Assessment of Geometric Similarity of Volunteered Geographic Information in Intra-city Roads Network

Fast and updated access to the information related to urban roads network is one of the requirements with various applications in the field of transportation. Accordingly, the present study tries to evaluate geometric similarities between volunteered geographical information through comparing them with the official dataset, in order to provide the researchers with more assured alternatives. To do so, a new solution is provided, which while identifying the corresponding features in volunteered geographical information dataset and official dataset, compares these features to each other in terms of length, distance, orientation, buffer overlapped area, complexity, size, and sinuosity. For this reason, District 6 of Tehran was selected as case of the study and the 1:2000 dataset (prepared by National Cartographic Center) was selected as the official dataset. The results were indicative of a 78% mean spatial similarity between the features in Open Street Map (OSM) dataset (a sample of volunteered geographical information) and those in the official dataset. Also, in order to evaluate improvement of geometric accuracy in volunteered geographic information dataset, urban roads networks were studied in years 2014-2017. The results showed that while public participation in creating the features has increased by 27.8%, the mean spatial similarity has also improved to 5.77%. Therefore, the researchers can use volunteered geographical information dataset as an alternative for the official dataset in many of the transportation related applications