International Journal of Mechatronics and Manufacturing Technology

Authors: Rohan T. Pawar , Meera S. Rathore

Abstract: Robot vision and AI-based recognition systems have emerged as critical technologies enabling artificial agents to perceive, understand, and interpret visual information from the environment. This paper reviews recent advancements in robot vision, deep learning models used in recognition tasks, and the integration of AI systems with robotics platforms. The review discusses core algorithms, sensing modalities, applications, and challenges faced by developers and researchers. Through comparing classical vision techniques with modern AI approaches, the paper highlights how convolutional neural networks (CNNs), transformer architectures, and sensor fusion methods improve accuracy and reliability. The paper also outlines future directions where robot vision can evolve more adaptively in real-world scenarios, especially under dynamic environmental conditions. Case studies and statistical comparisons are included to illustrate performance differences across popular models. The overall purpose of this review is to support researchers, students, and engineers looking for consolidated and practical insights.

Keywords: Robot Vision, Artificial Intelligence, Image Recognition, Deep Learning, Sensor Fusion, Object Detection, Visual Perception, Neural Networks, Real-Time Processing