Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
8145967 | Infrared Physics & Technology | 2018 | 22 Pages |
Abstract
Due to the weak feature and wide angle of long-distance aircraft targeting in the parking apron from front-looking infrared images, there are always false alarms in aircraft targeting detection. This leads to relatively poor reliability for detection results. In this paper, we present a scene-driven coarse-to-fine aircraft target detection method. First, we preprocess the image by combining the sharpened and enhanced images. Second, the region of interest (ROI) is segmented by using the local mean variance of the image and a series of subsequent processing. Then, target candidate areas are located by using the feature of local marginal distributions. Lastly, aircrafts can be detected accurately by a novel aircraft shape filter. Experiments on three infrared image sequences have shown that the presented method is effective and robust in detecting long-distance aircraft from front-looking infrared images and can also improve the reliability of the detection results.
Related Topics
Physical Sciences and Engineering
Physics and Astronomy
Atomic and Molecular Physics, and Optics
Authors
Jin Lin, Yihua Tan, Jinwen Tian,