The Design Of The Solder Joints Position Detection System Of Loudspeaker Based On Image Processing

Loudspeaker is an electro-acoustic device and used widely. With the increasing publicity of loudspeaker in daily life, people has higher requirement on the voice quality of loudspeaker. One of the effective methods to improve the outgoing quality of product is to strenghten the quality detection of loudspeaker. Currently, the frequently adopted artificial audition method fails to guarantee the loudspeaker detection result and efficiency due to its low efficiency and subjectivity. In the auto electro-acoustic test system, the operating procedure such as connecting the probe and the two solder joints is still by hand. In order to achieve fully automation of electro-acoustic detection, it is necessary that a solder joints positon detection system shall be designed. The detection system could identify the two solder joints positions automatically and then send this message to help the probe contact with the solder joints.The solder joints of loudspeaker have significant visual characteristics. With the development and maturity of digital image processing technique, image-based detection method has provided ideas for detecting the solder joints position of loudspeaker. This article designs the front-end system of electro-acoustic test system. With the comprehensive application of image processing technique and embedded technique, it has realized automatic real-time detection on solder joints of loudspeaker, made good preparation for electro-acoustic test, could improve enterprise’s automatic level and product quality.Firstly, the work of loudspeaker solder joints detection is represented and the performance requirements is discribed. Then a solder joints detection system of loudspeaker is designed which includes two parts: hardware and software. The hardware is composed of STM32F103ZET6 microcontroller, OV7670 image sensor, TFT LCD touch screen and other related circuits. Additionally, a ring-shaped LED light source device is build in order to decrease the light interference.Secondly, this paper focuses on the loudspeaker solder joints position detection algorithm. The louderspeaker image is preprocessing by using graying and median filtering method. And a new advanced median filtering algorithm is presented to increase the speed of traditional median filtering algorithm. After a variety of thresholding methods were compared, using a manual threshold method to extract the excircle and joints of the loudspeaker. After analysis and comparison of the Hough transform and the least square method, using the least square method to determin the center and radius of the excircle of the loudspeaker. According to the characteristics of the solder joints, this article has formulated three criteria, finally recognized the solder joints of loudspeaker from numerous “suspicious solder joints”, and calculated the required deflection angle θ.Thirdly, μC/OS-Ⅱ real-time operating system is applied to carry out multitask concurrency design, and detailed analysis and design has been conducted on image collecting module, image storing module, image processing module and man-machine interaction module of main task. Later, the C code of the project has been optimized in various aspects so that the image processing speed has increased by 1.45 times.Finally, systematic experimental platform has been set up to test different loudspeaker images. The experiment result shows that this system is characterized by accuracy, real-time, robustness and man-machine interaction. Hence, the research on the detection system of loudspeaker solder joints position is of certain application value. The implementation of this solution will improve the automatic level and product quality of loudspeaker enterprise.