This study focuses on noise reduction in neutron tomography, a non-destructive imaging technique with inherently low signal-to-noise ratio (SNR) due to the nature of neutron sources and detectors. This research explores the application of advanced denoising techniques, departing from iterative reconstruction methods, to improve image quality. Using a sample consisting of an aluminum cylinder filled with layers of copper balls and metal bars of various compositions, we reconstructed neutron tomography data at varying exposure times with the Filtered Backprojection (FBP) algorithm, followed by denoising using the Block Matching 4D (BM4D) algorithm. Results demonstrate that BM4D significantly enhances image quality by reducing normalized mean squared error (NMSE) and improving contrast-to-noise ratio (CNR). These findings suggest that BM4D can substantially reduce acquisition times in neutron tomography while maintaining detailed structural visibility and high image quality.