【正文】 級(jí)別特征的倒角和圓角。另一個(gè)缺點(diǎn)是，該系統(tǒng)可以與單個(gè)調(diào)校零件工作。如果基于規(guī)則的模式識(shí)別的系統(tǒng)具有未來(lái)的話，那么它應(yīng)該是混合方法。他們將構(gòu)成的常規(guī)方法優(yōu)勢(shì)結(jié)合起來(lái)并有許多成功應(yīng)用的例子，他們主要針對(duì)只適用于有限類(lèi)型的特征。更廣泛的特征的類(lèi)別成功的混合算法概括的缺少展示了這套自動(dòng)化特征識(shí)別方法的主要缺點(diǎn)。4. 總結(jié)AFR是在將CAD信息翻譯到一些適合于制造的某些指令過(guò)程中是第一步也是最重要的一步。對(duì)于一個(gè)完全自動(dòng)化的CAPP系統(tǒng)的開(kāi)發(fā)，消除對(duì)人類(lèi)在特征識(shí)別過(guò)程中的參與是必不可少的。和基于設(shè)計(jì)的特征相比，AFR的優(yōu)點(diǎn)是顯著的時(shí)間和人力資源的節(jié)省，以及所希望的零件的功能的保證，預(yù)定形狀特征庫(kù)的可能性的設(shè)計(jì)創(chuàng)意并沒(méi)有被限制。然而，盡管進(jìn)行了大量的研究，當(dāng)代AFR系統(tǒng)仍然很多缺點(diǎn)[64]：（i）識(shí)別算法的復(fù)雜性，尤其是在交叉特征的情況下，在特征集的確定上有一個(gè)巨大的問(wèn)題，對(duì)于這些產(chǎn)品，必須明確計(jì)劃，這些特點(diǎn)可以在識(shí)別其他特征的時(shí)候同時(shí)被識(shí)別出來(lái)。（ii）可識(shí)別的特征域是相當(dāng)有限的，大多數(shù)現(xiàn)代的識(shí)別系統(tǒng)主要是處理正交特征，但很少關(guān)注非正交和隨機(jī)特征。（ⅲ）所識(shí)別的特征附加的制造信息是不足夠豐富以方便后續(xù)生產(chǎn)計(jì)劃等的決定。有一些我們?cè)诳偨Y(jié)中要強(qiáng)調(diào)的基于規(guī)則的模式識(shí)別的AFR技術(shù)現(xiàn)狀的一些補(bǔ)充說(shuō)明：1. 零件特征模型的多種解釋的問(wèn)題沒(méi)有通用的解決方案。2. 大部分研究是在加工領(lǐng)域做的，雖然最近在其他生產(chǎn)領(lǐng)域有或多或少的成功實(shí)施AFR技術(shù)的報(bào)告，如鈑金沖壓，沖壓成型，成型及鑄造。3. 為了使在下游CAPP模塊的識(shí)別特征得到有效應(yīng)用，或者在檢查和裝配工藝規(guī)劃中的特征識(shí)別，AFR系統(tǒng)應(yīng)該能夠識(shí)別尺寸和連同其它要素屬性的幾何公差的。4. 雖然在這一領(lǐng)域的理論研究已經(jīng)投入了巨大的努力，但很少有實(shí)際應(yīng)用在商業(yè)軟件的證據(jù)，如KBM[65]或FeatureCAM[66]。 以規(guī)則為基礎(chǔ)的模式識(shí)別的當(dāng)代AFR系統(tǒng)的三大仍未解決的問(wèn)題： 。吳和劉[67]用一個(gè)嚴(yán)謹(jǐn)?shù)姆椒▽?duì)克服這些問(wèn)題的現(xiàn)有技術(shù)進(jìn)行了分類(lèi)，并分為了5組：（i）特征的通用表示。（二）伴有學(xué)習(xí)能力的識(shí)別器。（三）缺少信息的恢復(fù)。 （iv）與Brep混合的以特征為基礎(chǔ)的系統(tǒng)。（五）其他方面。然而，從所有如上所述的原因來(lái)看，提供完全自動(dòng)化的特征識(shí)別方法的整體算法仍然不存在。我們相信，在AFR領(lǐng)域的新研究中，我們應(yīng)著眼于通過(guò)人工神經(jīng)網(wǎng)絡(luò)等先進(jìn)的人工智能技術(shù)的實(shí)施尋找從形狀模式庫(kù)提取CAD模型和知識(shí)獲取的必要的幾何信息的更高效的算法。鳴謝所提出的研究工作得到了塞爾維亞共和國(guó)科學(xué)與環(huán)境保護(hù)部在6319項(xiàng)目名為“在金屬加工行業(yè)的制造系統(tǒng)和制造工藝的自動(dòng)化設(shè)計(jì)的實(shí)施”的框架的經(jīng)濟(jì)支持。參考文獻(xiàn)[1] M. Marquez, A. White, . Gill, Hybrid neural network – feature based manufacturability analysis of mould reinforced plastic parts, in: Proceedings of the Institution of Mechanical Engineers, vol. 215, Part B, 2001, pp. 1065–1079.[2] . Chang, Expert Process Planning for Manufacturing, AddisonWesleyPublishing Company, 1990.[3] . Zhang, L. Alting, Computerized Manufacturing Process PlanningSystems, Chapman amp。 Hall, 1994.[4] S. Subrahmanyam, M. Wozny, An Overview of Automatic Feature Recognition Techniques for ComputerAided Process Planning, Computersin Industry 26 (1995) 1–21.[5] . Tseng, . Joshi, Recognition of interacting rotational and prismatic machining features from 3D millturn parts, International Journal of Production Research 36 (11) (1998) 3147–3165.[6] . Han, M. Pratt, . Regli, Manufacturing feature recognition fromsolid models: a status report, IEEE Transactions on Robotics and Automation 16 (6) (2000) 782–796.[7] O. Owodunni, S. Hinduja, Evaluation of existing and new feature recognition algorithms. Part 1. Theory and implementation, Proceedings of the Institution of Mechanical Engineers 216 (Part B) (2002) 839–851.[8] . Miao, N. Sridharan, . Shah, CADCAM integration using machining features, International Journal of Computer Integrated Manufacturing 15 (4) (2002) 296–318.[9] B. Babic180。, Development of an intelligent CADCAPP interface, in: Proceedings of the International Conference on Intelligent Technologies in HumanRelated Sciences, 1996, pp. 351–357.[10] ISO 10303, parts 1, 11, 21, 42, 224, 324, ISO, Geneva。 1994–2000.[11] . Bhandarkar, R. Nagi, STEPbased feature extraction from STEPgeometry for agile manufacturing, Computers in Industry 41 (2000) 3–24.[12] R. Sharma, . Ghao, Implementation of STEP application protocol 224 in automated manufacturing planning system, Proceedings of the Institution of Mechanical Engineers 216 (Part B) (2002) 1277–1289.[13] J. Gao, . Zheng, N. Gindy, D. Clark, Extraction/conversion of geometric dimensions and tolerances for machining features,International Journal of Advanced Manufacturing Technology (2004) (Online).[14] J. Gao, . Zheng, N. Gindy, Extraction of machining features for CAD/CAM integration, International Journal of Advanced ManufacturingTechnology 24 (2004) 573–581.[15] . Jain, S. Kumar, Automatic feature extraction in PRIZCAPP, International Journal of Computer Integrated Manufacturing 11 (6) (1998)500–512.[16] N. Ismail, N. Abu Bakar, . Juri, Feature recognition patterns for form features using boundary representation models, International Journal of Advanced Manufacturing Technology 20 (2002) 553–556.[17] N. Ismail, N. Abu Bakar, . Juri, Recognition of cylindrical and conical features using edge boundary classification, International Journal of Machine Tools and Manufacture Design, Research and Application 45 (2005) 649–655.[18] V. Milacˇic180。, M. Urosˇevic180。, A. Veljovic180。, A. Miler, Race I. SAPT—xpert system based on hybrid concept of group technology, in: 19th CIRPInternational Seminar on Manufacturing Systems, Pennsylvania StateUniversity, USA, 1987.[19] . Henderson, . Anderson, Computer recognition and extraction of form features: a CAD/CAPP link, Computers in Industry 5 (1984) 329–339.[20] B. Babic180。, Z. Miljkovic180。, Feature recognition as the basis for integration of CAD and CAPP Systems, in: Proceedings of the Second WorldCongress on Intelligent Manufacturing Processes and Systems, 1997,pp. 596–01.[21] . Bouzakis, G. Andreadis, A featurebased algorithm for puter aided process planning for prismatic parts, International Journal of Production Engineering and Computers 3 (3) (2000) 17–2.[22] M. Sadaiah, . Yadav, . Mohanram, P. Radhakrishnan, A generativeCAPP system for prismatic ponents, International Journal ofAdvanced Manufacturing Technology 20 (2002) 709–19.[23] . Gavankar, . Henderson, Graphbased extraction of twoconnected morphological features from boundary representations, Journal of Intelligent Manufacturing 6 (1994) 401–13.[24] . Venuvinod, . Wong, A graphbased expert system approach to geometric feature recognition, Journal of Intelligent Manufacturing 6 (1994) 401–413.[25] . Yuen, . Venuvinod, Geometric feature recognition: coping with the plexity and infinite variety of features, International Journal of Computer Integrated Manufacturing 12 (5) (1999) 439–452.[26] . Yuen, . Wong, . Venuvinod, Development of a generic puteraided process planning support system, Journal of Materials processing Technology 139 (2003) 394–401.[27] M. Marefat, . Kashyap, Geometric reasoning for recognition of threedimensional object features, IEEE Transactions on Pattern Analysis and Machine Intelligence 12 (10) (1990) 949–965.[28] Q. Ji, . Marefat, A DempsterShafer approach for recognizing machine features from CAD models,