Abstract
Efficient aircraft detection in large-scale remote sensing images with complex scene has an important significance. In this paper, a novel aircraft detection method combined with ALLNet location and component matching is proposed. Firstly, during the suspected target locating stage, ALLNet with position estimation was designed to estimate the position of the suspected aircraft target quickly. Secondly, a component matching algorithm is designed in the target identification stage. Within the area of the suspected target position, the spatial component of aircraft is expanded to improve the accuracy of identification. A series experiments are conducted on the dataset from Google Earth, which includes images of airport. The experimental results show that the proposed method can achieve high performance in the aircraft target detection and identification.
Translated title of the contribution | A Remote Sensing Airport Aircraft Detection Method Combined with ALLNet Location and Part Matching |
---|---|
Original language | Chinese (Traditional) |
Pages (from-to) | 47-51 |
Number of pages | 5 |
Journal | Beijing Ligong Daxue Xuebao/Transaction of Beijing Institute of Technology |
Volume | 39 |
Publication status | Published - Oct 2019 |