Design and application of an automated test platform based on image recognition

Abstract

More complex and changeable software systems put forward higher requirements for software testing, traditional software testing methods have shown efficiency weakness in ensuring the quality of large-scale, high complexity software, automated testing methods have become the focus of research in the field. The currently proposed automated test method has insufficient reuse rate, low level of automation defects, and has not been widely promoted. This paper proposes an automated test platform and method based on image recognition for this problem, integrated image recognition algorithm, test software TestStand, graphical programming system LabVIEW and the system under test. Among them, LabVIEW can establish a model framework compatible with the system under test, which establishes a logical interaction relationship with the test sequence in TestStand; the image recognition algorithm can identify the desktop button of the current tested system, it also can recognize and click the button corresponding to the current test sequence of TestStand combined with the automation program, so as to replace manual operation and realize the whole process automation. The method is applied to the test project of an interlocking equipment, and the results prove that the automatic test method proposed in this article can shorten the test period to less than one-tenth, greatly reduce the test cost, and has good application and promotion value.