Raw water pretreatment can be divided into three stages: coagulation

sedimentation and filtration. The flocculation process directly affects the structure of alum and the turbidity of effluent. At present

the method of measuring the turbidity of the effluent is usually used to control the dosage of coagulant

but due to the time lag

it can not quickly reflect the coagulation effect and adjust the dosage of the feedback. With the rapid development of computer technology

the application of alum image processing technology can realize rapid

accurate and real-time detection of flocs state

so as to control the dosage more accurately and improve the coagulation effect.From the perspective of computer vision technology

this paper summarizes the technical characteristics of alum image acquisition and processing in flocs

introduces the method of real-time tracking and calculating the characteristic parameters of alum structure

such as equivalent particle size

fractal dimension

etc.

and some test results are also provided. Through these parameters the best coagulation effect can be judged

which provides the basis for coagulation control and dosing.