UTILIZATION OF COMPUTER VISION AND MACHINE LEARNING FOR APPLIED ENGINEERING: DATA ANALYSIS AND RECOGNITION

Abstract

Purpose. The primary purpose of this research article is to develop a ML model for applied engineering for varied actions, such as identifying sunflower seeds from other seed types and adjusting the screw press parameters accordingly for optimal oil extraction. Additionally, it aims to predict noise propagation and implement measures for noise reduction in relevant engineering applications.

Methodology. The research methodology involved analyzing scientific sources, experimental data, modeling, and machine learning. The ML model was trained using the Edge Impulse platform with a dataset of seed images, annotated for focus. After iterative training, validation, and testing, the model was embedded into an Arduino controller for real-time seed identification and automatic screw press operation regulation.

Findings. This article introduces an innovative applied engineering approach aimed at revolutionizing the seed oil extraction process through the automation of screw press operations using Machine Learning (ML) and Computer Vision (CV). Key findings include the successful differentiation between sunflower and pumpkin seeds and precise adjustment of screw press settings based on seed type identification. ML is also utilized to detect an empty seed feeder and halt press operations automatically, preventing equipment damage and ensuring efficiency. These results pave the way for enhanced automation and precision in seed oil extraction processes.

Originality. The application of ML and CV in seed oil extraction and screw press operation, as presented, underscores the transformative potential of these technologies in agricultural processes.

Practical value. Incorporating computer vision and machine learning into applied engineering streamlines processes, reduces errors, and enhances efficiency. This integration optimizes resource utilization, enables real-time decision-making, and boosts productivity across various engineering applications.