Parallelism and images: a parallelization experiment for image segmentation with an application for automatic classification of scenes obtained from orbital platforms

Segmentation is an important step in image analysis for remote sensing; it is responsible for splitting the image into distinct parts. Current segmentation methods used in images of this type are usually based on an analysis of the pixel values in all image bands. This type of analysis can be computationally expensive. In order to reduce the processing time in the image segmentation process, this paper introduces a parallel approach to image segmentation. This method is based on a region-growing technique. The parallelism is done using the SPMD (single process, multiple data) paradigm and the message-passing model under MPI (Message Passing Interface). The results have been compared with the output of a sequential system executing in the same environment.