Estimation of the contact point for nano-electronics

Working on the nanoscopic scale means that not all of your standard computer vision algorithms will work as you wish. There are a lot of challenges you need to face in order to adjust your methods to work in nanoscale. One of our clients had to confront some of these problems, as their solution required the estimation of the contact point of two nanometer-size objects based on an RGB image sequence. Given our experience in applying computer vision methods to unusual problems, we were able to find a suitable solution for the task.

We’ve developed a method for estimating the contact point of two objects in nanoscale based on RGB camera recordings.

Due to optical limitations occurring in nanoscale, we could not apply feature-based methods. Instead, we estimated the outer shape of one object by calculating the differences between the consecutive frames, applying machine learning algorithms and fused this information with the data from the sensor mounted on the servomotor.