Diaper Marker Visual Inspecting Method And System

Since diapers are baby hygiene products, the quality of their elastic waist and wa- ist label has a direct impact on users'amenity and safety. They are important symbols to determine the products'quality, so strict on-line detection is essential. Previously, the detection of diaper labels are mainly conducted by manual detection method, whi- ch is of high labor intensity and low efficiency. With the development of technology, visual inspection based on image processing has significantly increased automation d- egree of on-line inspection with high speed, high precision and high reliability.Taking diaper as study object, an automatic vision inspecting system is designed to make analysis and judgement of the diaper elastic waist's position and waist label's angle. Design label recognition and data storage module to make a complete system. This system can overcome the shortcomings of manual inspection with some advanc- ed characteristics, such as high accuracy, low-cost, fast and good stability. At the same time, the inspection principle and method of the system is generic and can be applied on product testing of the same type.A brightness-adjustable white LED is chosen as the illuminator of visual inspection system. Build visual sensing platform and take PC as processor. Grab production images with high-resolution colorful CCD industrial camera and transmit them by the 1394 bus to PC for processing. A visual inspection software is developed based on Visual C++6.0. Image processing algorithm is proposed to determine the elastic waist's position and waist label's angle. Inspection results will be stored and displaye- d on the interface.The subject focused on the design of visual inspection software. By the idea of modulization, software is divided into several parts including image collection module, image processing module, parameter configuration module, data display and storage module. Each module can work independently to make it easy for expansion and upgrade. Image processing module contains some widely used algorithms, such as image segmentation, morphological filtering and edge detection. More over, accor- ding to the characteristics of diaper, a color extracting algorithm is designed to effecti- vely reduce the external interference such as illumination changes on the system.Experimental results show that system can realize real-time detection for diaper surface marker with high accuracy and high speed. Detection algorithms can be selected by users through the parameter configuration module flexibly. The system can meet the requirements of diaper surface marker on-line inspection.