A Study On The Key Algorithms Of The Rmometor Automatic Reading System Based On Image Processing

Nowadays, manual methods which have many downsides are basically adopted in thermometer measurement and determination in our country. Therefore, it is a significant problem to realize the automation of thermometer measurement. In order to improve the accuracy and efficiency of thermometer measurement and faced with multi-objective measurement requirement and based on Image Processing Technology, this thesis comes up with and realizes the multi-objective thermometer dynamic real-time image reading processing algorithms.In the beginning, this thesis analyzes the researching status quo and the existing problems of image-based processing thermometer automation reading theory and technology with a combination of relevant foreign and domestic literature information. On the basis of concluding and absorbing the present researching results, combining engineering practices and applying Hough transformation algorithms and mathematical morphology theories, adopting image processing techniques like image binarization and image denoising, etc, this thesis proposes dynamic image segmentation algorithm for multi-target identification, non-deterministic standard line localization algorithm, non-deterministic mercury line position determination algorithm, non-deterministic thermometer scale line extraction algorithm and non-deterministic mercury stigma determination algorithm. On the basis of VS2008Development platform and SQL Server database and applying C++, C#programming technique, this thesis achieves liquid thermometer static and dynamic real-time reading algorithms and related technologies, and completes the thermometer automatic reading system.Through practical application by Institute of Measurement Technology in Zibo, the thesis carries out applying research with the proposed algorithms and established thermometer automatic reading system. The analysis of applying results demonstrate that the thermometer automatic reading algorithms proposed by this thesis have a strong capacity in adapting environmental changes and a high automatic reading accuracy and it can also run smoothly. Thus, thermometer testing efficiency has improved significantly and met the requirements of practical works.