A new method for marine oil spills detection is presented in this paper. It is shown that morphological images analysis techniques are powerful tools to extract accurately dark spots in SAR images, which are candidates to be oil spills. The main goal is that this segmentation can be carried out with no prior knowledge about ocean conditions during the image acquisition process. Moreover, these techniques can also be useful to extract some features used in the decision process of the candidates. Due to the homogeneity of the proposed algorithm it is suitable to be implemented in a quite efficient way. This work is focused to detect oil spills produced for cleaning the tanks of the oil tankers. Nevertheless, the main idea of the proposed algorithm is that it can be used with no restrictions.