摘要
提出了小波分析和方向滤波相结合的有云遥感图像目标检测方法。首先对原始遥感图像进行小波分解,采用比率-平均边缘检测法检测人造目标的边缘方向,在该方向上进行Frost滤波,有效抑制云层干扰的同时突出了人造目标及其边缘信息;然后进行阈值分割并剔除大块云层,根据区域的平均边缘梯度剔除小块云层,实现了有云遥感图像的舰船目标检测。实验证明,该方法能减少遥感图片云层对舰船检测的影响,并具有较高的目标检测率。
This paper presents a ship detection method combining wavelet analysis and directional filtering for cloudy remote sensing images. First, the original image is decomposed with wavelet, ratio edge detector is adopted to get the edge direction of manmade targets and Frost filter is performed in the direction, thus efficiently restraining cloud noise and strengthening the manmade target and its edge information at the same time. Then, the filtered image is segmented and the big cloud areas are wiped off, small cloud areas are eliminated according to their mean edge - grads. The experiments show the presented method can reduce the effect of cloud of remote sensing image on ship detection and has preferable target detection probability.
出处
《电讯技术》
2008年第1期90-93,共4页
Telecommunication Engineering
关键词
遥感图像
目标检测
小波分析
方向滤波
云层干扰
remote sensing image
target detection
wavelet analysis
directional filtering
cloud noise
