用户名: 密码: 验证码:
储粮害虫图像识别中的特征抽取研究
详细信息    本馆镜像全文|  推荐本文 |  |   获取CNKI官网全文
摘要
我国是世界上最大的粮食生产、储藏及消费大国,搞好粮食储藏是关系到国计民生的大事。近年来,我国粮食总储量高达5000亿公斤。为了确保粮食的安全储藏,每年国家用于粮食储备方面的补贴费用就有数百亿元,但仍有不少粮食因管理决策不善等原因而遭受损失,其中,国库储粮损失率在0.2%左右,损失十分惊人,而虫害是主要因素之一。我国《“十五”粮食行业科技发展规划》明确提出要实现粮仓虫害的自动化检测。目前国内外的扦样、声测、近红外等检测方法均不能准确地在线提供粮虫的种类、密度等信息。另外,随着储粮害虫抗药性的提高,它们的种类和密度近年来有上升的趋势,这给粮虫的自动检测提出了更高的要求。因此,开发科学实用、准确方便的储粮害虫在线检测系统是很有必要的,也是极为迫切的。
     利用图像识别的方法在线检测储粮害虫,具有准确度高、价格低廉、效率高、无污染、劳动量小、便于和粮库现有的计算机粮情检测系统相连接等优点,有助于粮库管理人员进行科学的决策,以及时采取合理的防治措施,达到粮食保质、保量、保鲜的目的。若使储粮损失再降低0.05%,每年可为国家挽回经济损失2.5亿元。
     本文在中科院模式识别国家重点实验室开放基金与河南省自然科学基金的资助下,对储粮害虫在线检测系统进行了研究,特别对粮虫特征抽取的三个组成部分:特征形成、特征选择和特征压缩进行了比较深入的研究。特征抽取环节是识别系统的关键,因此,该课题不仅具有重要的学术价值,而且有着广阔的应用前景,可创造可观的社会和经济效益。
     本文主要完成了以下工作:
     1.设计、制作了第三代粮虫取样装置 本工具式装置能定点、多层自动抽取粮食样本,提供均匀恒定的无影光照,保证粮食匀速单层通过,并获取到比较清晰的序列化粮虫图像,为后续的图像处理和识别打下良好的基础。
     2.粮虫图像增强与分割 运用灰度形态学图像平滑和自适应图像增强对粮虫灰度图像进行增强处理。用直方图高斯阈值法和相对熵阈值法提取粮虫图像的最优阈值,并进行分割以形成适于后续处理的二值化图像。对多种分割方法
    
     郑什】大学硕士学位论文 摘要
    比较分析,得到最优性能的分割方法。
    3.粮虫特征形成 提取出粮虫H值化图像的面积、圆形性、不变矩等17个
    形态学特征。针对粮虫的灰度图像,提取出基于灰度直方图、游程长度和灰度
    共生矩阵的27个纹理特征,并对提取的形态学特征进行归一化处理。
    4.粮虫特征选择 对模拟退火算法和遗传算法两种组合优化方法的算法提
    出、实现步骤、参数分析、具体实现进行了深入的探讨。对抽取出的44个特征
    进行分析,初步筛选掉可用于离线分析的纹理特征,并从17个原始粮虫形态学
    特征中,选择出面积、复杂度、等效圆半径等10个适于分类的特征。
    5.粮虫特征压缩 运用总体类内离散度矩阵K-L变换的特征压缩,包含在
    类平均向量中判别信息的最优特征压缩,基于距离可分性准则的特征压缩三种
    方法,将粮虫的10维特征向量压缩到5维,以提高识别系统的整体性能。
    6.12种9类粮虫识别分类 针对粮仓中危害严重的大谷盗、米象、谷蠢、
    锯谷盗、黑菌虫、长头谷盗等12种9类粮虫,根据135幅模拟现场的粮虫图像,
    建立九类粮虫均值、方差模板库,及相应的隶属函数,在模糊极大极小原则的
    基础上进行识别归类。以90幅粮虫图像进行识别检验,其识别率在95%以上。
    7.算法比较分析 分析比较图像增强、图像分割、特征选择、特征压缩等
    环节的多种算法,通过比较发现:灰度形态学图像平滑、相对嫡阈值法图像分
    割、模拟退火算法选择特征、基于距离可分性准则的特征压缩等,为各环节效
    果较好的算法,并将它们作为现场应用的识别系统的最终算法。
    8.在线检测系统实现 利用VisualC++6.0开发的粮虫识别系统软件包,
    与研制的取样装置相配合,能以86.5%的正确率在线识别出粮仓中危害严重的9
    类粮虫,为整套系统的产品化奠定了良好的基础。
     文中所设计的储粮害虫在线检测系统,在郑州、民权等国家粮食储备库进
    行了现场试验,得到了粮食储藏方面专家的好评。在第七届全国大学生“挑战
    杯”、河南省第一届大学生“挑战杯”、河南省大学生科技活动日等活动中,得
    到有关院士、专家学者的认可,并分别获得三等奖、一等奖、银奖等。由于时
    间和水平有限,还需要在粮虫的获取手段、藏匿于粮粒中的幼虫识别、种类扩
    展等方面进一步改进、完善和提高,以进一步提高系统的性能。
China is the large country of grain production, storage and consumption in the world. Doing well stored-grain management is a very important thing about the national economy and the people's livelihood. In recent years, the stored grain is more than 500 billions of kilograms in our country. The center government offers billions of RMB to the grain deports in order to managing storage well. But plenty of storage still was attainted because of the ill management. The stored-grain loss is very serious calculated by the current 0.2% loss ratio in the national grain deports. The stored-grain insect pest is one of the important facts. "The Fifth Grain Trade Science and Technology Development Programming " put forward definitely to realize automatic detection about stored-grain insect pests. The ways of the Sampling, the Sound Detecting, the Near Infrared and others in the word can't supply well and truly the grain pests' category, dense and other parameters. In addition, with the increase of the stored-grain pes
    ts' drug fastness, their category and dense are increasing in recent years. As a result, developing a kind of scientific, precision, simple detection technology for stored-grain pests is very necessary and imperious.
    There are a series of advantages making use of the image recognition technology to detect the stored-grain pests, such as high precise, low price, high efficiency, no pollution, less labor, convenient connection with the computer grain detection in grain deports, and so on. It can help the grain managers to make scientific decisions; in order to they can take rational prevention-measures in time, the storage can be managed in quality, quantity and greenness. If the storage loss ratio can decrease 0.05%, it may retrieve the lost 2500 millions RMB for China every year.
    This paper is supported by the Opening Foundation of National Laboratory of Pattern Recognition, CAS and Natural Science Foundation of Henan Province. It researched the stored-grain pests on-line detection system using of image recognition technology. The feature extraction composed of the feature forming, feature selecting and feature compressing was studied more profoundly. The aspect
    
    
    of feature extraction is the key of the recognition system, therefore, there is not only the important science value but also extensive utility foreground. It can create the considerable society and economical benefit. The main work is'as follows:
    1. Designing and making the third grain-pests sampling device. The implemental device can sample the grain automatically in the settled dot and the fixed layer, offer equality and invariable shadowless lamp, ensure the grain monolayer passing, obtain clear grain-pests image. It can settle a good foundation for the rear image enhancement and recognition.
    2. Enhancing and segmenting grain-pests image. The grain-pests image is smoothed by the gray morphology, enhanced by the adaptive method, segmented by the optimization thresholding offered by the histogram gauss method and the relative entropy method. Obtaining the best way in the familiar segmentation algorithms through the in-depth analysis and compare.
    3. Forming the stored-grain pests' features. Extracting 17 morphological features to stored-grain pests binary image, for example area, circle quality, invariable moment and so on. Extracting 27 texture features to stored-grain pests ' gray image. The morphological features extracted are normalized.
    4. Selecting the grain-pest features. The simulated annealing algorithm and genetic algorithm are studied in-depth. The two kinds of optimization algorithms' ideal offering, realization approach, parameter analysis and material realization. Through the compare of the 44 stored-grain pests' features, the texture features used in out-line analysis are taken out. They are used to select the 17 features normalized. There are 10 better features, for example, area, complexity, equivalent circle radius, and so on.
    5. Compressing the grain-pest featu
引文
[1]国家粮食局.“十五”粮食行业科技发展规划.2001.1.
    [2]万拯群.当前我国科学保粮问题之我见.粮食储藏,1999,28(1):25-29.
    [3]Wilkin D R,Fleurat-Lessard F.The detection of insects in grain using conventional samplings spears.Proc. 5 th Int. Wkg.Conf.on Stored-Product Protection,Bordeaux,France,1990:1445.
    [4]Vick,K.W.,Webb,J.C.,Weaver, B.A., et al.Sound detection of stored-product insects that feed inside kernels of grain.Econ.Entomol.1988,81:1489-1493.
    [5]Chambers,J.,Cowe I.A.,Van Wyk C.B.,et al. NIR analysis for the detection of insect pests in cereal grains.Proc. Int. Conf. on Diffuse Spectroscopy,MD USA,1992:96-100.
    [6]徐昉,邱道尹,沈宪章等.粮仓害虫的特征提取与分类研究.郑州工业大学学报,2000,21(4):62-65.
    [7]徐昉,邱道尹,白旭光等.图像识别在粮仓害虫检测方面的应用研究.郑州工程学院学报,2001,22(1):78-81.
    [8]徐防.基于图像识别的储粮害虫检测系统研究[硕士学位论文].郑州大学,2001.
    [9]Zayas,I.Y.,Flinn, P.W. Detection of insects in bulk wheat samples with machine vision.Trans. ASAE.,1998(3):883-888.
    [10]邱道尹,张红涛,陈铁军等.基于机器视觉的储粮害虫智能检测硬件系统.农业机械学报(已录用).
    [11]邱道尹,张红涛,陈铁军等.基于机器视觉的储粮害虫智能检测软件系统.农业机械学报(已录用).
    [12]章毓晋.图像工程(上册)——图像处理和分析.北京:清华大学出版社,2000.
    [13]吴一全,王厚枢.图像对比度增强处理方法(二).数据采集与处理,1999,5(1):37-50.
    [14]沈庭芝,方子文.数字图像处理与模式识别.北京,北京理工大学出版社,1998.
    [15]戴君.数学形态学在图像处理中的应用.佛山科学技术学院(自然科学版),1998,16(2):29-33.
    [16]陶德元,何小海,李舒平.两种有效的图像增强算法及其应用.数据采集与处理,1991,6(4):29-33.
    [17]吴一全,朱兆达.图像处理中阈值选取方法30年(1962—1992)(一).数据采集与处理,1993,8(3):193-201.
    [18]吴一全,朱兆达.图像处理中阈值选取方法30年(1962-1992)(二).数据采集与处理,1993,8(4):268-278.
    
    
    [19]王俊杰,黄心汉.一种对图像进行快速二值化处理的方法.电子技术应用,1998,(10):16-17.
    [20]Pal,N.K.,Pal, S.K. Entropy:a new definition and its application.IEEE Trans Syst Man Cybern,1991,21:1260-1270.
    [21]王建军,苑玮绮,张宏勋,一种基于相对熵的图象分割算法.信息与控制,1997,26(1):67-72.
    [22]黄凤岗,宋克欧.模式识别.哈尔滨:哈尔滨工程大学出版社,1998.
    [23]张远鹏,董海,周文灵.计算机图象处理技术基础.北京:北京大学出版社,1996.
    [24]Lu,C.S.,Cheng,P.C.,Chen C.F.Unsupervised texture segmentation via wavelet transform.Pattern Recognition,1996,30 (5):729-742.
    [25]崔屹.数字图像处理技术与应用.北京:电子工业出版社,1997.
    [26]邱道尹,张成花,张红涛等,神经网络在储粮害虫识别中的应用,农业工程学报(已录用)
    [27]边肇祺,张学工.模式识别(第二版).北京:清华大学出版社,2000.
    [28]章新华.一种特征选择的动态规划方法.自动化学报,1998,24(5):675-680.
    [29]向阳,龚新高.推广模拟退火方法及其应用.物理学进展,2000,20(3):319-334.
    [30]Sidelecki, W.,Slansky, J.A note on genetic algorithms for large-scale feature selection.1989,10 (5):335-347.
    [31]Roberto,B. Using mutual information for selection features in supervised neural net learning.IEEE Trans.on NN,1994,5(4):537-550.
    [32]Kirkpatrick,S.,Gellatt,C.D.,Vecchi, M.P.Optimization by simulated annealing.Science,1993, 220:671-680.
    [33]Gelfand, S.B.Mitter, S.K.Metropolis type annealing algorithm for global optimization in Rd.SIAM J.Control and Optim.,1993(1):111-131.
    [34]王丰,王兴泰,于万瑞.改进的模拟退火方法及其在电阻率图像重建中的应用.长春科技大学学报,1999,20(2):175-178.
    [35]Kyung, C.M.,Widder, J.,Mlynsk, D.A.Adaptive cluster growth:a new algorithm for circuit placement in. rectilinearregions.Computer Aided Design,1992,24(24):27-35.
    [36]杨若黎,顾基发.一种高效的模拟退火全局优化算法.系统工程理论与实践,1997(5):29-35.
    [37]李文勇,李泉永.基于模拟退火的全局优化算法.桂林电子工业学院学报,2001,21(2):33-37.
    [38]王小睿,吴信才,李军.模拟退火算法的改进策略在模板匹配中的应用.小型微型计算机系统,1997,18(8):32-37.
    
    
    [39]姚俊峰,梅炽,彭小奇等.混沌遗传算法及其应用.系统工程,2001,19(1):70-74.
    [40]段玉倩,贺家李.遗传算法及其改进.电力系统及其自动化学报,1998,10(1):39-52.
    [41]席欲庚,柴天佑,恽为民.遗传算法综述.控制理论与应用,1996,13(6):697-708.
    [42]Larranaga,P.,Kuijpers, C.M.,Murga,R.H.,et al.Learning Bagesian Network Structures by searching for the Best Ordering with Genetic Algorithms.IEEE trans on System,Man and Cybernetics-Part A:System and Human,1996,26(4):487-493.
    [43]Suzuki,J.A further result on the Markov chain model of genetic algorithms and its application to a simulated Annealing-Like strategy.IEEE Trans.on Systems,Man and Cybernetics-part B:Cybernetics,1998,28(1):95-102.
    [44]陈建安,郭大伟,徐乃平等.遗传算法理论研究综述.西安电子科技大学学报,1998,25(3):363—368.
    [45]潘美芹,贺国平,马学强.基于改进遗传算法的多维函数的优化计算.山东科技大学学报(自然科学版),2000,19(2):14-17.
    [46]石玉,陈小平,于盛林.利用排序对遗传算法的改进和自适应交叉概率.数据采集与处理,2000,15(2):185-190.
    [47]张晓绩,戴冠中,徐乃平.一种新的优化搜索算法——遗传算法.控制理论与应用,1995,12(3):265-273.
    [48]任平.遗传算法(综述).工程数学学报,1999,16(1):1-8.
    [49]Srinivas, M.,Patnaik,L.M.Adaptive Probabilities of Crossover and Mutation in Genetic Algorithms.IEEE Trans..Syst.Man and Cybernetics.1994, 24 (4):656-667.
    [50]Srinivas, M.,Pataik,L.M.Genetic algothrims: A survey.IEEE Trans. on Computer,1994,27(6):17-26.
    [51]张学良,黄玉美.遗传算法及其在机械工程中的应用.机械科学与技术,1997,16(1):47-52.
    [52]聂勋碧,施继承,王山山等.多参数油气预测系统.石油地球物理勘探,1997,32(3):357-364.
    [53]Yamashita,Y.,Ogawa, H.Relative Karhunen-Loeve transform.IEEE Tans.on Signal Processing,1996,44(2):371--378.
    [54]Hua,Y.Liu, W. Generalized Karhunen-Loeve transform.IEEE Signal Processing Letters,1998,5(6):141-142.
    [55]江铭虎,季文铎,林碧琴等.一种特征压缩及分类神经网络的研究.电路于系统学报,1997,2(3):18-23.
    
    
    [56]唐向宏,贺振华,杨绍国.利用小波变换进行地震数据压缩.成都理工学院学报,1999,26(2):183-186.
    [57]张永胜,郁可.基于小波神经网络的光谱数据压缩与分类研究.计算机研究与发展,1999,36(8):973-977.
    [58]Kung,S.Y.Adaptive principal component extraction(APEX) and application.IEEE Trans.Signal Processing,1994, 42 (5):1202-1217.
    [59]Defries, R.S.,Townshend, J.R.G. NDVI-derived L and Cover Classification ata Global Scale.International Journal of Re-mote Sensing,1994,15(17):3567-3586.
    [60]Running,S.W.,Loveland,T.R.,Pierce,L.L.,et al.A Remote Sensing Based Vegetation Classification Logic for Global LandCover Analysis.Remote Sensing of Environment, 1995, 51: 39-48.
    [61]周骏,曲云尧,周文涛等.煤与瓦斯突出模式识别预测软件的设计原理.山东矿业学院学报,1996,15(1):61-66.
    [62]田村秀行[日].计算机图像处理技术.北京;北京师范大学出版社,1986.
    [63]邵才瑞,李洪奇,张福明.句法模式识别及其在地层对比中的应用.中国海上油气(地质),2000,14(5):363-366.
    [64]王时标,姚振兴.数域上的模糊模式识别方法.模糊系统与数学,1995,9(4):32-38.
    [65]金连文,徐秉铮.基于多级神经网络结构的手写体汉字识别.通信学报,1997,18(5):21-27.
    [66]王耀南.卫星遥感图像的神经网络自动识别分类.湖南大学学报,1998,25(4):61-66.
    [67]张学工.关于统计学习理论与支持向量机,自动化学报,2000,26(1):32-42.
    [68]Scholkopf, B.,Sung, K.K.,Burges, C., et al.Comparing support vector machines with Gaussian kernels to radial basis function classifiers.IEEE Trans.on Signal Processing,1997,45(11):2758-2765.
    [69]文贡坚,王润生.基于模糊决策的快速识别多类目标的方法.模式识别与人工智能,1997,10(2):106—111.
    [70]邱道尹,张红涛,陈铁军等.模糊识别技术在储粮害虫检测中的应用.农业系统科学与综合研究(已录用).

© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号

地址:北京市海淀区学院路29号 邮编:100083

电话:办公室:(+86 10)66554848;文献借阅、咨询服务、科技查新:66554700