Batch to Batch Variation Identification in Production of Crystalline Products Based on Image Analysis and Particle Shape Classification

For crystalline particular products such as pharmaceuticals, shape, as a product quality variable, can be as important as size. This paper presents a methodology to derive shape descriptors from microscope images of particulate products and the use of the descriptors for classification and identification of batch to batch variations. Data from batch runs of laboratorial and industrial crystallisers are analysed.