Perception is one of the most important step in all modern robotics and automation tasks. It gives a robotic system the ability to recognize objects in its surroundings and have a better understanding of the environment. Having such perception capabilities enables the overall robotic or automated system to respond and interact more efficiently with its surroundings.
The perceived visual information from a perception system can also be used to augment other kinds of data. For example, depth information from depth cameras can augment distance measurements from lidars; frame comparisons from cameras can give a secondary measurement of displacement; readings from optical flow cameras can be coupled with accelerometer data to have more accurate measurement of velocities, etc.
The primary focus of this work is to develop, compare, and integrate novel perception systems and algorithms that have been crafted for robotics and automation applications.
Perception systems and algorithms to enhance the overall capabilities of robotics and automation systems have been explored for various domains, such as prosthetic limbs, drones, digital pathology and also for modern factory automation tasks.
This work provides a detailed explanation of the methodologies used, how they aid in their respective applications, and the research contributions.