基于视觉显著性的高分辨率遥感图像目标检测算法

Target detection algorithm for high resolution remote sensing image based on visual saliency

基于FT视觉显著性检测技术设计了高分辨率遥感图像目标检测方法。首先,对高分辨率遥感图像进行预处理。其次,应用FT显著性区域提取算法提取高分辨率遥感图像中的显著目标区域。然后,应用Otsu阈值分割算法对图像进行分割,将遥感图像中的显著性区域分割出来,再应用数学形态学处理对显著性区域进行区域连通。最后,实现对遥感图像中显著目标的定位。在多幅高分辨率遥感图像上的实验结果验证了该方法的有效性。

In this paper,a target detection method for high resolution remote sensing images is proposed based on FT visual saliency.First,high resolution remote sensing images are preprocessed,with the purpose of smoothing and denoising.Second,FT salient region extraction algorithm is used to extract salient target regions in high resolution remote sensing images.Then,Otsu threshold segmentation algorithm is utilized to segment the images,the salient regions in the images are segmented out,and the salient regions are connected by mathematical morphology processing.Finally,the positioning location of salient targets in remote sensing images is achieved.Experimental results on multiple high resolution remote sensing images verify the effectiveness of the proposed method.