In the brewing process, many factors affect the quality of products, such as raw material quality, craft, winemaking techniques, etc. In addition bottle defect (bottle day, the bottle and the bottle) and dirt also affect wine quality - a key element. Therefore, the wine must be empty before filling bottles for testing, remove the substandard goods, in order to carry out the wine bottling. At present, the majority of our winery or an empty glass bottles with artificial quality testing, inefficient, slow and labor intensive. Machine vision is perfect for high-volume production process measurement, inspection and identification, in a continuous linear CCD scan line measurement application is an advantage. Machine vision detection method can greatly improve the degree of automation of production, and machine vision is easy to realize information integration, can greatly improve product quality, increase production efficiency.

So, for glass bottles to detect this with a highly repetitive work, the application of machine vision to develop a theory of empty bottles to achieve automatic detection device detects the practice of great significance. I. Introduction Machine Vision Machine Vision is to use machines instead of human eyes do measure and judge. Machine vision systems is through the machine vision products (ie, the image pickup device, two kinds of CMOS and CCD) will be converted to an image pickup target signal sent to a dedicated image processing system, the pixel distribution and brightness, color and other information, change the digitized signal; imaging systems for various operations to these signals to extract the characteristics of the target, and then based on the results of discrimination to control the action of equipment.