A method of image correlation is presented to study sequential microscopic observations of human Haversian cortical bone. Imaging biological tissues is sometimes challenging owing to their complex microstructures in particular when microcracks appear. Bone microfractures can be studied in micro compression tests where the progressive growth of small cracks is imaged by light microscopy. The two-dimensional displacement field on the sample surface is then tracked by various digital image correlation methods based on cross-correlation formulation. Because of the potential high number of sequential observations, the method calculates the displacements at given growth steps obtained either by direct comparison of the studied step and the undeformed initial state, called 'direct correlation', or by iterative comparisons of successive pairs of observations, called 'gradual correlation'. In the gradual procedure, two cases are studied, referred to as 'invariant gradual correlation' and 'varying gradual correlation', when the correlation domain is transferred till the last observation or reinitialised for each image pairs. As bone is highly heterogeneous, two types of correlation procedures are considered with or without domain partition (WDP or WODP) delimiting material and strong discontinuities. The precision of the methods is specifically evaluated for experimental observations.