Commercial electron backscatter diffraction (EBSD) systems generally use interplanar angle matching for pattern indexing, and thus, they are unable to distinguish between some similar phases with close interplanar angles, such as Al and Si. The interplanar spacing is more diagnostic but generally difficult to apply in pattern indexing because it lacks precision. In this study, we proposed an efficient approach for accurately measuring interplanar spacing by correcting the reciprocal-lattice vector (RLV). The phase discrimination of Al and Si was performed by interplanar spacing matching. The Kikuchi bands were identified automatically by the self-developed method using pattern rotation combined with grey gradient recognition without the help of human eyes. The reliable RLV relationship was extracted by accurately drawing reciprocal-lattice vectors.The lengths of RLVs were corrected, and then the RLVs were used for evaluating lattice spacing. The results of five Kikuchi patterns with different clarity showed that this new method reduced the average error of interplanar spacings by 50.611% and achieved an average accuracy of 1.644% for lattice spacing calculation. The method could distinguish structures with a difference in lattice spacing of at least 3.3%. This method was also effective for fuzzy patterns and partially missing Kikuchi bands and might be used as a new strategy for improving the calculation accuracy of lattice spacing for fuzzy patterns. The method did not have additional requirements concerning the number of detected Kikuchi bands and poles. The accuracy of lattice spacing could be effectively improved by correcting the RLVs based on routine pattern recognition. This method might be used as an auxiliary approach to differentiate between similar phases and is well-adapted to the existing commercial EBSD system.