Research On Earphone Solder Joint Recognition And Positioning Based On Image Processing And Embedded System

As one of the main carriers of "vision" and "listening",earphones have become an essential daily electronic product for people.In the production process of earphones,in order to ensure the sound quality of finished earphones,i t is necessary to test the sound quality of semi-finished earphones before welding wires.The current test method is mainly based on manual operation of electro-acoustic device tester for testing.This method is not only inefficient,but also very different.Therefore,the automated test equipment for the sound quality of the earphones has begun to attract attention.The realization of earphone solder joint positioning is an essential step in the development process of this type of equipment.This article studies the problem that the round earphone cannot use mechanical devices to accurately locate the position of the earphone's welding point.Combined with embedded and image processing technologies,an embedded system designed for the identification and positioning of earphone's solder is designed.The identification and positioning of the earphone solder joints are realized.This article main tasks as follows:(1)Design the algorithm of earphone solder joint recognition and location based on OpenCV open source library.The grayscale,median filtering and other operations are used to preprocess the headphone image.Combined with the characteristics of the headphone image in this paper,in view of the shortcomings of the traditional median filtering time,a non-sorted median filtering suitable for this paper is proposed.On the basis of least squares circle fitting,the improvement is made,and the least square method based on constraints is proposed to obtain the parameters of the outer circle of the earphone,and the robustness is improved under the premise of guaranteeing the rate.The headphone solder joint is identified by the connected domain calibration algorithm,and the deflection angle of the headphone solder joint relative to the probe of the electroacoustic device tester is calculated using the obtained parameters.(2)The Realization of embedded system for earphone identification and positioning.Through analysis and comparison of commonly used processor chips and operating systems in the embedded field,the GEC6818 development board equipped with the S5P6818 processor was selected as the platform.By transplanting Uboot,Linux kernel and making root file system,the construction of Linux operating system on the development platform is completed.By cross-compiling and transplanting OpenCV3.2.0 and Qt5.9.6 source code,the software operating environment was built,and the working interface,browsing interface and setting interface were designed based on Qt to realize the function of human-computer interaction.(3)According to the test accuracy,real-time and stability of the system,design and test the experiment separately.The experimental results show that the test accuracy,real-time performance and stability of the system meet the application requir ements of the headphone sound quality automatic test equipment,and the system has certain practical application value.