Abstract—We present in this paper the analysis results of surface greenness patterns over some metropolitan areas in the East and Southeast Asia through remotely sensed measurement index obtained from the NOAA satellites. Remote sensing index used in our case studies is a vegetation health index that had been proven a proper proxy for Earth surface greenness assessment. This index had been computed and recorded as time series in a weekly timeframe. The greenness patterns of Bangkok, Beijing, and Ho Chi Minch City are used as demonstration cases of our analysis methodology, which is based on the time series clustering. The remotely observed vegetation health indices over Bangkok, Beijing, and Ho Chi Minch City during the years 1982 to 2016 had been clustered with three clustering algorithms: k-means, two-step, and Kohonen self-organizing network. The number of appropriate clusters is automatically determined from the Silhouette coefficient evaluation. Based on this coefficient, clustering results of greenness trends over Beijing and Ho Chi Minh City areas show the formation of three clusters, whereas the greenness pattern of Bangkok is more fluctuate with the formation of five clusters. The more number of clusters, the higher variation in greenness patterns.
Index Terms—Greenness patterns, NOAA remote sensing data, time series clustering, vegetation health index.
N. Kerdprasop is with the School of Computer Engineering and head of Data Engineering Research Unit, Suranaree University of Technology, Nakhon Ratchasima 30000, Thailand (e-mail: email@example.com).
K. Chansilp is with the School of Computer Engineering, Suranaree University of Technology (e-mail: firstname.lastname@example.org).
K. Kerdprasop is with the School of Computer Engineering and Knowledge Engineering Research Unit, Suranaree University of Technology, Nakhon Ratchasima 30000, Thailand (e-mail: email@example.com).
Cite: Nittaya Kerdprasop, Kacha Chansilp, and Kittisak Kerdprasop, "Greenness Pattern Analysis with the Remote Sensing Index Clustering," International Journal of Machine Learning and Computing vol. 7, no. 6, pp. 181-186, 2017.