Overview / Usage

We intend to measure the strain in a deformed ribbon on applying stress. For this we sample a ribbon into dense points. Each of the specific points is represented by a unique encoding of markers that is robust to inversions. The nearest neighbours of each marker is identified after the deformation along with their coordinates to decipher the encoding of each point and the new position. Then using statistical techniques, we can compute the change in ribbon position and hence the strain.

Methodology / Approach

First we sample points on the ribbon. We then encode each of the points using different encodings that are created by a group of markers using the minimum number of markers. Graph Theory algorithms are used to achieve this task.

The next step is to identify each of the unique markers. This is done using the ArUco library of OpenCV platform.

Then we identify the nearest neighbour of each of the markers. This is done using various Topology concepts on distance metric like Hausdorff Distance and then the best marker is selected. Finally we detect the markers in our ribbon, find their nearest neighbours, decipher their encoding and finally, measure the strain due to the deformation.

Technologies Used

OpenCV

Python

Repository

https://github.com/guptasamarth61/Nearest-Neighbour-Detection-and-Encoding-of-ArUco-Markers