, Tat]ana O. Backpropagation neural network for adaptive color image segmentation[J]. SPIE, 1997, 30(30): 70-74.
[65]粟海华,王睿斌,庄镇泉.一种基于判决神经网络彩色图像分色算法[J].模式识别与人工智能, 1999, 12(3): 54-58.
[66]张星明,李凤森.使用模糊竞争Hopfield网络进行图像分割[J].软件学报, 2000, 11(7): 953-956.
[67]熊联欢,胡汉平.用BP网络进行彩色图像分割和边缘检测[C].华中理工大学学报, 1999, 27(2): 10-14.
[68]赵雪春,戚飞虎.基于彩色分割的车牌自动识别技术[C].上海交通大学学报, 1998, 32(10): 4-9.
[69] Jia P, Evans MD, Ghate SR. Catfish Feature Identification via Computer Vision[J]. Transactions of the ASAE. 1999, 39(5): 1923-1931.
[70] Rafael C , Gonzalez, Richard E.阮秋琦,阮宇智译.数字图像处理(第二版)[M].北京:电子工业出版社, 2003: 537-548.
[71]席旭刚.智能机器人视觉系统研制[D].杭州:杭州电子工业学院, 2002.
[72]张瑞合.基于计算机视觉的空间目标定位系统研究[D].南京:南京农业大学, 2000.
[73]徐奕,周军,周源华.立体视觉匹配技术[J].计算机工程与应用, 2003, (15): 1-6.
[74]李介谷.计算机视觉的理论和实践[M].上海:上海交通大学出版, 1991: 21-30.
[75] Slama C. American Society of Photogrammetry[M], fourth edition, 1980.
[76] Faugeras O. There-Dimensional Computer Vision: a Geometric Viewpoint[M]. MITP ress, 1993.
[77] Brown. DC. Close-range camera calibration[J]. Photogrammetric Engineering, 1971, 37(8): 855-866.
[78] Zhengyou Zhang. A Flexible New Technique for Camera Calibration[M]. Microsoft Research Technical Report. December2, 1998.
[79]高锐.草莓收获机器人初步研究[D].北京:中国农业大学, 2004.
[80] Milan Sonka, Vaclav Hlavac, Roger Boyle.图像处理分析与机器视觉[M].艾海舟,武勃等译第二版.人民邮电出版社, 2003, 309-332.
[81] Z Zhang. A Flexible New Technique for Camera Calibration[J]. IEEE Transactions on Pattenr Analysis and Machine Intelligence, 2000, 22(11): 1330-1334.
[82] Milan Sonka, Vaclav Hlavac, Roger Boyle.图像处理分析与机器视觉[M].艾海舟,武勃等译第二版.人民邮电出版社, 2003, 349-352.
[83] Rafael C, Gonzalez, Richard E.数字图像处理[M].北京:电子工业出版社, 2005.
[84]隋靖,金伟其.双目立体视觉技术的实现及其进展[J].电子技术应用, 2004, (10): 4-6.
[85]潘华,郭戈.立体视觉研究的进展计算机测量与控制[J]. 2004, 12(12): 1121-1124.
[86]刘正东,杨静宇.自适应窗口的时间规整立体匹配算法[J].计算机辅助设计与图形学学报, 2005, 17(2): 291-295.
[87] Kuijper A. On the creations of critical points in scale space with applications to medical image analysis[R]. 2001.
[88] Clifton M, Schor. Unconstrained stereoscopic matching of lines[J]. Vision Research, 2000, 40: 151-162.
[89] Sups BJ, Kariquist WN. Patch-based stereo in a general binocular viewing[J]. Theory and experiment, 1997, 19(3): 247-252.
[90] Reed JN, Tillett RD. Picking mushrooms by robot In Proceedings of the Intenrational Symposium on Automation and Robotic in Bioproduction and Processing[J]. 1995, l: 27-34.