视觉认知理论在遥感影像道路网提取中的应用

Application of Visual Cognitive Theory in Road Network Extraction from Remote Sensing Images

在遥感影像上,道路被认为是颜色、纹理、形状相似的狭长线状目标,整个道路网在影像上会呈现非常显著的特征,极易引起人眼的注意,称之为感兴趣区域。它是场景中最能引起用户兴趣、体现图像主要内容的区域。视觉认知理论的研究表明:通过视觉注意机制可以模拟人眼的观察过程,找出遥感影像上的显著区域。文章提出应用视觉注意机制辅助遥感影像道路网提取的思想,通过对影像的显著区域进行分析和处理,得到最终的道路网。对比实验表明该算法可以有效地提高道路网提取的准确率和完整性。

In remote sensing images, roads are considered to be a long and narrow linear target which is similar in color, texture and shape. Based on these features, the entire road network in the image will show a very significant feature, which can easily excite the attention of human eyes, and is called the region of interest （ROI）. ROT in the scene can cause the most interest of users, and reflects the main content of the image area. Visual cognitive theory study shows that the visual atten- tion mechanism can simulate the observation processing of human eyes to identify the salient region of remote sensing images. This paper proposes the idea of using visual attention mechanisms to assist road network extraction by analyzing and processing the salient region, and gets the final road network. Comparative experiments show that the algorithm can effectively improve the accuracy and integrity of the road network extraction.