Research And Application On Flocs Detection And Sedimentation Velocity On-line Measurement Based On Computer Vision

Drinking water treatment process consists of coagulation, sedimentation, filtration anddisinfection generally, Dosing amount of coagulant is not only affect the effect of coagulationdirectly, but also relates to the water quality and the cost of preparing the water of watertreatment plant. The flocculation process is complex, it has different influential factors, thecharacteristics are nonlinearity and large time-delay, raw water quality, the types ofcoagulants, PH and other factors will affect the coagulation process. To control the dosingamount of coagulant accurately is a very difficult problem, the floc sedimentation velocitydetection (FVD) is a method of high efficient and more accurate detection, It takes thesedimentation velocity of flocs as main parameters, so don’t consider the complex physicaland chemical process. As for the FVD control system exists present problems such as insedimentation velocity, the measurement is inaccurate and the mathematical model is notperfect generally, so put forward a fusion model of multi-source data for control dosageamount, which takes sedimentation velocity as master control, and floc size, quantity asauxiliary control. Set up the hardware system for experiment; Do research on the flocdetection and segmentation algorithm, to improve the segmentation effect of flocs image, sothat to obtain more accurate characteristic parameters such as equivalent size of floc; Presentan improved algorithm of sedimentation velocity measurement based on particle filter; Makethe improved algorithm come true by programming, and the software realizes an onlineaccurate real-time detection system, which can implementate the parameters such as flocsnumber, area, and the equivalent size and the sedimentation velocity of flocs. Provides datasupport for the improved model. In this paper, the main work is as follows:(1)According to the requirements of detection and sedimentation velocity measuringsystem, design water treatment device, such as flocculation basin and sedimentation basin, atthe same time, according to the characteristics of the flocs, to choose the appropriate CCDcamera and lens and other device, then to form an detection and sedimentation velocitymeasurement hardware system in experiment environment, providing the experimentalconditions for the research of online measuring system.(2) Do pretreatment such as denoising for the high quality of image collected fromexperiments; Study different segmentation algorithm, then design an improved detectionalgorithm based on particle swarm optimization for enhanced Otsu, use the improvedalgorithm to detecte and do image segmentation, then process image by using mathematicalmorphology analysis method for further processing, Finally the algorithm of line scan connected domain identification is used to collect parameters such as the number of flocs,area, equivalent size.(3) Studying different multiple target tracking algorithms, to design an improved flocstracking algorithm based on particle filter and the compressed sensing, the improvedalgorithm tracks flocs effectively, meanwhile it solves phenomenons such as the intersection,block in the process of tracking, to draw the flocs trajectory successfully, and get thesedimentation velocity according to the coordinate computing method, improve the accuracyof the sedimentation velocity.(4) Putting forward suitable structure for the software of online measurement system forfloc, using the improved algorithms, based on the platform of VS2010, using C++programlanguage, write an software to detecte and do floc sedimentation velocity online measurement,realizing the real-time monitoring for various parameters of flocs and sedimentation velocity,the interface is friendly, and easy to use. At the same time using Matlab simulation techniquesto do some simulate experiments, to verify the accuracy and practicability of the improvedalgorithms further.