To achieve robust control of industrial crystallization processes, it is necessary to measure the sizes, shapes, and polymorphic forms (i.e., internal structures) of the developing crystal population. This paper describes a model-based object recognition algorithm designed to extract crystal size and shape information from noisy, in situ crystallization images. The effectiveness of the algorithm is demonstrated using in situ images obtained at low, medium, and high solids concentrations during an alpha-glycine cooling crystallization in water. With respect to measurements obtained through manual image analysis by human operators, the algorithm gives reasonably accurate size and shape measurements. The algorithm is sufficiently fast to enable real-time monitoring for typical cooling crystallization processes. (c) 2006 Elsevier Ltd. All rights reserved.