基于边缘敏感递归滤波的彩色航拍图像云检测

Cloud detection of color aerial images by edge-aware recursive filtering

为了有效检测出彩色航拍图像中的云区,提出一种基于边缘敏感递归滤波的彩色航拍图像云检测方法。通过基于自适应引导权的边缘敏感递归滤波算子构造航拍图像的差异映射,并结合基于类内方差的递归阈值选择方法,将输入的航拍图像区分为待定云区和非云区。利用边缘敏感递归滤波算子对原始航拍图像累进滤波,有效抽取分层的图像细节,构造航拍图像的细节映射以精化待定云区。利用客观性能指标对检测结果进行评价。实验结果表明,相较于其他相关研究工作,提出的方法有效易行,计算复杂度低。

In order to detect cloud areas of color aerial images effectively, a cloud detection method of color aerial images by edge-aware recursive filtering is proposed. The variation mapping is constructed using adaptive guided weight based edge-aware recursive filtering. Based on the variation mapping, the input aerial image is di- vided into pending cloud areas and non-cloud areas by intra-class variance based recursive threshold selection. The original aerial image is progressively filtered using the edge-aware recursive filtering and the hierarchical de- tails are effectively extracted. The detail mapping of the aerial image is constructed and used to refine the pend- ing cloud areas. The performance of the detection results is evaluated by the objective indicator. The experimen- tal results show that the proposed method is effective, easy, and of lower computational complexity compared with other similar methods.