Scraping of Vulnerable Physical Environment Images in Shodan Images in Turkey for Dataset

Abstract

In this study, a scraping software has been developed with python in order to use the physical space images in Shodan images section as a dataset for CNN based artificial intelligence applications. With this software, the images listed in the Shodan images section within the scope of the query "country:en port:554" are collected by web scraping technique. The scraping software scraped and stored the images obtained as a result of the query every 3 days for an average of two months. As a result, 2,220 images were obtained to be used as a dataset in CNN-based classification applications.