基于航摄影像的滇池捞鱼河湿地植物类型提取

Plant Types Extraction of Wetland at Laoyu River of Dianchi Lake Based on Aerial Photography

为了获取滇池湿地的植物类型分布信息,以捞鱼河湿地为示范区,采用分辨率为0.15 m的航摄影像为影像数据源,利用人机交互和面向对象相结合的影像解译方法,实现捞鱼河湿地植物类型分布信息的提取。在植物类型边界提取过程中,发现过高的影像分辨率(≤0.3 m)不利于成片具有空间异质性植物类型边界的正常提取,容易出现同一类型过度分割情况。通过邻域降分辨率重采样技术,将影像分辨率降至0.6 m时,能有效解决过度分割问题。当分辨率降至更低时,虽然能解决过度分割问题,但会导致地物边界出现不同程度的锯齿而与地物边界不符的后果。研究所应用的基于Feature Extaction模块的植物类型边界提取方法,能有效减少人机交互方法中人工勾绘边界的工作量,在保证分类精度的同时减少人力投入,对航摄影像在大范围湿地植物类型的提取中具有重要借鉴意义。

In this paper, in order to get the distribution information of Dianchi Lake wetland plant, considered the Laoyu river wetland as demonstration zone and aerial photography with 0.15 m resolution as image data source, realized the Laoyu river wetland plant types distribution information extraction by use of human-computer interaction and object-oriented image interpretation methods combination. In the plant type boundary extraction process, we found that high image resolution（ ~〈0.3 m） was not conducive to the plant type boundary extraction which distributed together and had spatial heterogeneity, and was prone to the same type of over segmentation. When reducing image resolution to 0.6 m, through the neighborhood resampling, the problem can be solved over segmentation effectively. When the resolution was lower, the over segmentation problem can be solved, but it lead the consequences of object boundary appeared different degree of serration and be inconsistent with the actual boundary. The application of plant types boundary extraction based on Feature Extaction module effectively reduced the workload in the artificial drew boundary of human-computer interaction method, and ensured the classification accuracy. This method had an important reference for the application of aerial photography at a wide range.