Research Of Thermometer Inspection And Processing Based On Machine Visio

Inspection of mercury thermometer is the key in the process of production. But most of the factories are still using the traditional method of artificial inspection in our country, this approach has some problems in large investment, easy fatigue, eye big error. A mercury thermometer inspection system based on HALCON is presented. The system includes image acquisition, image processing, thermometer reading recognition. The main work that is completed as follows:1. We put the mercury thermometer in the constant temperature water tank until mercury column height is stable, and temperature of the water tank is inspection temperature. Using CMOS camera to take the image of mercury thermometer and passing the image to the computer.2. The image is processed by the HALCON machine vision software. First of all, the original image gray image is converted to a level 256 image. Then the image is enhanced by the image enhance processing to make the scale and the mercury column prominent.3. We propose the templates matching and image pyramid methods in the process of scale and the mercury column recognition. This method can obtain the image that contains scale and the mercury column quickly and accurately. And presented to match the image of bot-hat transformation, so that the scale of the mercury column region more prominent, while eliminating background influence.4. The regions of scale and the mercury column are obtained with the method of threshold segmentation. We make open operation with the region of mercury column for three times to eliminate the burr and protrusions at the top of light and bot-hat transformation affects. Then according to the position of the mercury column scale and vertex position to calculate the thermometer readings and verification of temperature contrast to determine the thermometer whether qualified. The system with accuracy of 0.1℃ can meet the requirement of industry.