论文部分内容阅读
地面背景的红外仿真对虚拟战场环境具有重要意义。但是在基于遥感图像的红外地景仿真中,材质信息和边界信息的获取缺乏有效的方法。针对此问题,提出了一种遥感图像自动分类方法来进行材质信息和边界信息的提取。根据数据区特点,确定分类类别并建立样本库;对样本库进行特征提取形成样本特征库,利用样本特征库对分类器进行训练;对待分类图像进行特征提取后,输入分类器进行分类识别;对分类后的结果进行后处理和边界跟踪,进而得到边界信息和材质分类信息。实验表明,所提出的方法能有效地为红外地景仿真提供前端数据。
Infrared simulation of the ground background is of great significance to the virtual battlefield environment. However, in the infrared landscape simulation based on remote sensing images, there is no effective way to obtain material information and boundary information. Aiming at this problem, a remote sensing image automatic classification method is proposed to extract material information and boundary information. According to the characteristics of the data area, the classification categories are determined and the sample database is established. The sample database is extracted to form the sample database, and the sample database is used to train the classifier. After the classification of the image is extracted, Sorted results for post-processing and border tracking, and then get the boundary information and material classification information. Experiments show that the proposed method can effectively provide front-end data for infrared landscape simulation.