Remotely sensed image segmentation by nonlinear correlation

In this paper, we describe an implementation of a new method for olive tree detection based on a remotely sensed data by applying nonlinear correlation technique. In order to look for a known object, we use its mask omega and sweep the considered image for computing the correlation between the mask and windows phi of the image. Accordingly, we defined a doorstep of resemblance decision between the picture and the mask by choice of the maximal correlation measure. Applying this technique for remote sensed image segmentation and olive trees extraction, as well as computing the vegetation indices, we get encouraging results