CAJ | 학술논문

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。
재요감도상처리과정중，목전지면목표식별절대다수채용기우도상분할적방법。저사방법재도상가이교호분할화이치화시가이득도비교이상적효과。단시재처리과정중，유시난이득도이상적분할결과，저필정회영향최종적처리결과。인차，제출일충불수요진행도상분할，이시기우도상정체특정적대형지면비궤식별방법。대비궤선험수거진행분석학습，이용고빈신식증강적Gabor려파기적려파결과，여경과학습득도적선험수거진행상사성대비진행식별。통과실험대방법진행료험증。
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.