This dataset is the result of applying a Pipeline described in the paper "A GIS pipeline for the production of GeoAI datasets from drone imagery" in revision for publishing. The idea behind the dataset is to run deep learning segmentation models to compare which buffer size or buffer distance allows the model to learn better road characteristics to be extracted from drone aerial orthomosaics. In the problem a tradeoff between obtaining imbalance binary masks and masks overloaded with outliers is studied.