Design Of Automatic Test Platform For Camera Detection

In recent years,with the advancement of imaging technology,new media and other fields have become popular among the public.The imaging industry is much more active than before,and the functions of camera products as the carrier of recording media have become more abundant.At the same time,due to intensified market competition,the iteration speed of camera products has accelerated,resulting in a certain degree of instability in the listed machines.At present,in terms of testing methods,testers are mainly used to design test cases and test camera products manually.However,for some repetitive testing tasks,the labor cost is very high,and it is difficult to cover the test on a large scale,and it is prone to miss the test.At the same time,the research and data on the automatic testing technology and development of cameras are very scarce.Based on the above background,this paper researches and designs the core modules of the camera automatic detection platform,which can simulate various operation behaviors of users and perform functional tests on products before they are launched,which can improve the test efficiency to a certain extent and discover the potential of products within a limited time.problems,to ensure the stability of product quality and performance.The main work of this paper is: 1)First,a brief description of the related technologies of camera automation testing: including embedded software testing technology,camera testing technology,video scene classification technology and image similarity processing technology.According to the embedded software structure characteristics,camera test characteristics,etc.,the test type and test method are determined,and the key frame feature extraction,image pixel,feature and other matching algorithms are selected as the technical approach followed in this article.2)Secondly,design the core module scheme of the camera automatic test platform:including design background,automated test process,functional requirements analysis,test environment construction,core module function scheme design and feasibility demonstration.The main design idea is to convert the commands sent by the PC into corresponding control commands to simulate the operation of the camera,and use the video image processing technology to analyze and compare the test results.3)Then,design each key function module in detail.Terminal control is achieved by keyword-driven and wireless network remote control;Web-side button recognition uses SURF combined with FLANN algorithm to achieve rapid target identification;clustering method is used to achieve video key frame extraction,using histogram,perceptual hash and difference Methods such as hashing process the image similarity.4)Finally,verify the Web-side target recognition function and determine the Hessian matrix threshold,select two functional tests of the camera(user interface test and lens test)for experimental verification.The accuracy and time efficiency of the three image similarity processing algorithms are compared,and the algorithm that can meet the functional requirements of the platform is determined.At the same time,the efficiency of automated testing of user interface testing and lens testing was compared with manual testing,and the testing results were analyzed to verify the practicability of the automated testing platform.The automated test platform designed in this paper is more efficient in some test scenarios and test tasks,and has a wider coverage.It can improve the test efficiency to a certain extent,so as to achieve the purpose of ensuring product quality.It can also provide a reference for other companies to develop automated test platforms,and has certain application value.