Light polarizing-microscopy is a technique applied to identify different phases in liquid crystal analysis. With the linear increase of the temperature of the liquid crystal sample, in the image sequence acquisition stage, changes in the images, generated by the microscope, are observed. Such changes are related to the different phases and phase transitions. Based on Mathematical Morphology, this work aims to propose a new method for detection and quantification of phase and phase transitions. For this purpose lyotropic liquid crystal samples, exhibiting the classic sequence of phases: reentrant isotropic (Ir)→ discotic nematic (ND)→ biaxial nematic (NB)→ calamitic nematic (NC)→ isotropic (I) are studied. The connected morphological operators were shown to be very efficient to extract information from the complex structures. Once the image sequence is acquired, it may be processed by techniques designed to phase transition detection. From the image analysis point of view, texture changes are more evident when observed the all image sequence, and it allows to extract many features that may be applied to detect phase transitions and to characterize each phase as well. In this proposal, such features are extracted from the residue of a graylevel area opening operator.