The Study On Tank Liquid Level Detection Method Based On Image Processing

In production and daily life, people often use various types of tanks, such as: chemical storage tanks, the pouring cup of foundry industry, the boiler steam drum, reservoir, etc.Through the tank liquid level measurement, you can quickly adjust the capacity of liquid tank and make the production process smoothly. It is extremely important for the production and living security of people to measure level accurately, reliably, timely. There are several methods are used in level measurement, for example, float ball level gage, pressure pickup level gage and radar level gage,etc.They are often vulnerable to environmental factors, large errors. Liquid level detection based on image processing technology can effectively overcome these shortcomings. The method is being used widely. By reading a great deal of native and foreign literature, the author design a tank liquid level detection method based on image processing.This paper is mainly composed of the gray-scale transformation, noise filtering, image geometric transformation, the coarse location and fine location of liquid etc.The key techniques and new methods presented in this thesis are listed below:(1) Filtering, on the basis of traditional filtering methods, according to the characteristics of the level line and the noise, a denoising method is proposed.(2) Extracting interest region, the Integral projection is improved according to the characteristics of the level image. The liquid level area has been determined effectively.(3) Fine location of liquid, two methods is proposed. One is Hough Transform, the other is chain-code tracing + curve fitting. Two methods was optimized in the papers according to the characteristics of the level image. The ways can be selected according to the different environment.Simulation and programming on different tank are proposed in this article, results show that the algorithm can detect the liquid level effectively.