一种FFT的Hu矩的番茄形状特征提取

详细信息    查看全文 | 推荐本文 |

英文篇名：An Improved Hu Moment of Tomato Shape Feature Extraction Based on FFT 作者：王建潭 ; 木合塔尔·克力木 英文作者：WANG Jian-tan;MU He-tarer;School of Mechanical Engineering,Xinjiang University; 关键词：番茄 ; 机器视觉 ; 特征提取 ; 形状 ; 快速傅里叶变换 英文关键词：Tomato;;Machine Vision;;Feature Extraction;;Shape;;FFT 中文刊名：JSYZ 英文刊名：Machinery Design & Manufacture 机构：新疆大学机械工程学院; 出版日期：2019-05-08 出版单位：机械设计与制造 年：2019 期：No.339 基金：国家自然科学基金资助项目(51365052) 语种：中文; 页：JSYZ201905057 页数：4 CN：05 ISSN：21-1140/TH 分类号：235-238

摘要

目前基于机器视觉的番茄形状检测算子检测到的形状特征单一,相关的深入研究较少,为进一步探索合适的形状特征参数及检测算子,将Hu矩作为特征参数并利用支持向量机实现了番茄的形状检测分类,准确度较高。首先设计采用Laplacian算子与Sobel算子进行算法融合对采集到的番茄图像进行自适应增强处理;然后采用最小错误率贝叶斯决策算子对平番茄图像进行图像分割得到番茄目标二值化图像。然后对得到二值化后的番茄图像进行归一化处理,使得归一化后的图像具有平移、旋转、尺度缩放不变性的Hu参数,通过提取目标番茄的7个Hu矩特征值参数,最后将这些特征值输入支持向量机中,完成番茄无损分级检测。试验采用VS2010验证算法,对正常果形、轻度畸变果形、重度畸变果形共计1000个样本进行了训练测试,结果表明经过改进增强图像以及改进Hu矩算子对番茄正常果形的分级精度达到93.3%,符合实际番茄检测精度的要求。
        The tomato shape detection operator detection based on machine vision to the shape feature of a single,related research is less,in order to further explore the appropriate shape characteristic parameters and testing of the operator,this paper will be Hu moment as characteristic parameters and USES support vector machine(SVM)for detecting the shape of the tomato classification,accuracy is higher. First using Laplacian operator with Sobel operator to fusion algorithm for adaptive enhancement processing tomato images were collected;Then the minimum error bayes decision operator is adopted to flat tomato image get tomato target image binarization image segmentation. Then the screen image after binarization of tomato normalized processing,so that when the normalized image translation,rotation,scale and scaling invariance Hu parameters,by extracting the target of tomato of 7 Hu moment feature parameters,finally the characteristic values input support vector machine,the complete tomatoes nondestructive detection classification. Test the VS2010 verification algorithm,is adopted to normal shape,mild distortion shape,severe distortion shape and has trained a total of 100 samples tested,the results show that the improved to enhance the image and Hu moment operator on tomato normal shape classification accuracy of 93.3%,in line with the requirements of actual detection.

引文

[1]Tao Y,C.T.Morrow,P.H.Heinemann.Fourier-based separation technique for shape grading of potatoes using machine vision[J].Transactions of The ASAE,1995,38(3):949-957.
    [2]Heinemann,P.H,Varghese,ZA.Machine vision inspection of golden delicious apples[J].Transactions of The ASAE,1995,11(6):901-906.
    [3]Laykin,S,V.Alchanatis,E.Fallik,Y.Edan[J].Image-processing algorithm for tomato classifications[J].Transactions of The ASAE,2002,45(3):851-858.
    [4]李长勇,曹其新.基于深度图像的蔬果形状特征提取[J].农业机械学报,2012(S1):242-245.(Li Chang-yong,Cao Qi-xin.Based on depth image of fruits and vegetables shape feature extraction[J].Journal of Agricultural Machinery,2012(S1):242-245.)
    [5]郝敏,麻硕士,郝小冬.基于Zernike矩的马铃薯薯形检测[J].农业工程学报,2010(2):347-350.(Hao Min,Ma Shuo-shi,Hao Xiao-dong.Potato shape detection based on zernike moments[J].Transactions’of The Chinese Society of Agricultural Engineering,2010(2):347-350.)
    [6]杜永忠,平雪良,何佳唯.圣女果表面缺陷检测与分级系统研究[J].农业机械学报,2013(S1):194-199.(Du Yong-zhong,Ping Xue-liang,He Jia-wei.Cherry tomatoes surface defect detection and classification system research[J].Journal of Agricultural Machinery,2013(S1):194-199.)
    [7]李先锋,朱伟兴,纪滨.基于图像处理和蚁群优化的形状特征选择与杂草识别[J].农业工程学报,2010(10):178-182.(Li Xian-feng,Zhu Wei-xing,Ji Bin.Based on image processing and the shape of the ant colony optimization feature selection and weed identification[J].Journal of Agricultural Engineering,2010(10):178-182.)
    [8]Zunic,J.Hirota,K.Rosin,P.L.A Hu moment invariant as a shape circularity measure[J].Pattern Recognition,2010,43(1):47-57.
    [9]Otsu.A threshold selection method from gray-level histogram[J].IEEETrans on SMC29,1979,9(3):62-66.
    [10]Khotanzad A,Hongs Y H.Invariant image recognition by zernike moments[J].IEEE Trans.on Pattern Analysis and Machine Intelligence,1990,12(5):489-497.
    [11]Y.Ohta,T.Kanade.Takeo,T.Sakai.Color information for region segmentation[J].Computer Graphics and Image Processing,1980,13(3):222-241.
    [12]Li D.Analysis of moment invariants on image scaling and rotation[J].Innovations in Computing Sciences and Software Engineering,2010,1(1):415-419.
    [13]Hu M K.Visual pattern recognition by moment invariants[J].IEEE Transactions on Information Theory,1962,8(1):179-187.
    [14]公丕亮,张奇峰,张艾群.基于形态图的UVMS水下三维目标识别[J].机械设计与制造,2011(3):265-267.(Gong Pi-liang,Zhang Ji-feng,Zhang Ai-qun.UVMS underwater 3d object recognition based on shape figure[J].Journal of Mechanical Design and Manufacturing,2011(3):265-267)