Measurement of irregular particle size from the reconstructed shape images

A method for measuring the size of irregular particles is proposed. The smallest bounding rectangular box of the particle image can be acquired after obtaining interferometric out-of-focus image by using interferometric particle imaging technology and reconstructing particle shape through convolutional neural network. The number of image pixels of the particle can be obtained from the smallest bounding rectangular box. A new pixel size is defined by using the box counting method based on the analysis of the relationship between particle size and pixel size. The width and height of the particles are then obtained. The maximum relative errors in width and height are 1.47% and 1.16%, respectively. The results indicate that the proposed method can provide a research basis for the realization of three-dimensional size measurement of particles.