Astronomical imaging of a star cluster is one of the paramount ways to learn about stellar evolution, stellar dynamics. A large telescope is not generally accessible to all observers. In that context, small telescope observations with a proper denoising scheme can be an excellent alternative. This paperproposes a technique to denoise star cluster data using an undecimated wavelet transform, with a modified thresholding process. Our work aims to prove the effectiveness of such a wavelet-based technique on realtime data. We present drastically noise-infested observational data of the NGC 2301 star cluster, captured over five nights from Fr. Eugene Lafont Observatory, Kolkata. We observe that for highly noise-polluted data, the conventional methods of dark frame subtraction and flat frame division are inadequate to producethe desired quality of images due to functioning exclusively in the spatial domain. Thus, we take the wavelet-based multiresolution approach to ameliorate those raw images. We also introduce a modified thresholding function to modulate the image at different resolution levels. A standard star detecting software Daophot II quantifies the increment in the number of detected stars from raw images to theimages processed by our proposed method as: for red filter 397–903, for green filter 663–945, for blue filter 362–896. On the contrary, Daophot II can’t detect any star in the highly noise-polluted images processed by the conventional methods. Therefore, we hope our proposed processing methodology willmotivate others to initiate small telescope observations from any site restrained by its geographical location.