Automatic detection and assessment system of water turbidity based on image processing

The realization of automatic control system of releasing flocculating agent is one of the bottlenecks in the automation process of waterworks. For various reasons, most of waterworks in China decide the dosing amount just by artificial methods, which may lead to great subjectivity. It can result in direct economic losses and threaten residents' drinking water safety. To solve the problem, based on digital image processing and pattern recognition technology, the paper presents an advanced and practical automation scheme, trying to control the releasing of flocculating agent. According to the characteristics of alum images captured from waterworks, by comparing the effects of many vision algorithms, the paper selects a suitable combination and tries to implement it with EmguCV. On next step, the paper implements feature selection and classification with LibSVM. The experiments show that, the system designed in this paper is reasonable and feasible. © 2013 Universitas Ahmad Dahlan.