Abstraction and intelligent recognition of the shape feature of the machine part

An approach to recognize and extract features from a solid model of machine part is presented. It takes into account the engineering practice of designing a part by first constructing the chief elements and then successively refining the shape by adding subordinate details. The algorithm first filters out the minor structures from the part and builds the factor list with assigned weight to each entity. Through recognizing the main direction of part, choosing its optimum orientation and consulting the knowledge base the principal direction of part is finally determined to generate its front view without human intervention.