论文部分内容阅读
针对航空影像中目标提取的困难，提出了一种基于可变模板的航空影像中建筑物提取方法，并针对几种比较典型的建筑物，详细讨论了模板的设计、目标初始参数的获取及模板最优参数求解等几个关键性的问题。采用这种方法进行航空影像中目标的提取时，可以将影像中的多种信息和人的识别能力进行有效的融合，因而能够达到可靠地提取影像中目标的目的。作为一种有效的优化求解方法，模拟退火被用于目标提取的优化过程。为了验证这种方法提取航空影像中目标的能力，给出了利用可变模板提取航空影像中几种不同类型的建筑物的例子
Aiming at the difficulty of object extraction in aerial imagery, this paper proposes a method of building extraction in aerial image based on variable template. For several typical buildings, the design of template, acquisition of target initial parameters and template are discussed in detail. There are several key issues such as optimal parameter solving. When this method is used to extract objects in aerial images, multiple types of information in the images can be effectively combined with human recognition capabilities, so that the purpose of reliably extracting objects in the images can be achieved. As an effective optimization solution method, simulated annealing is used for the optimization process of target extraction. In order to verify the ability of this method to extract objects in aerial imagery, examples of the use of variable templates to extract several different types of buildings in aerial images are given.