Parking Lot Space Detection Based on Image Processing

Estimating the number of parking spaces, whether empty or occupied, in a given geographic area from aerial images can provide much useful information about where people live, work, and when. It is, therefore, an important tool for government and business in planning and responding to societal needs including the ability of first responders to react more confidently in the case of major events such as hurricanes or terrorist attacks. A starting point is to estimate the maximum capacity of parking lots, and then to estimate an occupancy factor at a given hour. In this paper, we aim to present the use of image processing to help detect the maximum capacity of a parking lot from aerial imagery, such as those available at Google maps. We use a variety of image processing techniques, especially morphological ones, to count the number of parking spaces that are demarked with white or yellow lines. The software is implemented in Matlab. We will show several examples for different types of parking lots such as fishbone and 45-degree patterns.