Parking lot delineation and object detection using a localized Convolutional Neural Network

Satellite imagery analysis is becoming increasingly important in a variety of fields. Automation of analysts' workflows has the potential to greatly increase their efficiency and effectiveness. Specifically, developing automatic parking lot extraction from overhead imagery has huge importance to predicting stores future sales. However, to the best of our knowledge, there is no prior work for the development of an end-to-end pipeline for this task. This paper outlines a method for automating parking lot extraction from overhead imagery. We address this challenge using the combination of a Convolutional Neural Network (CNN) and various morphological tools of analysis. Our pilot end-to-end pipeline has been shown to be effective even in complex environments, and exhibits promising results for the automation of imagery analysis.