Noise of image does not only reduce the quality of image but also interferes with the validity of correlative processing arithmetic seriously. Therefore, effective and robust methods of removing noise are very important for various signal processing. To improve quality of the actual distance remote control image in the paper, MeanShift algorithm of no parameter estimation is introduced and five methods of removing image noise are compared. Firstly, based on the reasonable assume to be noise model to remove image noise. Kalman filtering is used to remove noise. Then median filtering, mean filtering and Wiener filtering are performed separately. Finally, MeanShift algorithm is applied to remove noise. Experimental results show the five methods which are used in this paper reduce the noise in the night sky image to varying degrees. Moreover MeanShift algorithm in image denoising keeps the detail information and edge character of the image better. Compared with the five traditional filtering methods, MeanShift algorithm shows the advantage in image denoising of the actual night sky image background noise.