Comparison of Different Computer Vision Approaches for E-waste Components Detection to Automate E-waste Disassembly

Abstract

Electronic Waste (E-waste) is generated in a tremendous amount due to our increasing dependence on electronic devices and rapid upgrading in technological innovations. Environmental and health risks are posed because of e-waste toxic constituents. On the other hand, e-waste contains valuable recoverable materials that can gain economic benefits. Efficient e-waste recycling is optimally conducted when different components are separated and processed by their appropriate chemical techniques. If component separation were conducted by humans, it could put their health at risk and consume an unnecessary amount of time. Thus, Automation and robotics solutions are needed to carry out the recycling. At the heart of such solutions is a computer vision algorithm that can detect, localize and classify e-waste components. Different Computer vision approaches, both traditional and deep learning-based, were compared to know which approach will be more suitable for such a task.