大范围近海舰船近红外遥感检测方法

Near-infrared Remote Sensing Detection for Wide Area Offshore Ships

随着遥感技术的快速发展,应用遥感手段进行舰船目标检测也受到了国内外的普遍关注。为此,论文利用水体对近红外波段吸收性强的特性,提出了一种基于近红外遥感的大范围近海舰船目标检测方法。该方法先用最大类间方差对影像进行海陆阈值粗分割,再通过形态学处理得到海陆精分割影像;然后在此基础上,通过方差纹理滤波器减小海面背景的灰度变化区间,增强舰船的相对亮度,从而分割出疑似舰船目标;最后,经过特征分析排除非舰船目标,得到检测结果。实验结果表明,该方法能有效地检测的舰船目标,具有较高的精度。

With the rapid development of remote sensing technology,the application of remote sensing for ship target detec tion has also received widespread attention at home and abroad.Therefore,this paper proposes a wide area offshore ship target detec tion method on the basis of near-infrared remote sensing,considering the characteristics of water's strong absorption in the near-in frared region.The method firstly uses the maximum inter-class variance to perform rough sea-land segmentation on the image,and then obtains refined sea-land segmentation image through morphological processing.On this basis,the variance texture filter is used to reduce the gray-scale variation interval of the sea surface,and to improve the relative brightness of the ship,thereby seg menting the suspected ship target.At last,after the feature analysis,the false alarm is eliminated and the detection result is ob tained.The experimental results show that the method can effectively detect the ship target with high precision.