Description
Defence Research and Development Organisation is a premier R&D agency under the Government of India. I was selected as a Winter Intern for 2 months in the Solid State Physics Lab to work on one of their ongoing projects. The project I worked on was for inventory management and handling of optoelectronic devices using Machine Learning methods, which would ultimately be used to categorize devices from a live feed.
On a large scale industry level it becomes important to distinguish between such devices when there exist many different types of them. This could be done with the human eye but a faster, more reliable and efficient method would be to use a computer that can easily distinguish between images.
Specifically, I worked in the team that developed an image recognition model for accurately detecting and classifying these optoelectronic devices.
— to identify features, we drew keypoints for matching descriptors using distance calculation
— implemented ORB for feature selection
— Brute-Force matcher was used to find best descriptors
This project acted as my foundation to OpenCV and Image Processing. In fact, the practical knowledge I gained here was an instrumental step to my choice of opting for Image Processing as one of my electives in the subsequent year. This opened new doors in the world of AI, and allowed me to explore a vast variety of projects.