Abstract：The approach includes three parts:(1) image information processing, (2) body primitives separating,(3) 3-D representation of engineering object. Part (1) symbolizes the image in figural form of loops of linesand curves suitable for higher level processing. Part (2) separates all body primitives which compose theobject in accordance with special expert rules in the field of engineering drawing and relations among itsthree views by heuristic inferring method. Part (3) represents the whole body on a two-level model net.This approach has been succesfully used by assembly robot in block world.
1 迁三郎,江况正员,ロボツト工学ごの应用 2 Wesley M A, Markowsky G. Fleshing out projections. 1BM J Rcs Develop 25:6, 1981 3 Aldefcld B. On automatic recognition of 3D structures from 2D representations. CAD, 1983;15 (2) 4 王尚华,周远清.许万雍.图形数据的一种自动输人方法全国高校第九届人工智能学术讨论会,1985, 5 5 王尚华.智能机器人图形理解系统.中国科学院自动化研究所硕士学位论文,1988
