After collecting the images of our first tower, we had to start annotating them.

In our first draft of our Preliminary Design Report, we focused largely on background research of pre-exisitng softwares like ours. We found that a direct computer vision approach to this problem is much less common in this industry than a 3D scanning approach. However, the computer vision approach is more promising.

Therefore, we are starting to annotate our cell tower images with bounding boxes and finding the best way to develop our annotating protocol by testing it on open-source computer vision models. The first goal is to find which model is best at drawing bounding boxes on our antennas.