Research And Implementation Of Web GUI Automation Test Platform Based On Image Recognition Technology

With the rapid development of the Internet,Web application technology has been widely used,while the scale and complexity of Web application system is also gradually increasing,software testers pay more and more attention to how to ensure the correctness and reliability of the Web system function.At present,web application software has the characteristics of fast update,due to the changes of DOM structure and page elements,the proportion of GUI-level testing cost in the total cost of regression testing is increasing.Testers use the method of page control recognition based on DOM object to build automated test script,which has high maintenance cost of test script.Therefore,based on image recognition technology,this paper constructs an automated test platform according to the automated test execution process,which improves the test efficiency in regression test to a certain extent.The main work of this paper is as follows:(1)Combined with the technology of image recognition in the automated test,the traditional method based on DOM page element positioning is transformed into the method based on GUI image element matching positioning,and according to the characteristics of different display resolutions of the page to be tested,the matching algorithm is verified and the SIFT feature matching algorithm is selected,which can better adapt to the situation of scale change in GUI testing of the Web application.(2)According to the automated test execution process,the automated test platform is designed and implemented,which provides test project management,test case management,test script management,automated test execution,log and test report services and other functions.The platform is developed based on MTV mode of Django framework,at the same time,it integrates the open source test framework Airtest and Selenium,generates the test script based on the Img Page-Object mode combined with the operation keyword mapping table,and uses the Jinjia2 template engine to output the visual test report.The platform system has been deployed in the experimental environment,and its effectiveness has been evaluated.The platform system has been deployed in the experimental environment,and its effectiveness has been evaluated.The results show that the automated testing platform realizes the supplement of the traditional automated testing based on DOM element positioning through image recognition technology in the test execution,and also clearly shows the test report results.In this paper,four different versions of the same web application are selected for regression testing experiments,for the latest version,which improved the reuse pass rate and regression test efficiency of the original test cases.The experimental results show that with the increase of regression times,the test script based on GUI image element recognition and positioning can improve the recognition rate of web page controls,the reusability of test cases,and the test efficiency,which has a certain practical value.