Cerebral blood flow velocity is one of the most essential parameters related to brain functions and diseases. However, most existing mapping methods suffer from either inaccuracy or lengthy sampling time. In this study, we propose a particle-size-related calibration method to improve the measurement accuracy and a random-access strategy to suppress the sampling time. Based on the proposed methods, we study the long-term progress of cortical vasculopathy and abnormal blood flow caused by glioma, short-term variations of blood flow velocity under different anesthetic depths, and cortex-wide connectivity of the rapid fluctuation of blood flow velocities during seizure onset. The experimental results demonstrate that the proposed calibration method and the random-access strategy can improve both the qualitative and quantitative performance of velocimetry techniques and are also beneficial for understanding brain functions and diseases from the perspective of cerebral blood flow.