Machine Learning Approach for Converting Image Data/ Matrices (CSV) to 2D CAD Drawing.
Y. Yaswanth, K. Sai Surya, U. Nandini, P. Sathish Kumar, Dr. M. Vykunta Rao.
GMR Institute of Technology
Abstract:
The process of transforming the image-based engineering drawings into a well-structured format of Computer-Aided Design (CAD) drawings is considered to be a challenging process. It is considered that the traditional process of image-based engineering drawings recognition has been highly dependent on traditional drawing methodologies. Such a process has been considered a time-consuming and inefficient process of image-based engineering drawings recognition. In the proposed paper, a novel method of machine learning has been proposed for transforming the image-based engineering drawings and CSV matrix into accurate 2D CAD drawings by applying the Python and AutoLISP automation techniques. In the proposed method of machine learning for image-based engineering drawings recognition, the image-based engineering drawings preprocessing techniques, classification techniques, and CAD drawings automation techniques have been utilized for developing a novel and efficient method of transforming the image-based engineering drawings into a well-structured format of CSV matrix. In the image processing technique of machine learning for image-based engineering drawings recognition, edge detection, contour detection, and Hough transform are considered to be highly useful for image-based engineering drawings recognition. In the process of image-based engineering drawings recognition, different features are extracted from the image-based engineering drawings. Such features include lines, circles, arcs, and ellipses. The classification of features is considered to be highly useful for transforming the image-based engineering drawings into a well-structured format of CSV matrix. AutoLISP programming is considered to be highly useful for the automation of the process of generating accurate CAD drawings from the image-based engineering drawings. It has been considered that the proposed method of machine learning for image-based engineering drawings recognition is highly efficient and effective for generating accurate CAD drawings from the image-based engineering drawings without compromising the accuracy of the entities and annotation of drawing entities.
Keywords: CAD Automation, AutoLISP, Computational Geometry, 2D drawings, OpenCV, Feature Extraction, Engineering Drafting.
