Abstract—Remote sensing image clustering is the issue that is interested by remote sensing researchers. Remote sensing image can have multi bands and high space resolution. There are multi algorithms as K-Means, C-Means, Watersed... Wherein, KMeans is used and applied commonly to cluster remote sensing images about clustering quality and performance. However, when clustering large size remote sensing images, the converging speed of the algorithm is still slow. Moreover, this algorithm only points out clustered attribute of each pixel but do not show center near level of pixels in each cluster. This paper presents a technique which combines the algorithm Kmeans with Wavelet technique to execute remote sensing image with large and adding weights to adjust the center computing formula.

Index Terms—Clustering, remote sensing images, Kmeans, wavelet, weight.

Trung Nguyen Tu, Huy Ngo Hoang, Duc Dang Van are with the Institute of Information Technology, Vietnamese Academy of Science and Technology, Hanoi, Vietnam.
Thoa Vu Van is with the Post & Telecommunications Institute of Technology, Hanoi, Vietnam.
An Dang Duy is with the Institute of Industrial Machinery and Instruments Holding, Vietnam.

Cite:Trung Nguyen Tu, Huy Ngo Hoang, Duc Dang Van, Thoa Vu Van, An Dang Duy, "An Improvement of KMeans Algorithm Using Wavelet Technique to Increase Speed of Clustering Remote Sensing Images," International Journal of Computer and Electrical Engineering vol. 8, no. 2, pp. 177-184, 2016.