基于HSI的视觉注意力模型及其在船只检测中的应用

根据现有关于视觉心理学研究的相关成果和计算模型．提出了一种基于HSI颜色空间特征提取的视觉注意力模型．并应用于海上目标检测．首先把输入的RGB图像转换到HSI空间上．采用高斯金字塔和center-surround算子获得HSI三个分量下各自多尺度的视觉差异。通过对不同特征图的规格化和线性融合获得综合的显图．该方法应用于多种海上目标图像均取得较好效果．背景中的海浪杂波得到了有效抑制．提取得到的显区域包括了待检测的目标且范围较小。为后继的处理和分析提供了良好的基础．

Application of HSI Based on Visual Attention Model in Ship Detection

Visual attention analysis provides a mechanism to find the salient regions within the image.Based on this mechanism,a saliency detection method in HSI color space is proposed and applied to ship detection task.Firstly,an RGB image is transformed into HSI space.Gaussian pyramids are created for three features:hue,saturation and intensity.After calculating center-surround differences between different scales of the pyramid,some feature maps are obtained.Finally,these feature maps are combined into a saliency map.The experimental results indicate that this method is effective in ship detection task.Background noises are suppressed efficiently and the extracted salient regions are good for further processing.