Research On Pattern Recognition Methods Of GUI Objects In Automated Testing

GUI software testing is to test software with a graphical user interface.It can effectively detect potential errors or defects in the software.The emergence of GUI automated testing technology has effectively reduced the input of testing resources,improved the overall efficiency of the testing process,and reduced the software development iteration cycle on the basis of ensuring software quality.For all GUI automated test tools,how to identify the object to be tested is the primary problem and key technology.The traditional control identification method is based on the development framework and the API of the testing tool,which relies heavily on the openness of the API,and has the disadvantages of limited identification range,low flexibility and high cost,and the tester needs to understand part of the code implementation.In recent years,the technology of image pattern recognition has developed rapidly,and the automatic testing tools based on image recognition have performed well.In this paper,a new fast template matching algorithm is proposed in image processing technology and image pattern recognition template matching technology.And with the new fast template matching algorithm as the recognition driver,combined with other auxiliary functions,a GUI target quick recognition tool is designed to solve the existing tool control recognition of limited recognition range,low flexibility,image resolution sensitive and low recognition speed defects.The main research contents of this paper include the following three points.(1)Research the template matching algorithm in image pattern recognition technology and the similarity algorithm of template matching in Open CV,combining image processing technology and pixel comparison method,and propose a new fast template matching method,and apply it to GUI targets such as controls in software interfaces Identification of elements.(2)Aiming at the advantages and disadvantages of existing recognition methods,combined with the new fast template matching method proposed in this paper,a rapid recognition tool for GUI targets based on image pattern recognition is developed.Combined with actual test requirements,corresponding auxiliary functions have been added to the tool,such as automatic grabbing of the software interface diagram under test,automatic stretching of template diagrams,and storage of matching results to improve its adaptability,recognition efficiency,and recognition accuracy And ease of use.(3)The tool developed in this paper is verified in two aspects of recognition accuracy and recognition speed,and the improvement of the new fast template matching method proposed in this paper in the original algorithm is analyzed,and compared with the existing mature tool recognition methods.