Images denoising in case impulse noise using spline approximation

Analytical equations of a new spline approximation method for filtering impulse noise in images are obtained. The proposed method differs from the known ones: when filtering images, one-dimensional sequential spline functions are used for direct and inverse transformations, and the processing is performed in rows and columns. In this work, experimental studies based on computer simulation using special test images on the background of impulse noise were conducted. Experimental studies have shown the operability and high efficiency of the developed method, which allow to improve the quality of image filtering by up to 10 dB. In this case, the properties of spline functions make it possible to abandon the use of various masks, that is, to abandon inefficient linear methods of image filtering. The method can be used to create digital image processing systems in the industry, to create autonomous robots, under observation conditions that complicate the registration process, and in the absence of a priori information about the form of background noise