This paper reports our work on various aspects image processing and statistical analysis based on local texture and crystal object structure of biocrystallogram images. We built a modular test engine that executes objects of image data and analysis schemes, proposed a series of image processing and statistical analysis methods and have implemented a Gabor filter bank, a principal component analysis function capable of operating on high-dimensional data sets, and have our results summarized in a set of tests using multivariate statistics (MANOVA testing for grouping- or factor- effect by use of Wilk's , and group similarity measures based on Mahalanobis' distances between group centres). Finally we discuss issues of marginal interest to the central scope of the present study such as image registration errors in order to explain result deviations encountered, and the types of analysis tasks performed the requiring Laboratory's in order to devise appropriate statistical tests for the operational statistical analysis not implemented at present in the Laboratory's organization. In this project, structure has been approached as localization, orientation and size. In this study, the mentioned approach to structure took preference over the approach as de ned by a tree-shaped object located in the image by segmentation allocated a set of values for each "limb" and deploying graph theory to analyze the objects. Although the ultimate results of this pilot leave room for improvement, both of the image processing (filtering, segmentation, etc.) and classification, we do show that the methodologies presented have a promising potential for implementation in a future operational information management system run at the Laboratory.