A new operator for image processing based on Lp-Filter approximation

In this paper, we define and analyse some properties of a class of order filters. These filters can be regarded as adaptive L-Filters which can be tuned by setting only one parameter instead of N, where N is the filter size. Deterministic and statistical properties are discussed. Experimental results obtained on both synthetic and real images show that the noise reduction effect of the new filter is similar to that of optimal L-Filters and optimal Lp-Filters whereas the edge preservation is improved.