一种显著图分割的遥感油库检测方法

针对遥感影像中油罐检测识别率低、难度大的问题,提出了一种基于显著图分割分布式的目标检测识别方法,利用视觉显著模型得到油库疑似候选区域,采用多阀值Otsu方法分割出目标,以及利用油罐的似圆特征和分布式的空间分布规律对油库进行检测识别,通过油库场景分布的先验知识,提高检测识别效率、降低虚警率。大量实验表明,文中方法可有效实现对遥感影像中油库的检测识别。

A Saliency Map Segmentation Oil Tank Detection Method in Remote Sensing Image

Oil tank has been of great significance for both military and civilian applications.Its detection is an important application of remote sensing images.We propose a method based on saliency map segmentation distributed target detection and recognition for easier tank remote sensing image detection and higher recognition rate.The visual saliency model is adopted first to obtain suspected candidate depot area and the multi threshold Otsu method to segment the target;the circular and spatial distribution features of the depot are used for tank detection and identification;a priori knowledge of the scene by the distribution depot is employed to improve the detection and identification efficiency and reduce the false alarm rate.A large number of experiments show that this method can effectively realize the remote sensing images in the detection and recognition of the depot.