Rotation Sliding Window of the Hog Feature in Remote Sensing Images for Ship Detection

Ship detection plays a relatively vital role in the effect of the traditional of military. In remote sensing image, we combined Histograms of Oriented Gradients features and support vector machine for ship detection. However, hog feature does not have a rotation of invariant, ship can be in any direction. Consequently, in this paper, we proposes a measure of continuous interval rotating detection sliding window of hog feature. We extract and train hog feature of positive and negative samples. Then, continuous interval rotating sliding window of hog feature to improve the accuracy of detecting ship. The experiments reveal that the detection rate can reach high of 72.7% in the vertical direction of test ship. It is of practical significance for civil and military field.