Android Automated Testing Based On Image Comparison And Continuous Integration

In order to assure the quality of Android applications, a lot of theories and methods of Android testing have been put forward, among them, automated testing saves test time and cost. Automated testing tools and frameworks, such as Monkey Runner, Robotium, and so on, have been widely used. Based on research of these automated testing tools and frameworks, two common problems have been found: one is the need to write testing code with their APIs, which requires some programming ability, the threshold for the use of them is higher. The second is that they only focus on automated testing script execution process, other parts of the testing process still need manual operation.In this paper, some researches aiming at solving the above two problems have been carried out. Firstly, The method of image comparison has been applied to Android testing. Through acquiring the function page images of the application, and checking with benchmark sample images, to judge the correctness of the Android Activity, so as to realize the testing of Android applications. The design of Android testing system based on image comparison has been described in detail, in terms of image acquisition, image preprocessing, sample library design, image comparison algorithms, test report generation and so on. Secondly, by researching on continuous integration solution of Android, including Jenkins platform, unified code base, version control of source code under test, building of Android application, Android testing environment, a automated testing process has been designed, such as building testing environment automatically, launching testing system automatically, and automatic notification of test results. Then, the Android testing system based on image comparison and the Android continuous integration system has been realized respectively. And in the testing process of continuous integration system, testing based on image comparison is called to realize the automation of the Android testing. Finally through a case study to verify the feasibility of Android automated testing based on image comparison and continuous integration, and the availability of the testing system.The automated testing system use the method of image comparison to test Android applications, and use the method of continuous integration to realize the automatic testing process. With the testing system, new updating code can be test automatically in time during the project development process. Testers only need to design a simple test case, without writing any test code. The entire test process can run without any human intervention, and the quality of the Android applications can be intuitive and controllable.