基于模糊Renyi熵和区域增长的图像目标分割方法

为了实现快速精确的航空侦察图像目标分割,提出基于模糊Renyi熵和区域增长的分割方法。首先在Renyi最大熵分割的基础上,应用模糊隶属度函数,引入模糊Renyi熵,提高图像分割效果。然后为了获取种子点,提出了基于双金字塔和特征融合的显著性检测方法,并通过形态学重构开运算和区域极大值生成目标核心区域。最后,增长准则设计为将图像分割结果进行二值标记,然后选取与目标核心区域重叠最多的区域块为目标分割结果。实验结果表明,所提方法可实现复杂场景航空侦察图像舰船目标的快速和精确分割。

Image target segmentation method based on fuzzy Renyi entropy and region growing#br#

In order to achieve fast and accurate target segmentation of aerial reconnaissance images, a segmentation method based on fuzzy Renyi entropy and region growing is proposed. Firstly, based on the segmentation of Renyi maximum entropy, fuzzy membership function is applied, and fuzzy Renyi entropy is introduced to improve the image segmentation effect. Then, for obtaining the seed points, a saliency detection method based on dual pyramids and feature fusion is presented, and the target core region is generated by morphological open operation using reconstruction and region maximum. Finally, the growing criterion is designed to get the binary mark of image segmentation results, and the region block with the largest overlap on the target〖JP2〗 core region for the target segmentation results selected. The experimental results show that the proposed method can realize the fast and accurate segmentation of ship targets in the aerial reconnaissance images under various complex scenes.