High Speed Vision System For Battery’s Final Inspection

As dry battery consumed by human being has a large number and wide variety, it isurgent to achieve automation and high speed at all stages during the process of production.Dry battery owns a certain service life as daily consumer goods. In order to prompt messageto consumers, manufacturers always stamp a set of numbers at the end of the iron plates of thebattery cover. This set of numbers indicates the use of a battery cut-off date. This processexecutes before packing up battery tin. In the traditional production process, the dates in thebottom cover of battery have not been detected or detection approach relies on artificial visual.Workers’ ability to judge, reaction speed, these all factors will seriously affect the productsquality, so large manufacturers also hope this procedure can be automated detection. Thispaper studies the key technologies based on the visual detection of dry battery, and developeda detection system of high speed dry battery through image preprocessing, digital characterrecognition, remove synchronization and other methods.Firstly, based on the requirement analysis, the image processing system is consisted ofcamera, lighting and lens which are chosen by a lot of experiment. The system also is in a lowexpense but meets the basic demands that the image must be clear for the feature of thecharacter and colloid. As for the outputting part for the system, the IENC card controlsactuator to get rid of the bad production, getting an instruction from PC. Because of animprovement on the control strategy, the problem of the leaking on the images of battery issuccessfully overcome.Secondly, with the help of normal image preprocessing methods, for example, imagefiltering and segmentation, we could get those images that are obvious characters. The methodof image filtering is mean filter of which the template is cross. The methods for segmentationare wave analysis, synthetic method and moment preserving. Based on those processes, thepaper fits a circle by the method of least squares, which initially locate the characters on thebottom. Then searching for the feature of characters from the tangential and radial directionexactly located the characters. At the end of the process, refining image method is using theHilditch algorithm.Finally, with the facility of Visual studio2008, the paper finished the high speed visionsystem for battery’s final inspection, which includes several functions, such as image ofcapturing, showing and processing. The user also could look over the records by the database.