高分辨率遥感图像港口检测的复合线索视觉注意模型

发布时间：2018-01-06 09:15

本文关键词：高分辨率遥感图像港口检测的复合线索视觉注意模型　出处：《遥感学报》2017年02期 　论文类型：期刊论文

【摘要】：为提高高分辨率光学遥感图像港口自动检测的准确性,常需综合多类线索并进行复杂的特征提取、融合与分类推理,从而带来较高的计算复杂度。为此,仿生人类视觉注意机制,提出了一种复合线索视觉注意模型,综合利用高分辨率光学遥感图像港口多尺度底层特征和高层知识线索,实现了港口检测特征自然融合与综合分类推理。该方法在提高检测效果的同时较好地控制了计算量的增长,避免了复杂特征的大范围区域提取,采用多步快速算法降低了整个算法的计算复杂度,实现了计算资源受限条件下港口的快速定位与检测。同时,由于能将有限计算资源快速聚焦于最可能含有港口目标的区域,大大提高了目标检测方法响应的实时性。来自不同卫星的高分辨率光学遥感图像实验结果,验证了提出方法的有效性。
[Abstract]:In order to improve the accuracy of port automatic detection of high-resolution optical remote sensing images, it is often necessary to synthesize multiple clues and carry out complex feature extraction, fusion and classification reasoning, which brings high computational complexity. The visual attention mechanism of bionic human beings is studied. A visual attention model with complex cues is proposed, which synthesizes the multi-scale bottom features and high-level knowledge clues of high-resolution optical remote sensing images. The natural fusion of port detection features and comprehensive classification reasoning are realized. This method not only improves the detection effect but also controls the increase of computation and avoids the large area extraction of complex features. The multi-step fast algorithm is used to reduce the computational complexity of the whole algorithm and realize the fast location and detection of the port under the condition of limited computing resources. Because the limited computing resources can be quickly focused on the area where the port target is most likely, the real-time response of the target detection method is greatly improved. The experimental results of high-resolution optical remote sensing images from different satellites are obtained. The effectiveness of the proposed method is verified.
【作者单位】：海军工程大学电子工程学院;海军工程大学装备经济管理系;
【基金】：国家自然科学基金(编号:61372165,61501485)~~
【分类号】：TP751
【正文快照】： 引用格式:毛玲,张国敏.2017.高分辨率遥感图像港口检测的复合线索视觉注意模型.遥感学报,21(2):300 309Mao L and Zhang G M.2017.Complex cue visual attention model for Harbor detection in high-resolution remotesensing images.Journal of Remote Sensing,21(2):300

【相似文献】