What you see is what you test - Augmenting software testing with computer vision
|Titel||What you see is what you test - Augmenting software testing with computer vision|
|Buchtitel||Proceedings of the 10th IEEE International Conference on Software Testing, Verification and Validation Workshops (ICSTW 2017)|
|Note||Best Poster Award der ICST 2017|
The blind spot of software testing is the assessment of the actual behavior of the system under test in the real, physical world. In this paper we show how this inherent restriction of software testing to the “cyber world” can be overcome with the use of methods and techniques from computer vision. It augments conventional software testing and allows making observations about states and events in the physical world as well as the system’s real-world context. We implemented a demonstration scenario that shows how visual test automation can be combined with computer vision techniques to include observations of the physical properties of a mechatronic system in software testing. The successful application of the approach in a lab setting revealed several benefits and also some limitations. We discuss these benefits and limitations to highlight potential application scenarios in an industry setting and avenues for future research.