Machining features recognition from image of mechanical parts with the help of deep learning technic

In the production of industrial parts, machining is one of the most important operations in the field of manufacturing parts. The production of an industrial part takes place in three stages: design, process planning and manufacturing, and in all these stages, the computer is used as a powerful tool. In computer-aided process planning, the stage of identifying machining features is a prerequisite and an introduction to the next steps. Extracting information and identifying features from computer-aided design information has been continuously improved due to the increasing complexity of parts, but the research to find an optimal solution is endless. Over the past few decades, several methods have been introduced and applied by researchers to extract and identify machining features from design file information. In all the previous methods, the number and type of features are extracted as independent variables in the machining features identification pattern and from the part design file data. In this research, the charectrestics required to identify the machining features are extracted from the pixel values of the machining feature image by the artificial intelligence system automatically. The artificial intelligence system produced to identify the machining features in this research is able to identify all the information required for machining, including the name, the coordinates of the location of the feature relative to the part, and the dimensions required for the machining, by viewing the image of a part, and the information of the features present in the image the input to the system in a table.