In bubbling fluidized beds, important bubble properties can be estimated using electrical capacitance volume tomography (ECVT). The resolution of ECVT depends on sensor configuration and image reconstruction algorithm. At low resolution, the estimation of bubble size critically relies on threshold solids volume fraction. The majority of previous studies used a constant threshold; however, present study determined that the use of a constant threshold could fail to identify small bubbles and lead to an erroneous estimation of bubble size. A series of calibration experiments were conducted, and the data from them were used to develop an iterative method to determine the optimum threshold. The data from subsequent dynamic experiments were postprocessed using LBP algorithm and the optimum threshold method. The use of optimum threshold not only resolved the smaller bubbles but also estimated the increase in the bubble size more accurately as it moved up through the fluidized bed.