|In 2014, there were an estimated 245 million video surveillance cameras installed globally (HIS Markit). Real-time visual data have many applications. As researchers gain the ability to collect large amounts of visual data about the world, the true potential of data-driven research is recognized. Despite the large amount of publicly available real-time data, there are challenges that inhibit the true potential of analyzing real-time data from these cameras including identifying the cameras either statically or dynamically and identifying camera metadata once identification of individual cameras occurs.
Researchers at Purdue University have developed a method for identifying cameras on the internet; particularly, identifying public and closed circuit television cameras and the metadata associated with each in a static or dynamic manner. Examples of uses include monitoring traffic flow, viewing wildlife, detecting intruders or anomalies, and monitoring weather conditions and emergencies.
-Allows access to existing network of cameras
-Machine learning algorithm development
Mar 23, 2017
Purdue Office of Technology Commercialization
1801 Newman Road
West Lafayette, IN 47906
Phone: (765) 588-3475
Fax: (765) 463-3486