A novel operator based on grey relational analysis

A novel grey relational operator is introduced that can be viewed as an extension of the well known gradient operator. In this approach, the factors of the referential sequence are the ideal non-noise pixel and its around pixels, and the factors of the comparative sequence are the local image pixel and its around pixels. The grey relational degree can be calculated according to the grey relational analysis. The variation degree of the image pixels can be denoted by the grey relational degree. In contrast to the traditional gradient operator, this paper discusses the features of the novel grey relational operator, such as the directional features. The study of the novel operator is addressed in detail, and the results show the effectiveness of the novel operator. In all case, the novel operator can denote the relevant feature of the pixels to a variety of image analysis.