Image-Based Web UI Automated Testing Methods for Aerospace Business Software
Currently aerospace bisnesses are heavy,how to accomplish the testing of related software efficiently and accurately become particularly important.An image-based Web UI automated testing method for aerospace business software is designed.This method conducts automated testing from the user's perspective,and uses YOLOv8 model to detect the screenshot of tested web software page,and identifies three types of elements:ta-bles,line charts,and navigation buttons(including navigation menus,breadcrumbs,buttons,etc.).Then,Sele-nium is used to automatically test the accessibility of navigation button elements,and the text recognition tech-nology named PaddleOCR and Structural Similarity(SS1M)algorithm is used to test the accuracy of information for table elements and line chart elements respectively,and a test report is formed.This method is easy to port-able,it can provide the localization of elements for Selenium automatically and improve the issue that Selenium scripts are prone failure,and check the correctness of chart data automatically.Experimental results show that this method can cover 90％of the test cases,detect the location and information of errors clearly and reduce the workload of manual testing significantly and improve the testing efficiency.
Web UI automated testingYOLOv8structural similarity algorithmPaddleOCRSelenium
万方数据
维普
