Remote Sensing Image Ship Target Detection Method Based on Visual Attention Model

Sun Yuejiao,Lei Wuhu,Ren Xiaodong
DOI: https://doi.org/10.1117/12.2296260
2017-01-01
Abstract:The traditional methods of detecting ship targets in remote sensing images mostly use sliding window to search the whole image comprehensively. However, the target usually occupies only a small fraction of the image. This method has high computational complexity for large format visible image data. The bottom-up selective attention mechanism can selectively allocate computing resources according to visual stimuli, thus improving the computational efficiency and reducing the difficulty of analysis. Considering of that, a method of ship target detection in remote sensing images based on visual attention model was proposed in this paper. The experimental results show that the proposed method can reduce the computational complexity while improving the detection accuracy, and improve the detection efficiency of ship targets in remote sensing images.
What problem does this paper attempt to address?