Automated AutoCAD Drawing Assessment via Image Processing and Vector Transformation Techniques

Abstract

Conventional assessment practices in computer graphics courses, particularly those that utilize AutoCAD, often rely on manual grading or basic template-matching strategies. These methods are ineffective and biased, particularly when used for extensive evaluations. Intelligent evaluation methods and automated image processing must be integrated as educational technology continues to evolve. The purpose of the proposed effort is to develop and put into use an intelligent AutoCAD computer drawing evaluation system that uses image processing technologies. Enhancing assessment accuracy, automating scoring, and utilizing robotic technologies to combine virtual drawing analysis and actual drawing validation are the objectives. The system evaluates student drawings using MATLAB-based techniques, including vector transformation, grayscale conversion, binarization, and histogram similarity. It extracts components using DXF file parsing, performs geometric matching, and features extraction. A feedback-driven retransmission method ensures packet correctness. A servo motor-powered drawing computer duplicates input drawings, and performance is assessed using torque analysis, picture entropy, consistency, and smoothness criteria. The system could accurately reproduce student drawings with an accuracy of more than 0.1 cm and an average drawing speed of 1.75 cm/s. The system's dependability was confirmed when evaluation ratings for example drawings nearly matched hand grading. Within the robotic arm's torque limits, moment and motion analysis verified operational safety and accuracy. The proposed approach automates computer graphics analysis by combining hardware and software elements for perceptive evaluation. However, limitations on robot motion and image quality sensitivity were limitations, requiring future improvements.