Research Article Open Access

Adapting the LMF Temporal Splining Procedure From Serial to MPI/Linux Clusters

Sajia Akhter, Ipshita Sarkar, Kazi G. Rabbany, Nahid Akter, Shamim Akhter, Yann Chemin and Honda Kiyoshi


Remote Sensing (RS) provides images over large areas such as provincial or country level. During the last 20 years, it plays a vital role for developing many complex applications. However, RS image includes data noises due to influence of haze or cloud especially in the rainy season. It is thus necessary to remove the noise to recover the real ground variations of information studied. Local Maximum Fitting (LMF) is a combined procedure, which helps to remove the noisy data. When dealing with sufficiently large and such complex processing with RS data, single computers time processing extends to unacceptable limits. Such as, to remove the noise from RS image with 146 bands, 38 rows and 37 columns which mean 146 x 38 x 37= 205276 pixels, the LMF procedure requires 26 minutes approximately. So, 1000 x 1000 Remote Sensing Image with 146 bands is required approximately two weeks. It is necessary to reduce the time constraint and make the LMF process executable in a suitable time limit. This study deals with the design and implementation of a distributed LMF procedure. Especially, inside the LMF procedure, a consecutive amount of pixels (pixels in a column) is processed for each row. This behavior is used in this study to make the LMF parallel. For processing the RS image (146 bands, 38 rows and 37 columns), the execution time reduces to 16.13 minutes by adding distributed computing to the program (37 columns distributed to 3 computers).

Journal of Computer Science
Volume 3 No. 3, 2007, 130-133


Submitted On: 29 September 2006 Published On: 31 March 2007

How to Cite: Akhter, S., Sarkar, I., Rabbany, K. G., Akter, N., Chemin, Y. & Kiyoshi, H. (2007). Adapting the LMF Temporal Splining Procedure From Serial to MPI/Linux Clusters. Journal of Computer Science, 3(3), 130-133.

  • 0 Citations



  • LMF
  • MPI
  • remote sensing
  • distributed computing
  • cluster computers