|
Scientific Visualization
| Issue Year: | 2017 |
| Quarter: | 4 |
| Volume: | 9 |
| Number: | 4 |
| Pages: | 89 - 96 |
|
| Article Name: |
FOREIGN OBJECT DETECTION IN TV AND LWIR IMAGERY OF A RUNWAY BY DEEP CONVOLUTIONAL NEURAL NETWORKS |
| Authors: |
V.V. Kniaz (Russian Federation), V.V. Fedorenko (Russian Federation), V.A. Mizginov (Russian Federation), V.A. Knyaz (Russian Federation), W. Purgathofer (Austria) |
|
The paper is recommended by program committee of 27th International Conference on Computer Graphics and Vision GraphiCon’2017. |
| Address: |
V.V. Kniaz
State Res. Institute of Aviation Systems GosNIIAS, Moscow, Russian Federation
Moscow Institute of Physics and Technology MIPT, Russian Federation
V.V. Fedorenko
State Res. Institute of Aviation Systems GosNIIAS, Moscow, Russian Federation
V.A. Mizginov
State Res. Institute of Aviation Systems GosNIIAS, Moscow, Russian Federation
V.A. Knyaz
State Res. Institute of Aviation Systems GosNIIAS, Moscow, Russian Federation
Moscow Institute of Physics and Technology MIPT, Russian Federation
W. Purgathofer
TU-Wien – Vienna University of Technology, Wien, Austria |
| Abstract: |
The presence of foreign objects on airport runways poses a significant threat to the safety of air travel. Infrared camera based runway monitoring systems for automatic detection and visualization of foreign objects are highly demanded nowadays. Deep neural networks have recently became a powerful instrument for analysis of multispectral image sequences. This paper is focused on the development of an new deep neural network architecture for automatic detection of foreign objects on a runway. The architecture is based on the SqueezeNet network. The new network performs detection using a pair of images captured in visible and far infrared ranges. |
| Language: |
Russian |
| DOI: |
http://doi.org/10.26583/sv.9.4.09 |
|
|
|
