本論文將針對偵測待測系統的GUI元件提出新的方法，該方法將透過影像處理並與基本的通用輸入裝置進行協同識別，以偵測GUI元件；此方法具有高度與平台無關，並且無懼於平台的底層技術升級的特性。 ;In the field of software engineering, software testing is an important and major method to ensure software quality. In practical software testing, regression testing software tools have matured and varied. Regression test automation is to make sure that the system features that function correctly are not affected by software changes. On the other hand, “automated test case generation” remains to be very difficult in practice. The construction of the “GUI Model” is an important step toward GUI test case generation. If we can construct the GUI model of the system under test (such as application, webpage, etc.) accurately and quickly, it can be used to achieve GUI test case generation. Most attempts to construct the GUI model are done by reverse engineering the program on specific platform, environment, or even analyzing the source code of the program when there is poor support from platforms. In some modern platforms, they can use the API provided by the platforms to extract and operate GUI components (called “control-items”). However, these approaches introduce very high dependency and coupling to the underlying platform and the results are not applicable to another platform. For example, approaches that aim for web technology are not applicable to Windows or Linux. This paper proposes a new method for GUI component detection of the system under test using image processing and basic universal input devices. It is totally platform-independent. Finally, we will use this method to test different types of system under test to verify and evaluate the feasibility of the method.
