TVTAT - A Real Time Camera Imaging Testing Tool for Smart TVs: Preliminary Results

Test automation tools that can accurately control a smart TV device are rare to be found. The difficulty of creating a system that is able to control such a device generates specific needs. In addition, the Brazilian digital television system supports DTVPlay, a middleware that provides the ability to broadcast interactive applications written in HTML5, NCL and Lua, that must be fully implemented on at least 90% of televisions manufactured in Brazil. Thus, there is a standard to be followed and a set of tests that need to be performed with each new middleware release. This work presents an automation tool called TV Test Automation Tool (TVTAT), that performs non-invasive tests on smart TVs, mainly but not restricted to DTVPlay tests. TVTAT uses real-time computer vision techniques such as optical character recognition, image pattern matching and color verification to assert that the middleware’s implementation is according to the published specification. The results of some test scenarios are presented, demonstrating that there are trends that can be found either in application performance situations or in tests of availability of transmitted applications that depend on DTVPlay.