Intercellular glial calcium waves constitute a signaling pathway which can be visualized by fluorescence imaging of cytosolic Ca2+ changes. However, there is a lack of procedures for sensitive and reliable detection of calcium waves in noisy multiphoton imaging data. Here we extend multiscale vision model based on the undecimated wavelet transform for detection and extraction of calcium wave events in cerebellar cortex in vivo. In experimental data and validation studies using simulated data, the detection and characterization of glial calcium waves was significantly improved as compared to conventional analysis based on pixel thresholding. The described method allowed to visualize the underlying complexity of the glial networks as a prolonged Ca2+ elevation was observed at the origin of the wave, which could connect several recurrent waves separated by short time gaps, the data suggested a possible interplay between the waves and interneurons and the expanding front of the waves cold take variable forms and sometimes deviated from simple geometrical shapes, being jagged or making curved spurts when occurring near a blood vessel or a location of a previous wave. The calcium waves sprawled along blood vessels where propagation apparently occurred from end-foot to end-foot and was faster than away from vessels. Thus, small and large networks of molecular layer astrocytes and Bergmann glia can be activated in intercellular calcium waves, which are of potential importance for synaptic function and blood flow control.