A novel approach for adaptive threshold selection from gray level histogram using index of fuzziness is presented. The principle of threshold selection through minimizing index of fuzziness is described first

and then how to choose the membership function and window size is illustrated. Based on the object-background-contrast computed from the distance between the object and the camera

a new method for the automatic choice of window size of a predetermined membership function is provided. Also an algorithm to obtain object-background-contrast from the distance between the naval vessel and the camera is suggested. The experiment is conducted over three images of a target warship with different distance between the object and the camera and with different contrast. The results show that the approach proposed herein has better performance on adapting the scenario variance.