Research On Floc Recognition Method In Water Based On Image Processing

With the continuous development of society,people have higher and higher requirements for the quality of domestic water.Coagulation process is the most important part of water treatment.It refers to the process of adding coagulant in raw water and forming irregular flocs through full mixing.Improving the coagulation detection process can help to increase the production efficiency of waterworks,and also better guarantee the water quality to reach the standards.Coagulation detection used in most water works has the disadvantages of low efficiency and poor adaptability.For this reason,this paper studies a method of coagulation state analysis based on image processing,through enhanced pretreatment,image segmentation and feature extraction.And three steps are used to identify floc parameters in water and establish corresponding coagulant dosing rules.The main research work of this paper is as follows.1)Sharpening pretreatment of floc imageBased on the low clarity of the image collected in the initial step,the fractional order differential operator is used to sharpen the image to improve the picture definition.Compared with the integer order differential operator,the fractional order differential operator can enhance the texture details of the image while preserving the smooth area nonlinearly.The experimental results show that the fractional order differential operator used in the paper can effectively improve the visibility of the image.With the increase of the order,the texture information of the image is gradually enriched,and the image definition is gradually improved.However,when the order value is large,the image is prone to noise.2)Segmentation of floc imageBased on the preprocessed image,the improved Otsu algorithm is used to segment the image.As the traditional Otsu algorithm has the shortcoming of high computational complexity,the FOCPSO algorithm is used to optimize the Otsu algorithm: the maximum inter-class variance is used as the fitness function.The cloud model is introduced to divide the population into several sub-populations to increase the diversity of the population.The fractional order differential is added to the velocity renewal formula to characterize the state of particles in the past time,and the particles traces is more suitable for the state of particles.The experimental results show that the improved segmentation algorithm proposed in this paper improves the convergence speed of the algorithm under the premise of guaranteeing the segmentation effect to a certain extent,realizes fast and effective image segmentation.3)Study on feature extraction method of floc imageFor the segmented floc image,this paper studies the method of image feature extraction.The method repaires the image based on mathematical morphology to reduce the effect of voids on floc recognition,and the floc particles with little correlation are separated.Then,according to the floc images in a fixed visual angle of view and a unit time,the number of floc and the fractal dimension are calculated,and the judgement rules of dosage is established.Finally,the experimental results show the effectiveness of the feature extraction method.