Research On Soft Morphological Filters Optimization Algorithms And Their Application

Soft morphology is an important subclass of development of mathematical morphology. In 1991, L. Koskinen introduced soft morphological filters. The key idea of soft morphological filters is that the structuring set is divided into two parts: the "hard center" and the "soft boundary" where maximum and minimum are replaced by other order statistics. This makes the filter more tolerant to noise and makes the filter less sensitive to additive noise and small variations in the shapes of the objects to be filtered. In this paper, some researches of optimal algorithms and its application of soft morphological filters have been made as fellow.At first, different properties of soft morphological /filters are illustrated and analyzed by establishing the connection to stack filters through positive Boolean functions. It is shown that soft morphological filters are with many desirable properties. The experimental results of remote sensing image processing illustrated that soft morphological filters can preserve well details of image while filtering the noise.This paper focuses on the research of optimal algorithms for soft morphological filter design. Under MSE and MAE criteria, this paper proposes two effective and easy-to-be implemented algorithms for optimization of soft morphological filters, i.e. adaptive genetic algorithms and parallel annealing-adaptive genetic algorithms. What is more, in order to controlling the trade-off between noise attenuation and detail preservation in soft morphological filter design, a new method is presented by analyzing the two optimal method under breakdown probability constraint and sample selection probability constraint. The ratio between hard center and the size of soft boundary is used as constraint directly. Then, new error function (cost function) with respect to the traditional criteria MAE or MSE is established for optimization of softmorphological filers. The presented algorithms are applied in noisy image processing. Their feasibility and effectiveness are all proved by results of simulated experiments.Finally, edge detection and noise removal of subaqueous ultrasonic image are achieved successfully by using soft morphological operations. The experimental result illustrates that soft morphological filter can be used in image processing as a kind of filter with many desirable properties.