A Novel Adaptive Fuzzy Local Information C-Means Clustering Algorithm for Remotely Sensed Imagery Classification

Abstract

This paper presents a novel adaptive fuzzy local information c-means (ADFLICM) clustering approach for remotely sensed imagery classification by incorporating the local spatial and gray level information constraints. The ADFLICM approach can enhance the conventional fuzzy c-means algorithm by producing homogeneous segmentation and reducing the edge blurring artifact simultaneously. The major contribution of ADFLICM is use of the new fuzzy local similarity measure based on pixel spatial attraction model, which adaptively determines the weighting factors for neighboring pixel effects without any experimentally set parameters. The weighting factor for each neighborhood is fully adaptive to the image content, and the balance between insensitiveness to noise and reduction of edge blurring artifact to preserve image details is automatically achieved by using the new fuzzy local similarity measure.

Existing System

Expectation–Maximum, K-nearest-neighbor, Markov random field (MRF).

Proposed System

A novel adaptive FLICM (ADFLICM) clustering approach is proposed for remotely sensed imagery classification. In ADFLICM, a novel fuzzy local similarity measure is defined to replace the fixed parameter α in FCM_S. The new fuzzy local similarity measure Sir possesses several characteristics and advantages: Sir uses a pixel spatial attraction model to describe the relationships between pixels; Sir can be automatically determined by local spatial and gray level relationships between the center pixel and its neighboring pixels in a local window, and it is adaptive to the local image context without any artificial or empirical selection; using Sir, the clustered pixel is influenced by its neighboring pixels and its own features simultaneously, which is useful for retaining edges of regions and small patches when removing noise; and Sir makes the proposed algorithm relatively independent of the noise type, making it a promising choice for clustering in the absence of prior knowledge on noise. The proposed algorithm is able to overcome the drawbacks of the well-known FCM by incorporating local spatial and gray level information. The ADFLICM is effective in removing noise pixels and reducing the edge blurring artifact simultaneously. This advantage is based on the definition of a new local similarity measure, which can provide proper tradeoff between the center pixel and its neighboring pixels.

Architecture

BLOCK DIAGRAM