ROBUST K-MEANS TECHNIQUE FOR BAND REDUCTION OF HYPERSPECTRAL IMAGE SEGMENTATION

What is it about?

The presentwork is address about the latest techniques for segmentation of remotely sensed hyperspectral scenes, assembled through airborne or space-borne Earth observation instruments. The inter band cluster and intra band cluster techniques has examined for spectral un-mixing and hyperspectral image segmentation. The inter band clustering is carried out with K-Means, Fuzzy C-Means (FCM) and Robust K-Means (RKM) clustering mechanisms, whilst the Particle Swarm Clustering (PSC) mechanism ought to be used during intra band cluster parts. DB (Davies Bouldin) index be utilized to figure out the quantity of clusters. The hyperspectral bands have clustered besides a band that has predominant variance from every cluster has singled out, that makes diminished band. Moreover, PSC put forward the segmentation strategy on this reduced band. In PSC, the segmentation is put forward out by enhanced algorithm entitled as Enhanced Estimation of Centroid (EEOC). Performance of the above method is evaluated an assortment of scenarios in terms of pixels clustered and time complexity.

Featured Image

Photo by Jeremy Bishop on Unsplash

Photo by Jeremy Bishop on Unsplash

Why is it important?

This work has carried out in RKM for band reduction. The segmentation has performed with PSC