Oil depots detection from high resolution remote sensing images based on salient region extraction

The traditional methods of detecting oil depots usually use Hough transform and template matching, which often have lower detection rates and are difficult to implement. An efficient two-step detection framework is proposed in this paper to detect oil depots in high resolution remote sensing images. In the first stage, LC saliency model is used to detect the salient regions and shows a good performance on highlighting oil depots. In the second stage, task related targets from these salient regions are extracted by removing the irrelevant salient areas according to the special properties of the targets. According to the final shape, the area and distribution of oil depots, using image threshold segmentation and the graph-based clustering procedure, oil depots are detected with fairly good accuracy and efficiency.