| In many industries,including pharmaceuticals,and chemistry,crystal size distribution(CSD) of the production is one of the most important parameters. Several methods have been presented to get the CSD,among which Laser Particle Size Analyzer(LPSA),sieving and sedimentation are the most commonly used ones. The LPSA method gives an accurate result but the price is expensive;the sieving method is simple but the result is not accurate enough;the sedimentation method is time-consuming.Against the problem mentioned above,a method to estimate the CSD of PTA based on the micro-graph from a digital microscopy imaging system is studied.In this dissertation,the micro-graphs of the crystals are obtained by using an optic microscopy imaging system combined with a digital camera.A micrometer is utilized to calibrate the imaging system to obtain the relationship between pixel length and the actual length.Moreover,image segmentation,image information extraction,distribution estimation and hypothesis testing are applied to obtain the size distribution and the volume distribution.The main contributions of the dissertation are as follows.(1) An improved image segment algorithm based on matching concave points is proposed to segment the overlapped particles effectively.(2) A global comparison-search algorithm is proposed to compute the pixels of each particle which is utilized to convert to the diameter and the volume of the particle from the segmented image together with the depth-first search and the breadth-first search.(3) The particle size distribution(PSD) and the particle volume distribution(PVD) are estimated due to a few common used distributions and chi-square test is used to test the estimation results.The effectiveness of the proposed method is demonstrated by the applications on the PSD and the PVD estimation for PTA,which provides the consistent results with LSPA.Furthermore,the method can be extended to the size distribution estimation for other crystal products. |
PDF Full Text Request