一种基于整体区域特征的地面飞机识别方法

Approach of land aircrafts recognition based on whole region characters

在遥感图像处理过程中,目前地面目标识别绝大多数采用基于图像分割的方法。这些方法在图像可以较好分割和二值化时可以得到比较理想的效果。但是在处理过程中,有时难以得到理想的分割结果,这必定会影响最终的处理结果。因此,提出一种不需要进行图像分割,而是基于图像整体特征的大型地面飞机识别方法。对飞机先验数据进行分析学习,利用高频信息增强的Gabor滤波器的滤波结果,与经过学习得到的先验数据进行相似性对比进行识别。通过实验对方法进行了验证。

In the process of remote sensing images, most of present approaches about the detection and recognition of the targets on the land are based on the region segmentation. These approaches can work well when the images can be segmented well and get the ideal binary images. However, getting the ideal segmentation results is very hard. And this is the main obstacle which will affect the results. So without the image segmentation, this paper proposes an approach of the land aircrafts recognition based on the whole region characters. This paper analyzes the parameter of the interested planes. Then by using the Gabor filter, it recognizes the aircrafts by comparing the information between the candidate and the apriori infor-mation. The experiments validate the proposed approach.