Automatic Reading System for Thermometer Verification Based on Image Recognition

In order to improve the working efficiency and reading accuracy for glass liquid thermometer verification, an automatic reading system of thermometer was designed. On the one hand, the thermometer reading image was acquired after it automatically entered the reading window by the hardware; on the other hand, the image recognition and verification data processing were completed by the software. In the image reading recognition, an optimized image processing scheme was developed, the improved image tilt correction algorithm enhanced the reliability of thermometer tilt angle calculation; the gray stretch function and binarization algorithm were optimized and improved by the improved adaptive genetic algorithm to ensure to obtain the ideal binarization image; the combination of erosion and dilation algorithm got good image features while removing image noise, which provided ensure for the accurate reading; the verification efficiency was further improved by this optimized reading process. The experimental results show that the system can quickly and accurately recognize the thermometer verification reading, which also shows that the system can be effectively applied to the glass liquid thermometer verification.