Design Of On-line Particle Size Measurement System Based On Image

Particle image processing and analysis is an important research hotspot in the field of image research.In the process of industrial granulation,the real-time particle size information is an important reference element in the production control process.Applying image processing and analysis technology to the granulation production process can monitor and obtain the key information such as the average particle size,particle size distribution and particle shape in real time,which can effectively improve the quality and efficiency of product particle production,as well as the production of subsequent products Process automation provides key core technologies.Based on the particle image,this paper designs a particle size measurement algorithm and an online measurement software system.The main contents of this paper are as follows:Design of particle size measurement system.Firstly,the composition and selection of the hardware of the measurement system are introduced,including camera selection,telecentric lens selection and auxiliary light source system design.In order to meet the requirements of measurement in real environment,the camera lens is sealed.Then it introduces the design of the system software.Based on QT software,this paper designs a set of on-line particle size measurement system.The working process of the software mainly includes the steps of image acquisition,image algorithm processing and particle size analysis.Through the designed software interface,the interaction between the operator and the system can be realized.Particle image preprocessing.In view of the noise in the process of image acquisition,this paper compares two kinds of common denoising algorithms,and finally uses median filtering algorithm to denoise,and then compares a variety of image binarization algorithms.According to the binarization results of particle image,the local adaptive thresholding technology with better binarization effect under the condition of uneven illumination brightness is selected,and in order to avoid binarization In this paper,morphological filtering is used to remove the isolated small area and fill the holes.Particle image segmentation and contour reconstruction.In this paper,the distance image combined with h-minima transform is used to get the seed points.Then the seed points and background are marked on the distance image.Finally,the distance image is segmented by watershed.Due to the partial missing after segmentation of conglutinated particles,according to the prior knowledge that the shape of particles is similar to ellipse,this paper reconstructs the contour of the incomplete particles by ellipse fitting technology,so that they can be reduced to the actual contour,and then filters out the invalid particles in the image by fitting error and K-means clustering algorithm,which can ensure the validity and accuracy of particle measurement.On line measurement and result analysis of particle size.The measurement algorithm of particle shape and particle size characteristics is determined.The particle shape is defined by shape index,and the diameter of equal projection area circle is taken as the particle size.The histogram and line chart are used to show the particle size distribution and particle size distribution characteristics respectively.The calibration of the measurement system parameters is completed,so that the particle size can be transformed from the pixel representation to the real size.Finally,three groups of experiments are carried out for particle size measurement,and the error of particle size measurement is analyzed.