摘要
抽取了多个焊接金相组织的特征参数，设计了最小距离法分类器，建立了焊接金相组织识别系统，较好地实现了碳钢及低合金钢焊接金相组织的自动定量识别．通过实际评价表明，该系统对焊接接头组织识别的准确性和再现性符合定量金相的要求。这对于建立焊接金相组织和性能间的定量关系，选择合理的工艺参数，建立焊接ＣＡＤ技术具有重要的理论和实际意义。
Several feature parametres of the microstructures in welded joints is extracted, aminimum-distance-classifier is designed a microstructure recognition system for welded jointsis set up,thus the automatically puatitative recognition of the microstructures iswell reali-sed in the welded joints made of the mild steels and the low-alloy steels.The practical ev-aluation shows that the accuracy and reappearance of this system satisfies the repuiremets of the puantitative metallography for the recognition of microstructures in the welded jooints .This system is of great importance to theory and practice for the establishment of quantitat-ive relationship between the microstructures of welding and the mechanical properties of w-elded joints and for selecting the reasoaboe technological parametres,as well as rof creat-ing CAD system for welding.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
1996年第5期94-98,共5页
Journal of Mechanical Engineering
关键词
焊接接头
金相组织
特征抽取
图象识别
Welded joints
Microstructure
Feature extraction
Image recognition
