Automatic Machining Features Extraction from Two-Dimensional Image of Mechanical Parts with the Help of Artificial Intelligence

Extracting the required information from the design file is one of the main steps in the computer aided process planning. In previous methods of extracting machining features, various methods such as graph-based method, volume analysis method, logic rules method and other methods have been used. In all the previous methods, whether traditional methods or methods based on artificial intelligence, the input data to the machine feature identification system is the output information of a computer-aided design system. Converting the output information of a computer-aided design system to input data of a machining feature identification system is faced with limitations such as the variety of format and type of data arrangement, deleting some data from the design file due to geometric interference of features, slow extraction of features due to extensive information in the design file and the limitation of identifying different types of machining features by a unity feature identification system. In the present study, using artificial intelligence techniques based on deep learning, machining features are extracted directly from the two-dimensional image of a workpiece. The image may be prepared by a computer-aided design file, or it can be taken by a camera.